Adobe Flash Player.

  Adobe Flash Player

Handbook of Pharmaceutical Granulation Technology

 2005 by Taylor & Francis Group, LLC
Executive Editor
James Swarbrick
PharmaceuTech, Inc.
Pinehurst, North Carolina
Advisory Board
Larry L. Augsburger
University of Maryland
Baltimore, Maryland
Jennifer B. Dressman
Johann Wolfgang Goethe University
Frankfurt, Germany
Jeffrey A. Hughes
University of Florida College of
Gainesville, Florida
Trevor M. Jones
The Association of the
British Pharmaceutical Industry
London, United Kingdom
Vincent H. L. Lee
University of Southern California
Los Angeles, California
Jerome P. Skelly
Alexandria, Virginia
Geoffrey T. Tucker
University of Sheffield
Royal Hallamshire Hospital
Sheffield, United Kingdom
Harry G. Brittain
Center for Pharmaceutical Physics
Milford, New Jersey
Anthony J. Hickey
University of North Carolina School of
Chapel Hill, North Carolina
Ajaz Hussain
U.S. Food and Drug Administration
Frederick, Maryland
Hans E. Junginger
Leiden/Amsterdam Center
for Drug Research
Leiden, The Netherlands
Stephen G. Schulman
University of Florida
Gainesville, Florida
Elizabeth M. Topp
University of Kansas School of Pharmacy
Lawrence, Kansas
Peter York
University of Bradford School of Pharmacy
Bradford, United Kingdom
 2005 by Taylor & Francis Group, LLC
A Series of Textbooks and Monographs
1. Pharmacokinetics, Milo Gibaldi and Donald Perrier
2. Good Manufacturing Practices for Pharmaceuticals: A Plan for Total
Quality Control, Sidney H. Willig, Murray M. Tuckerman,
and William S. Hitchings IV
3. Microencapsulation, edited by J. R. Nixon
4. Drug Metabolism: Chemical and Biochemical Aspects, Bernard Testa
and Peter Jenner
5. New Drugs: Discovery and Development, edited by Alan A. Rubin
6. Sustained and Controlled Release Drug Delivery Systems,
edited by Joseph R. Robinson
7. Modern Pharmaceutics, edited by Gilbert S. Banker
and Christopher T. Rhodes
8. Prescription Drugs in Short Supply: Case Histories, Michael A. Schwartz
9. Activated Charcoal: Antidotal and Other Medical Uses, David O. Cooney
10. Concepts in Drug Metabolism (in two parts), edited by Peter Jenner
and Bernard Testa
11. Pharmaceutical Analysis: Modern Methods (in two parts),
edited by James W. Munson
12. Techniques of Solubilization of Drugs, edited by Samuel H.Yalkowsky
13. Orphan Drugs, edited by Fred E. Karch
14. Novel Drug Delivery Systems: Fundamentals, Developmental Concepts,
Biomedical Assessments, Yie W. Chien
15. Pharmacokinetics: Second Edition, Revised and Expanded, Milo Gibaldi
and Donald Perrier
16. Good Manufacturing Practices for Pharmaceuticals: A Plan for Total
Quality Control, Second Edition, Revised and Expanded, Sidney H. Willig,
Murray M. Tuckerman, and William S. Hitchings IV
17. Formulation of Veterinary Dosage Forms, edited by Jack Blodinger
18. Dermatological Formulations: Percutaneous Absorption, Brian W. Barry
19. The Clinical Research Process in the Pharmaceutical Industry,
edited by Gary M. Matoren
20. Microencapsulation and Related Drug Processes, Patrick B. Deasy
21. Drugs and Nutrients: The Interactive Effects, edited by Daphne A. Roe
and T. Colin Campbell
22. Biotechnology of Industrial Antibiotics, Erick J. Vandamme
23. Pharmaceutical Process Validation, edited by Bernard T. Loftus
and Robert A. Nash
 2005 by Taylor & Francis Group, LLC
24. Anticancer and Interferon Agents: Synthesis and Properties,
edited by Raphael M. Ottenbrite and George B. Butler
25. Pharmaceutical Statistics: Practical and Clinical Applications,
Sanford Bolton
26. Drug Dynamics for Analytical, Clinical, and Biological Chemists,
Benjamin J. Gudzinowicz, Burrows T.Younkin, Jr.,
and Michael J. Gudzinowicz
27. Modern Analysis of Antibiotics, edited by Adjoran Aszalos
28. Solubility and Related Properties, Kenneth C. James
29. Controlled Drug Delivery: Fundamentals and Applications,
Second Edition, Revised and Expanded, edited by Joseph R. Robinson
and Vincent H. Lee
30. New Drug Approval Process: Clinical and Regulatory Management,
edited by Richard A. Guarino
31. Transdermal Controlled Systemic Medications, edited by Yie W. Chien
32. Drug Delivery Devices: Fundamentals and Applications,
edited by Praveen Tyle
33. Pharmacokinetics: Regulatory  Industrial  Academic Perspectives,
edited by Peter G.Welling and Francis L. S. Tse
34. Clinical Drug Trials and Tribulations, edited by Allen E. Cato
35. Transdermal Drug Delivery: Developmental Issues and Research
Initiatives, edited by Jonathan Hadgraft and Richard H. Guy
36. Aqueous Polymeric Coatings for Pharmaceutical Dosage Forms,
edited by James W. McGinity
37. Pharmaceutical Pelletization Technology, edited by
Isaac Ghebre-Sellassie
38. Good Laboratory Practice Regulations, edited by Allen F. Hirsch
39. Nasal Systemic Drug Delivery, Yie W. Chien, Kenneth S. E. Su,
and Shyi-Feu Chang
40. Modern Pharmaceutics: Second Edition, Revised and Expanded,
edited by Gilbert S. Banker and Christopher T. Rhodes
41. Specialized Drug Delivery Systems: Manufacturing and Production
Technology, edited by Praveen Tyle
42. Topical Drug Delivery Formulations, edited by David W. Osborne
and Anton H. Amann
43. Drug Stability: Principles and Practices, Jens T. Carstensen
44. Pharmaceutical Statistics: Practical and Clinical Applications,
Second Edition, Revised and Expanded, Sanford Bolton
45. Biodegradable Polymers as Drug Delivery Systems,
edited by Mark Chasin and Robert Langer
46. Preclinical Drug Disposition: A Laboratory Handbook, Francis L. S. Tse
and James J. Jaffe
 2005 by Taylor & Francis Group, LLC
47. HPLC in the Pharmaceutical Industry, edited by Godwin W. Fong
and Stanley K. Lam
48. Pharmaceutical Bioequivalence, edited by Peter G.Welling,
Francis L. S. Tse, and Shrikant V. Dinghe
49. Pharmaceutical Dissolution Testing, Umesh V. Banakar
50. Novel Drug Delivery Systems: Second Edition, Revised and Expanded,
Yie W. Chien
51. Managing the Clinical Drug Development Process, David M. Cocchetto
and Ronald V. Nardi
52. Good Manufacturing Practices for Pharmaceuticals: A Plan for Total
Quality Control, Third Edition, edited by Sidney H. Willig and James R.
53. Prodrugs: Topical and Ocular Drug Delivery, edited by Kenneth B. Sloan
54. Pharmaceutical Inhalation Aerosol Technology, edited by
Anthony J. Hickey
55. Radiopharmaceuticals: Chemistry and Pharmacology, edited by
Adrian D. Nunn
56. New Drug Approval Process: Second Edition, Revised and Expanded,
edited by Richard A. Guarino
57. Pharmaceutical Process Validation: Second Edition, Revised
and Expanded, edited by Ira R. Berry and Robert A. Nash
58. Ophthalmic Drug Delivery Systems, edited by Ashim K. Mitra
59. Pharmaceutical Skin Penetration Enhancement, edited by
Kenneth A.Walters and Jonathan Hadgraft
60. Colonic Drug Absorption and Metabolism, edited by Peter R. Bieck
61. Pharmaceutical Particulate Carriers: Therapeutic Applications,
edited by Alain Rolland
62. Drug Permeation Enhancement: Theory and Applications,
edited by Dean S. Hsieh
63. Glycopeptide Antibiotics, edited by Ramakrishnan Nagarajan
64. Achieving Sterility in Medical and Pharmaceutical Products, Nigel A. Halls
65. Multiparticulate Oral Drug Delivery, edited by Isaac Ghebre-Sellassie
66. Colloidal Drug Delivery Systems, edited by Jorg Kreuter
67. Pharmacokinetics: Regulatory  Industrial  Academic Perspectives,
Second Edition, edited by Peter G.Welling and Francis L. S. Tse
68. Drug Stability: Principles and Practices, Second Edition, Revised
and Expanded, Jens T. Carstensen
69. Good Laboratory Practice Regulations: Second Edition, Revised
and Expanded, edited by Sandy Weinberg
70. Physical Characterization of Pharmaceutical Solids, edited by
Harry G. Brittain
 2005 by Taylor & Francis Group, LLC
71. Pharmaceutical Powder Compaction Technology, edited by
Goran Alderborn and Christer Nystrom
72. Modern Pharmaceutics: Third Edition, Revised and Expanded,
edited by Gilbert S. Banker and Christopher T. Rhodes
73. Microencapsulation: Methods and Industrial Applications,
edited by Simon Benita
74. Oral Mucosal Drug Delivery, edited by Michael J. Rathbone
75. Clinical Research in Pharmaceutical Development, edited by Barry Bleidt
and Michael Montagne
76. The Drug Development Process: Increasing Efficiency and Cost
Effectiveness, edited by Peter G.Welling, Louis Lasagna,
and Umesh V. Banakar
77. Microparticulate Systems for the Delivery of Proteins and Vaccines,
edited by Smadar Cohen and Howard Bernstein
78. Good Manufacturing Practices for Pharmaceuticals: A Plan for Total
Quality Control, Fourth Edition, Revised and Expanded, Sidney H. Willig
and James R. Stoker
79. Aqueous Polymeric Coatings for Pharmaceutical Dosage Forms:
Second Edition, Revised and Expanded, edited by James W. McGinity
80. Pharmaceutical Statistics: Practical and Clinical Applications,
Third Edition, Sanford Bolton
81. Handbook of Pharmaceutical Granulation Technology, edited by
Dilip M. Parikh
82. Biotechnology of Antibiotics: Second Edition, Revised and Expanded,
edited by William R. Strohl
83. Mechanisms of Transdermal Drug Delivery, edited by Russell O. Potts
and Richard H. Guy
84. Pharmaceutical Enzymes, edited by Albert Lauwers and Simon Scharpe
85. Development of Biopharmaceutical Parenteral Dosage Forms,
edited by John A. Bontempo
86. Pharmaceutical Project Management, edited by Tony Kennedy
87. Drug Products for Clinical Trials: An International Guide to Formulation 
Production  Quality Control, edited by Donald C. Monkhouse
and Christopher T. Rhodes
88. Development and Formulation of Veterinary Dosage Forms: Second
Edition, Revised and Expanded, edited by Gregory E. Hardee
and J. Desmond Baggot
89. Receptor-Based Drug Design, edited by Paul Leff
90. Automation and Validation of Information in Pharmaceutical Processing,
edited by Joseph F. deSpautz
91. Dermal Absorption and Toxicity Assessment, edited by Michael S. Roberts
and Kenneth A.Walters
 2005 by Taylor & Francis Group, LLC
92. Pharmaceutical Experimental Design, Gareth A. Lewis, Didier Mathieu,
and Roger Phan-Tan-Luu
93. Preparing for FDA Pre-Approval Inspections, edited by Martin D. Hynes III
94. Pharmaceutical Excipients: Characterization by IR, Raman, and NMR
Spectroscopy, David E. Bugay and W. Paul Findlay
95. Polymorphism in Pharmaceutical Solids, edited by Harry G. Brittain
96. Freeze-Drying/Lyophilization of Pharmaceutical and Biological Products,
edited by Louis Rey and Joan C. May
97. Percutaneous Absorption: DrugsCosmeticsMechanismsMethodology,
Third Edition, Revised and Expanded, edited by Robert L. Bronaugh
and Howard I. Maibach
98. Bioadhesive Drug Delivery Systems: Fundamentals, Novel Approaches,
and Development, edited by Edith Mathiowitz, Donald E. Chickering III,
and Claus-Michael Lehr
99. Protein Formulation and Delivery, edited by Eugene J. McNally
100. New Drug Approval Process: Third Edition, The Global Challenge,
edited by Richard A. Guarino
101. Peptide and Protein Drug Analysis, edited by Ronald E. Reid
102. Transport Processes in Pharmaceutical Systems, edited by
Gordon L. Amidon, Ping I. Lee, and Elizabeth M. Topp
103. Excipient Toxicity and Safety, edited by Myra L.Weiner
and Lois A. Kotkoskie
104. The Clinical Audit in Pharmaceutical Development, edited by
Michael R. Hamrell
105. Pharmaceutical Emulsions and Suspensions, edited by
Francoise Nielloud and Gilberte Marti-Mestres
106. Oral Drug Absorption: Prediction and Assessment, edited by
Jennifer B. Dressman and Hans Lennernas
107. Drug Stability: Principles and Practices, Third Edition, Revised
and Expanded, edited by Jens T. Carstensen and C. T. Rhodes
108. Containment in the Pharmaceutical Industry, edited by James P.Wood
109. Good Manufacturing Practices for Pharmaceuticals: A Plan for Total
Quality Control from Manufacturer to Consumer, Fifth Edition, Revised
and Expanded, Sidney H. Willig
110. Advanced Pharmaceutical Solids, Jens T. Carstensen
111. Endotoxins: Pyrogens, LAL Testing, and Depyrogenation, Second Edition,
Revised and Expanded, Kevin L. Williams
112. Pharmaceutical Process Engineering, Anthony J. Hickey
and David Ganderton
113. Pharmacogenomics, edited by Werner Kalow, Urs A. Meyer,
and Rachel F. Tyndale
114. Handbook of Drug Screening, edited by Ramakrishna Seethala
and Prabhavathi B. Fernandes
 2005 by Taylor & Francis Group, LLC
115. Drug Targeting Technology: Physical  Chemical  Biological Methods,
edited by Hans Schreier
116. DrugDrug Interactions, edited by A. David Rodrigues
117. Handbook of Pharmaceutical Analysis, edited by Lena Ohannesian
and Anthony J. Streeter
118. Pharmaceutical Process Scale-Up, edited by Michael Levin
119. Dermatological and Transdermal Formulations, edited by
Kenneth A.Walters
120. Clinical Drug Trials and Tribulations: Second Edition, Revised and
Expanded, edited by Allen Cato, Lynda Sutton, and Allen Cato III
121. Modern Pharmaceutics: Fourth Edition, Revised and Expanded,
edited by Gilbert S. Banker and Christopher T. Rhodes
122. Surfactants and Polymers in Drug Delivery, Martin Malmsten
123. Transdermal Drug Delivery: Second Edition, Revised and Expanded,
edited by Richard H. Guy and Jonathan Hadgraft
124. Good Laboratory Practice Regulations: Second Edition,
Revised and Expanded, edited by Sandy Weinberg
125. Parenteral Quality Control: Sterility, Pyrogen, Particulate, and Package
Integrity Testing: Third Edition, Revised and Expanded, Michael J. Akers,
Daniel S. Larrimore, and Dana Morton Guazzo
126. Modified-Release Drug Delivery Technology, edited by
Michael J. Rathbone, Jonathan Hadgraft, and Michael S. Roberts
127. Simulation for Designing Clinical Trials: A Pharmacokinetic-
Pharmacodynamic Modeling Perspective, edited by Hui C. Kimko
and Stephen B. Duffull
128. Affinity Capillary Electrophoresis in Pharmaceutics and Biopharmaceutics,
edited by Reinhard H. H. Neubert and Hans-Hermann Ruttinger
129. Pharmaceutical Process Validation: An International Third Edition,
Revised and Expanded, edited by Robert A. Nash and Alfred H.Wachter
130. Ophthalmic Drug Delivery Systems: Second Edition, Revised
and Expanded, edited by Ashim K. Mitra
131. Pharmaceutical Gene Delivery Systems, edited by Alain Rolland
and Sean M. Sullivan
132. Biomarkers in Clinical Drug Development, edited by John C. Bloom
and Robert A. Dean
133. Pharmaceutical Extrusion Technology, edited by Isaac Ghebre-Sellassie
and Charles Martin
134. Pharmaceutical Inhalation Aerosol Technology: Second Edition,
Revised and Expanded, edited by Anthony J. Hickey
135. Pharmaceutical Statistics: Practical and Clinical Applications,
Fourth Edition, Sanford Bolton and Charles Bon
136. Compliance Handbook for Pharmaceuticals, Medical Devices,
and Biologics, edited by Carmen Medina
 2005 by Taylor & Francis Group, LLC
137. Freeze-Drying/Lyophilization of Pharmaceutical and Biological Products:
Second Edition, Revised and Expanded, edited by Louis Rey
and Joan C. May
138. Supercritical Fluid Technology for Drug Product Development,
edited by Peter York, Uday B. Kompella, and Boris Y. Shekunov
139. New Drug Approval Process: Fourth Edition, Accelerating Global
Registrations, edited by Richard A. Guarino
140. Microbial Contamination Control in Parenteral Manufacturing,
edited by Kevin L. Williams
141. New Drug Development: Regulatory Paradigms for Clinical Pharmacology
and Biopharmaceutics, edited by Chandrahas G. Sahajwalla
142. Microbial Contamination Control in the Pharmaceutical Industry,
edited by Luis Jimenez
143. Generic Drug Product Development: Solid Oral Dosage Forms ,
edited by Leon Shargel and Izzy Kanfer
144. Introduction to the Pharmaceutical Regulatory Process ,
edited by Ira R. Berry
145. Drug Delivery to the Oral Cavity: Molecules to Market, edited by
Tapash K. Ghosh and William R. Pfister
146. Good Design Practices for GMP Pharmaceutical Facilities ,
edited by Andrew Signore and Terry Jacobs
147. Drug Products for Clinical Trials, Second Edition , edited by Donald
Monkhouse, Charles Carney, and Jim Clark
148. Polymeric Drug Delivery Systems, edited by Glen S. Kwon
149. Injectable Dispersed Systems: Formulation, Processing, and Performance,
edited by Diane J. Burgess
150. Laboratory Auditing for Quality and Regulatory Compliance,
Donald Singer, Raluca-Ioana Stefan, and Jacobus van Staden
151. Active Pharmaceutical Ingredients: Development, Manufacturing, and
Regulation, edited by Stanley H. Nusim
152. Preclinical Drug Development, edited by Mark C. Rogge and David Taft
153. Pharmaceutical Stress Testing: Predicting Drug Degradation, edited by
Steven W. Baertschi
154. Handbook of Pharmaceutical Granulation Technology: Second Edition,
edited by Dilip M. Parikh
 2005 by Taylor & Francis Group, LLC
Handbook of Pharmaceutical
Granulation Technology
Second Edition
edited by
Dilip M. Parikh
Synthon Pharmaceuticals Inc.
Research Triangle Park, North Carolina, U.S.A.
Boca Raton London New York Singapore
 2005 by Taylor & Francis Group, LLC
Published in 2005 by
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
 2005 by Taylor & Francis Group, LLC
No claim to original U.S. Government works
Printed in the United States of America on acid-free paper
10 9 8 7 6 5 4 3 2 1
International Standard Book Number-10: 0-8247-2647-2 (Hardcover)
International Standard Book Number-13: 978-0-8247-2647-8 (Hardcover)
This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with
permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish
reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials
or for the consequences of their use.
No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic, mechanical, or
other means, now known or hereafter invented, including photocopying, microfilming, and recording, or in any information
storage or retrieval system, without written permission from the publishers.
For permission to photocopy or use material electronically from this work, please access www.copyright.com
01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For
organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for
identification and explanation without intent to infringe.

Contributors . . . . xv
1. Introduction . . . . . . 1
Dilip M. Parikh
References . . . . 6
2. Theory of Granulation: An Engineering Perspective . . . . . . . . . . . . . 7
Bryan J. Ennis
1. Introduction . . . . 7
2. Wetting . . . . 19
3. Granule Growth and Consolidation . . . . 34
4. Granule Strength and Breakage . . . . 55
5. Controlling Granulation Processes . . . . 60
References . . . . 77
3. Drug Substance and Excipient Characterization . . . . . . . . . . . . . . 79
L. W. Chan and P. W. S. Heng
1. Introduction . . . . 79
2. Particle Shape, Size, and Surface Area . . . . 80
3. Solubility . . . . 89
4. Crystal Properties and Polymorphism . . . . 95
5. Other Physical Properties . . . . 100
6. Commonly Used Excipients in Granulation . . . . 102
7. Compatibility of Drug and Excipient . . . . 103
8. Conclusion . . . . 105
References . . . . 106
4. Binders and Solvents . . . . . . . . . . . . . . . 109
Ehab Hamed, Derek Moe, Raj Khankari, and John Hontz
1. Introduction . . . . 109
2. Types of Binders . . . . 109
3. Factors Influencing Binder Efficiency . . . . 115
 2005 by Taylor & Francis Group, LLC
4. Processing Parameters for Commonly Used Binders . . . . 119
References . . . . 125
5. Spray Drying and Pharmaceutical Applications . . . . . . . . . . . . . . 129
Metin C? elik and Susan C. Wendel
1. Introduction . . . . 129
2. Spray Drying Process Stages . . . . 130
3. Process Layouts . . . . 138
4. Theory of Spray Drying Fundamentals . . . . 139
5. Spray Drying Applications . . . . 146
6. Conclusion . . . . 154
References . . . . 155
6. Roller Compaction Technology . . . . . . . . 159
Ronald W. Miller
1. Introduction . . . . 159
2. Powder Granulation and Compaction . . . . 159
3. Background . . . . 160
4. Benefits of Roller Compaction . . . . 161
5. Compaction Theory . . . . 162
6. Design Features of Roller Compactors . . . . 166
7. Roll Configuration . . . . 167
8. Feed Screw Design . . . . 170
9. Future Trends in Granulation Technology . . . . 172
10. New Findings . . . . 173
11. Deaeration Theory . . . . 178
12. Roller Compaction and Near-Infrared
Spectroscopy . . . . 182
13. Roller Compaction and PAT . . . . 185
References . . . . 188
7. High-Shear Granulation . . . . . . . . . . . . . 191
Rajeev Gokhale, Yichun Sun, and Atul J. Shukla
1. Introduction . . . . 191
2. High-Shear Granulators . . . . 192
3. High-Shear Granulation Process . . . . 195
4. Mechanism of High Shear Wet Granulation . . . . 200
5. Factors Affecting the Granulation Process
and Granule Properties . . . . 204
6. Granulation End-Point Determination . . . . 212
7. Formulation Development (Optimization) . . . . 217
8. Process Scale-Up . . . . 219
9. Conclusion . . . . 224
References . . . . 224
x Contents
 2005 by Taylor & Francis Group, LLC
8. Low-Shear Granulation . . . . . . . . . . . . . 229
Tom Chirkot and Cecil Propst
1. Introduction . . . . 229
2. Mechanical Agitator Granulators . . . . 231
3. Rotating Shape Granulators . . . . 235
4. Scale-Up . . . . 240
5. End-Point Determination and Control . . . . 243
6. Conclusions . . . . 244
References . . . . 244
9. Batch Fluid Bed Granulation . . . . . . . . . 247
Dilip M. Parikh and Martin Mogavero
1. Introduction . . . . 247
2. Fluidization Theory . . . . 248
3. System Description . . . . 251
4. Particle Agglomeration and Granule Growth . . . . 261
5. Fluid Bed Drying . . . . 265
6. Process and Variables in Granulation . . . . 268
7. Process Controls and Automation . . . . 276
8. Process Scale-Up . . . . 282
9. Safety in Fluid Bed . . . . 286
10. Material Handling Options . . . . 291
11. Fluid Bed Technology Progress . . . . 294
References . . . . 304
10. Single-Pot Processing . . . . . . . . . . . . . . 311
Harald Stahl and Griet Van Vaerenbergh
1. Introduction . . . . 311
2. Typical Single-Pot Process . . . . 313
3. Drying Methods for Single-Pot Processors . . . . 315
4. Other Processes and Applications . . . . 321
5. Scale-Up of Drying Processes . . . . 324
6. Cleaning . . . . 326
7. Product Stability . . . . 326
8. Regulatory Considerations . . . . 326
9. Validation of Single-Pot Processors . . . . 327
10. Control Systems and Data Acquisition Systems . . . . 328
11. Safety . . . . 328
12. Conclusion . . . . 330
References . . . . 330
11. Extrusion/Spheronization as a Granulation Technique . . . . . . . . . 333
Ketan A. Mehta, Gurvinder Singh Rekhi, and Dilip M. Parikh
1. Introduction . . . . 333
2. Applications . . . . 335
3. General Process Description . . . . 336
Contents xi
 2005 by Taylor & Francis Group, LLC
4. Equipment Description and Process Parameters . . . . 337
5. Formulation Variables . . . . 352
6. Compression of Pellets . . . . 357
7. Conclusions . . . . 359
References . . . . 360
12. Effervescent Granulation . . . . . . . . . . . . 365
Guia Bertuzzi
1. Introduction . . . . 365
2. The Effervescent Reaction . . . . 366
3. Formulation . . . . 366
4. Raw Materials . . . . 367
5. Manufacturing of Effervescent Forms . . . . 371
References . . . . 382
13. Melt Granulation and Pelletization . . . . . 385
T. W. Wong, W. S. Cheong, and P. W. S. Heng
1. Introduction . . . . 385
2. Mechanism of Melt Agglomeration . . . . 389
3. Factors Affecting Melt Agglomeration . . . . 393
4. Control of Melt Agglomeration . . . . 395
5. Conclusions . . . . 400
References . . . . 401
14. Rapid Release Granulation . . . . . . . . . . . 407
P. W. S. Heng, Anthony Yolande, and Lee Chin Chiat
1. Introduction . . . . 407
2. Formulation-Related Factors . . . . 409
3. Granulation-Related Factors . . . . 417
4. Solid Dispersion . . . . 419
5. Conclusions . . . . 423
References . . . . 424
15. Continuous Granulation Technologies . . . . 431
Rudolf Schroeder and Klaus-Ju? rgen Steffens
1. Introduction . . . . 431
2. Comparison of Different Modes of Processing . . . . 431
3. Fluid Bed Systems . . . . 433
4. Mechanical Wet Granulation Systems . . . . 437
5. Roller Extrusion System . . . . 441
References . . . . 456
16. Scale-Up Considerations in Granulation . . 459
Y. He, L. X. Liu, and J. D. Litster
1. Introduction . . . . 459
2. General Considerations in Process Scale-Up: Dimensional Analysis
and the Principle of Similarity . . . . 460
xii Contents
 2005 by Taylor & Francis Group, LLC
3. Analysis of Granulation Rate Processes
and Implications for Scale-Up . . . . 462
4. Scale-Down, Formulation Characterization, and Formulation Design
in Pharmaceutical Granulation . . . . 473
5. Scale-Up of Fluidized Bed Granulators . . . . 475
6. Scale-Up of High-Shear Mixer Granulators . . . . 480
7. Concluding Remarks . . . . 487
Nomenclature . . . . 487
References . . . . 488
17. Sizing of Granulation . . . . . . . . . . . . . . . 491
Gurvinder Singh Rekhi and Richard Sidwell
1. Introduction . . . . 491
2. Theory of Comminution or Size Reduction . . . . 492
3. Properties of Feed Materials Affecting
the Sizing Process . . . . 493
4. Criteria for Selection of a Mill . . . . 494
5. Classification of Mills . . . . 495
6. Wet Milling . . . . 499
7. Variables Affecting the Sizing Process . . . . 502
8. Scale-Up . . . . 507
9. Case Studies . . . . 509
10. List of Equipment Suppliers . . . . 510
References . . . . 511
18. Granulation Characterization . . . . . . . . . 513
Raj Birudaraj, Sanjay Goskonda, and Poonam G. Pande
1. Introduction . . . . 513
2. Physical and Chemical Characterization
of Granules . . . . 513
3. Conclusion . . . . 531
References . . . . 531
19. Bioavailability and Granule Properties . . . 535
Sunil S. Jambhekar
1. Introduction . . . . 535
2. Bioavailability Parameters . . . . 536
3. Conclusion . . . . 541
References . . . . 542
Recommended Reading . . . . 543
20. Process Analytical Technology . . . . . . . . 545
D. Christopher Watts and Ajaz S. Hussain
1. Introduction . . . . 545
2. Background . . . . 546
3. PAT and Process Understanding . . . . 548
Contents xiii
 2005 by Taylor & Francis Group, LLC
4. PAT Tools and Their Application . . . . 549
5. Conclusion . . . . 552
References . . . . 552
21. Granulation Process Modeling . . . . . . . . 555
I. T. Cameron and F. Y. Wang
1. Modeling of Granulation Systems . . . . 555
2. Key Factors in Granulation Modeling . . . . 560
3. Representing Granulation Processes Through
Population Balances . . . . 562
4. Solving and Using Population Balances . . . . 570
5. Application of Modeling Techniques . . . . 577
6. Conclusion . . . . 590
References . . . . 591
22. Regulatory Issues in Granulation . . . . . . . 595
Prasad Kanneganti
1. Introduction . . . . 595
2. Pharmaceutical Quality Management . . . . 595
3. Postapproval Change Considerations . . . . 597
4. Validation of Granulation Processes . . . . 609
5. Conclusion . . . . 613
References . . . . 613
xiv Contents

Dilip M. Parikh
Synthon Pharmaceuticals Inc., Research Triangle Park, North Carolina, U.S.A.
Perrys Chemical Engineers Handbook (1) defines the granulation process as any
process whereby small particles are gathered into larger, permanent masses in which
the original particles can still be identified. This definition is of course particularly
appropriate to a pharmaceutical granulation where the rapid breakdown of agglomerates
is important to maximize the available surface area and aid in solution of the
active drug. The granulation process of size enlargement used within the pharmaceutical
industry has its roots in ancient times. The practice of delivering medicinal powder
by hand rolling into a pill by using honey or sugar has been used for centuries.
It is still the practice to deliver the botanical and herbal extract in homeopathic and
ayurvedic branches of medicine, which are still practiced in India along with allopathic
medicine. The term granulated material is derived from the Latin word
granulatum, meaning grained. The granulated material can be obtained by direct
size enlargement of primary particles, or size reduction from dry compacted material.
In modern times, granulation technology has been widely used by a wide range
of industries, such as coal, mining, and agrochemical. These industries employ
agglomeration techniques to reduce dust, provide ease of handling, and enhance
the materials ultimate utility.
The development of pharmaceutical granulation was driven by the invention of
the tablet press by W. Brockedon in 1843. Subsequent improvements in the tablet
machinery were patented in the United States by J. A. McFerran (1874), T. J. Young
(1874), and J. Dunton (1876). The demands on the granulation properties were further
enhanced in the 1970s as high-speed tablet and capsule filling machines with automated
controls were introduced. The continuous refinements in the regulatory requirements
such as low-dose products requiring blend uniformity/content uniformity
necessitated knowledge and technology to produce the required granule characteristics.
The high-speed compression and capsule filling machines require a uniform flow
of material to the dies or filling stations that produce pharmaceutical dosage form.
Granulation is an example of particle design. The desired attributes of the
granule are controlled by a combination of the formulation and the process.
Granulation methods can be divided into two major types: wet methods which
utilize some form of liquid to bind the primary particles, and dry methods which do
not utilize any liquid (Fig. 1).

A new approach in the 1990s was to use supercritical fluid technology to produce
uniform particles to replace crystallization. Even though super critical fluids
were discovered over 100 years ago, and the commercial plant was built over 20 years
ago in the United States, it is only now that the technology is used for a number of
pharmaceutical applications (25), so as to produce aspirin, caffeine, ibuprofen,
acetaminophen, etc. One of the major areas on which the research and development
of supercritical fluids is focused is particle design. There are different concepts such
as rapid expansion of supercritical solution, gas antisolvent recrystallization,
and supercritical antisolvent to generate particles, microspheres, microcapsules,
liposomes, or other dispersed materials.
When the supercritical fluid and drug solution make contact, a volume expansion
occurs leading to a reduction in solvent capacity, increase in solute saturation,
and then supersaturation with associated nucleation and particle formation. A number
of advantages are claimed by using this platform technology (6), such as particle
formation from nanometers to tens of micrometers, low residual solvent levels in
products, preparation of polymorphic forms of drug, etc.
The classical granulation process using either wet or dry methods is employed
in the process industries. Pharmaceutical granulation process is used for tablet and
sometimes capsule dosage forms; however, in some applications the process is used
to produce spherical granules for the modified release indications or to prepare granules
as sprinkles to be used by pediatric patients. In some countries like Japan, having
granulated product in a sachet is acceptable where a large dose of the drug
product is not suitable for swallowing. The reasons for granulating a pharmaceutical
compound are listed as follows:
1. To increase the uniformity of drug distribution in the product
2. To densify the material
3. To enhance the flow rates and rate uniformity
4. To facilitate metering or volumetric dispensing
Figure 1 Various granulation techniques.
2 Parikh
 2005 by Taylor & Francis Group, LLC
5. To reduce dust
6. To improve the appearance of the product.
Five primary methods exist to form an agglomerated granule. They are formation
of solid bridges, sintering, chemical reaction, crystallization, or deposition of
colloidal particles. Binding can also be achieved through adhesion and cohesion
forces in highly viscous binders.
Successful processing for the agglomeration of primary particles depends on
proper control of the adhesional forces between particles, which encourage agglomerate
formation and growth and provide adequate mechanical strength in the
product. Furthermore, the rheology of the particulate system can be critical to
the rearrangement of particles necessary to permit densification of the agglomerate
and the development of an agglomerate structure appropriate for the end-use
requirements. If the particles are close enough then the surface forces such as van
der Waals forces (short-range) and electrostatic forces can interact to bond particles.
Decreasing particle size increases surfacemass ratio and favors the bonding. van der
Waals forces are sevenfold stronger than electrostatic forces and increase substantially
when the distance between them is reduced, which can be achieved by applying
pressure as in dry granulation method.
The cohesive forces that operate during the moist agglomerates are mainly due
to the liquid bridges that develop between the solid particles. Electrostatic forces
keep particles in contact long enough for another mechanism to govern the agglomeration
The processing of drug substance with the excipients can be achieved without
going through the granulation steps. By simply mixing in a blender, a directly compressible
formulation can be processed and compressed in tablets or filled in the hard
gelatin capsules. In the 1970s, microcrystalline cellulose as a directly compressible
vehicle was introduced. The compressible formulation containing microcrystalline
cellulose is suitable for a number of products. This has several obvious advantages,
such as lower equipment cost, faster process time, and efficient operation involving
only two process steps. Sometimes excipient costs may have to be compared against
the savings in the processing steps and equipment by using alternate methods.
There are, however, a number of products that require low dose of drug
substance, where the blend uniformity and the content uniformity in the drug
product are critical. Traditionally, the assessment of the blend uniformity is done
after the blending process is complete. This required considerable delays in obtaining
results, and the sampling techniques and product discharge from the blender
required consistency to obtain satisfactory results. However, with the current interest
in process analytical technology (PAT) on-line measurement of ingredients is possible.
The U.S. Food and Drug Administration (FDA) has recently released guidance
for industry detailing the current thinking on PAT (7).
Other than content uniformity of a low-dose drug substance there are a number
of reasons why direct compression may not be suitable for a wide array of products.
These include the required flow properties; the amount of drug substance in a dosage
form may require it to be densified to reduce the size of the drug product, obtain the
required hardness, friability, disintegration/dissolution, and other attributes.
Another approach which is becoming popular is to use traditional spray-drying
process to produce drum to hopper granulation by-passing the conventional granulation
process. This process may be suitable for large-volume products such as overthe-
counter tablets or capsules.
Introduction 3
 2005 by Taylor & Francis Group, LLC
Dry compaction technique like roller compaction is experiencing renewed
interest in the industry. There are a number of drug substances which are moisture
sensitive and cannot be directly compressed. The roller compaction provides suitable
alternative technology for processing these products.
Early stages of wet granulation technology development, employed low-shear
mixers or the mixers/blenders normally used for dry blending such as ribbon mixers.
There are a number of products currently manufactured using these low-shear granulators.
The process control and efficiency has increased over the years; however, the
industry has embraced high-shear granulators for wet granulation because of its
efficient and reproducible process and modern process control capabilities. The
high-shear mixers have also facilitated new technologies, such as one-pot processing,
that use the mixer to granulate and then dry using vacuum, gas stripping/vacuum, or
microwave assist in the same vessel.
Fluid-bed processors have been used in the pharmaceutical industry for the last
35 years, initially only as a dryer, and now as a multiprocessor to granulate, dry,
pelletize, and coat particles. The most preferred method of granulation is to use
the high-shear mixer to granulate and use the fluid bed as a dryer in an integrated
equipment setup. This provides the best of both technologies: efficient controllable
dense wet granules and a fast drying cycle using fluid-bed dryer. Here again, the
choice of this approach will be dependent on the product being processed, and its
desired properties at the end of the granulation process. Extrusion/spheronization
is used to produce granulation for the tableting or pelletizing, which involves mixing,
extruding, spheronizing, and drying unit operations. These pellets can be produced
as matrix pellets with the appropriate polymer or are coated in fluid-bed unit to produce
modified release dosage forms. Table 1 illustrates various options to granulate a
pharmaceutical compound.
Many researchers studied the influence of material properties of the granulating
powder and process conditions on the granulation process in a rather empirical
way. In the 1990s a fundamental approach to research was started on various topics
in the granulation research, looking into more detailed aspects of particle wetting,
Table 1 Frequently Used Granulation Techniques and Subsequent Processing
Process Drying technique
Wet granulation Low-shear mixer Tray or fluid-bed dryer
High-shear mixer Tray or fluid-bed dryer
High-shear mixer Vacuum/gas stripping/microwave
Fluid-bed granulator/dryer Fluid-bed granulator/dryer
Spray dryer Spray dryer
Extrusion/spheronization Tray or fluid-bed dryer
Continuous mixer granulator
Fluid bedcontinuous or batch
Continuous fluid-bed granulator Fluid bed (continuous)
Dry granulation Process Further processing
Direct compression Blend and process further
Slugging Mill slugged tablets/blend/
recompress/process further
Roller compactor Compacts milled/blend/process
4 Parikh
 2005 by Taylor & Francis Group, LLC
mechanism of granulation, material properties, and influence of mixing apparatus on
the product. The overall hypothesis suggested that the granulation can be predicted
from the raw material properties and the processing conditions of the granulation
process. One of the major difficulties encountered in granulation technology is the
incomplete description of the behavior of powders in general. The ongoing
fundamental research on mixing, segregation mechanisms of powder, surface chemistry,
and material science are necessary to develop the theoretical framework of
granulation technology. An excellent review of the wet granulation process was
presented by Iveson and coauthors (8). The authors have advanced the understanding
of the granulation process by stating that there are three fundamental sets of rate
processes which are important in determining wet granulation behavior. These are
wetting and nucleation, consolidation and growth; and breakage and attrition. Once
these processes are sufficiently understood, then it will be possible to predict the
effect of formulation properties, equipment type, and operating conditions of granulation
behavior, provided these can be adequately characterized according to the
Efficient and cost-effective manufacturing of pharmaceutical products is being
evaluated by the scientists, engineers, and operational managers of pharmaceutical
companies worldwide. In the United States, where 49% of the world pharmaceutical
market is, pharmaceutical companies are under tremendous pressure from the managed
care organizations, politicians, and consumers. The pharmaceutical industry,
worldwide in general and in the United States in particular, faces a unique paradox
drive future innovation through substantial R&D investments and return competitive
margin to shareholders, while providing access to pharmaceutical products at low or
no cost. The industry has reached a critical juncture in its 100-year history. The industry
is impacted simultaneously by growing competition, declining market performance,
increasing regulation, escalating pricing pressures, and rapidly evolving innovations
for improving peoples health and quality of life. A new report (9) into pharmaceutical
R&D has identified an emerging trend favoring outsourcing of discovery, research,
clinical trials, and manufacturing of dosage forms, providing relief fromthe consistent,
high-growth financial return faced by the majority of pharmaceutical companies. Outsourcing
allows these companies to pursue potential new revenue streams outside of
their core focus areas, and to benefit from improved productivity, emerging technologies,
inlicensing opportunities, and increased growth. Consumers and local governments
in the United States are pressuring the FDA authorities and politicians to
allow importation of the drugs from other countries like Mexico and Canada where
costs are generally lower than in the United States. Demands for price control also
extend to Europe; government-backed pharmaceutical payment plans in Germany
and Italy, for example, have cut back reimbursements. Other European countries have
controls on the drug prices. As a result of these pricing pressures and to enhance the
drugs in the pipeline, mergers and acquisitions have accelerated. Acquisitions remain
the preferred route to quickly enhance a product portfolio. This trend of merging
of equals or takeover of the significant technological companies will continue. This
has created emergence of small niche technology companies as well. Major pharmaceutical
companies are witnessing the end of traditional research and development.
Drug delivery companies are becoming potential targets for mergers or strategic alliances.
During all of the upheaval that the industry is going through, it is becoming
obvious that the cost of development and production, and cost of goods, must be
controlled. Recently released draft guidance by the U.S. FDA for quality systems
approach for the current good manufacturing practices may help to streamline the
Introduction 5
 2005 by Taylor & Francis Group, LLC
compliance programs in the industry (10). The efficiencies in the research, development,
and manufacturing, which were not necessarily sought after, are becoming
the first priority of the pharmaceutical companies however small they may be in comparison
to the final cost of the product to consumer. The manufacturing of solid
dosage product is no exception.
The significant advances that have taken place in the pharmaceutical granulation
technology are presented in this book to provide the readers with choices that
are available. There is no substitute for good science. The characterization of the
drug substance along with the knowledge of granulation theory, process modeling
capability, in-line or on-line (PAT) tools, process scale-up approaches, and a good
definition of the end product required will prepare the reader to explore the various
options presented in this book. Each drug substance poses a unique challenge that
must be taken into consideration at the process selection stage by the scientists.
The various techniques presented in this book will further help the scientists in their
understanding and selection of the granulation process most appropriate for the
drug substance. For production engineering, validation, and quality professionals
in the industry, this book is intended to provide the fundamental understanding of
the technique of granulation, and the rationale behind the selection of each particular
technique. This will further enhance the ability to design the production plant, carry
out the technology transfer, scale-up, troubleshoot, and maintain the pharmaceutical
granulation operation, in accordance with regulatory compliance.
1. Ennis BJ, Litster JD. Particle enlargement. Perry RH, Greens D, eds. Perrys Chemical
Engineers Handbook. 7th ed. New York: McGraw Hill, 1997:20-5620-89.
2. Charoenchaitrakool M, Dehghani F, Foster NR. Micronization by RESS to enhance the
dissolution rates of poorly water soluble pharmaceuticals. Proceedings of the 5th International
Symposium on Supercritical Fluids, Atlanta, GA, April 812, 2000.
3. Matson DW, Fulton JL, Petersen RC, Smith RD. Rapid expansion of supercritical fluid
solutions: solute formation of powders, thin films, and fibers. Ind Eng Chem Res 1987;
4. Subra P, Boissinot P, Benzaghou S. Precipitation of pure and mixed caffeine and anthracene
by rapid expansion of supercritical solutions. Proceedings of the 5th Meeting on
Supercritical Fluids, Tome I, Nice, France, March 2325, 1998.
5. Gilbert DJ, Palakodaty S, Sloan R, York V. Particle engineering for pharmaceutical
applicationsa process scale up. Proceedings of the 5th International Symposium on
Supercritical Fluids, Atlanta, GA, April 812, 2000.
6. York P, et al. Supercritical fluids ease drug delivery. Manuf Chemist 2000; 2629.
7. Food and Drug Administration. Guidance for Industry PATA Framework for
Innovative Pharmaceutical Development, Manufacturing and Quality Assurance.
FDA, September 2004.
8. Iveson SM, Litster JD, Hopgood K, Ennis B. Nucleation, growth, and breakage phenomenon
in agitated wet granulation process: a review. Powder Technol 2001; 117:339.
9. Cambridge Healthcare Advisors (CHA) Report. Report identifies increasing outsourcing
by pharmaInpharma.com, September 29, 2004.
10. Guidance for Industry. Quality Systems Approach to Pharmaceutical Current Good
Manufacturing Practice Regulations, Draft Guidance, September 2004.
6 Parikh
 2005 by Taylor & Francis Group, LLC
Theory of Granulation: An Engineering
Bryan J. Ennis
E&G Associates, Inc., Nashville, Tennessee, U.S.A.
1.1. Overview
Wet granulation is a subset of size enlargement (15), which involves any process
whereby small particles are agglomerated, compacted, or otherwise brought together
into larger, relatively permanent structures in which the original particles can still be
distinguished. Granulation technology and size-enlargement processes have been
used by a wide range of industries, from the pharmaceutical industry to fertilizer
or detergent production to the mineral processing industries. Size enlargement generally
encompasses a variety of unit operations or processing techniques dedicated to
particle agglomeration. These processes can be loosely broken down into agitation
and compression methods.
Although terminology is industry specific, agglomeration by agitation will be
to a process vessel and is agglomerated, either batchwise or continuously, to form
a granulated product. Agitative processes include fluid bed, pan (or disk), drum,
and mixer granulators. Such processes are also used as coating operations for
controlled release, taste masking, and cases where solid cores may act as a carrier
for a drug coating. The feed typically consists of a mixture of solid ingredients,
referred to as a formulation, which includes an active or key ingredient, binders,
diluents, flow aids, surfactants, wetting agents, lubricants, fillers, or end-use aids
(e.g., sintering aids, colors or dyes, taste modifiers). A closely related process of spray
drying is also included here, but is discussed more fully elsewhere [Refs. 6,7 and
carrier cores, or spray dried product consisting of agglomerated solidified drops.
An alternative approach to size enlargement is agglomeration by compression,
or compaction, where the mixture of particulate matter is fed to a compression device
which promotes agglomeration due to pressure, as depicted
continuous sheets of solid material or solid forms such as briquettes or tablets are
produced. Compaction processes range from confined compression devices such as
referred to as granulation. As depicted in Figure 1, a particulate feed is introduced
in Figure 2. Either
Chapter 5]. Product forms generally include agglomerated or layered granules, coated
tableting to continuous devices such as roll presses (Chapter 6), briqueting machines,
 2005 by Taylor & Francis Group, LLC
such as ram extrusion. Capsule filling operations would be considered as low-pressure
compaction processes.
At the level of a manufacturing plant, the size-enlargement process involves
several peripheral, unit operations such as milling, blending, drying, or cooling,
addition, more than one agglomeration step may be present as in the case of a pharmaceutical
process. In the case of pharmaceutical granulation, granulated material is
almost exclusively an intermediate product form, which is then followed by tableting.
In the context of granulation, therefore, it is important to understand compaction
processes to establish desirable granule properties for tableting performance.
Numerous benefits result from size-enlargement processes as summarized in
tion in the case of pharmaceutical processing is to create free-flowing, nonsegregating
blends of ingredients of controlled strength, which may be reproducibly metered in
Figure 2 The unit operation of compressive agglomeration or compaction. (From Refs. 15.)
Figure 1
8 Ennis
and classification, referred to generically as an agglomeration circuit (Fig. 3). In
and extrusion (Chapter 11). Some processes operate in a semicontinuous fashion
Table 1. A wide variety of size-enlargement methods are available; a classification of
these is given in Table 2 with key process attributes.Aprimary purpose of wet granula-
The unit operation of agitative agglomeration or granulation. (From Refs. 15.)
 2005 by Taylor & Francis Group, LLC
subsequent tableting or for vial or capsule filling operations. The wet granulation
process must generally achieve the desired granule properties within some prescribed
range. These attributes clearly depend on the application at hand. However, common
to most processes is a specific granule size distribution and granule voidage. Size
distribution affects flow and segregation properties, as well as compaction behavior.
Granule voidage controls strength, and impacts capsule and tablet dissolution
behavior, as well as compaction behavior and tablet hardness.
Control of granule size and voidage will be discussed in detail throughout this
chapter. The approach taken here relies heavily on attempting to understand interactions
at a particle level, and scaling this understanding to bulk effects. Developing
an understanding of these microlevel processes of agglomeration allows a rational
approach to the design, scale-up, and control of agglomeration processes. Although
Figure 3 A typical agglomeration circuit utilized in the processing of pharmaceuticals
Table 1 Objectives of Size Enlargement
Production of a useful structural form
Provision of a defined quantity for dispensing, with improved flow properties for metering
Improved product appearance
Reduced propensity to caking
Increased bulk density for storage
Creation of nonsegregating blends of powder ingredients
Control of solubility
Control of porosity, hardness, and surface-to-volume ratio and particle size
Source: From Refs. 15.
Theory of Granulation 9
involving both granulation and compression techniques. (From Refs. 15.)
 2005 by Taylor & Francis Group, LLC
Table 2 Size-Enlargement Methods and Application
Product size
(mm) Granule density
Scale of
operation Additional comments Typical applications
Tumbling granulators:
0.520 Moderate 0.5800
Very spherical granules Fertilizers, iron ore,
nonferrous ore,
agricultural chemicals
Mixer granulators
Handles very cohesive
materials well, both batch
and continuous
Continuous high shear
(e.g., Shugi mixer)
0.12 Low Up to 50
Chemicals, detergents, clays,
carbon black
Batch high shear
(e.g., paddle mixer)
0.12 High Up to 500 kg,
Pharmaceuticals, ceramics
Fluidized granulators:
Fluidized beds
Spouted beds
Wurster coaters
0.12 Low (agglomerated),
moderate (layered)
100900 kg, batch;
50 ton/hr,
Flexible, relatively easy to
scale, difficult for cohesive
powders, good for coating
Continuous: fertilizers,
inorganic salts, detergents;
batch: pharmaceuticals,
agricultural chemicals,
nuclear wastes
10 Ennis
 2005 by Taylor & Francis Group, LLC
Centrifugal granulators 0.33 Moderate to high Up to 200 kg,
Powder layering and coating
Pharmaceuticals, agricultural
Spray methods
Spray drying 0.050.5 Low Morphology of spray dried
powders can vary widely
Instant foods, dyes,
detergents, ceramics,
Prilling 0.72 Moderate Urea, ammonium nitrate0
Pressure compaction
Very narrow size
distributions, very sensitive
to powder flow and
mechanical properties
Pharmaceuticals, catalysts,
inorganic chemicals,
organic chemicals, plastic
preforms, metal parts,
ceramics, clays, minerals,
animal feeds
Extrusion >0.5 High to very high Up to 5 ton/hr
Roll press >1 Up to 50 ton/hr
Tablet press 10 Up to 1 ton/hr
Molding press
Pellet mill
Theory of Granulation 11
Source: From Refs. 15.
 2005 by Taylor & Francis Group, LLC
the approach is difficult, qualitative trends are uncovered along the way, which aid in
formulation development and process optimization, and which emphasize powder
characterization as an integral part of product development and process design
1.2. Granulation Mechanisms
Four key mechanisms or rate processes contribute to granulation, as originally outlined
by Ennis (3,4), and later developed further by Ennis and Litster (1,5). These
include wetting and nucleation, coalescence or growth, consolidation, and attrition
or breakage (Fig. 4). Initial wetting of the feed powder and existing granules by
the binding fluid is strongly influenced by spray rate or fluid distribution as well
as feed formulation properties, in comparison with mechanical mixing. Wetting promotes
nucleation of fine powders, or coating in the case of feed particle size in excess
of drop size. Often wetting agents such as surfactants are carefully chosen to enhance
poorly wetting feeds. In the coalescence or growth stage, partially wetted primary
particles and larger nuclei coalesce to form granules composed of several particles.
The term nucleation is typically applied to the initial coalescence of primary particles
in the immediate vicinity of the larger wetting drop whereas the more general term of
Figure 4 The rate processes of agitative agglomeration, or granulation, which include
powder wetting, granule growth, granule consolidation, and granule breakage. These processes
combine to control granule size and porosity, and they may be influenced by formulation
or process design changes.
12 Ennis
 2005 by Taylor & Francis Group, LLC
coalescence refers to the successful collision of two granules to form a new, larger
granule. In addition, the term of layering is applied to the coalescence of granules
with primary feed powder. Nucleation is promoted from some initial distribution
of moisture, such as a drop, or from the homogenization of a fluid feed to the
bed, as with high-shear mixing. The nucleation process is strongly linked with the
wetting stage. As granules grow, they are consolidated by compaction forces due
to bed agitation. This consolidation stage controls internal granule voidage or granule
porosity, and therefore end-use properties such as granule strength, hardness, or
dissolution. Formed granules may be particularly susceptible to attrition if they are
inherently weak or if flaws develop during drying.
These mechanisms or rate processes can occur simultaneously in all granulation
units, ranging from spray drying to fluidized beds to high-shear mixers. However,
certain mechanisms may dominate in a particular manufacturing process. For example,
fluidized bed granulators are strongly influenced by the wetting process, whereas
mechanical redispersion of binding fluid by impellers and particularly high-intensity
choppers diminishes the wetting contributions to granule size in high-shear mixing.
On the other hand, granule consolidation is far more pronounced in high-shear
mixing than in fluidized-bed granulation. These simultaneous rate processes taken
as a wholeand sometimes competing against one anotherdetermine the final granule
size distribution and granule structure and voidage resulting from a process, and
therefore the final end-use or product quality attributes of the granulated product.
1.3. Compaction Mechanisms
Compaction is a forming process controlled by mechanical properties of the feed in
relationship to applied stresses and strains. Microlevel processes are controlled by
particle properties such as friction, hardness, size, shape, surface energy, and elastic
Key steps in any compaction process include (1) powder filling, (2) stress application
and removal, and (3) compact ejection. Powder filling and compact weight
variability is strongly influenced by bulk density and powder flowability (1,2), as well
as any contributing segregation tendencies of the feed. The steps of stress application
Powders do not transmit stress uniformly. Wall friction impedes the applied load,
causing a drop in stress as one moves away from the point of the applied load
(e.g., a punch face in tableting or roll surface in roll pressing.) Therefore, the applied
load and resulting density are not uniform throughout the compact, and powder frictional
properties control the stress transmission and distribution in the compact (8).
The general area of study relating compaction and stress transmission is referred to
as powder mechanics (1,2,8,9). For a local level of applied stress, particles deform at
their point contacts, including plastic deformation for forces in excess of the particle
surface hardness. This allows intimate contact at surface point contacts, allowing
cohesion/adhesion to develop between particles, and therefore interfacial bonding,
which is a function of their interfacial surface energy. During the short time scale
of the applied load, any entrapped air must escape, which is a function of feed
permeability, and a portion of the elastic strain energy is converted into permanent
plastic deformation. Upon stress removal, the compact expands due to the remaining
elastic recovery of the matrix, which is a function of elastic modulus, as well as any
expansion of the remaining entrapped air. This can result in loss of particle bonding,
and flaw development, and this is exacerbated for cases of wide distributions in
Theory of Granulation 13
and removal consist of several competing mechanisms, as depicted in Figure 5.
 2005 by Taylor & Francis Group, LLC
compact stress due to poor stress transmission. The final step of stress removal
involves compact ejection, where any remaining radial elastic stresses are removed.
If recovery is substantial, it can lead to capping or delamination of the compact.
These microlevel processes of compaction control the final flaw and density
distribution throughout the compact, whether it is a roll pressed, extruded, or tableted
product, and as such, control compact strength, hardness, and dissolution
behavior. Compaction processes will not be discussed further here, with the remainder
of the chapter focusing on wet granulation, agitative processes. For further
1.4. Formulation vs. Process Design
The end-use properties of granulated material are controlled by granule size and
internal granule voidage or porosity. Internal granule voidage egranule and bed
voidage ebed, or voidage between granules, are related by
rbulk ? rgranule?1  ebed? ? rs?1  ebed??1  egranule? ?1:1?
where rbulk, rgranule, and rs are bulk, granule, and skeletal primary particle density,
respectively. Here, granule voidage and granule porosity will be used interchangeably.
Granule structure may also influence properties. To achieve the desired product
quality as defined by metrics of end-use properties, granule size and voidage may be
manipulated by changes in either process operating variables or product material
further by Ennis and Litster (1,2,5). The first approach is the realm of traditional
process engineering, whereas the second is product engineering. Both approaches
are critical and must be integrated to achieve the desired end point in product
quality. Operating variables are defined by the chosen granulation technique and
Figure 5 The microlevel processes of compressive agglomeration, or compaction. These
processes combined control compact strength, hardness, and porosity.
14 Ennis
variables (Figs. 4 and 5) as initially outlined by Ennis (3,4), and later developed
discussion regarding compaction, see Chapter 6 and Refs. 1, 2, and 10.
 2005 by Taylor & Francis Group, LLC
peripheral processing equipment, as illustrated for a fluidized-bed and mixer granulator
in Figure 6. In addition, the choice of agglomeration technique dictates the
mixing pattern of the vessel. Material variables include parameters such as binder
viscosity, surface tension, feed particle size distribution, powder friction, and the
adhesive properties of the solidified binder. Material variables are specified by the
choice of ingredients, or product formulation. Both operating and material variables
together define the kinetic mechanisms and rate constants of wetting, growth,
consolidation, and attrition. Overcoming a given size-enlargement problem often
requires changes in both processing conditions and product formulation.
The importance of granule voidage to final product quality is illustrated in
granule attrition to increase, and dissolution rate to increase with an increase in
granule voidage. Bulk density is clearly a function of both granule size distribution,
which controls bed voidage or porosity between granules, and the voidage within the
granule itself. The data of Figure 7 are normalized with respect to its zero intercept,
or its effective bulk density at zero granule voidage. The granule attrition results of
fines passing a fine mesh size following attrition in a tumbling apparatus. Granules
weaken with increased voidage. The dissolution results of Figure 9 measure the
length required for granule dissolution in a long tube, or disintegration length (also
based on the CIPAC test method). Increased granule voidage results in increased
dissolution rate and shorter disintegration length. All industries have their own
specific quality and in-process evaluation tests. However, what they have in common
are the important contributing effects of granule size and granule voidage.
An example of the importance of distinguishing the effects of process and
formulation changes can be illustrated with the help of Figures 8 and 9. Let us
assume that the particular formulation and current process conditions produce a
granulated material with a given attrition resistance and dissolution behavior
(indicted as current product). If one desires instead to reach a given target,
either the formulation or the process variable may be changed. Changes to the
process, or operating variables, generally readily alter granule voidage. Examples
to decrease voidage might include increased bed height, increased processing time,
Figure 6
Theory of Granulation 15
Figures 79 for a variety of formulations. Here, bulk density is observed to decrease,
Figure 8 are based on a CIPAC test method, which is effectively the percentage of
Typical operating variable for pharmaceutical granulation processes. (From Ref. 4.)
 2005 by Taylor & Francis Group, LLC
or increased peak bed moisture. However, only a range of such changes in voidage,
are due instead to changes in formulation properties. Therefore, it may not be
Figure 7 Impact of granule density on bulk density: Normalized bulk density as a function
Figure 8 Impact of granule density on strength and attrition: Illustration of process changes
vs. formulation changes. Kc is the fracture toughness. (From Ref. 4.)
16 Ennis
and therefore attrition resistance and dissolution, is possible and is illustrated by
moving along the given formulation curve. The various curves in Figures 8 and 9
of granule voidage. (From Ref. 4.)
 2005 by Taylor & Francis Group, LLC
possible to reach a target change in dissolution without changes in formulation or
material variables. Examples of key material variables affecting voidage would
include feed primary particle size, inherent formulation bond strength, and binder
solution viscosity, as discussed in detail in the following sections. Understanding this
crucial interaction between operating and material variables is critical for successful
formulation, and requires substantial collaboration between processing and formulation
groups, and a clear knowledge of the effect of scale-up on this interaction.
1.5. Key Historical Investigations
A range of historical investigations has been undertaken involving the impact of
cal variables have included the effects of bed hydrodynamics and agitation intensity,
pan angle and speed, fluid-bed excess gas velocity, mixer impeller and chopper
speeds, drum rotation speed, spray method, drop size, nozzle location, and binder
and solvent feed rates. While such studies are important, their general application
and utility to studies beyond the cited formulations and process conditions can be
severely limited. Often the state of mixing, moisture distribution and rates, and material
properties such as formulation size distribution, powder frictional properties,
and solution viscosity are insufficiently defined. As such, these results should be used
judiciously and with care. Often even the directions of the impact of operating variables
on granule properties are altered by formulation changes.
Two key pieces of historical investigation require mention, as the approach
developed here stems heavily from this work. The first involves growth and breakage
mechanisms, which control the evolution of the granule size distribution (14), as
Figure 9 Impact of granule density on disintegration: Gc is the strain energy release rate.
Theory of Granulation 17
operating variables on granulation behavior. As a review, see Refs. 35,1113. Typi-
(From Ref. 4.)
 2005 by Taylor & Francis Group, LLC
illustrated in Fig. 10. These include the nucleation of fine powder to form initial
primary granules, the coalescence of existing granules, and the layering of raw material
onto previously formed nuclei or granules. Granules may simultaneously be
compacted by consolidation and reduced in size by breakage. There are strong interactions
between these rate processes. In addition, these mechanisms in various forms
have been incorporated into population-balance modeling to predict granule size in
the work of Kapur and Sastry (1419) (Chapter 23 for details). Given the progress
made in connecting rate constants to formulation properties, the utility of population-
balance modeling has increased substantially.
The second important area of contribution involves the work of Rumpf (20,21)
and Ausburger and Vuppala (22), which studied the impact of interparticle force H
on granule static tensile strength, or
sT ?
1  e
e  H
a2 ? A
1  e
e  g cos y
A ? 9=4 for pendular state
A ? 6 for capillary state 
Forces of a variety of forms were studied, including viscous, semisolid, solid,
electrostatic, and van der Waals forces. Of particular importance was the contribution
of pendular bridge force arising from surface tension to granule tensile strength.
Capillary pressure deficiency due to the curvature of the pendular bridge in addition
(here, interparticle velocity U?0). This force summed over the granule area results
in a granule static strength, which is a function of pore saturation S as experimentally
plotted. The states of pore filling have been defined as pendular (single bridges),
Figure 10
18 Ennis
to a contact line force results in an interparticle force, as highlighted in Figure 11
Growth and breakage mechanisms in granulation processes. (From Ref. 14.)
 2005 by Taylor & Francis Group, LLC
funicular (partially complete filling and single bridges), capillary (nearly complete
filling S80100%) followed by drop formation, and loss of static strength. This
approach will be extended in subsequent sections to include viscous forces and
dynamic strength behavior (U 6? 0).
The approach taken, in this chapter, follows this same vein of research as originally
established by Rumpf and Kapur, namely, relating granule and particle level
interactions to bulk behavior through the development of the rate processes of
wetting and nucleation, granule growth and consolidation, and granule breakage
and attrition. Each of these will now be dealt with in turn in the following sections.
2.1. Overview
The initial distribution of binding fluid can have a pronounced influence on the size
distribution of seed granules, or nuclei, which are formed from fine powder. Both the
final extent of and the rate at which the fluid wets the particulate phase are
important. Poor wetting results in drop coalescence, and in fewer, larger nuclei with
ungranulated powder and overwetted masses, leading to broad nuclei distributions.
Granulation can retain a memory, with nuclei size distribution impacting final granule
size distribution. Therefore, initial wetting can be critical to uniform nuclei formation
and often a narrow, uniform product. Wide nuclei distributions can lead to a
wide granule size distribution. When the size of a particulate feed material is larger
than drop size, wetting dynamics controls the distribution of coating material which
has a strong influence on the later stages of growth. Wetting phenomena also influence
redistribution of individual ingredients within a granule, drying processes, and
redispersion of granules in a fluid phase. Other granule properties such as voidage,
Figure 11
Theory of Granulation 19
nomenclature.) (From Refs. 2022.)
Static yield strength of wet agglomerates vs. pore saturation. (See Section 3.5 for
 2005 by Taylor & Francis Group, LLC
strength, and attrition resistance may be influenced as well. Preferential wetting of
certain formulation ingredients can cause component segregation with granule size.
An extensive review of wetting research may be found in Litster and Ennis (5), Parfitt
2.2. Mechanics of the Wetting Rate Process
As outlined previously, wetting is the first stage in wet granulation involving liquid
binder distribution to the feed powder. There are two extremes: (1) liquid drop size is
large compared to unit or primary particle size of the feed and (2) particle size is
large compared to the drop size.
the wetting process consists of several important steps. First, droplets are formed
related to spray distribution, or spray flux defined as the wetting area of the bed
per unit time. Important operating variables include nozzle position, spray area,
spray rate, and drop size. Second, droplets impact and coalesce on the powder
bed surface if mixing or wet-in time is slow. Third, droplets spread and penetrate into
the moving powder bed to form loose nuclei, again coalescing if wet-in is slow. In the
case of high-shear processes, shear forces break down overwet clumps, also producing
For the second case of small drop size compared to the primary particle size,
collisions between the drop and the particle followed by spreading of the liquid over
the particle surface. If the particle is porous, then liquid will also be sucked into the
pores by capillary action. The wetting dynamics control the distribution of coating
material, which has a strong influence on the later stages of growth as well as coating
2.3. Methods of Measurement
Methods of characterizing the rate process of wetting include four approaches as
powder. This approach involves the measurement of a contact angle of a drop on
a powder compact, given by the YoungDupre? equation, or
gsv  gsl ? glv cos y ?2:1?
where gsv, gsl, glv are the solidvapor, solidliquid, and liquidvapor interfacial energies,
respectively, and y is measured through the liquid. In the limit of gsvgslglv,
the contact angle equals 0 and the fluid spreads on the solid. The extent of wetting is
controlled by the group glv cosy, which is referred to as adhesion tension. Sessile drop
studies of contact angle can be performed on powder compacts in the same way as
on planar surfaces. Methods involve the (1) direct measurement of the contact angle
from the tangent to the airbinder interface, (2) solution of the LaplaceYoung
equation involving the contact angle as a boundary condition, or (3) indirect calculations
of the contact angle from measurements of, e.g., drop height. The compact can
either be saturated with the fluid for static measurements, or dynamic measurements
granulation processes, the dynamics of wetting are often crucial, requiring that
powders be compared on the basis of a short time scale, dynamic contact angle.
20 Ennis
As depicted in Figure 12 for the first case of fine feeds compared to drop size,
the liquid will coat the particles as depicted in Figure 13. Coating is produced by
may be made through a computer imaging goniometer, as depicted in Figure 14. For
(23), and Hapgood (24), with a summary of references given in Table 3.
illustrated in Table 4. The first considers the ability of a drop to spread across the
 2005 by Taylor & Francis Group, LLC
Table 3 Summary of Wetting Research Contributions
Workers (Year) Materials Equipment Investigated
Schaefer and Worts
Lactose and maize
starch powder with
gelatine, PVP, and
MCC binders
Fluid bed Drop size
Schaafsma et al.
Lactose powder with
water and 38%
PVP solution
Packed bed
petri dish
Drop size
Waldie (1991) Lactose and ballotini
powders with water
and 5% PVP
Fluid bed Drop size
Watano et al. (1994) Lactose and
cornstarch with
Hybrid fluid
Drop size, binder
Holm et al. (1983,
Lactose and calcium
hydrogen phosphate
with water and 10
15% PVP solution
Drop size, binder
distribution, and
shear forces
Aulton and Banks
Lactose and salicylic
acid powder with
Fluid bed Contact angle
Gluba et al. (1990) Talcum, chalk, and
kaolin powder
combinations with
Drum Contact angle
Jaiyeoba and Spring
Lactose, boric acid,
suphanilamide, and
salicylic acid
powders with water.
Fluid bed Contact angle
Krycer and Pope
Paracetamol powder
with HPMC, PVP,
and water
Fluid bed Spreading coefficients
Zajic and Buckton
powder with
HPMC, PVP, and
Spreading coefficients
Rankell et al. (1964) Aluminium hydroxide
and sucrose powder
with water
Fluid bed Binder addition rate,
powder flux, spray
Heim and Antkowiak
Talc powder and
Drum Binder addition rate
Crooks and Schade
5% Phenylbutazone in
lactose powder mix
with 10% solution
Fluid bed Binder addition rate
Kristensen and
Schaefer (1994)
Lactose powder with
PEG300 and
High shear Viscosity, binder
dispersion, shear
Theory of Granulation 21
 2005 by Taylor & Francis Group, LLC
Important factors are the physical nature of the powder surface (particle size, pore
size, porosity, environment, roughness, pretreatment). The dynamic wetting process
is therefore influenced by the rates of ingredient dissolution and surfactant adsorption
and desorption kinetics (25).
The second approach to characterize wetting considers the ability of the fluid
of the extent and rate of fluid rise by capillary suction into a column of powder, better
known as the Washburn test. Considering the powder to consist of capillaries of
radius R, the equilibrium height of rise he is determined by equating capillary and
gravimetric pressures, or
he ?
2glv cos y
DrgR ?2:2?
Figure 12 Stages of wetting for fine powder compared to drop size. (From Ref. 5.)
Table 3 Summary of Wetting Research Contributions (Continued )
Workers (Year) Materials Equipment Investigated
Schaefer et al. (1992) Lactose powder with
PEG300 and
High shear Viscosity, binder
dispersion, shear
forces, powder flux
Cartensen et al. (1976) Lactose, sucrose, and
starch powder
mixed with water
High shear Binder dispersion
Kokubo et al. (1996) Lactosecornstarch
Powder flux
Davies and Gloor
Lactose, comstarch,
magnesium stearate,
gelatin, and benzene
powder with gelatin
Fluid bed Spray zone
22 Ennis
to penetrate a powder bed, as illustrated in Figure 15. It involves the measurement
Source: From Ref. 5.
 2005 by Taylor & Francis Group, LLC
where is Dr the fluid density with respect to air, g is gravity, and glv cos y is the adhesion
tension as before.
In addition to the equilibrium height of rise, the dynamics of penetration are
particularly important. Ignoring gravity and equating viscous losses with the capillary
pressure, the rate (dh/dt) and dynamic height of rise h are given by
dt ?
Rglv cos y
or h ? ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Rglv cos y
2m  t s ?2:3?
where t is time and m is binder fluid viscosity (23). The grouping of terms in brackets
involve the material properties which control the dynamics of fluid penetration,
namely, average pore radius, or tortuosity R (related to particle size and void distribution
of the powder), adhesion tension, and binder viscosity. The rate of capillary
fluid rise, or the rate of binding fluid penetration in wet granulation, increases with
increasing pore radius (generally coarser powders with larger surfacevolume average
particle size), increasing adhesion tension (increased surface tension and
decreased contact angle), and decreased binder viscosity.
The contact angle of a binderparticle system is not itself a primary thermodynamic
quantity, but rather a reflection of individual interfacial energies (Eq. 2.1),
which are a function of the molecular interactions of each phase with respect to
one another. An interfacial energy may be broken down into its dispersion and polar
components. These components reflect the chemical character of the interface, with
the polar component due to hydrogen bonding and other polar interactions and the
dispersion component due to van der Waals interactions. These components may be
determined by the wetting tests described here, where a variety of solvents are chosen
as the wetting fluids to probe specific molecular interactions (26). Interfacial energy
is strongly influenced by trace impurities which arise in crystallization of the active
ingredient, or other forms of processing such as grinding. It may be modified by judicious
selection of surfactants (27). Charges may also exist at interfaces, as characterized
by electrokinetic studies (28). The total solidfluid interfacial energy (i.e., both
dispersion and polar components) is also referred to as the critical solid surface
energy of the particulate phase. It is equal to the surface tension of a fluid which just
wets the solid with zero contact angle. This property of the particle feed may be
determined by a third approach to characterize wetting, involving the penetration
of particles into a series of fluids of varying surface tension (27,29), or by the variation
of sediment height (30).
The last approach to characterizing wetting involves chemical probing of properties
which control surface energy. As an example, inverse gas chromatography
(IGC) uses the same principles and equipment as standard gas chromatography.
In IGC, however, the mobile phase is composed of probe gas molecules that move
Figure 13
Theory of Granulation 23
 2005 by Taylor & Francis Group, LLC
Stages of wetting for coarse powder compared to drop size. (From Ref. 5.)
Table 4 Methods of Characterizing Wetting Dynamics of Particulate Systems
Mechanism of wetting Characterization method
Spreading drops on powder surface Contact angle goniometer:
Contact angle
Drop height or volume
Spreading velocity
Kossen, Heertjes. Chem Eng Sci 1965;
Pan et al. Dynamic Properties of Interfaces
and Association Structure. New York:
American Oil Chem Soc Press, 1995
Penetration of drops into powder bed Washburn test:
Rate of penetration by height or volume
Bartell cell:
Capillary pressure difference
Parfitt, ed. Dispersion of Powders in
Liquids. Washburn: Elsevier Applied
Science Publishers Ltd., 1986, and Phys Rev
1921; 17:273
Bartell, Osterhof. Ind Eng Chem 1927;
Penetration of particles into fluid Flotation tests:
Penetration time
Sediment height
Critical solid surface energy distribution
Ayala R. Ph.D.Thesis, Chem. Eng.,
Carnegie Mellon University, 1985
Fuerstaneau et al. Coll Surf 1991; 60:127
Vargha-Butler. In: Botsaris, Glazman, eds.
Interfacial Phenomena in Coal Technology,
Chap. 2, 1994
Chemical probing of powder Inverse gas chromatography:
Preferential adsorption with probe gases
Zeta potential and charge
Surfactant adsorption:
Preferential adsorption with probe
Lloyd et al., eds. ACS Symposium Series.
Vol. 391. Washington, DC: ACS, 1989
Aveyard, Haydon. An Introduction to the
Principles of Surface Chemistry. Cambridge
University Press, 1973
Shaw. Introduction to Colloid & Surface
Chemistry. London: Butterworths & Co.
Ltd., 1983
24 Ennis
Source: From Refs. 1,2,5.
 2005 by Taylor & Francis Group, LLC
through a column packed with the powder of interest, which is the stationary phase.
As the probe molecules travel through the packed column, they adsorb to and desorb
from the powder. The rate and degree of this interaction is determined by the surface
chemistry of the powder and the probe molecules. Since the surface chemistry of the
probe molecules is known, this allows calculation of the surface chemistry (surface
Figure 15
Figure 14
Theory of Granulation 25
Characterizing wetting by dynamic contact angle goniometry. (From Refs. 4,25.)
Characterizing wetting by Washburn test and capillary rise. (From Refs. 4,23.)
 2005 by Taylor & Francis Group, LLC
energies) of the powder with the help of a series of plots of both alkane and various
polar probes. A distinct advantage of IGC is the reproducible measurements of physical
chemical surface properties which control adhesion tension.
The strength of the solid/liquid interactions determines the average retention
time of a probe. Retention time data for each probe is converted into net retention
volumes, VN. The free energy of desorption is given by
DG ? RT ln VN ? c ? 2Na ffiffiffiffiffiffiffiffiffiffiffi gDS
q ? ffiffiffiffiffiffiffiffiffi gPS
L q  ? c ?2:4?
where R is the universal gas constant, T is the column temperature, c is a system constant,
N is Avagadros number, and a is the surface area of a probe molecule. As illustrated
in Figure 16, a plot of RT ln VN vs. a ffiffiffiffiffi gDL
p should give a straight line for a
series of alkanes, the slope of which allows determination of the solids dispersive
surface energy gDS
. Plotting RT ln VN vs. a ffiffiffiffiffi gDL
p for the polar probes will give a point
that is generally somewhere above the alkane reference line. The polar solid energy
is then found from a plot of these deviations.
2.4. Granulation Examples of the Impact of Wetting
Wetting dynamics has a pronounced influence on initial nuclei distribution formed
from fine powder. As an example, the influence of contact angle on the average granangle
was varied by changing the percentages of hydrophilic lactose and hydrophobic
salicylic acid (31). Note that granule size in this study is actually nuclei size, since
little growth has taken place in the process. Nuclei size is seen to improve with contact
angle. In addition, the x-coordinate would more appropriately be replaced with
adhesion tension. Aulton et al. (32) also demonstrated the influence of surfactant
concentration on shifting nuclei size due to changes in adhesion tension.
drop profiles are imaged as a drop wets in to a formulation tablet. Note that the time
Figure 16
26 Ennis
ule size formed in fluid-bed granulation is illustrated in Figure 17. Water contact
Figure 18(A) illustrates an example of dynamic wetting, where a time series of
IGC retention times for glass powder. (From Refs. 4,5.)
 2005 by Taylor & Francis Group, LLC
scale of wetting in this case is 2 sec, with nearly complete wet-in occurring in 1 sec.
This particular formulation was granulated on a continuous pan system in excess
a second lotreferred to as problem technicalexperiences significantly degraded
granule strength and requires production rates to be substantially reduced. This is
associated with nearly twice the initial contact angle (120) and slower spreading
velocity when compared with the good technical. Poor wetting in practice can translate
into reduced production rates to compensate for increased time for drops to
work into the powder bed surface. Weaker granules are also often observed, since
poor wet phase interfacial behavior translates in part to poor solid bond strength
and high granule voidage. Note that differences in the lots are only observed over
the first 0.250.5 sec, illustrating the importance of comparing dynamic behavior
of formulations, after which time surfactant adsorption/desorption reduces the
contact angle.
As an example of Washburn approaches as
of fluid penetration rate and the extent of penetration on granule size distribution for
drum granulation was shown by Gluba et al. (Powder Hand. & Proc 1990; 2:323).
Increasing penetration rate, as reflected by Eq. 2.3 increased granule size and
decreased asymmetry of the granule size distribution.
2.5. Regimes of Nucleation and Wetting
ders consists of several important steps. First, droplets are formed related to spray
distribution, or spray flux defined as the wetted area of the bed per unit time.
Figure 17 The influence of contact angle on nuclei size formed in fluid-bed granulation of
lactose/salicylic acid mixtures. Powder contact angle determined by goniometry and percent
Theory of Granulation 27
of 2 tons/hr. Figure 18(B) compares differences in lots of the formulation. Note that
illustrated in Figure 19, the effect
As previously depicted in Figure 12, the wetting and nucleation process for fine powlactose
of each formulation are given in parentheses. (From Ref. 31.)
 2005 by Taylor & Francis Group, LLC
Figure 18 Dynamic imaging of drop wetting, and its impact on continuous pan granulation:
(A) Dynamic image of a drop wetting into a formulation, with good active ingredient.
(B) Comparison of surface spreading velocity and dynamic contact angle vs. time for good
and bad active ingredients or technical. Bad technical required reduced production rates.
28 Ennis
(From Ref. 4.)
 2005 by Taylor & Francis Group, LLC
Second, droplets impact and coalesce on the powder bed surface if mixing or wet-in
time is slow. Third, droplets spread and penetrate into the moving powder bed to
form loose nuclei, again coalescing, if wet-in is slow. In the case of high-shear processes,
shear forces break down overwet clumps, also producing nuclei.
Two key features control this wetting and nucleation process. One is the time
required for a drop to wet-in to the moving powder bed, in comparison to some circulation
time of the process. As discussed previously, this wet-in time is strongly
influenced by formulation properties (e.g., Eq. 2.3). The second is the actual spray
rate or spray flux, in comparison to solids flux, or mixing rates. Spray flux is strongly
influenced by manufacturing and process design.
One can envision that drop penetration time and spray flux define regimes of
nucleation and wetting. If the wet-in is rapid and spray fluxes are low, individual
drops will form discrete nuclei somewhat larger than the drop size, defining a droplet
controlled regime. At the other extreme, if drop penetration is slow and spray flux is
large, drop coalescence and pooling of binder material will occur throughout the
powder bed, which must be broken down by mechanical dispersion. In the mechanical
dispersion regime of nucleation, shear forces control the breakdown of wetting
clumps, independent of drop distribution.
Falling between these two extreme regimes, drop overlap and coalescence
occur to varying extent defining an intermediate regime of nucleation, being a function
of penetration time and spray flux. To better define wetting, particularly in the
sense of process engineering and scale-up, we will consider drop penetration or wet-in
time and spray flux in greater detail.
Beginning with penetration time, Eq. 2.3 defines key formulation properties
controlling capillary rise in powder beds. From considering a distribution of macrodetermined
a total drop penetration time tp of
tp ? 1:35
Reff g cos y   ?2:5?
As shown previously, drop wet-in time decreases with increasing pore radius
Reff, decreasing binder viscosity and increasing adhesion tension. In addition, drop
penetration time decreases with decreasing drop size Vd and increasing bed porosity
eeff. Effective pore radius Reff is related to the surface-volume average particle size
d32, particle shape, and effective porosity of packing eeff by
Reff ?
1  eeff   ?2:6?
To remain with a droplet controlled regime of nucleation, the penetration time
given by Eq. 2.5 should be less than some characteristic circulation time tc of the
granulator in question. Circulation time is a function of mixing and bed weight,
and can change with scale-up.
powder bed of width B moves past a flat spray of spray rate (dV/dt) at a solids velocity
of w. For a given spray rate, the number of drops is determined by drop volume,
which in turn defines the drop area a per unit time which will be covered by the
spray, giving a spray flux of:
Theory of Granulation 29
and micropores in the moving powder bed, as shown in Figure 20, Hapgood (24)
Now turning our attention to spray flux, consider Figure 21 where an idealized
 2005 by Taylor & Francis Group, LLC
Figure 19 Influence of capillary penetration on drum granule size: Increasing penetration
rate, as reflected by Eq. 2.3 increases granule size and decreases asymmetry of the granule size
Figure 20
30 Ennis
Drop penetration in a moving powder bed. (From Ref. 5.)
 2005 by Taylor & Francis Group, LLC
dt ?
4  ?
dV=dt ? ?
dd ?2:7?
As droplets contact the powder bed at a certain rate, the powder moves past
the spray zone at its own velocity, or at solids flux given for this simple example by
dt ? Bw ?2:8?
The ratio of the droplet spray flux to the solids flux defines a dimensionless
spray flux given by
ca ?
da=dt ? ?
dA=dt ? ??
dV=dt ? ?
dd dA=dt ? ? ?2:9?
The dimensionless spray flux is the ratio of the rate at which wetted area is covered
by droplets to the area of flux of powder through the spray zone, and is a measure
of the density of drops falling on the powder surface. As with drop penetration
time, it plays a role in defining the regimes of nucleation as illustrated in Figure 22
(5,23). For small spray flux (ca  1), drops will not overlap on contact and will form
separate discrete nuclei for fast penetration time. For large spray flux (ca1), however,
significant drop overlap occurs, forming nuclei much larger than drop size, and
in the limit, independent of drop size.
Figure 22 Monte-Carlo simulations of drop coverage. (From Ref. 5.)
Figure 21
Theory of Granulation 31
Idealized flat spray zone in a spinning riffle granulator. (From Ref. 5.)
 2005 by Taylor & Francis Group, LLC
For the case of random drop deposition as described by a Poisson distribution,
Hapgood (23) showed the fraction of surface covered by spray was given by
fsingle ? 1  exp?ca? ?2:10?
In addition, the fraction of single drops forming individual nuclei (assuming
rapid drop penetration) vs. the number of agglomerates formed was given by
fsingle ? exp?4ca? ?2:11?
fagglom ? 1  exp?4ca? ?2:12?
Examples of the above as applied to nucleation are depicted in Figures 23 and
tion of drop size and spray flux. Lactose was sprayed with a flat spray in a spinning
spray flux of , ca?0.22, a clear relationship is seen between nuclei size and spray
distribution, with nuclei formed somewhat larger than drop size as shown in Figure
23. However, as the speed of the riffles is slowed down (i.e., solids velocity and solids
flux are decreased, and dimensionless spray flux increased), the nuclei distribution
The spray flux captures the impact of equipment operating variables on nucleation,
and as such is very useful for scale-up if nucleation rates and nuclei sizes are to
be maintained constant. The overall impact of dimensionless spray flux on nucleation
and agglomerate formation is illustrated in with agglomerates
increasing with increased spray flux as clearly governed by Eq. 2.12 for the case of
rapid drop penetration.
Figure 23 Effect of spray drop distribution (low spray flux) on nuclei distribution: Lactose
32 Ennis
feed powder in spinning granulator. (From Ref. 33.)
24 and described by Litster et al. (33). Here, nuclei distribution was studied as a funcwidens
with the formation of agglomerates as depicted in Figure 24.
riffle granulator, mimicking the geometry of Figure 21. For a moderate, intermediate
 2005 by Taylor & Francis Group, LLC
the help of dimensionless drop penetration time tp and spray flux ca, or
tp ?
tc ?
penetration time
circulation time
and ca ?
da=dt ? ?
dA=dt ? ??
spray flux
solids flux ?2:13?
A droplet controlled nucleation regime occurs when there is both low spray
fluxrelatively few drops overlapand fast droplet penetrationdrops wet into
Figure 24 Effect of powder velocity (variable spray fluxwater) on nuclei distribution:
Figure 25 Agglomerate formation vs. spray flux: Lactose powder with water and HPLC
Theory of Granulation 33
The regimes of nucleation may now be defined, as depicted in Figure 26 with
Lactose feed powder in spinning granulator. (From Ref. 33.)
solutions. (From Ref. 5.)
 2005 by Taylor & Francis Group, LLC
the bed completely before bed mixing allows further drop contact. Nuclei will be
formed of the order of drop size.
A mechanical dispersion regime occurs at the other extreme of high spray flux
giving large drop overlap and coalescence, and large drop penetration times, promoted
by poor wet-in rates and slow circulation times and poor mixing. In this
regime, nucleation and binder dispersion occur by mechanical agitation. Viscous,
poorly wetting binders are slow to flow through pores in the powder bed in the case
of poor penetration time. Drop coalescence on the powder surface occurs (also
known as pooling) creating very broad nuclei size distributions. Binder solution
delivery method (drop size, nozzle height) typically has minimal effect on the nuclei
size distribution, though interfacial properties may affect nuclei and final granule
An intermediate regime exists for moderate drop penetration times and moderate
spray flux, with the resulting nuclei regime narrowing with decreases in both.
There are several implications with regard to the nucleation regime map of
Figure 26 with regard to trouble shooting of wetting and nucleation problems. If
drop penetration times are large, making adjustments to spray may not be suffi-
ciently narrow granule size distributions if remaining in the mechanical regime. Significant
changes to wetting and nucleation only occur if changes take the system
across a regime boundary. This can occur in an undesirable way if processes are
not scaled, with due attention to remaining in the drop controlled regime.
3.1. Overview
bution of a particulate feed in a granulation process is controlled by several mechanisms,
including wetting, nucleation, and coating; granule coalescence and
consolidation; and granule breakage. Nucleation of fine powders and coating of
existing granules by the fluid phase have been discussed in the previous section.
Figure 26 A possible regime map of nucleation, relating spray flux, solids mixing (solids flux
34 Ennis
As previously outlined in Figures 4 and 10, the evolution of the granule size distriand
circulation time), and formulation properties. (From Refs. 5,24.)
 2005 by Taylor & Francis Group, LLC
Breakage mechanisms will be treated in the following section. Here, we focus particularly
on growth and consolidation mechanisms.
Granule growth includes the coalescence of existing granules as well as the
layering of fine powder onto previously formed nuclei or granules. The breakdown
of wet clumps into a stable nuclei distribution can also be included among coalescence
mechanisms. As granules grow by coalescence, they are simultaneously compacted
by consolidation mechanisms, which reduce internal granule voidage or
There are strong interactions between these rate processes, as illustrated in
Figure 27 for the case of drum granulation of fine feed. Here, granule size is illustrated
to progress through three stages of growth, including rapid, exponential
growth in the initial nucleation stage, followed by linear growth in the transition
stage, finishing with very slow growth in a final balling stage. Simultaneously with
growth, granule porosity or voidage is seen to decrease with time as the granules
are compacted. Granule growth and consolidation are intimately connected;
increases in granule size are shown here to be associated with a decrease in granule
porosity. This is a dominant theme in wet granulation.
As originally outlined in Ennis (3), these growth patterns are common throughout
fluidized-bed, drum, pan, and high-shear mixer processes for a variety of formulations.
Specific mechanisms of growth may dominate for a processsometimes to
the exclusion of others (4,5,11,12). However, what all processes have in common
is that the prevailing mechanisms are dictated by a balance of critical particle level
properties, which control formulation deformability, and operating variables which
control the local level of shear, or bed agitation intensity.
3.2. Mechanics of the Growth Rate Process
In order for two colliding granules to coalesce rather than break up, the collisional
Figure 27 Granule porosity and mean (pellet) size: Typical regimes of granule growth and
Theory of Granulation 35
kinetic energy must first be dissipated to prevent rebound as illustrated in Figure 28.
consolidation. (Adapted from Refs. 1519.)
 2005 by Taylor & Francis Group, LLC
In addition, the strength of the bond must resist any subsequent breakup forces in
the process. The ability of the granules to deform during processing may be referred
to as the formulations deformability, and deformability has a large effect on growth
rate. Increases in deformability increase the bonding or contact area thereby dissipating
and resisting breakup forces. From a balance of binding and separating forces
and torque acting within the area of granule contact, Ouchiyama and Tanaka (34)
derived a critical limit of size above which coalescence becomes impossible, or a maximum
growth limit given by
Dc ? ?AQ3B=2K3=2sT?1=?4?3=2?Z? ?3:1?
Here, K is deformability, a proportionality constant relating the maximum
compressive force Q to the deformed contact area, A is a constant with units of
[L3/F], which relates granule volume to impact compression force, and sT is the tensile
strength of the granule bond. Granules are compacted as they collide. This expels
pore fluid to the granule surface thereby increasing local liquid saturation in the contact
area of colliding granules. This surface fluid (1) increases the tensile strength of
the liquid bond sT and (2) increases surface plasticity and deformability K.
Dc represents the largest granule that may be grown in a granulation process,
and it is a harmonic average granule size. Therefore, it is possible for the collision of
two large granules to be unsuccessful, their average being beyond this critical size,
whereas the collision of a large and small granule leads to successful coalescence.
The growth limit Dc is seen to increase with increased formulation deformability
(which will be shown to be a strong function of moisture and primary particle size
distribution), increased compressive forces (which are related to local shear levels
in the process), and increased tensile forces (which are related to the interparticle
forces previously discussed.) The parameters z and Z depend on the deformation
mechanism within the contact area. For plastic deformation, z?1, Z?0, and K /
1/H where H is hardness. For elastic, Hertzian deformation, z?2/3, Z?2/3, and
K / (1/E)2/3 where E is the reduced elastic modulus. Granule deformation is initially
dominated by inelastic behavior of contacts during collision (35).
Figure 28 Mechanisms of granule coalescence for low- and high-deformability systems:
Rebound occurs for average granule sizes greater than the critical granule size Dc.
36 Ennis
K?deformability. (From Refs. 1,2.)
 2005 by Taylor & Francis Group, LLC
3.3. Types of Granule Growth
The importance of deformability to the growth process depends on bed agitation
intensity. If little deformation takes place during granule collisions, the system is
referred to as a low deformability or low agitation intensity process. This generally
includes fluid bed, drum, and pan granulators. Growth is largely controlled by the
extent of any surface fluid layer and surface deformability, with surface fluid playing
a large role in dissipating collisional kinetic energy. Growth generally occurs at a faster
time scale than overall granule deformation and consolidation. This is depicted in
Figure 29, where smaller granules can still be distinguished as part of a larger granule
structure, or a pop-corn-type appearance as often occurs in fluid-bed granulation.
(Note that such a structure may not be observed if layering and nucleation alone
dominate with little coalescence of large granules.) In this case, granule coalescence
and consolidation have less interaction than they do with high deformability systems,
making low deformabilitylow agitation systems easier to scale and model.
For high shear rates or bed agitation intensity, large granule deformation
occurs during granule collisions, and granule growth and consolidation occur on
the same time scale. Such a system is referred to as a deformable or high agitation
intensity process, and it generally includes continuous pin and plow shear-type mixers,
as well as batch high-shear pharmaceutical mixers. In these cases, substantial
collisional kinetic energy is dissipated with deformation of the wet mass composing
the granule. Rather than a sticking together process as often occurs in the low
deformability process of fluid beds, granules are smashed together as with a highshear
mixer, and smaller granules are not distinguishable with the granule structure,
as depicted in Figure 29. High-agitation, high-deformable processes generally produce
denser granules than those with low deformablity and low agitation intensity. In addition,
the combined and competing effects of granule coalescence and consolidation
make high-agitation processes difficult to model and scale. Both coalescence and
consolidation increase with increase in shear and deformability, whereas as granules
densify, they become less deformable, which works to lower coalescence in the later
stages of growth.
Figure 29 Granule structures resulting from (A) low- and (B) high-deformability systems,
Theory of Granulation 37
typical for fluid-bed and high-shear mixer granulators, respectively. (From Ref. 4.)
 2005 by Taylor & Francis Group, LLC
Bedagitation intensity is controlled bymechanical variables of the process such as
fluid-bed excess gas velocity, or mixer impeller and chopper speed. Agitation intensity
controls the relative collisional and shear velocities of granules within the process and
therefore growth, breakage, consolidation, and final product density. Figure 30 summarizes
typical characteristic velocities, agitation intensities,andcompaction pressures,
and product relative densities achieved for a variety of size-enlargement processes.
Last, it should be noted that the process or formulation itself cannot define
whether it falls into a low- or high-agitation intensity process. As discussed more
fully in the following sections, it is a function of both the level of shear as well as
the formulation deformability. A very stiff formulation with low deformability may
behave as a low deformability system in a high-shear mixer, or a very pliable formulation
may act as a high-deformable system in a fluid-bed granulator.
3.4. Granule Deformability
Granule deformabilityand hence Dcis a strong function of moisture, as illu-
K is related to both the yield strength of the material sy, i.e., the ability of the material
to resist stresses, and the ability of the surface to be strained without degradation
or rupture of the granule, with this maximum allowable critical deformation strain
of agglomerates during compression as a function of liquid saturation, with strain
denoted by DL/L. In general, high deformability K requires low yield strength sy
and high critical strain (DL/L)c. For this formulation, increasing moisture increases
deformability by lowering interparticle frictional resistance, which also increases
mean granule size (Fig. 31).
In most cases, granule deformability increases with increasing moisture,
decreasing binder viscosity, decreasing surface tension, decreasing interparticle
Figure 30 Classification of agglomeration processes by agitation intensity and compaction
38 Ennis
strated in Figure 31 by the marked increase in average granule size. Deformability
denoted by (DL/L) . Figure 32 illustrates the low shear rate stressstrain behavior
pressure. (From Refs. 1,2,5.)
 2005 by Taylor & Francis Group, LLC
friction, and increasing average particle size (specifically dsv, or the surfacevolume
mean size), as well as increasing bed agitation intensity. The important contributions
of binder viscosity and friction to granule deformability and growth are illustrated
Note that 60% of the energy is dissipated through viscous losses, with the majority of
the remainder through interparticle friction. Very little is loss due to capillary forces
controlled by surface tension. Therefore, modern approaches to granule coalescence
rest in understanding the impact of granule deformability on growth, rather than the
original capillary framework alone, as put forth by Rumpf (20,21) regarding pendular
and funicular forces due to interparticle liquid bridges.
3.5. Interparticle Forces Due to Pendular Bridges
Interstitial fluid and the resulting pendular bridges play a large role in both granule
growth and granule deformability. Pendular bridges between particles of which a
granule is composed, give rise to capillary and viscous interparticle forces. These
forces develop interparticle pressure, which also allows friction to act between point
deformability, warrant further attention. Note that capillary forces for small contact
Figure 31 Effect of granule saturation on mean granule diameter, indicating the marked
increase in granule deformability. Mean granule diameter is a measure of the critical limit
of size Dc. Granulation of calcium hydrogen phosphateaqueous binder solutions. Fielder
Theory of Granulation 39
by fractions of energy dissipated during a granule collision as depicted in Figure 33.
contacts (Fig. 34). Interparticle forces due to pendular bridges, and their impact on
PMAT 25 VG mixer. (From Ref. 36.)
 2005 by Taylor & Francis Group, LLC
angle attract particles (but repel for y>90), whereas viscous and frictional forces
act to resist the direction of motion.
separated by a gap distance 2ho approaching one another at a velocity U. The two
particles may represent two primary particles within the granule, in which case we
are concerned about the contribution of interparticle forces to granule strength
and deformability. Or they may represent two colliding granules, in which case we
are concerned with the ability of the pendular bridge to dissipate granule kinetic
energy and resist breakup forces in the granulation process. The two spheres are
bound by a pendular bridge of viscosity m, density r, and surface tension g. The
pendular bridge consists of the binding fluid in the process, which includes the added
solvent and any solubilized components. In some cases, it may also be desirable to
include very fine solid components within the definition of the binding fluid, and,
therefore, consider instead a suspension viscosity and surface tension. These material
parameters vary on a local level throughout the process, and are also time dependent
and a function of drying conditions.
For the case of a static liquid bridge (i.e., U?0) with perfect wetting, surface
tension induces an attractive capillary force between the two particles due to a threephase
contact line force and a pressure deficiency arising from interfacial curvature.
The impact of this static pendular bridge force on static granule strength has been
studied and reported extensively (3,2022). It is important to recognize that in most
processes, however, the particles are moving relative to one another and, therefore,
the bridge liquid is in motion. This gives rise to viscous resistance forces which can
contribute significantly to the total bridge strength. The strength of both Newtonian
Figure 32 The influence of sample saturation S on granule deformability: Deformation
strain (DL/L) is measured as a function of applied stress, with the peak stress and strain
denoted by tensile strength sy and critical strain (DL/L)c of the material. Dicalcium phosphate
with 15 wt% binding solution of PVP/PVA Kollidon VA64. Fifty percent compact porosity.
40 Ennis
As shown in Figures 11 and 34, consider two spherical particles of radius a
(From Ref. 37.)
 2005 by Taylor & Francis Group, LLC
and non-Newtonian pendular bridges have been studied extensively (3,39,40). For
Newtonian fluids (39), the dynamic strength was shown to be given by
pga ? Fcap ? Fvis ? Fo ? 3Ca=e where
Fcap ? ?2  2Ho? sin2 j
Fvis ? 3Ca=e
Ca ? mU=g
Here, all forces have been made dimensionless with respect to a measure of the
capillary force, or pga. Note that the bridge strength consists of two components.
Fcap is the strength of a static capillary bridge, and is a function of curvature of
Figure 33 Distribution of energy dissipation during agglomerate collisions, with granular
simulations of wall impact for 128 ms duration for invisicid and viscous binder agglomerates.
Figure 34 Interparticle forces and granule deformability: Interparticle forces include
Theory of Granulation 41
(From Ref. 11.)
capillary forces, viscous lubrication forces, and frictional forces. (From Ref. 4.)
 2005 by Taylor & Francis Group, LLC
the interface Ho and the filling angle. In dimensional form, it is given by
F cap ? pga?2  2Ho? sin2 j ?3:3?
Fvis is the strength of a viscous, dynamic bridge, and is equivalent to the force
between two spheres approaching one another in an infinite fluid. This force is a
function of binder viscosity m, and the collision velocity U. Here, e?2ho/a is gap distance,
and not granule voidage. In dimensional form, the viscous force is given by
F vis ? 6pmUa=e ?3:4?
From Eq. 3.2, one finds that the dynamic bridge force begins with the static
bridge strengthwhich is a constant independent of velocity (or Ca), and then
increases linearly with Ca, which is a capillary number representing the ratio of viscous
(pmUa) to capillary (pga) forces, and is proportional to velocity. This is con-
firmed experimentally as illustrated in Figure 35 for the case of two spheres
approaching axially. Extensions of the theory have also been conducted for
non-Newtonian fluids (shear thinning), shearing motions, particle roughness, wettability,
and time-dependent drying binders. The reader is referred to Ennis (3,39)
for additional details. Curves for non-Newtonian fluids are also included in
Figure 35.
For small velocities, small binder viscosity, and large gap distances, the
strength of the bridge will approximate a static pendular bridge, or Fcap, which is
proportional to and increases with increases in surface tension. This force is equivalent
to the static pendular force H previously given in Eq. 1.2 as studied by Rumpf
(20,21) and Augsburger and Vuppala (22). On the other hand, for large binder viscosities
and velocities, or small gap distances, the bridge strength will approximately
be equal to Fvis, which is proportional to and increases with increases in binder viscosity
and velocity. This viscous force is singular in the gap distance and increases
dramatically for small separation of the particles. It is important to note that as
Figure 35 Maximum strength of a liquid bridge between two axial moving particles as
42 Ennis
a function of Ca for Newtonian and shear thinning fluids. (From Ref. 3.)
 2005 by Taylor & Francis Group, LLC
granules are consolidated, resulting in decreases in effective interparticle gap distance,
and as binders dry, resulting in large increases in binder viscosity, the dynamic
bridge strength can exceed the static strength by orders of magnitude.
3.6. Dynamic Wet Mass Rheology and Deformability Revisited
the case of slow yielding. However, the dependence of interparticle forces on shear
rate clearly impacts wet mass rheology and therefore deformability. Figure 36 illustrates
the stressstrain response of compacts, demonstrating that the peak flow or
yield stress increases proportionally with compression velocity (41). In fact, in a similar
fashion to dynamic liquid bridge forces, the peak flow stress of wet unsaturated
compacts (initially pendular state) can be seen to also increase with Ca as follows
g=a ? so ? ACa
so ? 5:0  5:3
A ? 280  320; B ? 0:58  0:64
Ca ? m_ea=g
There are several important issues worth noting with regard to these results.
First is the similarity between the strength of the assembly or compact given by
Eq. 3.5 and the strength of the individual dynamic pendular bridge given by Eq.
3.2; both curves are similar in shape with a capillary number dependency. As with
the pendular bridge, two regions may be defined. In region 1, for a bulk capillary
number of Ca < 104, the strength or yield stress of the compact depends on the static
pendular bridge, and therefore on surface tension, particle size, and liquid load-
Figure 36 Typical compact stress response for fast compression vs. crosshead compression
velocity for glass ballotini (d32?35 mm) and compact diameter 20 mm, length 25 mm. (From
Theory of Granulation 43
The yield stress of powder compacts has previously been reported in Figure 32 for
(Fig. 37):
Ref. 41.)
 2005 by Taylor & Francis Group, LLC
ing. In region 2, for Ca > 104, the strength depends on the dynamic pendular
bridge, and therefore on binder viscosity and strain rate, in addition to particle size.
Second, the results of Figure 37 do not clearly depict the role of saturation and
compact porosity, though these properties are known to affect strength. Decreases
in compact porosity generally increases compact strength through increases in inter-
(36,37,42). Hence, the curve of Figure 37 should be expected to shift with these variables,
particularly since the viscous force for axial approach is singular in the interparticle
gap distance (Eqs. 3.4).
Third, the static granule or assembly strength (Eq. 1.2), as originally developed
by Rumpf (20,21) and Augsburger and Vuppala (22), is captured by the constant,
low Ca value of the yield stress or
y  so g=a ? ?
1  e
e  H
1  e
e  g cos y
for Ca < 104 ?3:6?
Fourth, the mechanism of compact failure also depends on strain rate.
function of capillary number. At low Ca, compacts fail by brittle fracture with
macroscopic crack propagation, whereas at high Ca, compacts fail by plastic flow.
Within the context of granulation, small yield stresses at low Ca may result in
unsuccessful growth when these stresses are compared with large breakup forces.
With increased yield stress comes stronger granules but also decreased deformability.
Therefore, high strength might imply a low-deformability growth mechanism for
low-shear processes such as a fluid bed. On the other hand, it might imply smaller
growth rates for high-shear processes which are able to overcome this yield stress
and bring about kneading action and plastic flow in the process. Therefore, it is
important to bear in mind that increased liquid saturation may initially lower yield
stress, allowing more plastic deformation during granule collisions. However, as
granules grow and consolidate and decrease in voidage, they also strengthen and rise
Figure 37
44 Ennis
Dimensionless peak flow stress of Fig. 36 vs. bulk capillary number, for various
particle friction, whereas increases in saturation lower strength (e.g., Figure 32)
Figure 38 illustrates schematically the crack behavior observed in compacts as a
binder solutions. (From Ref. 41.)
 2005 by Taylor & Francis Group, LLC
in yield stress, becoming less deformable with time and withstanding shear forces in
the granulator. Hence, the desired granule strength and deformability is linked in
a complex way to granulator shear forces and consolidation behavior.
3.7. Low-Agitation IntensityLow-Deformability Growth
For those low-agitation processes or formulations which allow little granule deformation
during granule collisions, consolidation of the granules occurs at a much
slower rate than growth, and granule deformation can be ignored to a first approximation.
The growth process can be modeled by the collision of two nearly stiff granules
each coated by a liquid layer of thickness h. A pictorial depiction of a granule
collision is illustrated in Figure 38. For the case of zero plastic deformation, the
probability of successful coalescence as developed by Ennis and coworkers (3,43)
is governed by a dimensionless energy, or Stokes number St given by
St ?
9m ?3:7?
where u0 is the relative collisional velocity of the granules, r is granule density, a is
the harmonic average of granule diameter, and m is the solution phase binder viscosity.
The Stokes number represents the ratio of initial collisional kinetic energy to the
energy dissipated by viscous lubrication forces, and it is one measure of normalized
bed agitation energy. Successful growth by coalescence or layering requires that
St < St where St ? 1 ?
er  ln?h=ha? ?3:8?
where St is a critical Stokes number representing the energy required for rebound.
The binder layer thickness h is related to liquid loading, er is the coefficient of restitution
of the granules, and ha is a measure of surface roughness or asperities. The
critical condition given by Eq. 3.8 controls the growth of low deformability systems.
This criteria has also been extended to capillary coalescence (3) and for cases
of plastic deformation (44).
Binder viscosity is a function of local temperature, collisional strain rate (for
non-Newtonian binders), and binder concentration dictated by drying rate and local
Figure 38 Collisions between surface wet granules, beginning with approach, and ending
Stokes model.
Theory of Granulation 45
with separation. (From Refs. 3,5,43,44.) Note that no deformation takes place in the original
 2005 by Taylor & Francis Group, LLC
mass transfer and local bed moisture. It can be controlled through judicious selection
of binding and surfactant agents and measured by standard rheological techniques
(45). The collisional velocity is a function of process design and operating variables,
and is related to bed agitation intensity and mixing. Possible choices of u0 are sum-
Note that u0 is an interparticle collisional velocity, which is not necessarily the local
average granular flow velocity.
Three regimes of granule growth may be identified for low-agitation intensity
low-deformability processes (3,43), as depicted for fluid-bed granulation in Figure 39,
and outlined as follows:
Noninertial regime: For small granules or high binder viscosity lying within a
noninertial regime of granulation, all values of St will lie below the critical
value St and therefore all granule collisions result in successful
growth provided binder is present. Growth rate is independent of granule
kinetic energy, particle size, and binder viscosity (provided other rate
processes are constant). Distribution of binding fluid and degree of mixing
then control growth and, and this is strongly coupled with the rate
process of wetting (see previous section). As shown in Figure 39, both
binders have the same initial growth rate for similar spray rates, independent
of binder viscosity. Increases in bed moisture (e.g., spray rate, drop
rate) and increases in granule collisions in the presence of binder will
increase the overall rate of growth. It must be borne in mind, however,
that there is a 100% success of these collisions, since dissipation of energy
far exceeds collisional kinetic energy.
Inertial regime: As granules grow in size, their collisional momentum increases,
leading to localized regions in the process where St exceeds the critical
value St. In this inertial regime of granulation, granule size, binder
Figure 39 Median granule diameter for fluid-bed granulation of ballotini with binders of dif-
46 Ennis
marized in Figure 30, and discussed more fully regarding scale-up in Chapter 18.
ferent viscosity indicating regimes of growth. (From Refs. 3,43.)
 2005 by Taylor & Francis Group, LLC
viscosity, and collision velocity determine the proportion of the bed in
which granule rebound is possible. Increases in binder viscosity and
decreases in agitation intensity increase the extent of granule growth
i.e., the largest granule that may be grown [e.g., Dc of Eq. 3.1]. This is
confirmed in with the CMC binder continuing to grow,
whereas the PVP system with lower viscosity slows down in growth. Note
that the rate of growth, however, is controlled by binder distribution and
mixing, and not binder viscosity. For example, increasing binder viscosity
will not affect growth rate, or initial granule size, but it will result in
an increased growth limit.
Coating regime: When the spatial average of St exceeds St, growth is balanced
by granule disruption or breakup, leading to the coating regime of granulation.
Growth continues by coating of granules by binding fluid alone.
The PVP system with lower viscosity is seen to reach its growth limit and
therefore coating regime in Figure 39.
Transitions between granulation regimes depend on bed hydrodynamics. As
demonstrated by Ennis et al. (3,4,43), granulation of an initially fine powder may
exhibit characteristics of all three granulation regimes as time progresses, since St
increases with increasing granule size. Implications and additional examples regarding
the regime analysis are highlighted by Ennis (3,4,43). In particular, increases in
fluid-bed excess gas velocity exhibit a similar but opposite effect on growth rate
to binder viscosity; namely, it is observed to not affect growth rate in the initial
Noninertial regime of growth, but instead lowers the growth limit in the inertial
Figure 40 (A) Exponential growth in drum granulation reaching a growth limit dmax, or
maximum extent of growth (kt)max, which are functions of moisture saturation (1519). (B)
Maximum extent of noninertial growth (kt)max as a linear function of saturation of the powder
feed and binder viscosity. (C) Maximum extent normalized for differences in binder viscosity,
Theory of Granulation 47
Figure 39
drum speed, granule density by Stokes number. (From Ref. 46.)
 2005 by Taylor & Francis Group, LLC
3.8. Example: Extent of Noninertial Growth
Growth by coalescence in granulation processes may be modeled by population
nel which describe growth. For fine powders within the noninertial regime of growth,
all collisions result in successful coalescence provided binder is present. Coalescence
occurs via a random, size independent kernel which is only a function of liquid loading
y, as well as mixing, or
b?u; v? ? k ? kf ?y? ?3:9?
For random growth and in the presence of sufficient binding fluid, it may be
rigorously proven that the average granule size increases exponentially with time, or
d ? d0 ekt ?3:10?
This exponential increase in size with time is confirmed experimentally in
granule saturation S is connected to liquid loading y and porosity.) Based on the
regime analysis work above, growth will continue in a process while the conditions
of Eq. 3.8 are met, i.e., dissipation exceeds collisional kinetic energy. Examples of
these growth limits are seen in the drum granulation work of Kapur (1519) in
the maximum extent of granulation (kt)max occurring within the noninertial regime is
given by
Figure 41 Granule diameter as a function of time for high-shear mixer granulation, illustrating
the influence of deformability on growth behavior. Directions of increasing viscosity and
impeller speed indicated by arrows. (A) Ten-liter melt granulation of lactose with 15 wt% binder
and impeller speed of 1400 rpm for two different viscosity grades of polyethylene gylcol
binders. (B) Ten-liter vertical high-shear mixer granulation of dicalcium phosphate with
15 wt% binder solution of PVP/PVA Kollidon VA64, liquid loading of 16.8 wt%, and chop-
48 Ennis
Figure 40, where increases in liquid loading f (y) increase growth rate. (Note that
Figure 40, as well as fluid beds (Fig. 39) and mixers (Fig. 41). It may be shown that
balances (Chapter 21). It is necessary to determine both the mechanism and the kerper
speed of 1000 rpm for varying impeller speed. (From Refs. 47,48.)
 2005 by Taylor & Francis Group, LLC
?kt?max ? 6 ln?St=St0? f ?y? / ln
ru0d0   ?3:11?
St0 is the Stokes number based on initial nuclear diameter d0 (46). Extent
(kt) max is taken as the logarithm of the growth limit in the first random stage of
growth, or dmax. The growth limits dmax
in Figure 40(B). Here, (kt)max is observed to depend linearly on liquid loading y.
Therefore, the maximum granule size depends exponentially on liquid loading, as
From Eq. 3.11, it is possible to scale or normalize a variety of drum granulation
data to a common drum speed and binder viscosity. Maximum granule size dmax
and extent (kt)max depend linearly and logarithmically, respectively, on binder
viscosity and the inverse of agitation velocity. This is illustrated in Figure 40(B),
where the slope of each formulation line depends linearly on binder viscosity.
Figure 40 illustrates the normalization of extent (kt)max for the drum granulation
of limestone and fertilizers, correcting for differences in binder viscosity, granule
density, and drum rotation speed, with the data collapsing onto a common line
following normalization.
3.9. High-Agitation IntensityDeformable Growth
For high-agitation processes involving high-shear mixing or for readily deformable
formulations, granule deformability, plastic deformation, and granule consolidation
can no longer be neglected as they occur at the same rate as granule growth. Typical
growth are evident, which reveal the possible effects of binder viscosity and impeller
speed, as shown for data replotted vs. impeller speed in Figure 42. The initial, nonequilibrium
stage of growth is controlled by granule deformability, and is of most
Figure 42 Granule diameter as a function of impeller speed for both initial nonequilibrium
Theory of Granulation 49
of Figure 40(A) are replotted as extents
observed experimentally (Fig. 31).
growth profiles for high-shear mixers are illustrated in Figure 41. Two stages of
and final equilibrium growth limit for high-shear mixer granulation (From Ref. 10,48.)
 2005 by Taylor & Francis Group, LLC
practical significance in manufacturing for high-shear mixers. Increases in St due to
lower viscosity or higher impeller speed increase the rate of growth, since the system
becomes more deformable and easier to knead into larger granule structures (e.g.,
based on rigid, low-deformability granules, where high viscosity and low velocity
increase the growth limit.
Growth continues until disruptive and growth forces are balanced in the process.
This last equilibrium stage of growth represents a balance between dissipation
and collisional kinetic energy, and so increases in St decrease the final granule size, as
expected from the Stokes analysis. Note that the equilibrium granule diameter
decreases with the inverse square root of the impeller speed, as it should based on
St?St, with u0?a(du/dx)?oa.
The Stokes analysis is used to determine the effect of operating variables and
binder viscosity on equilibrium growth, where disruptive and growth forces are
balanced. In the early stages of growth for high-shear mixers, the Stokes analysis
in its present form is inapplicable. Freshly formed, uncompacted granules are easily
deformed, and as growth proceeds and consolidation of granules occur, they will
surface harden and become more resistant to deformation. This increases the importance
of the elasticity of the granule assembly. Therefore, in later stages of growth,
older granules approach the ideal Stokes model of rigid collisions. In addition, the
Stokes number controls in part the degree of deformation occurring during a collision
since it represents the importance of collision kinetic energy in relation to viscous
dissipation, although the exact dependence of deformation on St is presently
The Stokes coalescence criteria of Eq. 3.8 developed by Ennis must be generalized
to account for substantial plastic deformation in order to treat the initial nonequilibrium
stages of growth in high-agitation systems such as high-shear mixers. In
this case, granule growth and deformation are controlled by a generalization of St,
or a deformation Stokes number Stdef, as originally defined by Tardos et al. (49,50):
Stdef ?
sy ?
sy ?3:12?
Viscosity has been replaced by a generalized form of plastic deformation controlled
by the yield stress sy, which may be determined by compression experiments.
As shown previously, yield stress is related to deformability of the wet mass, and is a
function of shear rate, binder viscosity, and surface tension, primary particle size and
Critical conditions required for granule coalescence may be defined in terms of
the viscous and deformation Stokes number, or St and Stdef, respectively. These
represent a complex generalization of the critical Stokes number given by Eq. 3.9,
and are discussed in detail elsewhere (5,44).
An overall view of the impact of deformability on growth behavior may be
in a regime map, and yield stress has been measured by compression experiments
(51). Growth mechanism depends on the competing effects of high-shear promoting
growth by deformation on the one hand and the breakup of granules giving a growth
limit on the other. For high velocities, growth is not possible by deformation due to
high shear, and the material remains in a crumb state. For low pore saturation,
50 Ennis
Fig. 29). These effects are contrary to what is predicted from the Stokes analysis
friction, and saturation and voidage as previously given by Eq. 3.5 and Figures 32
gained from Figure 43, where types of granule growth are plotted vs. deformability
 2005 by Taylor & Francis Group, LLC
and 37.
growth is only possible by initial wetting and nucleation, with the surrounding
powder remaining ungranulated. At intermediate levels of moisture, growth occurs
at a steady rate for moderate deformability, but has a delay in growth for low
deformability. This delay, or induction time, is related to the time required to
work moisture to the surface to promote growth. For high moisture, very rapid
and potentially unstable growth occurs.
The currrent regime map as presented requires considerable development.
Overall growth depends on the mechanics of local growth, as well as the overall mixing
pattern and local/overall moisture distribution. Levels of shear are poorly understood
in high-shear processes. In addition, growth by both deformation and the rigid
growth model is possible. Lastly, deformability is intimately linked to both voidage
and moisture. They are not a constant for a formulation, but depend on time and the
growth process itself through the interplay of growth and consolidation. Nevertheless,
the map provides a starting point, and is discussed in additional detail in
3.10. Determination of St
The extent of growth is controlled by some limit of granule size, either reflected by
the critical Stokes number St or by the critical limit of granule size Dc. There are
several possible methods to determine this critical limit. The first involves measuring
the critical rotation speed for the survival of a series of liquid binder drops during
drum granulation (3). A second refined version involves measuring the survival of
granules in a couette-fluidized shear device (49,50). Both the onset of granule deformation
and the complete granule rupture are determined from the dependence of
granule shape and the number of surviving granules, respectively, on shear rate.
The critical shear rate describing complete granule rupture defines St, whereas
the onset of deformation and the beginning of granule breakdown defines an
Figure 43 Regime map of growth mechanisms, based on moisture level and deformabilty of
Theory of Granulation 51
Chapter 18.
formulations. (From Ref. 51.)
 2005 by Taylor & Francis Group, LLC
additional critical value (Fig. 44). The third approach is to measure the deviation in
the growth rate curve from random exponential growth (52). The deviation from
random growth indicates a value of w, or the critical granule diameter at which
noninertial growth ends. This value is related toDc
Figure 44 Determination of the onset of granule deformation and complete granule breakdown
with the fluidized couette constant shear device. Y is a yield number, Fsur is the fraction
of surviving granules, andDdef
Figure 45 Determination of critical granule diameter, or growth limit, from the evolution of
52 Ennis
(Fig. 45). (See Chapter 21 regarding
is the average degree of granule deformation. (From Refs. 49,50.)
the granule size distribution. (From Ref. 51.)
 2005 by Taylor & Francis Group, LLC
modeling for further discussion.) The last approach is through the direct measurement
of the yield stress through compression experiments.
3.11. Example: High-Shear Mixer Growth
An important case study for high-deformability growth was conducted by Holm et al.
(42) for high-shear mixer granulation. Lactose, dicalcium phosphate, and dicalcium
phosphate/starch mixtures (15% and 45% starch) were granulated in a Fielder
PMAT 25 VG laboratory scale mixer. Granule size, porosity, power level, temperature
rise, and fines disappearance were monitored during liquid addition and wet
massing phases. Impeller and chopper speeds were kept constant at 250 and 3500
rpm, respectively, with 7.07.5 kg starting material. Liquid flow rates and amount
of binder added were varied according to the formulation.
Figure 46 illustrates typical power profiles during granulation, whereas
wet massing time (as opposed to total process time) is defined as the amount of time
following the end of liquid addition, and the beginning of massing time is indicated
in Figure 47.
Clear connections may be drawn between granule growth, consolidation,
power consumption, and granule deformability. From comparing Figures 46 and
47 for the case of lactose, one may note that there is no further rise in power following
the end of water addition (beginning of wet massing), and this corresponds to no
further changes in granule size and porosity. In contrast, dicalcium phosphate continues
to grow through the wet massing stage, with corresponding continual
increases in granule size and porosity. Last, the starch formulations are noted to
have power increase for approximately 2 min into the wet massing stage, corresponding
to 2 min of growth; however, growth ceases when power consumption levels off.
Therefore, power clearly tracks growth and consolidation behavior.
Further results connecting power and growth to compact deformability are
provided in Holm (42). The deformability of lactose compacts, as a function of
saturation and porosity, is shown to increase with moisture in a stable fashion.
Figure 46 Power consumption for lactose, dicalcium phosphate, and dicalcium phosphate/
Theory of Granulation 53
Figure 47 illustrates the resulting granule size and voidage (or porosity). Note that
starch mixtures (15% and 45% starch) granulated in a Fielder PMAT 25 VG. (From Ref. 42.)
 2005 by Taylor & Francis Group, LLC
Therefore, growth rates and power rise do not lag behind spray addition, and growth
ceases with the end of spraying. Dicalcium phosphate compacts, on the other hand,
remain undeformable until a critical moisture is reached, after which they become
extremely deformable and plastic. This unstable behavior is reflected by an inductive
lag in growth and power after the end of spray addition, ending with unstable growth
and bowl sticking as moisture is finally worked to the surface.
In closing, a comment should be made with regard to using power for control
and scale-up. While it is true the power is reflective of the growth process, it is a
dependent variable in many respects. Different lots of an identical formulation,
e.g., may have different yield properties and deformability, and a different dependence
on moisture. This may be due to minute particle property changes controlling
the rate processes. Therefore, there is no unique relationship between power and
growth. However, power measurements might be useful to indicate a shift in formation
Lastly, specific power is required for scale-up, where power is normalized by
the active portion of the powder bed. The impact of scale-up on mixing and distribution
of power in a wet mass, however, is only partly understood at this point.
3.12. Granule Consolidation
Consolidation of granules determines granule porosity or voidage, and hence granule
density. Granules may consolidate over extended times and achieve high densities if
there is no simultaneous drying to stop the consolidation process. The extent and
rate of consolidation are determined by the balance between the collision energy
and the granule resistance to deformation. The voidage e may be shown to depend
on time as follows:
e  emin
e0  emin ? exp?bt? where b ? fn?y; St; Stdef? ?3:13?
Here, y is liquid loading, e0 and emin are the beginning and final minimum porosity,
respectively (53). The effect of binder viscosity and liquid content are complex
and interrelated. For low-viscosity binders, consolidation increases with liquid
Figure 47 Granule size and porosity vs. wet massing time for lactose, dicalcium phosphate,
and dicalcium phosphate/starch mixtures (15% and 45% starch) granulated in a Fielder
54 Ennis
PMAT 25 VG. (From Ref. 42.)
 2005 by Taylor & Francis Group, LLC
content as shown in Figure 48 (53). This is the predominant effect for the majority of
granulation systems, with liquid content related to peak bed moisture on average.
Increased drop size and spray flux are also known to increase consolidation. Drying
effects peak bed moisture and consolidation as well through varying both moisture
level as well as binder viscosity. For very viscous binders, consolidation decreases
with increasing liquid content. As a second important effect, decreasing feed particle
size decreases the rate of consolidation due to the high specific surface area and low
permeability of fine powders, thereby decreasing granule voidage. Last, increasing
agitation intensity and process residence time increases the degree of consolidation
by increasing the energy of collision and compaction. The exact combined effect
of formulation properties is determined by the balance between viscous dissipation
and particle frictional losses, and therefore the rate is expected to depend on the
viscous and deformation Stokes numbers (53).
4.1. Overview
Dry granule strength impacts three key areas of pharmaceutical processing. These
include the physical attrition or breakage of granules during the granulation and
drying processes, the breakage of granules in subsequent material handling steps
such as conveying or feeding, and last, the deformation and breakdown of granules
in compaction processes such as tableting. Modern approaches to granule strength
rely on fracture mechanics (54). In this context, a granule is viewed as a nonuniform
physical composite possessing certain macroscopic mechanical properties, such as a
generally anisotropic yield stress, as well as an inherent flaw distribution. Hard materials
may fail in tension, with the breaking strength being much less than the inherent
tensile strength of bonds because of the existence of flaws. Flaws act to concentrate
or notches have been added to the tablets, which were subsequently broken under
three-point bend loading (see later). In all cases, the tablets break at the razor
scorewhich acts as a sharp flaw to concentrate stressrather than at the tableted
original indentation notch.
Figure 48 Effect of binder liquid content and primary feed particle size on granule porosity
for the drum granulation of glass ballotini: Decreasing granule porosity corresponds to
Theory of Granulation 55
stress, as depicted in Figure 49 for commercial metamucil tablets. Here, razor scores
increasing extent of granule consolidation. (From Ref. 53.)
 2005 by Taylor & Francis Group, LLC
Bulk breakage tests of granule strength measure both inherent bond strength
trates granule attrition results for a variety of formulations. Granule attrition clearly
increases with increasing voidage; note that this voidage is a function of granule consolidation
discussed previously. Different formulations fall on different curves, due
to inherently differing interparticle bond strengths. It is often important to separate
the impact of bond strength vs. voidage on attrition and granule strength. Processing
influences flaw distribution and granule voidage, whereas inherent bond strength is
controlled by formulation properties.
The mechanism of granule breakage is a strong function of material properties
of the granule itself as well as the type of loading imposed by the test conditions (55).
Ranking of product breakage resistance by ad hoc tests may be test specific, and in
the worst case differ from actual process conditions. Instead, material properties
should be measured by standardized mechanical property tests which minimize the
effect of flaws and loading conditions under well-defined geometries of internal
stress, as described in the following section.
4.2. Mechanics of the Breakage Rate Process
Fracture toughness Kc defines the stress distribution in the body just before fracture
and is given by
Kc ? Ysf ffiffiffiffiffi pc p ?4:1?
where sf is the applied fracture stress, c is the length of the crack in the body, and Y
The elastic stress is increased dramatically as the crack tip is approached. In practice,
however, the elastic stress cannot exceed the yield stress of the material, implying a
region of local yielding at the crack tip. Irwin (57) proposed that this process zone
size rp be treated as an effective increase in crack length dc. Fracture toughness is
then given by
Kc ? Ysf ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi p?c ? dc? p with d c  rp ?4:2?
The process zone is a measure of the yield stress or plasticity of the material in
comparison to its brittleness. Yielding within the process zone may take place either
plastically or by diffuse microcracking, depending on the brittleness of the material.
Figure 49 Breakage of metamucil tablets by under three-point loading with razor scoring.
(A) Upper left: unscored; upper right: notched scored; bottom: scored 90 to notch. (B) Break-
56 Ennis
and granule flaw distribution and voidage (3,4). Figure 8 presented previously illusis
a calibration factor introduced to account for different body geometries (Fig. 50).
age results. (From Ref. 4.)
 2005 by Taylor & Francis Group, LLC
For plastic yielding, rp is also referred to as the plastic zone size. The critical strain
energy release rate Gc is the energy equivalent to fracture toughness, first proposed
by Griffith (58). They are related by
Gc ? K2
c =E ?4:3?
4.3. Fracture Measurements
In order to ascertain fracture properties in any reproducible fashion, very specific
test geometry must be used since it is necessary to know the stress distribution at
predefined, induced cracks of known length. Three traditional methods are: (1) the
three-point bend test, (2) indentation fracture testing, and (3) Hertzian contact
cal geometry and force response for the case of a three-point bend test. By breaking a
series of dried formulation bars under three-point bend loading of varying crack
Figure 50
Figure 51 Three-point bend and indentation testing for fracture properties. (From Ref. 56.)
Theory of Granulation 57
compression between two spheres of the material. Figures 51 and 52 illustrate a typi-
Fracture of a brittle material by crack propagation. (From Ref. 56.)
 2005 by Taylor & Francis Group, LLC
length, fracture toughness is determined from the variance of fracture stress on crack
In the case of indentation fracture, one determines hardness H from the area of
the residual plastic impression, and fracture toughness from the lengths of cracks
propagating from the indent as a function of indentation load F (59). Hardness is
a measure of the yield strength of the material. Toughness and hardness in the case
of indentation are given by ture toughness Kc is seen to range from 0.01 to 0.06MPa m1/2, less than that typical
for polymers and ceramics, presumably due to the high agglomerate voidage. Critical
strain energy release rates Gc are from 1 to 200 J/m2, typical for ceramics but less
than that for polymers. Process zone sizes dc are seen to be large and of the order
of 0.11 mm, values typical for polymers. Ceramics on the other hand typically have
process zone sizes less than 1 m m. Critical displacements required for fracture may
be estimated by the ratio Gc/E, which is an indication of the brittleness of the material.
This value was of the order of 107108 mm for polymerglass agglomerates,
similar to polymers, and of the order of 109 mm for herbicide bars, similar to ceramics.
In summary, granulated materials behave not only similar to brittle ceramics,
which have small critical displacements and yield strains, but also similar to ductile
polymers, which have large process or plastic zone sizes.
4.4. Mechanisms of Breakage
The process zone plays a large role in determining the mechanism of granule
breakage (56), with mechanisms such as those previously presented in
Figure 52 Typical forcedisplacement curve for three-point bend semistable failure. (From
58 Ennis
Table 5 compares typical fracture properties of agglomerated materials. Frac-
Table 1.
Ref. 56.)
length, as given by Eq. 4.2. (For details, see Ref. 56.)
 2005 by Taylor & Francis Group, LLC

Agglomerates with process zones small in comparison to granule size break by a brittle
fracture mechanism into smaller fragments, called fragmentation or fracture. On
the other hand, for agglomerates with process zones of the order of their size, there is
insufficient volume of agglomerate to concentrate enough elastic energy to propagate
gross fracture during a collision. The mechanism of breakage for these materials is
one of wear, erosion, or attrition brought about by diffuse microcracking. In the
limit of very weak bonds, agglomerates may also shatter into small fragments or primary
Each mechanism of breakage implies a different functional dependence of
breakage rate on material properties. Granules generally have been found to have
large process zones, which suggests granule wear as a dominant mechanism of breakage
or attrition. For the case of abrasive wear of ceramics due to surface scratching
by loaded indentors, Evans and Wilshaw (60) determined a volumetric wear rate V of
V ?
c H1=2
P5=4l ?4:5?
where di is indentor diameter, P is applied load, l is wear displacement of the indentor,
and A is apparent area of contact of the indentor with the surface. Therefore,
wear rate depends inversely on fracture toughness. For the case of fragmentation,
Yuregir et al. (61) have shown that the fragmentation rate of organic and inorganic
crystals is given by
ru2a ?4:6?
where a is crystal length, r is crystal density, and u is impact velocity. Note that hardness
plays an opposite role for fragmentation than for wear, since it acts to concentrate
stress for fracture. Fragmentation rate is a stronger function of toughness as well.
Drawing on analogies with this work, the breakage rates by wear Bw and fragmentation
Bf for the case of fluid-bed granulation and drying processes should be of
the forms:
Bw ?
c H1=2
b ?U  Umf? ?4:7?
r?U  Umf ?2a ?4:8?
where d is granule diameter, d0 is primary particle diameter, (UUmf) is fluid-bed
excess gas velocity, and hb
sion rate on material properties for bars and granules undergoing a wear mechanism
of breakage, as governed by Eqs. 4.5 and 4.7.
5.1. An Engineering Approach to Granulation Processes
Future advances in our understanding of granulation phenomena rest heavily in
engineering process design. A change in granule size or voidage is akin to a change
60 Ennis
is bed height. Figure 53 illustrates the dependence of ero-
 2005 by Taylor & Francis Group, LLC
in chemical species, and so analogies exist between granulation growth kinetics and
chemical kinetics and the unit operations of size enlargement and chemical reaction.
considered for successful process design. Let us begin by considering a small-volume
either the molecular or the primary particle/single-granule scale. On the granule scale
of scrutiny, the design of chemical reactors and granulation processes differs conceptually
in that the former deals with chemical transformations, whereas the latter
deals primarily with physical transformations controlled by mechanical processing
 Here, the rate processes of granulation are controlled by a set
of key physicochemical interactions. These rate processes have been defined in the
proceeding sections, including wetting and nucleation, granule growth and consolidation,
and granule attrition and breakage.
We now consider a granule volume scale of scrutiny, returning to our smallvolume
element of material A of Figure 54. Within this small volume for the case
of chemical kinetics, we generally are concerned with the rate at which one or more
chemical species is converted into a product. This is generally dictated by a reaction
rate constant or kinetic constant, which is in turn a local function of temperature,
pressure, and the concentration of feed species, as was established from previous
physicochemical considerations. These local variables are in turn a function of overall
heat, mass, and momentum transfer of the vessel controlled by mixing and
heating/cooling. The chemical conversion occurring within a local volume element
may be integrated over the entire vessel to determine the chemical yield or extent
of conversion for the reactor vessel; the impact of mixing and heat transfer is generally
considered in this step at the process volume scale of scrutiny. In the case of a
granulation process, an identical mechanistic approach exists for design, where
 This approach was pioneered by Hans Rumpf (62) and others in the early 1960s at the Universita
?t Karlsruhe, leading to development of mechanical process technology within chemical
engineering in Germany, or Mechanische Verfahrenstechnik. This was key to the founding of
powder technology, an area in which the United States has traditionally lagged behind Europe
and Japan in the areas of chemical engineering and pharmaceutical processing (6163).
Figure 53 Bar wear rate and fluid-bed erosion rate as a function of granule material
properties. Kc
Theory of Granulation 61
These analogies are highlighted in Figure 54, where several scales of analysis must be
element of material A within a mixing process as shown in Fig. 55, and consider
is fracture toughness and H is hardness. (From Ref. 56.)
[Refs. 6264] .
 2005 by Taylor & Francis Group, LLC
chemical kinetics is replaced by granulation kinetics. The performance of a granulator
may be described by the extent of granulation of a species. Let (x1, x2, . . . , xn) represent
a list of attributes such as average granule size, porosity, strength, and any other
generic quality metric and associated variances. Alternatively, (x1, x2, . . . , xn) might
represent the concentrations or numbers of certain granule size or density classes, just
as in the case of chemical reactors. The proper design of a chemical reactor or a granulator
then relies on understanding and controlling the evolution (both time and spatial)
of the feed vector X to the desired product vector Y. Inevitably, the reactor or
granulator is contained within a larger plant scale process chain, or manufacturing circuit,
with overall plant performance being dictated by the interaction between individual
unit operations. At the plant scale of scrutiny, understanding interactions
between unit operations can be critical to plant performance and product quality.
These interactions are far more substantial with solids processing, than with liquid
gas processing. Ignoring these interactions often leads processing personnel to misdiagnose
sources of poor plant performance. Tableting is often affected by segregation or
poor mixing. Segregation becomes vital for preferential wetting and drug assay
Figure 54 Changes in state as applied to granulator kinetics and design.
62 Ennis
 2005 by Taylor & Francis Group, LLC
variation per size class. This is often effected by trace impurites in the production of
drug or excipients (1,2,64).
There are several important points worth noting with regard to this approach.
First, the engineering approach to the design of chemical reactors is well developed
and an integral part of traditional chemical engineering education [e.g. (65)]. At
present, only the most rudimentary elements of reaction kinetics have been applied
to granulator design (1,2,5). Much more is expected to be gleaned from this approach
over the coming decade.
Examples might include staged addition of ingredients, micronization of processes,
and tailored process designs based on specific formulation properties. Second,
an appreciation of this engineering approach is absolutely vital to properly scale-up
granulation processes for difficult formulations. Last, this perspective provides a
logical framework to approach and unravel complex processing problems, which
often involve several competing phenomena. Significant progress has been made
with this approach in crystallization (66) and grinding (67).
Figure 55 Granulation within a local volume element, as a subvolume of a process
granulator volume, which controls local size distribution.
Theory of Granulation 63
 2005 by Taylor & Francis Group, LLC
Many complexities arise when applying the results of the previous sections
detailing granulation mechanisms to granulation processing. The purpose of this section
is to summarize approaches to control these rate processes by placing them
within the context of actual granulation systems and granulator design. Additional
5.2. Scale: Granule Size and Primary Feed Particles
When considering a scale of scrutiny of the order of granules, we ask what controls
the rate processes, as presented in detail in the previous sections. This key step links
formulation on material variables to the process operating variables, and successful
granulator design hinges on this understanding. Two key local variables of the
volume element A include the local bed moisture and the local level of shear (both
shear rate and shear forces). These variables play an analogous role of species concentration
and temperature in controlling kinetics in chemical reaction. In the case
of chemical reaction, increased temperature or concentration of a feed species generally
increases reaction rate. For the case of granulation considered here, increases
in shear rate and moisture result in increased granule/powder collisions in the presence
of binding fluid, resulting in an increased frequency of successful growth events
and increases in granule growth rate. Increases in shear forces also increase the granule
consolidation rate, and aid growth for deformable formulations. In the limit of
very high shear (e.g., due to choppers), they promote wet and dry granule breakage,
or limit the growth at the least. Last, in the case of simultaneous drying, bed and gas
phase moisture and temperature control heat and mass transfer and the resulting
drying kinetics.
5.3. Scale: Granule Volume Element
Next, consider a scale of scrutiny of the level of a small bulk volume of granules, or
granule size distribution controlled by the local granulation rate processes as shown
with fine powder to form initial nuclei, either directly or through mechanical breakdown
of pooled overwetted regions. It is generally useful to consider the initial powder
phase and drop phases as independent feed phases to the granule phase. In
addition, the granule phase can be broken down into separate species, each species
corresponding to a particular granule mesh size cut. Nucleation therefore results
in a loss of powder and drop phases, and the birth of granules. Granules and initial
nuclei collide within this volume element with each other and with the surrounding
powder phase, resulting in both granule growth, and consolidation due to compaction
forces. Granule growth by coalescence results in the discrete birth of granules to
a new granule size class or species, as well as loss or death of granules from the
originating size classes. On the other hand, granule growth by layering and granule
consolidation result in a slow differential increase and decrease in granule size,
respectively. Granule breakage by fracture and attrition (or wear) act in a similar,
but opposite fashion to granule coalescence and layering, increasing the powder
phase and species of smaller granules. Last, this volume element of granules interacts
with the surrounding material of the bed, as granulated, powder, and drop
phases flow to and from the surrounding volume elements. The rate processes of
granulation and the flows or exchanges with the surrounding elements combine to
64 Ennis
volume element of material A in Figure 55. This volume element has a particular
pictorially in Figure 56. In the wetting and nucleation rate process, droplets interact
details of modeling and granulator design can be found in Chapters 16 and 21.
 2005 by Taylor & Francis Group, LLC
control the local granule size distribution and growth rate within this small-volume
As illustrated in Figure 56, conducting an inventory of all granules entering
and leaving a given size class n(x) by all possible granulation mechanisms leads to
a microlevel population balance over the volume element given by
@t ?
@xi ?naui? ? Ga ? Ba  Da ?5:1?
where n(x,t) is the instantaneous granule size distribution, which varies with time and
position. G, B, and D are growth, birth, and death rates due to granule coalescence
and granule fracture. The second left-hand side term reflects contributions to the
distribution from layering and wear, as well as interchanges of granules from the
surrounding volume elements. Nucleation rate would be considered a boundary
condition of Eq. 5.1, providing a source of initial granules. Eq. 5.1 governs the local
granule growth of volume element A.
Although actual processes would require specific examination, some general comments
are warranted. Beginning with nucleation, in the case of fast drop penetration
into fine powders, and for small spray flux, new granules will be formed of the order
of the drop size distribution, and contribute to those particular size cuts or granule
species. If spray is stopped at low moisture levels, one will obtain a bimodel distribution
of nuclei size superimposed on the original feed distribution. Very little growth
may occur for these low moisture levels. This should not be confused with induction
type growth, which is a result of low overall formulation deformability. In fact, the
moisture level of the nuclei themselves will be found to be high and nearly saturated.
Moisture, however, is locked up within these nuclei, surrounded by large amounts of
Figure 56 The population balance over a sieve class, over specific granule size class.
Theory of Granulation 65
Solutions to this population balance are described in greater detail in Chapter
21, as well as (5,46,52). Analytical solutions are only possible in the simplest of cases.
 2005 by Taylor & Francis Group, LLC
fine powder. Therefore, it is important not to confuse granule moisture, local moisture,
and the overall average peak bed moisture of the process; they are very much
not the same and are influenced by proper vessel design and operation. As moisture
levels increase, and the concentration of the ungranulated powder phase decreases,
the portion of the granule phase increases. As granules begin to interact more fully
due to decreased surrounding powder and greater chances of achieved wet granule
interaction, granule coalescence begins to occur. This in turn results in a decrease
in granule number, and a rapid often exponential increase in granule size, as previously
demonstrated. Coalescence generally leads to an initial widening of the granule
size distribution until the granule growth limit is reached. As larger granules
begin to exceed this growth limit, they can no longer coalesce with granules of similar
size. Their growth rate drops substantially as they can only continue to grow by coalescence
with fine granules, or by layering with any remaining fine powder. At this
point, the granule size distribution generally narrows with time. Note that this provides
a local description of growth, whereas the overall growth rate of the process
depends greatly on mixing described next, as controlled by process design.
5.4. Scale: Granulator Vessel
The local variables of moisture and shear level vary with volume element, or position
in the granulator, which leads to the kinetics of nucleation, growth, consolidation,
and breakage being dependent on position in the vessel, leading to a scale of scrutiny
and nucleation will be high at position D. Significant growth will occur at position B
due to increased shear forces and granule deformation, as well as increased contacting.
Significant breakage can occur at position C in the vicinity of choppers. Each of
these positions or volume elements will have theirown specific granule size distribution.
Solids mixing (2,68) impacts overall granulation in several ways. First, it
controls the local shear. Local shear rates and forces are functions of shear stress
transfer through the powder bed, which is in turn a function of mixer design and
bed bulk density, granule size distribution, and frictional properties. Local shear
rates determine granule collisional velocities. This first area is possibly one of the
least understood areas of powder processing, and requires additional research to
establish the connection between operating variables and local shear rates and
Second, solids mixing controls the interchange of moisture, powder phase, and
droplet phases among the local volume elements. Third, it controls the interchange
of the granulated phase. Within the context of reaction kinetics (65), one generally
considers extremes of mixing between well-mixed continuous and plug flow continuous,
or well-mixed batch processes. The impact of mixing on reaction kinetics is
well understood, and similar implications exist for the impact of mixing on granulation
growth kinetics. In particular, well-mixed continuous processes would be expected to
provide the widest granule size distribution (deep continuous fluidized beds are an
example), whereas plug flow or well-mixed batch processes should result in narrower
distributions.y In addition, it is possible to narrow the distribution further by
y All else being equal, plug flow continuous and batch well mixed processes should produce
identical size distributions. It is very difficult to achieve uniform mixing in practice, with
properly operating fluidized beds possibly coming the closest.
66 Ennis
of the vessel size. As shown in Figure 55, moisture levels, drop phase concentration,
forces. It is also a very important scale-up consideration, as discussed in Chapter 18.
 2005 by Taylor & Francis Group, LLC
purposely segregating the bed by granule size,z or staging the addition of ingredients,
though this is a less explored area of granulator design. Last, it should be possible to
predict the effects of dispersion, backmixing, and dead/stagnant zones on granule size
distribution, based on previous work regarding chemical reaction kinetics.
Equation 5.1 reflects the evolution of granule size distribution for a particular
volume element. When integrating this equation over the entire vessel, one is able to
predict the granule size distribution vs. time and position within the granulator. Last,
it is important to understand the complexities of scaling rate processes on a local
level to overall growth rate of the granulator. If such considerations are not made,
misleading conclusions with regard to granulation behavior may be drawn. Wide distributions
in moisture and shear level, as well as granule size, and how this interacts
with scale-up must be kept in mind when applying the detailed description of rate
processes discussed in the previous sections. With this phenomenological description
of granulation in place, we will now discuss controlling wetting, growth and consolidation,
and breakage in practice, as well as the implications for two of the more
common pharmaceutical granulation processes, namely fluid-bed and high-shear
mixer granulation.
5.5. Controlling Processing in Practice
mixer granulation. From a processing perspective, we begin with the uniformity of
the process in terms of solids mixing. Approaching a uniform state of mixing as previously
described will ensure equal moisture and shear levels and, therefore, uniform
granulation kinetics throughout the bed; on the other hand, poor mixing will lead to
differences in local kinetics. If not accounted for in design, these local differences will
lead to a wider distribution in granule size distribution and properties than is necessary,
and often in an unpredictable fashionparticularly with scale-up.
Increasing fluid-bed excess gas velocity (UUmf) will increase solids flux and
decrease circulation time. This can potentially narrow nuclei distribution for intermediate
drop penetration times. Growth rates will be minimally affected due to
increased contacting; however, the growth limit will decrease. There will be some
increase in granule consolidation, and potentially a large increase in attrition. Lastly,
initial drying kinetics will increase. Impeller speed in mixers will play a similar role in
increasing solids flux. However, initial growth rates and granule consolidation are
likely to increase substantially with an increase in impeller speed. The growth limit
will decrease, partly controlled by chopper speed.
Fluidized beds can be one of the most uniform processes in terms of mixing
and temperature. Powder frictional forces are overcome as drag forces of the fluidizing
gas support bed weight, and gas bubbles promote rapid and intensive mixing. In
the case of mixers, impeller speed in comparison to bed mass promotes mixing, with
choppers eliminating any gross maldistribution of moisture and over growth.
With regard to bed weight, forces in fluid beds, and therefore consolidation and
granule density, generally scale with bed height. As a gross rule of thumb, ideally, the
power input per unit mass should be maintained with mixer scale-up, related in part
z Pan granulation is a specific process promoting segregation by granule size. Since large granules
interact less with smaller granule size classes, layering can be promoted at the expense of
coalescence, thereby narrowing the granule size distribution.
Theory of Granulation 67

In summary, for the case of fluid-bed granulation, growth rate is largely controlled
by spray rate and distribution and consolidation rate by bed height and peak bedmoisture.
For the case of mixers, growth and consolidation are controlled by impeller and
chopper speed. From a formulation perspective, we now turn to each rate process.
5.6. Controlling Wetting in Practice
summarizes typical changes in material and operating variables which
improve wetting uniformity. Also listed are appropriate routes to achieve these
changes in a given variable through changes in either the formulation or the processing.
Improved wetting uniformity generally implies a tighter granule size distribution
and improved product quality. Eqs. 2.5, 2.9, and 2.13 provide basic trends of
the impact of material variables on wetting dynamics and extent, as described by
the dimensionless spray flux and drop penetration time.
Since drying occurs simultaneously with wetting, the effect of drying can substantially
modify the expected impact of a given process variable and this should not
be overlooked. In addition, simultaneous drying often implies that the dynamics of
wetting are far more important than the extent.
Adhesion tension should be maximized to increase the rate and extent of both
binder spreading and binder penetration. Maximizing adhesion tension is achieved
by minimizing contact angle and maximizing surface tension of the binding solution.
These two aspects work against one another as surfactant is added to a binding fluid,
and in general, there is an optimum surfactant concentration for the formulation (27).
Surfactant type influences adsorption and desorption kinetics at the three-phase contact
line. Inappropriate surfactants can lead to Marangoni interfacial stresses which
slow the dynamics of wetting (25). Additional variables which influence adhesion tension
include (1) impurity profile and particle habit/morphology typically controlled in
the particle formation stage such as crystallization, (2) temperature of granulation,
and (3) technique of grinding, which is an additional source of impurity as well.
Decreases in binder viscosity enhance the rate of both binder spreading and binder
penetration. The prime control over the viscosity of the binding solution is through
binderconcentration.Therefore, liquidloadinganddryingconditionsstrongly influence
binder viscosity. For processes without simultaneous drying, binder viscosity generally
decreases with increasing temperature. For processes with simultaneous drying, however,
the dominant observed effect is that lowering temperature lowers binder viscosity
and enhances wetting due to decreased rates of drying and increased liquid loading.
Changes in particle size distribution affect the pore distribution of the powder.
Large pores between particles enhance the rate of binder penetration, whereas they
decrease the final extent. In addition, the particle size distribution affects the ability
of the particles to pack within the drop as well as the final degree of saturation (69).
The drop distribution and spray rate of binder fluid have a major influence on
wetting. Generally, finer drops will enhance wetting as well as the distribution of binding
fluid. The more important question, however, is how large may the drops be or how
high a spray rate is possible. The answer depends on the wetting propensity of the feed.
If the liquid loading for a given spray rate exceeds the ability of the fluid to penetrate
and spread on the powder, maldistribution in binding fluid will develop in the bed. This
maldistribution increases with increasing spray rate, increasing drop size, and decreasing
spray area (due to, e.g., bringing the nozzle closer to the bed or switching to fewer
nozzles). Themaldistribution will lead to large granules on the one hand and fine ungranulated
powder on the other. In general, the width of the granule size distribution will
70 Ennis
Table 7
 2005 by Taylor & Francis Group, LLC
Table 7 Controlling Wetting in Granulation Processes
Typical changes in material or operating
variables which improve wetting uniformity
Appropriate routes to alter variable
through formulation changes
Appropriate routes to alter variable
through process changes
Increase adhesion tension Alter surfactant concentration or type to maximize
adhesion tension and minimize Marangoni effects
Control impurity levels in particle formation
Maximize surface tension Alter crystal habit in particle formation
Minimize contact angle Precoat powder with wettable monolayers,
e.g., coatings or steam
Minimize surface roughness in milling
Decrease binder viscosity Lower binder concentration Raise temperature for processes without
simultaneous drying Change binder
Decrease any diluents and polymers which act
as thickeners
Lower temperature for processes with
simultaneous drying since binder
concentration will decrease due to
increased liquid loading
Increase pore size to increase rate of
fluid penetration
Modify particle size distribution of feed ingredients Alter milling, classification, or formation
conditions of feed if appropriate to modify
particle size distribution Decrease pore size to increase extent of
fluid penetration
Improve spray distribution (Related to
dimensionless spray flux, given by ratio
of spray to solid fluxes)
Improve atomization by lowering binder fluid viscosity Increase wetted area of the bed per unit mass
per unit time by increasing the number of
spray nozzles, lowering spray rate,
increasing air pressure or flow rate of two
fluid nozzles
Increase solids mixing (Related to
dimensionless spray flux)
Improve powder flowability of feed Increase agitation intensity (e.g., impeller
speed, fluidization gas velocity, or rotation
Minimize moisture buildup and losses Avoid formulations which exhibit adhesive characteristics
with respect to process walls
Maintain spray nozzles to avoid caking and
nozzle drip; avoid spray entrainment in
process air streams, and spraying process
Theory of Granulation 71
Source: Refs. 1,2,5.
 2005 by Taylor & Francis Group, LLC
Table 8 Controlling Growth and Consolidation in Granulation Processes
Typical changes in material or operating
variables which maximize growth and
Appropriate routes to alter variable
through formulation changes
Approprziate routes to alter variable
through process changes
Rate of growth (low deformability):
Increase rate of nuclei formation Improve wetting properties (see Wetting subsection);
Increase binder distribution
Increase spray rate and number of drops
Increase collision frequency Increase mixer impeller or drum rotation
speed or fluid-bed gas velocity
Increase residence time Increase batch time or lower feed rate
Rate of growth (high deformability):
Decrease binder viscosity Decrease binder concentration or change binder; decrease
any diluents and polymers which act as thickeners
Decrease operating temperature for systems
with simultaneous drying; otherwise
increase temperature
Increase agitation intensity Increase mixer impeller or drum rotation
speed or fluid-bed gas velocity
Increase particle density
Increase rate of nuclei formation,
collision frequency and residence
time, as above for lowdeformability
72 Ennis
 2005 by Taylor & Francis Group, LLC
Extent of growth:
Increase binder viscosity Increase binder concentration, change binder, or add
diluents and polymers as thickeners
Increase operating temperature for systems
with simultaneous drying; otherwise
decrease temperature
Decrease agitation intensity Decrease mixer impeller or drum rotation
speed or fluid-bed gas velocity Decrease particle density
Increase liquid loading Extent observed to increase linearly with
Rate of consolidation:
Decrease binder viscosity As above for high deformability systems As above for high deformability systems; in
addition, increase compaction forces by
increasing bed weight, or altering mixer
impeller or fluid-bed distributor plate
Size is controlled in milling and particle
Increase agitation intensity
Increase particle density
Increase particle size
Particle size and friction strongly interact with binder
viscosity to control consolidation; feed particle size
may be increased and fine tail of distribution
Source: From Refs. 1,2,5.
Theory of Granulation 73
 2005 by Taylor & Francis Group, LLC
Table 9 Controlling Breakage in Granulation Processes
Typical changes in material or operating
variables which minimize breakage
Appropriate routes to alter variable through
formulation changes
Appropriate routes to alter variable
through process changes
Increase fracture toughness Increase binder concentration or change binder; bond
strength strongly influenced by formulation and
compatibility of binder with primary particles
Decrease binder viscosity to increase
agglomerate consolidation by altering
process temperatures (usually decrease for
systems with simultaneous drying)
Maximize overall bond strength
Minimize agglomerate voidage
Increase bed agitation intensity (e.g.,
increase impeller speed, increase bed
height) to increase agglomerate
Increase granulation residence time to
increase agglomerate consolidation, but
minimize drying time
Increase hardness to reduce wear Increase binder concentration or change binder; binder
plasticity strongly influenced by binder type
See above effects which decrease
agglomerate voidage Minimize binder plasticity
Minimize agglomerate voidage
Decrease hardness to reduce fragmentation Change binder; binder plasticity strongly influenced by
binder type
Reverse the above effects to increase
agglomerate voidage Maximize binder plasticity
Maximize agglomerate voidage Apply coating to alter surface hardness
74 Ennis
 2005 by Taylor & Francis Group, LLC
Decrease load to reduce wear Lower formulation density Decrease bed agitation and compaction
forces (e.g., mixer impeller speed, fluid-bed
height, bed weight, fluid-bed excess gas
Decrease contact displacement to reduce
Decrease contacting by lowering mixing and
collision frequency (e.g., mixer impeller
speed, fluid-bed excess gas velocity, drum
rotation speed)
Decrease impact velocity to reduce
Lower formulation density Decrease bed agitation intensity (e.g., mixer
impeller speed, fluid-bed excess gas
velocity, drum rotation speed)
Also strongly influenced by distributor plate
design in fluid beds, or impeller and
chopper design in mixers
Source: From Refs. 1,2,5.
Theory of Granulation 75
 2005 by Taylor & Francis Group, LLC
increase and generally the average size will decrease. Improved spray distribution can be
aided by increases in agitation intensity (e.g., mixer impeller or chopper speed, drum
rotation rate, or fluidization gas velocity) and by minimizing moisture losses due to
spray entrainment, dripping nozzles, or powder caking on process walls.
5.7. Controlling Growth and Consolidation in Practice
imize granule growth and consolidation. Also listed are appropriate routes to
achieve these changes in a given variable through changes in either the formulation
or the processing. Growth and consolidation of granules are strongly influenced by
rigid (especially fluid beds) and deformability (especially mixers) Stokes numbers.
Increasing St increases energy with respect to dissipation during deformation of
granules. Therefore, the rate of growth for deformable systems (e.g., deformable formulation
or high-shear mixing) and the rate of consolidation of granules generally
increases with increasing St. St may be increased by decreasing binder viscosity or
increasing agitation intensity. Changes in binder viscosity may be accomplished by
formulation changes (e.g., the type or concentration of binder) or by operating temperature
changes. In addition, simultaneous drying strongly influences the effective
binder concentration and viscosity. The maximum extent of growth increases with
decreasing St and increased liquid loading, as reflected by Eqs. 3.11. Increasing
particle size also increases the rate of consolidation, and this can be modified by
upstream milling or crystallization conditions.
5.8. Controlling Breakage in Practice
minimize breakage. Also listed are appropriate routes to achieve these changes in a
given variable through changes in either the formulation or the processing. Both
fracture toughness and hardness are strongly influenced by the compatibility of the
binder with the primary particles, as well as the elastic/plastic properties of the binder.
In addition, hardness and toughness increase with decreasing voidage and are influenced
by previous consolidation of the granules. While the direct effect of increasing
gas velocity and bed height is to increase breakage of dried granules, increases in these
variables may also act to increase consolidation of wet granules, lower voidage, and
therefore lower the final breakage rate. Granule structure also influences breakage rate,
e.g., a layered structure is less prone to breakage than a raspberry-shaped agglomerate.
However, it may be impossible to compensate for extremely low toughness by changes
in structure. Measurements of fracture properties help define expected breakage rates
for a product and aid product development of formulations.
This chapter is the result of many collaborative efforts. Support for initial granulation
research was provided by the International Fine Particle Research
Institute, G. Tardos and R. Pfeffer of The City College of the City University of
New York, and E.I Du Pont de Nemours & Company. Based on earlier versions
with A. Maraglou of Du Pont, the material was developed into a training course
by E&G Associates, and later refined in collaboration with J. Litster of the Univer-
76 Ennis
Table 8 summarizes typical changes in material and operating variables which max-
Table 9 summarizes typical changes in material and operating variables necessary to
 2005 by Taylor & Francis Group, LLC
sity of Queensland. All of these collaborations, as well as discussions with P. Mort,
A. Adetayo, J. Seville, S. Pratsinis, S. Iveson, K. Hapgood, J. Green, P. C. Kapur, T.
Schaefer, and H. Kristensen are acknowledged with great appreciation. Last, to the
countless course participants over the last decade goes a special thank you.
1. Ennis BJ, Litster J. Size enlargement and size reduction. In: Green D, Perry R, eds.
Perrys Chemical Engineers Handbook. 7th ed. New York: McGraw-Hill, 1994:
Section 21.
2. Ennis BJ, ed. Solids-solids processing. In: Green D, Perry R, eds. Perrys Chemical Engineers
Handbook. 8th ed. New York: McGraw Hill, 2005:Section 19.
3. Ennis BJ. On the Mechanics of Granulation. Ph.D. thesis, The City College of the City
University of New York, University Microfilms International, 1990.
4. Ennis BJ. Design & Optimization of Granulation Processes for Enhanced Product
Performance. Nashville, TN: E&G Associates, 19902004.
5. Litster J, Ennis BJ. The Science & Engineering of Granulation Processes. Dordrecht, The
Netherlands: Kluwer Academic, 2004.
6. Masters K. Spray Drying Handbook 3rd ed. New York: Wiley & Sons, 2002.
7. Masters K. Spray Drying in Practice. ApS, Denmark: SprayDryConsult International,
8. Ennis BJ, Mehos G, Wu S, Patel P. Characterizing the impact of flow aids on flowability
of pharmaceutical excipients by automated shear cell. AAPS Annual Meeting.
Baltimore: MD, 2004.
9. Brown RL, Richards. Principles of Powder Mechanics. Oxford: Pergamon Press, 1970.
10. Stanley-Wood N. Enlargement and Compaction of Particulate Solids. London: Butterworth
& Co. Ltd, 1983.
11. Ennis BJ. Powder Technol 1996; 88:203.
12. Turton R, Tardos G, Ennis BJ. Fluidized Bed Coating and Granulation. In: Yang WC,
Fluidization, Solids Handling & Processing, New Jersey: Noyes Publications, Westword,
13. Pietsch W. Size Enlargement by Agglomeration. Chichester: John Wiley & Sons Ltd.,
14. Sastry K, Fuerstenau D. In: Sastry, ed. Agglomeration 77. New York: AIME, 1977:381.
15. Kapur PC. Adv Chem Eng 1978; 10:55.
16. Kapur PC. Chem Eng Sci 1971; 26:1093.
17. Kapur PC, Fuerstenau DW. Ind Eng Chem Eng (Proc Des Dec) 1966; 5:5.
18. Kapur PC. Chem Eng Sci 1971; 26:1863.
19. Kapur PC. Chem Eng Sci 1971; 26:1093.
20. Rumpf H. The strength of granules and agglomerates. In: Knepper WA, ed. Agglomeration.
New York: Interscience, 1962: 379414.
21. Rumpf H. Particle adhesion. In: Sastry KVS, ed. Agglomeration 77, New York: AIME,
1997: 97129.
22. Augsburger L, Vuppala M. Chap. 2, In: Parikh D, Handbook of Pharmaceutical Granulation
Technology. 1st ed., New York: Marcel-Dekker, 1999.
23. Parfitt G, ed. Dispersion of Powders in Liquids. Amsterdam: Elsevier Applied Science
Publishers Ltd., 1986.
24. Hapgood K. Nucleation and Binder Dispersion in Wet Granulation, Ph.D. Thesis,
University of Queensland, 2000.
25. Pan R, Green J, Ennis BJ, Maderelli C. Dynamic Properties of Interfaces & Association
Structure. New York: American Oil Chemical Society Press, 1995.
26. Zisman WA. Contact angle, wettability, and adhesion. Adv Chem Ser 1964; 43:1.
27. Ayala R. Ph.D. Thesis, Chem. Eng., Carnegie Mellon University, 1985.
Theory of Granulation 77
 2005 by Taylor & Francis Group, LLC
28. Shaw R. Introduction to Colloid and Surface Chemistry. London: Butterworths & Co.
Ltd, 1983.
29. Fuerstaneau DW, Diao J, Williams MC. Coll Surf 1991; 60:127.
30. Vargha-Butler EI. In: Botsaris GD, Glazman YM, eds. Interfacial Phenomena in Coal
Technology. Chap. 2. New York: Marcel-Dekker, 1989.
31. Aulton ME, Banks M. Proceedings of Powder Technology in Pharmacy Conference:
Powder Advisory Centre, Basel, Switzerland, 1979.
32. Aulton ME et al. J Pharm Pharmacol 1977; 29:59.
33. Litster JD, Hapgood KP, Kamineni SK, Hsu T, Sims A, Roberts M, Michaels J. Powder
Technol 2001; 124:272.
34. Ouchiyama N, Tanaka T. I and EC. Proc Des Dev 1982; 21:29.
35. Johnson KL. Contact Mechanics. Cambridge University Press, 1985.
36. Ritala M, Holm P, Schaefer T, Kristensen HG. Drug Dev Ind Pharm 1988; 14(8):1041.
37. Holm P, Schaefer T, Kristensen HG. Powder Technol 1985; 43:213.
38. Adams M, Thornton C, Lian G. Agglomeration & size enlargement. In: Ennis BJ, ed.
Proceedings of the First International Particle Technology Forum. Vol. 1. Denver,
CO: AIChE, 1994:155286.
39. Ennis BJ, Li J, Tardos G, Pfeffer R. Chem Eng Sci 1990; 45(10):3071.
40. Mazzone D, Tardos G, Pfeffer R. J Coll Interface Sci 1986; 113:544.
41. Iveson SM, Breathe JA, Page NW. Powder Technol 2002; 127:149.
42. Holm P, Schaefer T, Kristensen HG. Part V and VI. Powder Techol 1985; 43:213233.
43. Ennis B, Tardos G, Pfeffer R. Powder Technol 1991; 65:257.
44. Liu LX, Litster JD, Iveson SM, Ennis BJ. AIChE J 2000; 46(3):529.
45. Bird RB, Armstrong R, Hassager O. Dynamics of Polymeric Liquids. Vol. 1. New York:
John Wiley & Sons, Inc., 1977.
46. Adetayo AA, Lister JD, Pratsinis SE, Ennis BJ. Powder Technol 1995; 82:37.
47. Schaefer T, Holm P, Kristensen HG. Drug Dev Ind Pharm 1990; 16(8):1249.
48. Schaefer T, Holm P, Kristensen HG. Pharm Ind 1990; 52(9):1147.
49. Tardos G, Khan MI. AIChE Annual Meeting, Miami, 1995.
50. Tardos GI, Khan MI, Mort PR. Powder Technol 1997; 94:245.
51. Iveson SM, Wauters PAL, Forrest S, Litster JD, Meester GMH, Scarlett B, Ennis B.
Powder Technol 2001; 117:83.
52. Adetayo AA, Ennis B. AIChE J 1997; 43(4):927.
53. Iveson SM, Litster JD, Ennis BJ. Powder Techol 1996; 88:15.
54. Lawn B. Fracture of Brittle Solids. 2nd ed. Cambridge University Press, 1975.
55. Benbow JJ, Bridgwater J. Powder Technol 1987; 49:97.
56. Ennis BJ, Sunshine G. Tribology International 1993; 26:319.
57. Irwin J Appl Mech 1957; 24:361.
58. Griffith AA. Philos Trans T Soc 1920; A221:163.
59. Johnsson L, Ennis BJ. In: Ennis BJ, ed. Proceedings of the First International Particle
Technology Forum. Vol. 2. Denver, CO: AIChE, 1994:178.
60. Evans AG, Wilshaw TR. Acta Metall 1976; 24:939.
61. Yuregir KR, Ghadiri M, Clift R. Chem Eng Sci 1987; 42:843.
62. Rumpf H. Translated FA Bulletin. Particle Technology. New York: Chapman & Hall, 1990.
63. Ennis BJ, Green J, Davies R. The US Legacy of Neglect. Chem Eng Prog 1994; 90(4):32.
64. Ennis BJ, Green J. Visualizing Pharmaceutical Manufacturing as an Integrated Series of
Particle Processes. New York: Interphex, 1995.
65. Levenspiel O. Chemical Reaction Engineering. 2nd ed. New York: Wiley, 1972.
66. Randolph AD, Larson MA. Theory of Particulate Processes. 2nd ed. San Diego: Academic
Press Inc., 1991.
67. Prasher CL. Crushing and Grinding Process Handbook. New York: Wiley, 1987.
68. Weineko? tter R, Gericke H. Mixing of Solids. Dordrecht, The Netherlands: Kluwer
Academic, 2000.
69. Waldie B. Chem Eng Sci 1991; 46:2781.
78 Ennis
 2005 by Taylor & Francis Group, LLC
Drug Substance and Excipient
L. W. Chan and P. W. S. Heng
National University of Singapore, Singapore
Characterization of drug substances and excipients is a very important step at the
preformulation phase of product development. Although testing will involve additional
time and cost, failure to carry out the appropriate characterization tests can
be even more costly to manufacturers if the product made is not within specifi-
cations. Preformulation characterization of raw materials creates a body of information
which is very useful in the development of products. The lack of such
information leaves the formulator with little leeway for remediation action when
a problem arises from the production process or from the quality of the finished
product. It is important to eliminate the possible influences of the raw material characteristics
before venturing into investigation of processing variables. The knowledge
derived from the characterization of raw materials can also serve to enable better
specifications to be drawn up for procuring materials with the aim of either reducing
cost or improving product quality. In addition, a review of material characterization
results can provide an excellent database for the assessment of suppliers who can
provide materials of consistent quality.
Materials from reputable companies may be supplied with detailed speci-
fications and their methods of determination may be obtained, if requested. The
information on specifications such as purity or content is very often available.
Nevertheless, it is prudent to confirm such information. The information provided
by different suppliers may vary. The type of tests carried out or the techniques used
for the characterization of a particular physical property, for example, the particle
size distribution, may be different. Comparison of materials from different suppliers
can therefore be difficult. Sometimes, the analytical result supplied by the manufacturer
is given as falling within a certain range and this gives virtually no information
about batch to batch variation of the material.
It is therefore important to have a system of in-house characterization of raw
materials alongside the stability and functional tests for the finished product. Whenever
possible, tests carried out should yield quantitative results rather than a pass/
fail or present/absent assessment. Retrospective studies of the finished product test
 2005 by Taylor & Francis Group, LLC
results together with well-documented production process validation and characterization
tests of raw materials can provide directions for refinement of the production
process and improve the specifications for raw materials.
The method of material characterization varies considerably as it depends on
the nature and form of material used as well as the process involved in the conversion
of the raw materials to the finished products. The desirable characterization test
for each material depends on the material itself and the likely use or influence of
a particular material property on the process or product. For instance, detailed
information of the particle size distribution of a drug material may be less important
when the end product is a solution, more important when the material is to be granulated
and very critical when preparing an inhalant. As unnecessary testing can be
translated into additional cost, careful consideration of the type and number of tests
to be carried out on the raw materials should be weighed against the usefulness of
tests to give information on the identity and quality of raw materials, their effects
during processing and manufacturing and the functionality and esthetics of the
finished products.
The task of building up a body of information on materials is indeed difficult
given the wide spectrum of drugs and excipients used in pharmaceutical granulation.
Many compendial tests are concerned with the chemical aspects of testing and rarely
address the physical characterization of excipient materials. The physical aspects of
raw materials are more likely than their chemical properties to exert a greater
influence on the granulation process as well as the quality and functionality of the
finished products. The ability to define excipients using the correct functionality tests
would undoubtedly benefit the formulator greatly as better-defined excipients could
help to eliminate many processing problems. It is obvious that some material
characterization tests such as determination of identity and purity are important
but discussion on them is not included as they are better dealt separately.
2.1. Particle Shape
Particle shape is an important parameter which can have a significant effect on the
bulk properties of a powder. It is well known that spherical particles flow better,
pack better, and have a lower surface to volume ratio. Despite the well-recognized
importance of particle shape, the method of shape determination has not been
clearly defined owing to the complexity and variability of the three-dimensional
particles. In general, shape measurement methods are only able to define accurately
the shape if the shape of constituent particles can be correctly predicted based on a
two-dimensional model.
Shape of particles may be assessed descriptively by terms such as spherical,
elongated, acicular, angular, or a host of other terms. Although these are descriptive
terms, if accurately used, they can convey a general idea of the particle shape. However,
they reveal little about the degree to which the particles take upon a particular
shape. Without a comparable quantitative measure, it may be difficult to assess the
effects of particle shape on a process or product.
From the linear dimensional description of breadth, length, and height, some
shape data can be derived. Breadth is usually defined as the minimum distance
between two parallel lines bracketing the particle while length is the maximum
distance between two parallel lines enclosing the particle and is perpendicular to
80 Chan and Heng
 2005 by Taylor & Francis Group, LLC
the breadth. The height is the thickness of the particle resting in its most stable
orientation. The measurement of height is generally difficult for small particles as
they are usually viewed through a microscope.
Direct microscopic measurement of particle dimensions is very tedious. The
particle dimension is obtained using a linear eyepiece graticule. A camera lucida
attachment may be used to trace out the outline of particles onto a paper. The perimeter
of particle tracing may be obtained by using either a string or a planimeter.
Projected area can be obtained by using paper with grids or by cutting out tracing
of particle and weighing. By determining the area per unit weight of the paper,
weight of tracing may be converted into an area measurement. The popularity of
this laborious manual mode of dimensional measurement has declined significantly
with the introduction of image analyzers. When a video camera is attached to a
microscope, the image obtained on a high-resolution monitor may be digitized
and analyzed using a computer as shown in Figure 1. Information on the breadth,
length, perimeter, and area of particles can be determined very rapidly. The degree
of accuracy of such measurements depends on the clarity of image available and
the separation of particles from one another.
There are a large number of transformation models which can be used to
analyze the image dimensional parameters in the determination of particle shape.
Some methods require additional information of particle surface area, volume or
thickness to give better estimates of shape but for most purposes, a simple approach
is often preferred. The common treatment of data, namely, breadth, length, perimeter,
and area to reflect the particle shape (1,2) is given as follows:
Aspect or elongation ratio ?
Breadth ?2:1?
Figure 1 Image analysis system.
Drug Substance and Excipient Characterization 81
 2005 by Taylor & Francis Group, LLC
Bulkiness factor ?
?Length ? Breadth? ?2:2?
Formfactor ?
?Perimeter?2 ?2:3?
The elongation ratio is very useful for assessing deviation from a spherical
shape to an elongated form. The formfactor, on the hand, gives a measure of
sphericity and a perfect sphere has a formfactor value of unity. The bulkiness factor
gives an indication of solidity and large indentations on the particles give rise to low
values (3).
For many users of image analyzers, a wider array of mathematical models
is usually available on the software. The problem that confronts many users is
not the mathematical treatment of image analysis data but rather the selection of an
appropriate transformation model. Procedures for Fourier shape analysis and fractal
measurement (47) are also available to allow further treatment of image analysis
data. Indirect methods using two or more measurement techniques like sedimentation
and sieving have also been used for particle shape determination (8). Vendors
for image analysis systems include Data Translations, Malboro, MA; Edax,
Mahwah, NJ; and Synoptics, Cambridge, U.K.
2.2. Particle Size
There is much literature on the measurement of particle size and size distribution.
With the abundance of information on particle sizing and sizing methods, a less
experienced operator can find it daunting as to which method is the best. Upon
completion of a measurement, the task of data interpretation and validation of
the accuracy and reliability of measurement can be difficult. Often, standards are
used for comparison and the information obtained from these standards are used
for calibration of the size measuring system. Standards used are very often ideal
particles with a high degree of sphericity and a narrow size distribution. The real
samples for size analysis are often not spherical and can have a considerably wide
size distribution. It is therefore important to regard all results with some suspicion
and in a comparative perspective, which depends on the standard calibrator used
and the method of measurement. Proper and stringent development of the method
for size measurement should be carried out to ensure reliability, reproducibility,
and sensitivity of measurement method.
The present discussion on the particle sizing methods does not attempt to
establish yet another theory and practice guide on particle sizing but to review the
popular methods of particle sizing, their problems and usefulness in providing
information valuable to a formulator. Little attempt will be made to explain the theories
of various sizing methods as they are dealt with in many other comprehensive
publications on this subject (810).
2.2.1. Microscopy
Microscopy is a very old technique capable of sizing fine powders accurately. It is
nowadays regarded as an unpopular method to use due mainly to the tedium in measurement
and the availability of more sophisticated and easy-to-use particle sizers.
82 Chan and Heng
 2005 by Taylor & Francis Group, LLC
Nevertheless, microscopy still presents as the most direct method of particle size
analysis. However, as a particle sizer, a minimum of at least 625 particles must be
measured for computation of reasonable size statistics (11). In a particle sizing
laboratory, the value of the microscope is usually not as a primary equipment for
particle sizing but rather for the preliminary examination of the size distribution
of a powder sample prior to the use of a less direct particle size analyzer. The estimation
of particle size characteristics obtainable from the microscope would certainly
provide a rough guide for the setting of an automated particle sizer to the correct
sizing range or checking if the size analysis data obtained are within an expected
Unlike other particle sizers, the microscope gives the operator a clear visual
outline of each individual particle being measured. In addition, the material may
be presented in its normal form, dry or in a liquid medium. Particles that are aggregated
can be identified and will not be measured as a single entity. Problems associated
with microscopic sizing lie not only in the tedious nature of the technique
but also in the slide sample preparation, optical resolution, and operator bias.
Sample preparation for microscopy must ensure a representative distribution
of particles on the slide to be used for sizing. When powder material is to be made
as a suspension, there should be neither loss of smaller particles through dissolution
nor loss of large particles due to sedimentation prior to transferring onto a slide. It is
best to wet a sample on the slide itself. Presentation of the area of slide specimen for
sizing must be carried out in a systematic and orderly manner by moving the slide to
view different areas of the sample. There is an intrinsic orientation problem as particles
when presented always find their most stable orientation. Microscopy involves
measuring a particle from only the top view and measurement is two dimensional.
Thus, flakes or discoids will tend to be oversized. Decision on the dimension(s) to
measure needs to be made should a linear eyepiece graticule be used. An area graticule
may also be employed for projected area determination. Orientation problems
may be overcome by dispersing powders onto adhesive before viewing or setting particles
in plastic or wax, which is then sectioned for viewing. These techniques will
definitely make an already tedious technique even more so.
Clarity is always a problem as sizes of particles approach the lower limits of
light microscope optics, which is about 1 mm. The quality of lenses may also play
a part. There is often a tendency to oversize slightly due to fringe effects around
particles. Various microscope accessories and lighting techniques can improve the
viewed image resolution to a varying extent. In some cases, the use of dyes can help
improve the contrast between particle and background.
An operator with good technique is necessary in order to obtain reliable results
with minimal systematic errors. Operator technique can influence accuracy of microscopic
determination of particle size. There is a natural tendency for an operator
to pay greater attention to large particles as he is less likely to miss them. Not only
do small particles tend to be missed, they also pose problems by being hidden by
the larger particles or appearing as clumps. However, in volumetric terms, small
particles have a lesser influence.
Automation of the microscopy technique began with an image shearing
technique (12) and progressed to image projection and measurement and, finally,
computer image digitization and image analysis (13). The preceding discussion on
particle shape has covered the use of image analysis. Information for particle size
and size distribution is easily obtained from the set of data for deriving shape factor
of the particles. An extension of light microscopy is scanning electron microscopy.
Drug Substance and Excipient Characterization 83
 2005 by Taylor & Francis Group, LLC
Measurements may be made directly from the video image or via photomicrographs.
Attached image analysis systems may also be utilized for sizing. There have also been
improved methods to incorporate image processing systems to microscopic methods.
By taking images at varying focal distances, a composite image with highly
sharpened edges can be digitally produced.
2.2.2. Sedimentation
Sedimentation technique for particle sizing and classification has been in use for
a long time. In recent years, however, the success of light scattering particle sizers
has effectively eroded away the market for sedimentation-based particle sizers.
Sedimentation technique for sizing is based on the settling of particles under gravity
described by Stokes law. For a particle of diameter, d, and density, r1, under the
force of gravity, g, in a fluid of viscosity, Z, and density, r2, at its terminal velocity,
n, the accelerating force due to gravity is balanced by the viscous drag and
n ?
d2g?r1  r2?
18Z ?2:4?
The Andreasen pipette introduced in the 1920s is perhaps the most popular
manual apparatus for sampling from a sedimenting suspension. Determination of
the change in density of the sampled particle suspension with time enables the calculation
of size distribution of the particles. As Stokes law applies only to spherical
particles, the nonspherical particles give a mean diameter referred to as Stokes
equivalent diameter. The size range measurable by this method is from 2 to
60 mm (8). The upper limit depends on the viscosity of liquid used while the lower
limit is due to the failure of very small particles to settle as these particles are kept
suspended by Brownian motion.
An improvement of the Andreasen pipette method is to use a pan attached to a
sensitive balance which records the changes in weight of the pan as an increasing
amount of suspending particles settle on it. Later, sedimentation techniques using
light extinction by changes in turbidity of the suspension and x-ray were introduced
for more sensitive and rapid measurements.
The introduction of centrifugal sedimentation makes the technique capable of
determining the distribution of particles below 5 mm. The lower limit depends on the
centrifugal velocity. The time of analysis is reduced drastically and multiple samples
in cells can be analyzed simultaneously.
2.2.3. Sieving
Sieving is probably the oldest method of sizing, used initially for particle classification
rather than size analysis. The introduction of high-quality standardized woven-wire
sieves in a ffiffiffi 2 p progression, starting from 75 mm has helped to establish sieving as a
widely used particle sizing method, especially for the larger particles. Conceptually,
particle sizing by sieving is easily understood as the different meshes classify the particles
to different weight-based size fractions giving rise to the frequency distribution.
The process of sieving involves a nest of sieves, usually five to eight, arranged
from the largest apertures to the finest followed by a receiving pan. The sieves with a
lid and receiving pan are then placed on a sieve shaker, which may gyrate, oscillate,
or vibrate the sieves. Most commonly used shakers are vibrators. It is important to
determine the time required for completion of sieving. This is commonly taken as the
time when there is no further change in the weight of material retained on each sieve
84 Chan and Heng
 2005 by Taylor & Francis Group, LLC
after additional sieving time is given. The load for sieve analysis should be sufficient
in order that the amount present after sieving on each mesh can be accurately
In size analysis using sieves, it must be borne in mind that the aperture size of a
given sieve is not an absolute cutoff value for the size of particles permitted to pass
through. The permitted aperture tolerance is much wider and following the maximum
tolerance allowed by the British Standard 410 (1976), particles from 55 to
95 mm may go through or be arrested by a 75 mm aperture size sieve. Particles tend
to pass through based on their narrower cross-sectional area. An elongated particle,
such as a rod-shaped particle, will pass through an aperture just larger than its minimum
cross-sectional profile. Woven sieves can be made down to about 30 mm. Finer
sieves of much higher precision from 100 mm down to a few micrometers can be made
using etching methods and these sieves are often referred to as electroformed micromesh
sieves. Sieves with bases that have accurately drilled or punched circular holes
instead of woven mesh are available for sizing larger particles, generally about
500 mm and larger.
Size analysis by sieving is a relatively slow process and there may be problems
with dust pollution. For many drugs, the safety of the operator needs to be considered.
In addition, wire mesh stretches with repeated use. Apertures may be blanked
by improper or inadequate washing. For fine powders, aggregation of powders due
to cohesive or electrostatic charges can give inaccurate results. Inadequate sieving
time will also produce unreliable data. It is also important to ensure that the sieving
process itself does not bring about size reduction.
As sieving fine powder in a dry state may be a problem due to the cohesive nature
of such a powder and a long sieving time is required, wet sieving by suspending
the powder in a suitable liquid can improve sieving efficiency. However, the procedure
becomes more tedious. It now requires the additional drying of size fractions.
It is recommended that sieving starts with the finest mesh to remove the fines with
a volume of liquid, followed by classifying the powder retained with the largest to
the smallest mesh.
Air-jet sieving is a much more popular method for sizing fine powders below
75 mm than wet sieving. It involves the use of a vacuum pump to remove air from
the underside of a sieve. Air current is also supplied from the underside of the sieve
through a rotating arm of jets, which helps to unclog the mesh. A collecting cyclone
may be attached in the vacuum line to collect the fines. In-line filters may also be used
to collect the fines. Air-jet sieving is usually used as a one-mesh sieving. For information
of size distribution, composite size distribution may be obtained from separate
air-jet sieving operations using different meshes for samples of the same powder.
A common point of discontent with size analysis using sieves is that the process
requires quite a bit of preparatory work, weighing, and subsequent washing. Yet,
a typical analysis would yield only seven to eight points on the size distribution
and this may not be sufficiently discriminating. Nevertheless, sieving is a straightforward
and robust technique suitable for a wide variety of fine to very coarse powders.
Material properties such as density, optical property, water solubility, or conductivity
are not required for computation of the particle size.
2.2.4. Electrical Sensing
The electrical sensing zone principle, which is more commonly known as the Coulter
principle, is based on a simple electrical property that the electrical resistance
Drug Substance and Excipient Characterization 85
 2005 by Taylor & Francis Group, LLC
between two compartments containing an electrolyte and connected by an aperture
is proportional to the electrical conducting area of the aperture. Figure 2 illustrates
the basic Coulter principle. By drawing electrolyte from one compartment to the
other, particles streaming through will decrease the conducting area of the aperture.
By fast time-based tracing of the resistance, resistance pulses coinciding with the passage
of particles through the aperture will be obtained. The amplitude of the pulse is
proportional to the volume of the particle. With a pulse analyzer, the Coulter counter
can analyze a large number of particles within a short time.
Particle size range detectable depends on the aperture tube used. Each aperture
tube is effective over a size range of about 240% of its nominal diameter. Apertures
of sizes from 15 to 4000 mm are available. Before use, it is necessary to calibrate the
equipment with a standard latex containing monosize spherical particles of mean size
within 520% of the aperture diameter.
When working with small apertures, aperture blockage may be a problem.
With large apertures for sizing large particles, settling of the large particles may give
rise to sizing errors. It is important that the material for sizing is nonconductive and
nonporous. For porous particles, sizing values obtained may be much smaller than
those derived by visual inspection. Prior to addition of the powder for sizing, it is
necessary to ensure a low background count. When dispersed in the electrolyte,
the powder particles must not be flocculated and do not dissolve in the electrolyte
solution. Care must be taken to ensure proper dispersion of powder. The concentration
of particles must be within the range acceptable. Higher concentration will result
in higher errors due to coincidence while low concentration will necessitate a longer
counting time.
2.2.5. Light Scattering
There has been a tremendous growth in the application of light scattering technique
for particle sizing in recent years and light scattering particle sizers have taken a
Figure 2 The Coulter principle.
86 Chan and Heng
 2005 by Taylor & Francis Group, LLC
lion share of the market for particle sizers. The large number of manufacturers of
instruments using light scattering technique has also a hand in convincing many
laboratories of their need for such a sizer. It is however unfair to ascribe the recent
popularity of light scattering sizing to just its marketing. Light scattering particle
sizers for both wet and dry samples are indeed much easier to use and are highly
efficient. The measurement time is short and the method is able to produce detailed
and reproducible particle size information.
Without dwelling on the theories of light scattering, particle sizers using this
principle can be roughly divided into three groups, two for determining particle size
down to about 1 mm using light obscuration and laser diffraction and the third, for
submicron particle size using photon correlation spectroscopy. A schematic diagram
of the various techniques of light scattering is shown in Figure 3.
Light obscuration or blockage technique involves measuring particles singly.
The passage of a particle across the light beam produces a reduction in the amount
of transmitted light which is detected by a sensor directly opposite the incident light.
The pulses are then classified, giving the frequency distribution. The degree of light
diffraction, opacity, and orientation of the particle as it passes the light beam can
affect the extent of light blockage and may affect measurement accuracy.
As a small particle passes through a beam of light in a laser diffraction sizer, it
will scatter light, which will be directed onto a diode array detector directly opposite
the incident light. The detector has a series of photodiodes arranged outward from
a central photodiode detector. Since the intensity of the light scattered decreases as
the scatter angle increases, photodiode elements are generally larger as they are
further from the center. Calculations for particle size and size distribution involve
rather complex mathematics. Simply put, sizing of a particle is based on the angle
of diffracted light, with small particles diffracting at wider angles than larger
particles. Thus, from the light scattering pattern, information on the size distribution
of the particles can be obtained through a series of complex calculations.
Sample presentation for light obscuration sizer involves the dispersion of the
particles in a liquid medium like in the case of the electrical sensing sizer. The main
difference is that light obscuration sizer may operate in the absence of an electrolyte.
Figure 3 Schematic diagram for light-scattering particle sizing.
Drug Substance and Excipient Characterization 87
 2005 by Taylor & Francis Group, LLC
For laser diffraction, sizing may be carried out on both particles dispersed in liquid
as well as dry powders using dry powder feeders. Laser diffraction is also very useful
for sizing aerosol particles and sprays. Increasingly, there has been an attempt to use
information from diffraction to elucidate shape parameters.
The possibility of measuring the particle size and size distribution of powders
in their dry state with an acceptable level of accuracy and in a very short timespan
has attracted many laboratories. For many users, the laser diffraction sizer is an effi-
cient equipment for producing detailed and accurate size distribution information of
the powder particles. However, the sizing of some powders using a dry powder feeder
with the laser diffraction sizer can be fraught with problems. Possible causes of nonreproducible
results are the poor control of ambient humidity, cohesive nature of the
powder, powder particles that fragmentate easily, large size span of the particles,
variable rate of introduction of particles during measurement period, possible segregation
of powders during introduction, and stray powder particles depositing on the
lens. The sizing of a powder composed of very large and very small particles can be
problematic as a portion of the small particles will adhere onto their larger counterparts.
The complete dislodging of the smaller particles can be an extremely difficult
task and with different feed rates of particles for sizing, different results may be
obtained. In addition, humidity of the atomizing air can also affect sizing results.
2.2.6. Photon Correlation Spectroscopy
The measurement of submicrometer particles had been difficult until the introduction
of photon correlation spectroscopy for particle sizing. This technique enables
particles from a few nanometers to a few micrometers to be measured. The measurement
principle involves the determination of fluctuations in the scattered light intensuspended
particles in the liquid. Large particles will diffuse more slowly than smaller
ones and therefore, the rate of decay in intensity of scattered light at a particular
measuring point will depend on the size of the particle. The particle size distribution
is computed using complex calculations and approximations from the intensities of
scattered light (normally at 90 to the incident beam) and their rate of decay. Multiple
angle measurements are sometimes applied to improve the quality of the size
parameters obtainable.
Although the photon correlation sizer represents a very interesting instrument
for the sizing technologist, powders for granulation do not generally fall into the
submicrometer range. The few possible uses of this instrument are in the evaluation
of polymeric materials employed for binding and coating. Suppliers for both laser
diffraction and scattering particle sizers include Beckman Coulter, Fullerton, CA,
USA; Hach Ultra Analytics, Grant Pass, OR, USA; Malvern Instruments, Worcs,
U.K.; and Horiba, Kyoto, Japan.
2.3. Particle Surface Area
Surface area measurement is usually carried out by either gas permeability or
adsorption. The technique of gas permeability depends on measuring the resistance
to gas flow through a packed bed of particles. It is important that packing of the bed
is uniform and from the volumetric flow rate of the gas through and pressure drop
across the bed, the specific surface area of the powder can be calculated. The measurement
of specific surface area by gas permeability does not take into account
88 Chan and Heng
sity at an angle (Fig. 3). These fluctuations are the result of Brownian motion of
 2005 by Taylor & Francis Group, LLC
the very small pores or fissures since the flow of gas is not hindered as it passes over
them. More accurate measurements can be made by measuring gas flow under
reduced pressure but still, the accuracy cannot match that obtainable by gas adsorption
if the total area to be determined includes those of the fine pores. Although gas
permeability gives a lower specific area for a powder compared with gas adsorption,
the value obtained is sometimes more useful in explaining factors like lubricity and
flow, which would not involve the micropores present in the particles.
Gas adsorption is carried out by placing a powder sample in a chamber and
evacuating the air within. The latter process is commonly referred to as degassing.
Upon achieving a very high vacuum, known volumes of an adsorbing gas are introduced.
From the knowledge of pressures and temperatures before and after introduction
of the adsorbing gas, usually nitrogen, calculations of total sample surface area
can be made. The surface area determination by gas adsorption is based on a simple
principle. From Avogadros number, a known volume of air at a certain temperature
and pressure contains a determinable number of molecules. When varying volumes
of gas are introduced to a degassed sample, the small pressure changes in the chamber
are recorded and using a calculation technique known as the BET method, the
initial amount of gas molecules which are adsorbed onto the surface forming
a monolayer can be calculated. Thus, the surface area covered by the gas molecules
can be determined by multiplying the number of molecules needed with the surface
area occupied per molecule. Samples are usually cooled to a low temperature using
liquid nitrogen. There are variations in the technique for gas adsorption by different
instrument manufacturers (14).
In addition to determining the specific surface area, pores below 50nm
may also be characterized by gas adsorption. Distribution of larger pores, 0.003
0.004 mm, can be determined by mercury intrusion porosimetry technique, where
the volume of mercury intruded under pressure represents the volume of pores whose
entrant diameter can be calculated from the applied pressure (15). The main suppliers
of equipment for surface and porosimetry include Quantachrome, Boynton, FL,
USA, and Micromeritics, Norcross, GA, USA.
The solubilities of drugs and excipients are an important physicochemical property
as they affect the bioavailability of the drug, the rate of drug release into the dissolution
medium, and consequently, the therapeutic efficacy of the pharmaceutical product.
It must be borne in mind that a drug must first be in solution in order to be
absorbed into the blood circulation. If the solubility of the drug is less than desirable,
steps must be taken to improve its solubility or to use another more soluble drug
form. Excipients which are poorly soluble in water might retard the release of drug
into the dissolution medium. Hence, the determination of drug and excipient solubilities
constitutes an important aspect of formulation study.
The solubility of a material is usually determined by the equilibrium solubility
method, which employs a saturated solution of the material. The saturated solution
is obtained by stirring an excess of the material in the solvent for a prolonged period
of time at a constant temperature until equilibrium is attained. As a guide, stirring
the mixture overnight is usually adequate for achieving equilibrium. The saturated
solution can also be obtained by warming the solvent with an excess of the material
and allowing the mixture to cool to the required temperature. This, however, may
Drug Substance and Excipient Characterization 89
 2005 by Taylor & Francis Group, LLC
produce a supersaturated solution for some materials and therefore, heating should
be applied with caution. A portion of the saturated solution obtained by either
method is then removed with the aid of a syringe through a membrane filter at different
time intervals. The determination is completed only if at least two successive
samples have the same results, indicating that equilibrium is attained. The final value
thus obtained is the solubility of the material. The material present in the sample of
saturated solution may be assayed by a variety of methods such as UV spectrophotometry,
electrical conductivity measurement, gravimetric or volumetric analysis,
and chromatographic methods.
The following precautions should be observed in order to obtain accurate and
reproducible solubility values:
 The material and solvent must be pure
 The temperature must be properly controlled
 It is essential that some undissolved material is present in the solution to
ensure that the solution obtained is saturated
 The saturated solution used for assay must be free from undissolved
 The method of assay must be reliable
A good understanding of the factors affecting solubilities of materials is pertinent
in explaining the changes in solubility under different conditions. These factors
can therefore be employed to improve the solubility and bioavailability of various
3.1. Nature of Solvent
Some materials dissolve very readily in a solvent while others dissolve sparingly. The
solubility of the material in a given solvent depends on the ability of the solvent to
overcome the forces that bind the atoms or molecules of the material. Studies have
shown a definite correlation between the solubility and the molecular structures of
the material and solvent. It is noted that the greater the similarity in molecular structure,
the higher would be the solubility of the material in the solvent. As a rule, polar
materials dissolve readily in polar solvents and nonpolar materials in nonpolar
solvents. Materials with both polar and nonpolar groups in their molecules may
dissolve in polar solvents but their solubilities tend to decrease as the proportion
of nonpolar groups in the molecule increases.
A large number of materials have poor solubilities in water and often pose
a problem in the formulation of pharmaceutical products. Addition of another
solvent, in which these materials are more readily soluble, will increase the concentration
of the materials in the solution. This additional solvent is known as a
cosolvent and common examples include glycerin, sorbitol, propylene glycol, and
polyethylene glycols. The proportion of cosolvent required varies from system to
3.2. Temperature
The solubilities of most materials increase with rising temperature due to their
endothermic dissolution process. Similarly, the solubilities of materials which exhibit
exothermic dissolution decrease with rising temperature. The relationship between
90 Chan and Heng
 2005 by Taylor & Francis Group, LLC
solubility and temperature is expressed by solubility curves. Three typical solubility
curves are shown in Figure 4.
A material whose solubility increases with rising temperature exhibits a continuous
curve with a positive slope (Curve A) while that whose solubility decreases with
rising temperature exhibits a negative slope (Curve B). Some solubility curves show
an abrupt change in slope at certain temperatures (Curve C). This phenomenon is
attributed to the change in nature of the material at the temperature where the slope
changes direction. For example, the solubility curve C is derived from a material
which can exist in two forms. The curve shows that Form I is converted to
Form II at temperature t. The dissolution of Form I in water is endothermic, which
explains the increasing solubility of the material with rising temperature until t.
Above this temperature, Form I is converted to Form II, which exhibits exothermic
dissolution. The slope of the solubility curve therefore changes from positive to
negative as the temperature exceeds t.
3.3. Crystal Characteristics
Materials may exist as amorphous or crystalline substances. Some materials such as
cortisone, tetracycline, sulfathiazole, and chloramphenicol palmitate can exist in
more than one crystalline form and this property is described as polymorphism.
The different crystalline forms which are known as polymorphs exhibit different
degrees of stability. The lattice structure of the crystalline substances may be altered
by the incorporation of molecules of the solvent from which crystallization occurs.
The resultant crystals obtained are called solvates. If the solvent is water, the crystals
are said to be hydrated.
The different forms of a material have varying solubilities. The amorphous
substance is more soluble than the crystalline counterpart. Among the crystalline
forms, the metastable polymorphs are generally more soluble than the stable polymorphs.
Hydrated crystals tend to exhibit a lower solubility in water compared with
their anhydrous form. On the contrary, the aqueous solubilities of the nonaqueous
solvates are often greater than those of the unsolvated forms.
Figure 4 Typical solubility curves.
Drug Substance and Excipient Characterization 91
 2005 by Taylor & Francis Group, LLC
3.4. Particle Size
It is important to distinguish equilibrium solubility from intrinsic solubility of
a material. Unlike intrinsic solubility, equilibrium solubility is not affected by the
particle size of the material. The solubilities of materials reported in the literature
generally refer to the equilibrium solubilities. The method of determining equilibrium
solubility is given in an earlier section.
The intrinsic solubility of a material is dependent on the particle size of the
material. Smaller particles have a higher intrinsic solubility compared with their larger
counterparts. This is aptly explained by the existence of a higher interfacial free
energy on smaller particles resulting in a thermodynamically unstable system which
is corrected by greater dissolution of the particles and production of a supersaturated
solution. The increase in intrinsic solubility with decrease in particle size, however,
ceases when the particles have a very small radius and any further decrease in particle
size causes a decrease in intrinsic solubility.
3.5. pH
The solubility of a material will be affected by the pH of the liquid medium if the
material is acidic or basic. For example, a weakly acidic drug is more soluble in
an alkaline solution while a weakly basic drug is more soluble in an acidic solution.
This phenomenon is due to the formation of more soluble salts as a result of acid
base reaction. Conversely, the weakly acidic drug will precipitate from the solution
if the pH is lowered by the addition of an acid while the weakly basic drug will
precipitate from the solution if the pH is raised by the addition of an alkali. The
precipitation is a result of the conversion of the drug in solution to the less soluble
unionized form.
The relationship between pH and solubility of a material is given by Eqs. 3.1
and 3.2. These equations which are modified from the HendersonHasselbalch
equation are useful in the estimation of the solubility of materials under different
pH conditions.
For acidic materials,
pH ? pKa ? log
S  S0
S0 ?3:1?
For basic materials,
pH ? pKa ? log
S  S0 ?3:2?
where S is the overall solubility of the drug and S0 is the solubility of its unionized
3.6. Additives
Additives refer to other substances incorporated into the solvent in which the drug
is dissolved. The drug in the solution may exist as ionized or nonionized forms.
Drugs that dissociate in the solvent to form ions are described as ionizable while
those that do not dissociate are nonionizable. Like the drugs, some additives ionize
readily in the solvent while others do not. Among the additives that ionize, some
produce a similar ion to the drug. The effect of additives on the solubility of a
92 Chan and Heng
 2005 by Taylor & Francis Group, LLC
drug depends on the nature of the drug as well as the additives. For ease of understanding,
the additives will be classified as common ions, indifferent electrolytes, and
3.6.1. Effect of Common Ions on the Solubility of Ionizable Drugs
The solubility of a sparingly soluble drug is decreased by an additive that possesses a
similar ion to the drug. This phenomenon which is a result of common ion effect can
be explained by the Law of Mass Action.
The equilibrium of a saturated solution in the presence of undissolved drug is
represented as follows:
AB?s? , AB , A? ? B
Undissolved Undissociated Ions
drug molecule
If the material is sparingly soluble, the concentration of dissolved drug is sufficiently
small to assume complete dissociation into ions. The overall equilibrium may therefore
be expressed as:
AB?s? , A? ? B ?3:4?
From the Law of Mass Action, the equilibrium constant (K) for this reversible reaction
is given by the following equation:
K ? ?A??B
?AB?s? ?3:5?
where the square brackets indicate the concentrations of the respective components.
Since the concentration of a solid may be regarded as being constant, Eq. 3.5 can be
rewritten as:
K0s ? ?A??B ?3:6?
where K0s
is a constant which is known as the solubility product of the drug AB.
Some drugs have molecules which contain more than one ion of each type. The equilibrium
and solubility product of these drugs are similarly expressed as:
AxBy?s? , xA? ? yB ?3:7?
K0s ? ?A?x?By ?3:8?
If the additive dissociates to produce either A? or B, the concentrations of these
ions in the solvent will increase. As a result, the product [A?][B] or [A?]x[B]y will
also increase. It should be recalled that K0s is a constant. Therefore, if K0s
is exceeded
by the product of the concentrations of ions, the equilibrium will move toward the
left in order to restore the equilibrium and the drug will be precipitated. This
explains the decrease in the solubility of a drug by common ion effect.
3.6.2. Effect of Indifferent Electrolytes on the Solubility of Ionizable Drugs
Additives which dissociate to form ions different from those of the drug are known
as indifferent electrolytes. Unlike common ions, indifferent electrolytes may increase
the solubility of a sparingly soluble drug.
Drug Substance and Excipient Characterization 93
 2005 by Taylor & Francis Group, LLC
The solubility product defined by Eq. 3.6 is only an approximation from the
more exact thermodynamic relationship expressed by the following equation:
Ks ? aA?  aB ?3:9?
where Ks is the solubility product of drug AB and aA? and aB are the activities of the
respective ions. The activity of an ion is defined as the effective concentration of the
ion in solution. It generally has a lower value than the actual concentration because
some of the ions are taken out of play by strong association with oppositely
charged ions.
At infinite dilution, the wide separation of ions prevents interionic association
and the actual concentration and activity of the ion are equal. This situation is
applicable to a sparingly soluble drug where the concentrations of ions produced
are so small that the ions are completely unassociated. Therefore, the solubility product
can be expressed by Eq. 3.6. However, if the concentration of ions increases, the
effects of interionic association are no longer negligible and the activity becomes less
than the actual concentration. The activity coefficient, which is the ratio of activity
to actual concentration (Eq. 3.10), indicates the extent of interionic association. An
activity coefficient of unity shows that the ions are completely unassociated while
smaller values show greater interionic association:
?A? ? fA? or aA? ? fA?  ?A? ?3:10?
where fA? is the activity coefficient of ion A.
Eq. 3.9 can thus be expressed as:
Ks ? fA?  ?A?  fB  ?B ? fA?  fB  ?A?  ?B ?3:11?
According to Eq. 3.6, the product of the concentrations is the constant K0s. The
product of the activity coefficients of the respective ions may be equated to f 2
A?B ,
where fA?B is the mean activity coefficient of the drug. Hence,
Ks ? K0sf 2
A?B ?3:12?
The value of the activity coefficient decreases from unity as the overall concentration
of ions of the solution increases. Since Ks is a constant, it follows that K0s
will increase
with the ionic strength and become larger than Ks. The increase in the solubility of a
drug by the addition of indifferent electrolytes is attributed to the increase in the
ionic strength of the solution due to the indifferent electrolytes.
3.6.3. Effect of Nonelectrolytes on the Solubility of Ionizable Drugs
The solubility of an ionizable drug depends on the dissociation of the drug into ions.
The degree of dissociation is affected by the dielectric constant of the solvent. Solvents
with a high dielectric constant, being polar in nature, are able to reduce the
forces that attract the oppositely charged ions produced by dissociation of the drug.
Addition of a nonelectrolyte such as alcohol will lower the dielectric constant of the
solvent. This will decrease the dissociation and, subsequently, solubility of the drug.
3.6.4. Effect of Electrolytes on the Solubility of Nonionizable Drugs
Nonionizable drugs do not dissociate into ions in solution. They exist as single
molecules and their solubility in the solvent depends on the formation of weak
94 Chan and Heng
 2005 by Taylor & Francis Group, LLC
intermolecular bonds with the molecules of the solvent. For example, the solubility
of the drug in water is dependent on the formation of hydrogen bonds between
the molecules of the drug and water. When an electrolyte is added, it will dissociate
to form ions which have a high affinity for water. These ions will compete with the
molecules of the drug for water and reduce the solubility of the drug in water.
3.6.5. Effect of Surfactants
Surfactants are solutes that cause a marked decrease in the surface tension of the solvent.
These substances are commonly employed as solubilizing, wetting, and emulsifying
agents. They are composed of a lipophilic group that has little affinity for water
and a hydrophilic group that has strong affinity for water. At a specific concentration
known as the critical micellar concentration, the surfactant molecules exist as
large aggregates called micelles. In an aqueous system, the hydrophilic group is
orientated on the exterior while the lipophilic groups are on the interior of the
micelles. Drugs which are poorly soluble in water may be taken into the interior
of these micelles, resulting in more of the drug being able to go into solution. The
enhanced solubility obtained as a result of the solubilization phenomenon is known
as the apparent solubility of the drug.
3.6.6. Effect of Complex Formation
The amount of drug that can go into solution may be altered by the addition of a
substance which interacts with the drug to form a complex. The solubility of the
complex will determine the apparent solubility of the drug. If the complex is more
soluble than the drug, a larger amount of the drug will dissolve to form the complex.
Thus, the drug will show a higher solubility in the solvent. Similarly, if the complex is
less soluble, some of the drug will be precipitated in the form of the complex. The
drug will therefore show a lower solubility in the solvent. It should be borne in mind
that the modified solubility obtained is not the solubility but the apparent solubility
of the drug.
Materials may occur as amorphous substances without any definite structure or as
crystalline particles with a definite structure and shape. Some materials may exist
in more than one crystalline form (polymorph) and are described as exhibiting polymorphism.
The type of crystal formed depends on the conditions, such as temperature
and type of solvent, under which crystallization is induced. At a specific
temperature or pressure, more than one polymorph can exist but only one will be
thermodynamically stable. The less stable or metastable forms will be converted to
the stable form with time. Studies show that it may take from minutes to years to
revert to the stable lattice structure.
The different crystalline forms of a material generally differ in many physical
characteristics, such as solubility, melting point, optical and electrical properties,
density, hardness, and stability. The use of metastable polymorphs frequently results
in higher solubility and dissolution rates while the stable polymorphs are often more
resistant to chemical degradation. It is obvious that any change in the crystalline
form will affect the therapeutic efficacy of a pharmaceutical product. Therefore,
Drug Substance and Excipient Characterization 95
 2005 by Taylor & Francis Group, LLC
knowledge of the crystalline form of a drug in the formulation of pharmaceutical
products is very important and steps should be taken to ensure that the crystals
do not convert from one form to another during production and storage of the product.
Some problems which may arise if the crystal properties of the drug are not
properly characterized are precipitation, low stability, and poor bioavailability of
drug. The significance of polymorphism in pharmacy is discussed in detail by
Haleblian and McCrone (17).
A number of techniques may be used to identify the crystalline form of a material.
It is advisable to employ more than one method in the analysis as the use of only
one method has been found to be unreliable at times.
4.1. Dissolution Study
An amount of the material in excess of its solubility is added to the dissolution
medium and aliquot samples are removed and assayed at appropriate time intervals.
The concentration of the material in solution as a function of time is then plotted.
The crystalline form which constitutes the material is reflected by the shape of the
dissolution curve. The typical dissolution profile of metastable polymorph which
readily reverts to the stable form is shown in Figure 5A.
The concentration of the metastable polymorph is noted to increase much
more rapidly at the initial period of the dissolution study and then drop to that
of the stable polymorph. For the stable polymorph, the dissolution profile just
increases gradually to a plateau. The solubility of the metastable form is indicated
by the peak of its dissolution curve. In some cases, the metastable polymorph
does not revert readily to the stable form. The dissolution curve of such a metastable
form lies above that of the stable form, indicating that the former is more
soluble (Fig. 5B). The plateau of each curve indicates the solubility of the respective
Figure 5 Typical dissolution profiles.
96 Chan and Heng
 2005 by Taylor & Francis Group, LLC
4.2. X-Ray Diffraction
Crystalline materials in powder form exhibit characteristic X-ray diffraction patterns
with peaks of varying heights and in different positions. These diffraction patterns
are also known as X-ray diffractograms. A typical X-ray diffractogram is illustrated
in Figure 6.
The polymorphs of a material have different crystal packing arrangements and
thus produce differences in their diffractograms from which the crystalline form of
the material is identified. This method of analysis is nondestructive and requires a
very small sample of the material which can be examined without further processing.
X-ray diffraction studies are especially useful for investigations on the changes of
crystal form during processing. In some cases, the extent of the conversion of a crystalline
drug to the amorphous form can be determined.
Conventional X-ray diffraction method cannot distinguish polymorphs which
have relatively small crystallites because of the low resolution of diffractometers.
Synchrotron sources have been employed to obtain high-resolution electron diffraction
patterns which enable differentiation of such polymorphs (18). The development
of very sensitive charge-coupled detectors (CCD) has allowed electron diffraction
patterns to be recorded in a few seconds using very low electron currents.
4.3. Infrared Analysis
As mentioned earlier, the polymorphs of a material show varying crystal packing
arrangements and produce different X-ray diffractograms. The crystal packing
arrangement also affects the energy of molecular bonds and results in different IR
spectra for the polymorphs of a material. Identification of the crystalline form of
a material is based on the spectrum derived. Infrared (IR) analysis can be used for
both qualitative and quantitative identification. It is important to use only materials
in the solid form as the polymorphs of a material in solution have identical IR
In the area of substance identification, near IR is increasingly being used
for confirming the identity of a chemical substance. Chemometric methods using
Figure 6 Typical X-ray diffractogram.
Drug Substance and Excipient Characterization 97
 2005 by Taylor & Francis Group, LLC
near IR have also been developed for source identification of pharmaceutical
4.4. Thermal Analysis
In this method, the polymorphs are identified by their thermal behaviors. The change
in energy of the polymorph as it undergoes transformation when it is heated is
recorded as a thermogram. An example of a thermogram is given in Figure 7. The
thermogram consists of characteristic peaks, including melting point (Tm) and glass
transition temperature (Tg). The peaks pointing downward indicate endothermic
changes such as melting, sublimation, and desolvation. The different polymorphs
of a material will exhibit different thermograms which allow them to be identified.
Differential scanning calorimetry (DSC) and differential thermal analysis
(DTA) are two methods of thermal analysis commonly used for these studies. In
DSC, the energy resulting from the crystalline transformation is recorded as a function
of temperature. In DTA, the energy is expressed by differential temperature
(sample vs. inert substance).
Conventional DSC has a major limitation: if a Tg occurs in the same temperature
range as another transition, for example, water or solvent loss, the two events
cannot be separated. This limitation may be overcome by employing modulated temperature
DSC (MTDSC), where the measurements are conducted using sine wave
temperature programs defined by underlying heating rate, amplitude, and period.
The heat capacity change associated with the Tg can be separated from the heat flow
changes caused by melting, drying, and solvent loss. By use of the phase angle curve
produced from the MTDSC data analysis, very small changes in specific heat can be
detected, thereby increasing the sensitivity of the method. Based on thermal behavior,
MTDSC is able to differentiate the amorphous and polymorphic forms
of a material with much greater clarity (19). One of the disadvantages of this
Figure 7 Example of a thermogram.
98 Chan and Heng
 2005 by Taylor & Francis Group, LLC
method is that the data analysis and interpretation are more difficult than for DSC.
In addition, the experiment process can be long as much lower heating rates
are used.
4.5. Hot Stage Microscopy
The polarizing microscope fitted with a hot stage is very useful for identifying
the crystalline forms of a material. In this method, the polymorph is heated to a
temperature at which it undergoes a change in birefringence and/or appearance
which is characteristic of the polymorph.
4.6. Raman Spectroscopy
Raman spectroscopy provides molecular information about the crystalline, as well as
the amorphous forms of a material. In this method, the material is subjected to
a laser beam and a spectrum of the scattered light obtained. The spectrum shows
vibrational bands of the material at their characteristic frequencies. The amorphous
and polymorphic forms of a material can be distinguished by their characteristic
Raman spectroscopy and IR spectroscopy complement each other. The former
measures a change in polarization whereas the latter measures a change in dipole
moment. IR-inactive vibrations can be strong in Raman spectra and vice versa.
For example, vibrations in the region of 10400cm1 are more easily studied by
Raman than by IR spectroscopy.
One advantage of the Raman spectroscopy method is that no sample preparation
is required, thus the likelihood of inducing phase changes through processing is
reduced. However, representative sampling is critical for quantitative analysis (20).
The results are affected by the particle size of the material. The use of Fourier transform
Raman spectrometers with a longer wavelength laser of 1064nm eliminates the
problem of any fluorescent background. With the utilization of fiber optics, real-time
crystallization can be monitored (21). Thus, this method is useful for on-line analysis
of pharmaceutical processes.
4.7. Solid-State Nuclear Magnetic Resonance Spectroscopy
Solid-state nuclear magnetic resonance (SSNMR) spectroscopy is a more advanced
method for differentiating the polymorphs of a material. The substance is placed in
a strong magnetic field and subjected to radiofrequency radiation. The individual
nuclei experience different magnetic environments and thus show different changes
in resonant frequency characterized by chemical shift. SSNMR spectra show sharp
resonance at chemical shifts characteristic of the molecular and crystal structure.
The polymorphs are differentiated by their characteristic spectra.
This method is suitable for the study of molecular motions in a wide range of
time scales (1021010 sec) and thus complements the X-ray diffraction method,
which is limited to motions that are relatively fast (1018 sec). It has a few
advantages. It can be used for characterization of solid-state forms that cannot be
crystallized and studied by the X-ray diffraction method. It is also useful for
quantifying components of heterogenous mixtures. In contrast to IR and Raman
spectroscopy, the results are less affected by the particle size of the test material.
Drug Substance and Excipient Characterization 99
 2005 by Taylor & Francis Group, LLC
It is undoubted that the type of physical characterization tests for a drug or excipient
depends very much on the material concerned as well as the processing involved.
Material testing can be broadly divided into two types, namely physical testing
and functionality testing. Physical testing, which is used to determine properties such
as size, solubility, and crystal form is generally more direct and the procedures are
better established. Functionality testing which evaluates, for example, lubricity, flow,
creep, and tack is less well established but, if carried out, may yield useful information
about the raw materials and their potential effects on the processing.
With the introduction of automated helium pycnometer, true densities of raw
materials can be determined with ease and at a high degree of accuracy. The true
density can serve to assure the formulator the identity of the material and sometimes
reveal the state of raw materials, like partial changes of a powder from anhydrous to
hydrated form.
Packing studies of the powder can reveal its rheological properties. These
studies are carried out by filling a volumetrically calibrated cylinder with powder
and tapping it. From the weight of powder filled and the changes in volume of the
powder on tapping, changes in apparent densities can give information about the
powder flowability (22). A small change in apparent density before and after tapping
indicates good flow properties. The Hausner ratio, the ratio of tapped density to
poured density has been found to be extremely sensitive to particle shape (23).
The poured density is the undisturbed packing density of a powder in a calibrated
cylinder after filling, usually done by passing powder through a sieve. The tapped
density is the packing density after tapping a bed of powder until no change in
the packing density is seen. An alternative expression used to predict powder flowability
is the compressibility index or Carr index, which is the ratio of the difference
between the tapped and poured densities to the tapped density, expressed as
a percentage.
Flowability of a powder may be determined directly using a flow cup, which is
often vibrated, or indirectly by the angle of repose. The angle of repose measurement
usually involves forming a conical powder heap and measuring the maximum
inclination of heap. The powders that can flow well give small angles of repose. A
variation of the heap forming measurement is to use a loosely packed powder bed
or transparent cylinder partially filled with powder and slowly rotating the powder
bed away from the initial horizontal plane of the powder surface. The angle of
inclination when sliding of powder just occurs is used to characterize the powder
flowability. Determination of powder avalanches is a simple and efficient way to
evaluate the flow properties of pharmaceutical powders. Powder flow properties
are assessed by comparing their flow indices using a powder flowability analyzer
(TSI Incorporated, USA). The flow analyzer consisted of a transparent acrylic drum
The drum could be programmed to rotate at user-defined rates. A metal mesh
collar was fitted onto the inner circumference of the drum to prevent the sliding of
powder particles down the circumferential wall instead of avalanching. Test powders
should be sieved before being poured into the sample drum.
Analysis should be performed at a range of drum speeds in order to determine
the flow index, as calculated by the equation described as follows:
100 Chan and Heng
fitted on a vertical bayonet-type mount (Fig. 8).
 2005 by Taylor & Francis Group, LLC
Flow index ?
Si ?5:1?
where n is the total number of speeds tested and Si is the standard deviation of the
tested speed.
Strange attractor diagrams are generated by plotting the time difference
between each avalanche and the next formed throughout the test duration. These
plots provide a visual comparison of the flow properties of powders. Free-flowing
powders will generate strange attractor diagrams that are dense and closer to the
origin. In addition, their calculated flow indices will be correspondingly small.
Shear studies of powder bed can also be carried out to evaluate the cohesiveness
of powders. There are many variations of shear cells for evaluation of powder
The change in the packing properties of powders under pressure is used to
study the compactability of powders. This study can be made using compression
punches attached to pressure transducers (24). The force exerted and the displacement
can be utilized to calculate useful parameters for assessing the compactability
of a powder or powder mix.
For polymers, mechanical testing such as creep testing of films formed can provide
information on the suitability of the polymer or the additives added for their
film forming function (25). When polymers are used as a binder, adhesive properties
in addition to polymer viscosity may be determined. The adhesive property can be
evaluated by the measurement of tack or stickiness. This measurement involves
determining the force required to detach two platens held together by the polymer
solution. Other tests that may be carried out include measurements for surface
Figure 8 Avalanche powder flowability tester.
Drug Substance and Excipient Characterization 101
 2005 by Taylor & Francis Group, LLC
activity, glass transition temperature, cloud point, and adhesion strength of dried
Excipients for granulation can be largely divided into two categories, as bulking
agents and as functional additives. It is true that bulking agents or fillers also serve
a function in that they form the core or structure of a dosage form. Nevertheless,
bulking agents generally differ from the functional additives in that they are usually
inert materials that are relatively inexpensive. Functional additives include binders,
disintegrants, lubricants, colorants, and stabilizing agents. Besides the pharmacopoeia,
several recently published handbooks can provide a compilation of commonly
used pharmaceutical excipients (26,27).
The choice of excipients depends on a number of factors, namely, the drug
used, the process involved, the formulator, and the cost of excipient. Differences will
be seen in the choice of excipients by innovator companies and generic companies
as they have different cost considerations. Some granulation processes like fluid
bed granulation would require tighter control of drug and excipient specifications
compared with wet granulation using a paddle mixer. In fluid bed, besides the
control of particle size distribution, particle density of excipient should not be too
greatly different from that of the drug. Extrusionspheronization would generally
require microcrystalline cellulose.
A very common filler is lactose, although other sugars, dicalcium phosphate,
starch, pregelatinized starch, and microcrystalline cellulose are also used. The
starches and microcrystalline cellulose are also disintegrants in tablets. For tablets,
other commonly used disintegrants include sodium starch glycolate, croscarmellose,
crospovidone, and low-substituted hydroxypropylcellulose. Lubricants are usually
not added till prior to filling or tableting of the granules. The most commonly used
lubricant is magnesium stearate. Other lubricants used include calcium stearate,
stearic acid, wax, hydrogenated vegetable oil, talc, and starch. Much work has been
done on lubricants and it is well established that the physical characterization of the
lubricant is very important to ensure consistency in functionality, especially when
a change is made in the supplier (28).
Binders used in granulation consist of a wide variety of sugars and polymers
natural, semisynthetic, and synthetic. Sugars used include sucrose, glucose, and sorbitol.
Examples of natural polymers are acacia, alginic acid, sodium alginate, gelatin,
and starch. There is an inherent variability in the natural polymers of different
batches and this sometimes gives rise to problems in production. It is essential to
characterize binders for their viscous properties, at the least, to minimize potential
processing problems. The semisynthetic binders include ethylcellulose, sodium carboxymethylcellulose,
methylcellulose, hydroxypropylcellulose, and hydroxypropylmethylcellulose.
By comparison with the natural binders, variability between
batches of the semisynthetic binders from the same supplier can be expected to be
much less. For each polymer type, a number of viscosity grades are available. The
two types of synthetic binders in use are polyvinylpyrrolidone and polyethylene glycol.
Polyvinylpyrrolidone is widely used in both wet massing and fluid bed granulation.
The main application of polyethylene glycol as a binder is in melt granulation.
Other excipients employed in pharmaceutical products include colorants,
coating aids, stabilizers, pH modifiers, and release rate modifiers. They play very
102 Chan and Heng
 2005 by Taylor & Francis Group, LLC
important roles and would require specific characterization tests. In the making of a
product, the functional excipients to be added must be carefully selected to ensure
that the final product is bioavailable and esthetically pleasing. As all raw materials
have to undergo a processing schedule for conversion of the materials to the finished
product, it is likely that the processing can be very unpredictable if standardizing of
raw materials has not been carried out.
In the formulation of any pharmaceutical product, it is imperative to ensure that
the ingredients used are compatible with one another. Incompatibilities can occur
between drug and excipient as well as between the excipients themselves. Incompatibilities
may be manifested through many modes, such as acidbase interaction and
complex formation, resulting in lower potency and/or stability and eventually poor
therapeutic efficacy of the product. It is therefore essential to avoid incompatibilities
and this is achieved by carrying out studies to detect potential interactions between
the components used in the formulation.
7.1. Stability Study
This is the traditional method of detecting incompatibilities. Mixtures of the drug
and excipient are prepared and stored under exaggerated conditions of heat, light,
and humidity. A detailed discussion on the realistic proportions of drug and excipient
to be used in the investigation was reported by Akers (29). The mixtures are
examined for any physical change and aliquot samples are withdrawn for assay of
the intact drug at varying time intervals. Incompatibility is reflected by various signs
such as appearance of precipitate and decrease in the concentration of intact drug.
7.2. Thermal Analysis
A relatively simple approach for the investigation of potential interaction between
a drug and an excipient can be carried out using differential scanning calorimetry.
The drug, individual excipients, and binary mixtures of the drug and excipient are
separately scanned at a standard rate over a temperature range that encompasses
all the thermal features of the drug and excipients. Each mixture consists of 50%
drug and excipient in order to maximize the likelihood of an interaction. The thermograms
of the mixtures and the appropriate individual components are compared.
Interaction is deduced by changes in the thermal features such as elimination or
Changes in shape, onset, maximum temperature, and relative height of the peaks
may also indicate interaction. However, it should be cautioned that these changes
could also arise from physical mixing of the components.
A big advantage of differential scanning calorimetry over the traditional
stability test is the speed of determination. However, like all methods, differential
scanning calorimetry has its own limitations. It is not applicable if the test materials
exhibit properties that make data interpretation difficult, such as eutectic formation,
coincident melting, and dissolution of one component in the melt of the other. It is
not advisable to rely on differential scanning calorimetry alone to determine incompatibility.
Chrzanowski et al. (30) reported that differential scanning calorimetry
Drug Substance and Excipient Characterization 103
appearance of a peak in the thermogram of the mixture. This is illustrated in Figure 9.
 2005 by Taylor & Francis Group, LLC
indicated no incompatibilities in mixtures of fenretinideexcipient and mefenidil
excipient, whereas the traditional stability study showed some incompatibilities.
Hence, differential scanning calorimetry should only be used to supplement the
stability test by eliminating the incompatible excipients and reducing the number
of test samples.
7.3. Chromatographic Methods
Chromatography was first used for the separation of colored leaf pigments. The
operation of chromatography is based on the distribution of a material between a
stationary phase and a mobile phase. The stationary phase can be a solid or a liquid
supported on a solid while the mobile phase can be a gas or a liquid which flows continuously
around the stationary phase. The different components in a mixture can be
separated and identified as a result of differences in their affinity for the stationary
In addition to its application in the separation and identification of materials,
chromatography is also employed to detect potential interactions between materials.
Both thin-layer chromatography and liquid chromatography are commonly employed
in this area of study. In thin-layer chromatography, the stationary phase
consists of a powder adhered onto a glass, plastic, or metal plate. The powders commonly
used are silica, alumina, polyamides, cellulose, and ion-exchange resins. Solutions
of the drug, excipient, and drugexcipient mixture are prepared and spotted on
the same baseline at one end of the plate. The plate is then placed upright in a closed
chamber containing the solvent, which constitutes the mobile phase. As the solvent
moves up the plate, it carries with it the materials. Those materials which have
a stronger affinity for the stationary phase will move at a slower rate. The material
is identified by its Rf value, which is defined as the ratio of the distance which the
material has moved to the distance the solvent front has moved. The position of
the material on the plate is indicated by spraying the plate with certain reagents or
Figure 9 Thermograms indicating drugexcipient interaction.
104 Chan and Heng
 2005 by Taylor & Francis Group, LLC
exposing the plate to ultraviolet radiation. If there is no interaction between the
drug and excipient, the mixture will produce two spots whose Rf values are identical
with those of the individual drug and excipient. If there is interaction, the complex
formed will produce a spot whose Rf value is different from those of the individual
In liquid chromatography, the distribution of the material between the solid
stationary phase in a column and the liquid mobile phase is determined. The material
is identified by the time taken for it to elute from the column. Solutions of the drug,
excipient, and drugexcipient mixture are prepared and injected into the column.
The materials will elute from the column at different speeds, depending on their affi-
nity for the column. The concentration of the material that elutes from the column is
detected and plotted against time to give to a chromatogram. If there is interaction
between the drug and excipient, the complex formed will exhibit an elution time different
from those of the individual components. This is illustrated in Figure 10,
which shows the chromatograms of the drug, excipient, and drugexcipient mixture
plotted together. Similarly, gas chromatography may be used.
The greatest difficulty for any process technologist is to decide on the type and extent
of material characterization to be undertaken such that it is cost-effective in the long
run. Often, it is the problem from the production run that necessitates further material
characterization to be carried out either for the purpose of resolving the problem
or to prevent future occurrences. This chapter serves to identify the more common
material characterization methods that can be carried out and the potentially useful
information that can be inferred from the tests. It is hoped that the discussion of the
many methods of material characterization could help in the choice of characterization
methods for material testing.
Figure 10 Chromatograms illustrating drugexcipient interaction.
Drug Substance and Excipient Characterization 105
 2005 by Taylor & Francis Group, LLC
1. Heywood H. The evaluation of powders. J Pharm Pharmac 1963 (suppl 15):56T.
2. Wadell H. Volume, shape and roundness of rock particles. J Geol 1932; 40:443.
3. Stanley-Wood NG. Particle characterisation by size, shape and surface for individual
particles. Stanley-Wood NG, ed. Enlargement and Compaction of Particulate Solids.
London: Butterworths, 1983:18.
4. Kaye BH. Multifractal description of ragged fine particle profile. Particle Char 1984;
5. Kaye BH. Specification on the ruggedness and/or texture of a fine powder profile by its
fractal dimension. Powder Technol 1978; 21:1.
6. Flook AG. The use of dilation logic on the Quantimet to achieve dimension characterisation
of textured and structured profiles. Powder Technol 1978; 21:295.
7. Meakin P. Simulations of aggregation processes. Avnir D, ed. The Fractal Approach to
Heterogeneous Chemistry. Chichester: Wiley, 1990:131160.
8. Allen T. Particle Size Measurement. 4th ed. London: Chapman and Hall, 1990.
9. Svarosky L. Characterization of powders. Rhodes MJ, ed. Principles of Powder
Technology. Chichester: John Wiley, 1990:35.
10. Washington C. Particle Size Analysis in Pharmaceutics and Other Industries. U.K: Ellis
Horwood, 1992.
11. British Standard 3406, Part 4, 1963.
12. Timbrell V. A method of measuring and grading microscopic spherical particles. Nature
1952; 170:318.
13. Russ JC. Computer-Assisted Microscopy. New York: Plenum Press, 1990.
14. Sing KSW. Adsorption methods for surface area determinations. Stanley-Wood NG,
Lines RW, eds. Particle Size Analysis. Cambridge, U.K: The Royal Society of Chemistry,
15. Lowell S, Shields JE. Particle Surface Area and Porosity. 2nd ed. London: Chapman and
Hall, 1983:205.
16. Buckley HE. Crystal Growth. New York: Wiley, 1951:29.
17. Haleblian J, McCrone W. Pharmaceutical applications of polymorphism. J Pharm Sci
1969; 58:911.
18. Li ZG, Harlow RL, Foris CM, Li H, Ma P, Vickery RD, Maurin MB, Toby BH. Polymorph
determination for the GP IIb/IIIa antagonist, roxifiban, using a combination of
electron diffraction and synchrotron x-ray powder diffraction techniques. J Pharm Sci
1999; 88(3):297.
19. Bottom R. The role of modulated temperature differential scanning calorimetry in the
characterization of a drug molecule exhibiting polymorphic and glass forming tendencies.
Int J Pharm 1999; 192:47.
20. Taylor LS, Zografi G. The quantitative analysis of crystallinity using FT-Raman spectroscopy.
Pharm Res 1988; 15(5):755.
21. Findlay WP, Bugay DE. Utilization of Fourier transform-Raman spectroscopy for the
study of pharmaceutical crystal forms. J Pharm Biomed Anal 1998; 16:921.
22. Yamashiro M, Yuasa Y, Kawakita K. An experimental study on the relationships
between compressibility, fluidity and cohesion of powder solids at small tapping numbers.
Powder Technol 1983; 34:225.
23. Kostelnik MC, Beddow JK. New techniques for tap density. In: Hausner HH, ed. Modern
Developments in Powder Metallurgy IV, Processes. Proceedings of the 1970 International
Powder Metallurgy Conference. New York: Plenum Press, 1970.
24. Watt PR. Tablet Machine Instrumentation in Pharmaceutics: Principles and Practice.
Chichester: Ellis Horwood, 1988.
25. Okhamafe AO, York P. Interaction phenomena in pharmaceutical film coatings and
testing methods. Int J Pharm 1987; 39:1.
106 Chan and Heng
 2005 by Taylor & Francis Group, LLC
26. Handbook of Pharmaceutical Excipients. Washington, DC: American Pharmaceutical
Association, London: Pharmaceutical Society of Great Britain, 1986.
27. Brittain HG, ed. Analytical Profiles of Drug Substances and Excipients. San Diego: Academic
Press, 1993.
28. Chowhan ZT. Excipients and their functionality in drug product development. Pharm
Tech 1993; 17:72.
29. Akers MJ. Preformulation testing of solid oral dosage form drugs. Can J Pharm 1976;
30. Chrzanowski FA, Ulissi LA, Fegely BJ, Newman AC. Preformulation excipient compatibility
testing. Application of differential scanning calorimetric method versus a wet
granulation simulating isothermal stress method. Drug Dev Ind Pharm 1986; 12:703.
Drug Substance and Excipient Characterization 107
 2005 by Taylor & Francis Group, LLC
Binders and Solvents
Ehab Hamed, Derek Moe, and Raj Khankari
CIMA Labs Inc., Eden Prairie, Minnesota, U.S.A.
John Hontz
Biovail, Chantilly, Virginia, U.S.A.
Binders are adhesives that are added to solid dosage formulations. The primary role
of binders is to provide the cohesiveness essential for the bonding of the solid particles
under compaction to form a tablet. In a wet-granulation process, binders promote
size enlargement to produce granules and thus improve flowability of the blend during
the manufacturing process. Binders may also improve the hardness of the tablets
by enhancing intragranular as well as intergranular forces. In a direct compression
process, binders often act as fillers and impart compressibility to the powder blend.
The cohesive properties of binders may reduce friability of the tablets and thus aid
in their durability and elegance. Although the purpose of using binders in a tablet formulation
is not to influence its disintegration and dissolution rate, these properties
may be modified due to the altered wettability of the formulation.
Binders are usually natural polymers, synthetic polymers, or sugars. The selection of
the type of binder for a particular system is quite often empirical and dependent on
the previous experience of the formulator, in conjunction with examination of excipient
compatibility studies. Selection of the quantity of binder required in a particular
system can be determined by optimization studies, using parameters such as
granule friability, tablet friability, hardness, disintegration time, and the drug dissolution
rate. Some commonly used binders in wet granulation with their usual concentration
range, along with the granulating system are listed in Basic
properties of some widely used binders with their method of incorporation will be
discussed in this section (14).
Table 1.
 2005 by Taylor & Francis Group, LLC
2.1. Natural Polymers
2.1.1. Starch
Starch is a polymeric carbohydrate obtained from various plant sources, such as
potato, wheat, maize, rice, and tapioca. It is a GRAS-listed material and traditionally
was one of the most widely used tablet binders. It is insoluble in cold water and in
alcohol but it gelatinizes in hot water to form a paste. Starch paste can be prepared
by dispersing starch in 11.5 parts of cold water for initial wetting followed by addition
of two to four times as much boiling water with continuous stirring until a translucent
paste is obtained. This paste is further diluted by cold water to the desired
concentration. Alternatively, starch paste can also be prepared by heating the cold
water suspension of starch to boiling in a steam-jacketed kettle with constant stirring.
Freshly prepared starch paste is used at a concentration of 525% (w/w) in a
tablet granulation. Relatively soft and friable granules are produced when starch
paste is used as a binder. Consequently, it yields tablets that disintegrate readily.
During the wet-massing process, high viscosity of the starch paste can sometime
make it difficult to evenly distribute the binder in the powder blend.
Table 1 Commonly Used Granulating Systems
Method of
used in
formula Solvent
Percentage used
in granulating
Natural polymers
Starch Wet mixing 25 Water 525
Pregelatinized starch Wet mixing 25 Water 1015
Dry mixing 510 Water
Gelatin Wet mixing 13 Water 510
Acacia Wet mixing 35 Water 1015
Alginic acid Dry mixing 15 Water
Sodium alginate Wet mixing 13 Water 35
Synthetic polymers
PVP Wet mixing 0.55 Water or
Dry mixing 510
Methyl cellulose Wet mixing 15 Water 215
Dry mixing 510
HPMC Wet mixing 25 Water or
Dry mixing 510
Na-CMC Wet mixing 15 Water 515
Dry mixing 510
Ethyl cellulose Wet mixing 15 Ethanol 210
Dry mixing 510
Glucose Wet mixing 225 Water 2550
Sucrose Wet mixing 225 Water 5067
Sorbitol Wet mixing 210 Water 225
110 Hamed et al.
Source: Modified from Ref. 3.
 2005 by Taylor & Francis Group, LLC
2.1.2. Pregelatinized Starch
Pregelatinized starch is a modified starch used in tablet formulations as a binder,
diluent, and disintegrant. It is obtained by chemically and mechanically processing
the starch to rupture all or part of the starch granules. This process renders starch
flowable, directly compressible, and soluble in warm water without boiling. As a
binder in a wet-granulation process, pregelatinized starch can be used either as a solution
reconstituted in water or as a dry blending followed by wetting with water. The
latter process requires two to four times more binder to achieve the same binding effect.
Pregelatinized starch is available in fully or partially pregelatinized forms. The
degree of pregelatinization determines its solubility in cold water. Cold-water soluble
fraction for a partially pregelatinized starch is 1020%. Starch 1500 is partially pregelatinized
starch containing 20% maximum cold-water soluble fraction, which makes
it useful for wet granulation. The water-soluble fraction acts as a binder, while the
remaining fraction facilitates the tablet disintegration process.
2.1.3. Gelatin
Gelatin is a mixture of purified protein fractions obtained by partial acid hydrolysis
(Type A gelatin) or alkali hydrolysis (Type B gelatin) of animal collagens. It is insoluble
in cold water and in alcohol, but is soluble in hot water. In hot water, gelatin forms
a gel on cooling to 3540C. At temperatures >40C, the system exists as a solution.
Therefore, the gelatin solutions must be used when warm to avoid gel formation.
During the preparation of gelatin solution, gelatin must be wetted in cold water
and then heated with gentle agitation to ensure dissolution. The agitation intensity must
be kept low to prevent air entrapment in the viscous solution. Use of gelatin as a binder
is limited in general-purpose tablets because it produces tablets characterized by high
hardness and slow disintegration. However, these properties of gelatin along with its
smooth mouth feel can prove to be advantageous in a lozenge formulation.
Gelatin reacts with aldehydes, aldehydic sugars, anionic and cationic polymers,
electrolytes, metal ions, plasticizers, preservatives, and surfactants. In a wet-granulation
process of a formulation containing color, the migration of dyes toward the upper
surface of the static bed during the drying operation is often amplified by the presence
of gelatin due to its high affinity for dyes. The gelatin solutions are susceptible to microbial
contamination upon storage, therefore freshly prepared solutions should be used.
2.1.4. Acacia
Acacia is a natural gum obtained from the acacia trees. It is a complex, loose
aggregate of sugars and hemicelluloses. It is commercially available in a powdered
form, a granular form, or as a spray-dried product. As a tablet binder, it is used
in an aqueous solution or added in dried form prior to moistening with water. Acacia
forms very hard tablets, which disintegrate slowly. Aqueous solutions are susceptible
to bacterial and enzymatic degradation. It is incompatible with amidopyrine, cresol,
phenol, ethanol, ferric salts, and a number of other substances. Acacia, which was
widely used in the past as a tablet binder, is rarely used today, in favor of one of
the many synthetic polymers.
2.1.5. Tragacanth
Tragacanth is a naturally occurring dried gum. It poses similar problems as those of
acacia. Dry addition to the blend followed by addition of water works better than
addition of the solution, because it is difficult to prepare and use the mucilage.
Binders and Solvents 111
 2005 by Taylor & Francis Group, LLC
2.1.6. Alginic Acid
Alginic acid is a polymannuronic acid extracted from seaweed. It is used as a binder
and disintegrating agent at concentrations between 1% and 5%. It slowly hydrolyzes
at room temperature and is insoluble in water. Therefore, it is best incorporated in a
dry state. It is incompatible with strong oxidizing agents. It forms insoluble alginates
with the alkaline earth metals and group III metals, with the exception of magnesium.
These alginates may delay disintegration of the tablets due to their gelling
2.1.7. Sodium Alginate
Sodium alginate slowly dissolves in water to form a viscous solution. A 35% solution
typically is used in wet-granulation processes. It has also been used in sustained
release formulations because it delays the dissolution of a drug from tablets. It is
hygroscopic and its aqueous solution is susceptible to microbial contamination.
2.2. Synthetic Polymers
2.2.1. Polyvinylpyrrolidone (Povidone)
Polyvinylpyrrolidone (PVP) is versatile and one of the most commonly used binders.
It is readily soluble in water and freely soluble in alcohol and many other organic
solvents. It is available in a variety of grades of different molecular weights. PVP
is generally used in the form of a solution; however, it can be added to the blends
in the dry form and then granulated in situ. It is frequently used as a binder in effervescent
and chewable tablets because the tablets manufactured using PVP as a binder
generally harden with age. Aqueous or hydroalcoholic solutions of PVP are used to
granulate insoluble materials and alcoholic solutions are used for granulating soluble
materials. It is used as a binder at concentrations between 0.5% and 5%. Low- to
medium-viscosity grades are preferred since the high-viscosity grades of PVP have
been known to cause dissolution problems. It is highly hygroscopic and picks up significant
amounts of moisture at low relative humidities and can deliquesce at high
relative humidities.
2.2.2. Methyl Cellulose
Methyl cellulose (MC) is a long-chain substituted cellulose in which 2732% of the
hydroxyl groups are in the form of methyl ether. It is available in a variety of grades
of different degrees of substitution and average molecular weight. Therefore, it offers
considerable latitude in binding strength. Efficiency of MC as a binder improves with
increasing molecular weight. Low- or medium-viscosity grades are preferred when
used as a binder. It may be added as a dry powder or as a solution. Although an
aqueous solution of 15% can be used to granulate soluble or insoluble excipients,
it is a better binder for soluble excipients such as lactose and mannitol.MC produces
granulations that compress easily. Granulation produced using 5% MC solution is
equivalent in hardness to 10% starch paste. It produces robust tablets with moderate
hardness, which does not increase with age.
MC is practically insoluble in hot water, ethanol, chloroform, ether, and saturated
salt solutions. In cold water, it swells and disperses slowly to form a clear to
opalescent, viscous dispersion. To produce an aqueous solution, the appropriate
quantity of MC powder is suspended in 25% of the required amount of water at
112 Hamed et al.
 2005 by Taylor & Francis Group, LLC
80C. The remaining amount of water is added cold, or ice water is added to the hot
slurry with vigorous stirring in order to cool it down to 20C. A clear aqueous solution
of MC is obtained. MC can also be mixed as a dry powder with another powder
prior to mixing with cold water or may be moistened with an organic solvent such as
95% ethanol prior to the addition of water.
2.2.3. Hydroxypropyl Methyl Cellulose (Hypromellose)
Hydroxypropyl methyl cellulose (HPMC) is a propyleneglycol ether of MC. HPMC
is available in a variety of viscosity grades. Its binding properties are comparable to
those of MC. Concentrations of 25% (w/w) may be used as a binder in either wetor
dry-granulation processes. HPMC is soluble in cold water and forms a viscous
colloidal solution. To prepare an aqueous solution, HPMC is first hydrated in
2030% of required water at 8090C with vigorous stirring. Cold water is added
to produce the required volume. Hydroalcoholic solutions or mixtures of water
and other water-miscible solvents such as glycol can also be used to dissolve HPMC.
HPMC is first dispersed in the organic solvent, at a ratio of five to eight parts of
solvent to one part of HPMC. Cold water is then added to produce the required
volume. HPMC is incompatible with some oxidizing agents.
2.2.4. Sodium Carboxymethyl Cellulose
Sodium carboxymethyl cellulose (Na-CMC) is a sodium salt of carboxymethyl ethers
of cellulose. It is available in a variety of molecular weights which influence the
viscosity of the solution and its swelling properties. It easily disperses in water at
all temperatures to form a clear colloidal solution. Its aqueous solubility varies with
its degree of substitution, which is the average number of hydroxyl groups substituted
per anhydroglucose unit. A 515% solution may be used for the granulation
of soluble as well as insoluble powders. The granulations produced using Na-CMC
as a binder are softer, but have good compressibility. It forms tough tablets of
moderate hardness. Na-CMC is a highly hygroscopic material. It can adsorb a large
quantity (>50%) of water at high relative humidities. Therefore, the tablets using
Na-CMC as a binder have a tendency to harden with age. Na-CMC is incompatible
with strongly acidic solutions and with metal salts of iron, aluminum, zinc, etc.
2.2.5. Ethyl Cellulose
Ethyl cellulose, an ethyl ether of cellulose, is available in a variety of grades, which
differ in their viscosity. Low-viscosity grades are used as binders in concentrations of
210% in ethanol. Ethyl cellulose produces softer granules, which compress into
tablets that easily disintegrate. However, the dissolution of the active ingredient from
these tablets may be slower because ethyl cellulose is insoluble in water. Ethyl cellulose
may be used in a dry form or in an ethanolic solution for wet granulation. Ethyl
cellulose is a good nonaqueous binder for water sensitive formulations.
2.2.6. Polyethylene Glycol
Polyethylene glycols (PEGs), by themselves, have limited binding action; however,
they can enhance the effectiveness of tablet binders and impart plasticity to granules.
They can also be used in thermoplastic granulations. In this process, a powder blend
containing 1015% (w/w) of PEG-6000 is heated to 7075C to obtain a paste-like
Binders and Solvents 113
 2005 by Taylor & Francis Group, LLC
mass, which forms granules if stirred while cooling. This technique is used in lozenge
2.2.7. Polymethacrylates (Eudragit)
Polymethacrylates (Eudragit NE 30D and Eudragit RS 30D) can be used as binders
in aqueous or nonaqueous wet-granulation processes. They are supplied as 30% aqueous
dispersions. Dilution with water prior to use is recommended. Eudragit RS 30D
is incompatible with magnesium stearate.
2.2.8. Polyvinyl Alcohol
Polyvinyl alcohols (PVAs) are available in a variety of viscosity grades. Viscosity
ranges from 10 to 100 cps lend themselves for tablet granulations. PVAs are
water-soluble polymers. They form softer granulations, which yield tablets that do
not harden with age.
2.3. Sugars
2.3.1. Glucose (Dextrose)
Glucose, when applied as a syrup in concentrations >50% in wet-granulation processes,
exhibits good bonding properties. It produces moderately strong granules
and tablets, which are hard and brittle. Glucose is also used as a direct compression
tablet diluent and binder, primarily in chewable tablets. Anhydrous dextrose adsorbs
significant amounts of moisture at 25C and 85% relative humidity to form a monohydrate.
The monohydrate also absorbs moisture at 85% relative humidity. Dextrose
is a reducing sugar and in its aldehyde form, can react with amines, amides, amino
acids, etc. Brown coloration may occur in tablets containing dextrose and strong
alkali or amines.
2.3.2. Sucrose
Sucrose is commercially available in several forms such as granular, fine granular,
fine, superfine, and confectioners sugar. The confectioners sugar is most commonly
used in wet-granulation formulations. Sucrose syrup, containing 5067% (w/w)
sucrose, is used as a binder in wet-granulation processes. It can also be used as a
dry binder where it is granulated with water or hydroalcoholic solutions. Like glucose,
sucrose produces strong, but hard and brittle tablets. The amount of binder
determines the tablet hardness. Softer granules can be obtained by using hydroalcoholic
mixtures as granulating solutions. Finely divided sugar is hygroscopic. Tablets
containing large amounts of sucrose may harden with age, which may result in
slower disintegration. In systems where quick overwetting occurs, the amount and
the rate of addition of the sucrose syrup must be carefully monitored. Use of sucrose
with starch paste may improve the tablet quality. Sucrose is incompatible with aluminum.
It hydrolyses in the presence of acids. Contamination of powdered sucrose
with heavy metals may lead to incompatibility with substances such as ascorbic acid.
2.3.3. Sorbitol
Sorbitol is an optical isomer of mannitol. It is highly hygroscopic at relative humidities
of 65% and above. Therefore, it used as a humectant in pharmaceutical formulations.
114 Hamed et al.
 2005 by Taylor & Francis Group, LLC
Whenever this property of sorbitol as a moisture control agent is desirable, it can be
used as a binder. Up to 220% of sorbitol can be added as 1025% aqueous solution
in wet-granulation formulations.
2.4. New Binders
Recently, several new natural and synthetic binders have been investigated for potential
applications in pharmaceutical formulations. Nearly all the natural binders
investigated were gums obtained from different plant origins, including Khaya
gum (5), Leucaena leucocephala seed gum (6), Anacardium occidentale gum (7), Gellan
gum (8), and a combination of detarium gum and veegum (9). While some of
these materials have been approved by the U.S. FDA for use as food additives,
the issues of compendial requirements, lack of literature support, and potential
reproducibility problems associated with materials obtained from plant origin can
limit their applicability in the pharmaceutical industry. New synthetic binders that
have been investigated include maltrodextrins (10) and chitosan derivatives (11).
The function of binders in a tablet formulation is to impart strength and to reduce
the friability of granules and tablets. Several factors influence the effectiveness of
a binder in a formulation. Some of these factors are related to the drug and other
excipients in the formulation and others are related to the binder and solvent
system used. Storage conditions as well as granulation process parameters can also
significantly affect the binder efficiency.
3.1. Properties of the Drug and Other Excipients in the Formulation
3.1.1. Particle Size
The primary particle size of the drug or excipients can affect granule strength, porosity,
and consolidation rate during high-shear granulation. Smaller particles with
higher surface areas have more contact points to allow the formation of stronger
granules than larger particles when granulated with enough binder (12,13). However,
because of the greater surface area, the solvent and binder requirements increase as
the primary particle size decreases (1416). Therefore, to better anticipate the effect
of reducing the excipients primary particle size on granule properties, the degree of
liquid saturation and binder level must be considered. At the same binder and solvent
level, excipients with larger particle sizes consolidate easier and produce less
porous granules as compared to finer particles (17,18).
Another factor that can be related to the excipients primary particle size is the
distribution of drug through different size granules. When the drug particles were
smaller than the filler, larger granule size fractions exhibited the highest drug content
and when the drug particles were larger, the highest drug concentration was found in
the smallest granule size fraction (12). The results can be explained in terms of the
ability of small drug particles to form stronger granules that resist breakdown during
high-shear granulation and grow to a further extent than those formed primarily by
larger filler particles.
Binders and Solvents 115
 2005 by Taylor & Francis Group, LLC
3.1.2. Solubility
The solubility of drugs and other excipients in the granulating solvent can signifi-
cantly affect the granules properties. Increasing the excipients solubilities in the
granulating solvent decreased the solvent requirement and led to the formation of
granules of tighter particle size distribution and reduced friability (19,20). Accordingly,
lactose-based granules are formed at a lower degree of water saturation compared
to insoluble fillers such as dicalcium phosphate (21,22) and calcium hydrogen
phosphate (23). The findings can be attributed to the increased plasticity of the moist
lactose agglomerates owing to its solubility in water, which increased the propensity
for particle coalescence and growth during high-shear granulation processes. Holm
et al. (24) described lactose granulation as less sensitive to moisture content and
kneading intensity than calcium hydrogen phosphate.
Changing the formulation proportions of water-soluble excipients can dramatically
alter the granules properties. The addition of microcrystalline cellulose to a
lactose-based formulation increased the solvent requirement and produced larger
granules for both PVP and gelatin-based granulations (25). Increasing the starch
content in a lactose:starch-based formulation leads to the formation of smaller granules
with a wider particle size distribution (26). The results were explained in terms of
water absorption by starch particles and the formation of weaker granules that could
not grow in size to the same extent as when starch was absent (27).
Drug solubility in the granulating solvent can affect its distribution in different
granule size fractions. Drugs with high solubility in the granulating solvent have a
higher tendency to migrate during drying creating a drug-rich crust. Abrasion during
subsequent milling leads to the formation of highly concentrated fines relative to
larger granules (13).
3.2. Properties Related to the Binder and Solvent System
3.2.1. Mechanical Properties of the Binder
The mechanical and film-forming properties of a binder determine the strength and
deformation behavior of a binder matrix. These properties of the binder matrix primarily
determine the effectiveness of the binder. The tensile strengths of the films of
acacia, gelatin, methylhydroxyethyl cellulose, PVP, and starch prepared with varying
moisture levels by equilibration at different relative humidities were reported (28).
mentioned earlier. The results show that acacia and PVP form weak films while gelatin
films possess the highest tensile strength. PVP was also reported to exhibit low
values of Youngs modulus showing that it is the most deformable binder (28). This
high deformability of PVP aids in consolidation during compaction. Consequently,
PVP is considered the most preferred wet binder by many (29).
3.2.2. BinderSubstrate Interactions
As mentioned earlier, significant determinants of the granule and tablet strength
are wettability of the substrate by the binder solution, binder cohesion, and
bindersubstrate adhesion (30). Rowe (3133), in a series of publications, presented
theoretical approaches to predict the bindersubstrate interactions. According to
Rowe, dense nonfriable granules are expected when the spreading coefficient of binder
over substrate is positive while negative spreading coefficient leads to the formation
of porous granules (31). For a low-polarity substrate, such as griseofulvin, either
116 Hamed et al.
Figure 1 shows the effect of moisture content on the tensile strength of the binders
 2005 by Taylor & Francis Group, LLC
PVP or starch would be the optimal binder while for high-polarity substrates, such as
theophylline, acacia or HPMC would be the optimal binder (31). PVP and HPMC
were reported to have positive spreading coefficient over lactose, acyclovir, and pentoxyfilline
with the latter having the highest spreading coefficient and acyclovir having
the lowest spreading coefficient (34). When granulated in fluid bed with both
binders, pentoxyfilline produced the least friable granules followed by lactose and
then Acyclovir. The results also showed that PVP is more efficient than HPMC
due to the lower work of cohesion and adhesion of the latter (34). This approach
is based solely on the hypothesis that optimum spreading of the binder is the main
criterion for successful formulation. It does not take into account other equally
important factors, such as disintegration, dissolution, and flow properties. However,
the studies show the potential of using the theoretical approach in binder selection
and formulation optimization.
PVP-based granulation of detergent-cleaned hydrophilic glass beads produced
larger granules with more crushing resistance and lower friability than those of
dimethylsilane surface treated hydrophobic glass beads (35). The results can be
attributed to the better adhesion of PVP to hydrophilic surfaces.
3.2.3. Binder Solution Viscosity and Surface Tension
To better understand the importance of binder solution viscosity and surface tension,
the mechanism of granule formation must be considered. When two particles
collide during granulation, they either coalesce or spring back from each other. It
is the balance of the kinetic energy of the colliding particles against the other forces
acting to bring the particles together that decides the outcome of the collision process
Figure 1 The effect of moisture content on the tensile strength of binder films. &, gelatin;
G, methylhydroxyethylcellulose; , starch; &, acacia; , PVP. The vertical bars show the
Binders and Solvents 117
limits of error of the means at P ? 0.95. (From Ref. 28.)
 2005 by Taylor & Francis Group, LLC
(36,37). The forces that help dissipate the collision energy and keep the colliding particles
together include the static pendular bridge (surface tension effect), viscous and
interparticle friction forces (17).
Increasing binder solution viscosity increased the granule size and decreased
the amount of binder required to initiate granule growth in both high-shear (37)
and fluid-bed (36) granulation processes. However, when the binder solution viscosity
is very high, problems with binder spreading and distribution arise. The highest
binder solution viscosity that can be used depends on the excipients and drug properties
and the processing parameters, such as the ability of existing pump and/or spray
systems to handle viscous liquids. Experiments in high-shear granulators with initial
binder solution viscosities exceeding 100 mPa.s have been reported (38).
When different molecular weights of PVP and HPMC were dissolved in water,
reductions in surface tension were observed (39). The decrease in surface tension was
higher for HPMC but independent of the concentrations of both binders and inversely
dependent on the molecular weight of PVP (39). Decreasing binder solution surface
tension decreased the capillary suction pressure and decreased the friction
resistance to consolidation and, therefore, increased the granule consolidation rate,
but also increased the minimal attainable granule porosity, i.e., granules consolidated
fast but not as far (18). Decreasing the binder solution surface tension was also
reported to decrease the liquid requirement to attain overwetting (40).
3.2.4. Solvent Properties
Granule properties can be significantly affected by changing the granulating solvent.
The effect stems from the change in excipient solubility and wettability, as well as the
mechanism of granule consolidation. In the pharmaceutical industry, only aqueous,
hydroethanolic, and ethanolic solvent systems are widely used. Lactose granulations
prepared using ethanolic PVP solution had higher porosity and friability compared
to those prepared using water as a granulating solvent (41). For hydroethanolic solvent
systems, increasing the ethanol content has been shown to increase tablet
strength (41). The results were attributed to the increased fragmentation tendency
of granules when ethanol contents were increased (42). Changing the solvent system
can affect the formulation excipients wettability and influence binder distribution.
For example, PVP distribution in low-shear granulation process improved when
water was replaced with a hydroalcoholic solution (43).
3.3. Storage Conditions
Storage conditions can severely alter the tablet properties through their effect on the
binders physical characteristics. For example, the hardness and disintegration time
of Ranitidine tablets prepared using the PVP wet granulation method decreased
when the tablets were stored under elevated moisture levels (44). PVP can absorb
a significant quantity of water when exposed to elevated humidity, which decreases
the polymer glass transition temperature (45,46). PVP is proposed to stay in a glassy
stage at room temperature at a relative humidity 55% and completely convert
to the rubbery state at room temperature at a relative humidity >75% (45). Such
changes in the physical state of a binder can dramatically change a tablets hardness,
friability, disintegration, and drug release. Moreover, the presence of other additives
in the formulation can augment the plasticizing effect of moisture and cause the granule
and tablet properties to change at a lower moisture level than those used in accelerated
stability protocols.
118 Hamed et al.
 2005 by Taylor & Francis Group, LLC
HPC was reported not to have a glass transition temperature, but rather melt
at a temperature of 190195C (46). When compared to PVP, tablets prepared using
HPC maintained its dissolution profile after storage for 24 weeks under different
moisture/temperature conditions while those prepared using PVP showed a reduction
in pyrydazine release when stored at 40C/75% relative humidity (46).
4.1. Polyvinylpyrrolidone
4.1.1. High-Shear Granulation
The granulation process in high-shear equipment is influenced by a score of variables
including impeller speed, amount of binder and granulating solvent, liquid addition
rate, wet massing time, method of binder incorporation (wet vs. dry), method of
solution addition (dripping vs. atomization), equipment design, and chopper speed.
These parameters influence the binder performance through their combined effects
on the binder distribution/dissolution, the degree of saturation, and the extent to
which granules are allowed to grow and densify before they are broken down by
the high-speed chopper. Impeller Speed. Impeller speed is the most significant parameter
affecting the binder performance in high-shear granulation. Increasing the impeller
speed in PVP and PVPPVA copolymer-based formulations increases the granule
size (21,22,4749) and the bulk and tapped density (21,49), but decreases the percent
fines (47,49) and the granule porosity (21,49). However, the effect of impeller speed
on binder performance is dependent on the degree of solvent saturation, the binder
amount, the stage at which the impeller speed is varied (liquid addition vs. wet massing),
and the properties of the formulation materials. At a low degree of solvent
saturation or with an insufficient amount of binder, the effect of increasing impeller
speed is in favor of breaking up the weak nuclei/agglomerates and hindering any
further granule growth (26,50). PVP Concentration.
dicalcium phosphate granules vs. binder concentration for PVP in comparison to
other binders (51).
As the binder concentration increases, the crushing strength of the granules
increases. The plot also shows that starch and gelatin can produce much stronger
granules with lower concentrations compared to acacia, PVP, or PEG 4000 (51).
As expected, the particle size of granules also increases as the binder concentration
concentration (25).
The intragranular bonds formed during granule drying that contribute to the
tablet strength include cohesion of binder film and binder substrate adhesion. Jarosz
and Parrott (52) showed that the radial and axial tensile strengths of dicalcium
phosphate tablets increased with increasing concentrations of PVP in the formula-
High molecular weight PVP produces larger calcium hydrogen phosphate granules
than hydrolyzed gelatin and low-viscosity HPMC (38). When compared to PVP,
gelatin-based granules start growing at a lower degree of liquid saturation and the
dependence of granule size on binder concentration was more prominent (38). Cutt
et al.(53) compared glass ballotini granules prepared using PVP, hydrolyzed gelatin,
and HPMC. PVP produced the least friable granules at lower concentrations and
Binders and Solvents 119
Figure 2 shows a plot of crushing strength of
increases. Figure 3 shows a plot of mean particle size of lactose granules vs. binder
tion (Fig. 4).
 2005 by Taylor & Francis Group, LLC
HPMC produced the most friable and smallest granules owing to its lower work of
adhesion and cohesion (53). Gelatin produced the strongest granules that needed
more work to compress due to their higher resistance to deformation (53). Liquid Addition Rate and Wet Massing. Liquid addition rate can significantly
influence binder distribution and granular uniformity. Lower water addition
rate improves PVPPVA copolymer distribution and reduces the liquid
requirement to obtain an acceptable granulation (26,54). Hydroalcoholic and alcoholic
solutions can behave differently due to heat generation and faster evaporation
in high-shear mixers. The effect of liquid addition rate is dependent on the solubility
of the formulation excipients in the solvent used. Slower PVP solution addition rate
enhances granule growth and decreases porosity of the insoluble dicalcium phosphate
granulations (21); lactose-based formulations, on the other hand, are less sensitive
to liquid addition rate (22,49). Increasing the wet massing time of PVP-based
formulations increases granules size and density, while decreasing the percentage of
fines and granule porosity (16,21,49,54). Method of Binder Incorporation. PVP can be incorporated in the
granulation either by dry or by wet addition methods. The method of binder incorporation
affects the granules properties through its effect on binder dissolution and
Figure 2 Crushing strength of wet granulated dicalcium phosphate. Binders: , gelatin;
120 Hamed et al.
& , potato starch mucilage; , acacia; D, povidone; and , PEG 4000. (From Ref. 51.)
 2005 by Taylor & Francis Group, LLC
distribution. However, given the good hydration and high solubility in combination
with the use of high impeller and chopper speed, PVP is usually uniformly distributed
when incorporated by either method using water, hydroalcoholic, or alcoholic
solvents. Nevertheless, a higher PVP level may be needed with the dry addition
method to obtain similar granulation results.
4.1.2. Fluid-Bed Granulation PVP Concentration and Molecular Weight. Binder solution concentration
is one of the most crucial parameters controlling granule properties in fluid-bed
granulations. Increasing PVP concentration increases the granule size, narrows the
granule size distribution, and improves granule flowability and strength (55,56).
The increase in granule size can be related to the increase in binder solution viscosity
with a subsequent increase in droplet size together with an increase in the
force binding the particles at the higher binder concentration.
However, granule size enlargement with increasing PVP level reaches a plateau
at a critical PVP concentration (25,34). Increasing the PVP level beyond this critical
concentration does not lead to further growth of granules but rather widens the size
distribution and decreases the friability of the granules (25).
4.1.3. Roller Compaction PVP Concentration. The effect of binder concentration on the
properties of granules and tablets prepared by roller compaction is dependent on
other processing parameters including feed rate, roll pressure, and speed. Increasing
the binder level increases tablet hardness and decreases their friability due to the
Figure 3 Plot of the mean granule size as a function of binder concentration for the granu-
Binders and Solvents 121
lation of lactose. (From Ref. 25.)
 2005 by Taylor & Francis Group, LLC
imparted plasticity of the formed granules (57). However, at high binder level,
increasing the roll pressure increases the tablet friability and decreases the tablets
hardness due to the formation of less friable granules with higher fragmentation
resistance during compression (25). When the PVP level was reduced to 1%,
the granules properties were dependent on the drug properties (58). Increasing
the roll pressure in this case had no effect on tablet hardness or friability.
However, increasing the roll and feed rate can lead to a slight decrease in tablet
friability (58).
4.1.4. Extrusion/Spheronization
Several processing factors can affect the binder performance during extrusion and
spheronization processes including hydration level, extruder speed, shape, hole size,
spheronizer speed, and residence time. While the use of microcrystalline cellulose
alone or in combination with lactose is widely adopted, the addition of binders can
aid in the formation of a plastic mass than can be easily extruded and spheronized.
Increasing the PVP or starch/gelatin level increases the pellet size and decreases their
friability (59).
Figure 4 The influence of povidone on the tensile strength of tablets of dibasic calcium phosphate
dihydrate compressed at 2268 kg (______) and at 4536 kg (- - - - -). D, axial; , radial.
122 Hamed et al.
(From Ref. 52.)
 2005 by Taylor & Francis Group, LLC
When microcrystalline cellulose was spray dried with PVP, HPC, HPMC, Na-
CMC, and pregelatinized starch, the properties of the produced pellets were
improved in terms of yield, sphericity, and surface smoothness (60). PVP and
HPC were described as superior to other binders in terms of water requirements,
process sensitivity, and yield (60).
4.2. Cellulosic Polymers
4.2.1. High-Shear Granulation Polymer Concentration. The effect of cellulosic polymer concentration
on granule properties is dependent on their method of addition. In the wet addition
method, increasing the concentration of HPC, different viscosity grades of
HPMC and MC in lactosecornstarchmicrocrystalline cellulose systems increases
the percentage of binder-rich oversize granules (61). The effect can be attributed
to the increase in binder solution viscosity and heterogenous binder distribution.
In the dry addition method, the effect of increasing polymer concentration on a
granules properties depends on the nature of the polymer and its molecular weight/
viscosity. Increasing the concentration of HPC and low-viscosity HPMC (3 cP)
increases the median granule size (61). Increasing the concentration of mediumand
relatively high-viscosity HPMC (6 and 15 cP) and MC is accompanied by a
slight decrease in granule median size and increase in the percentage of oversized
granules (61). The different behavior can be related to the extent and rate of binder
dissolution as well as binder adhesiveness at different degrees of hydration.
4.2.2. Fluid-Bed Granulation Polymer Concentration. Similar to a high-shear granulation process,
the effect of increasing HPMC and HPC concentrations on granule properties is
dependent on their method of addition in fluid-bed granulation. The polymers are
more uniformly distributed when added by the wet addition method and the granule
size and strength increase with increasing the polymer concentrations (62,63). In the
dry addition method, HPMC is more concentrated in the larger granules and
the granule size is independent of the concentration of HPMC (62,64). Agitating
the fluid bed, however, showed no difference in HPMC and HPC distributions
between the wet and dry addition methods (65). In general, HPC produced stronger
granules than HPMC when added in the dry state since it maintains its flexibility and
adhesiveness at low moisture content, which allows more binding during drying (65).
Increasing the level of MC, on the other hand, is not accompanied by an increase in
granule size or strength when it is included by both dry and wet addition methods
(65). MC has the lowest thermal gelation temperature and thus it loses its adhesiveness
at the low moisture level applied in fluid-bed granulation.
4.3. Gelatin
Gelatin was one of the most commonly used binders during the 1960s and 1970s.
Recently, the use of gelatin in pharmaceutical formulations has significantly reduced.
Many factors contributed to the lower popularity of gelatin including its high
Binders and Solvents 123
 2005 by Taylor & Francis Group, LLC
binding ability and the resultant slow tablet disintegration, and perhaps most importantly,
the introduction of new semisynthetic binders whose methods of preparation
are easier and exhibited better reproducibility and reliability than gelatin.
In gelatin-based formulations, granule properties are highly sensitive to the
level of gelatin used (54,66). Gelatin produces stronger granules than CMC, PEG
(67,68), MC (69), acacia, and PVP (25) due to the high tensile strength of the gelatin
film compared to other binders. In fluid-bed granulation processes, increasing the
gelatin concentration increases the granular size up to a certain critical concentration
beyond which the granule size becomes independent of gelatin concentration (25).
The granulation method can affect the properties of granules and tablets prepared
using gelatin through their effect on gelatin distribution within the granules.
Acetaminophen granules and tablets prepared by wet massing, roller compaction,
and spray drying of the substrate-binder slurry were compared (7072). Wet massing
produced a sponge-like binder matrix embedding substrate particles while the roller
compaction produced a distribution of discrete binder particles in the substrate
particles. Spray drying coated the granules with an outer shell of the binder and
produced superior tablets when compared to wet massing and roller compaction
4.4. Starch
Despite the reproducibility problems associated with a variety of starch sources and
the method of paste formation, starch paste is still a relatively popular binding agent
in the pharmaceutical industry owing to its superior compatibility with many drugs.
The granule properties are greatly affected by the method of starch paste preparation,
in particular the gelatinization temperature. Starches obtained from different
sources start gelatinization at different temperatures (corn starch starts at 75C
and potato starch starts at 60C) and yield pastes with different concentration gradient
viscosities (73). Increasing the temperature at which starch is gelatinized
increases the degree of gelatinization with a subsequent increase in tablet hardness
(73). Variability in ascorbic acid tablet properties decreases when starch paste was
prepared and pumped under precise temperature control (74).
Starch paste concentration can also affect granule properties. When compared
with diluted paste (5%, w/w) with good fluidity, a concentrated paste (15%, w/w)
with high viscosity produced lactose granules with a sharper particle size distribution
when using a high-shear mixer (75). Nevertheless, high impeller speed and longer
processing time was needed for deaggregation and a better distribution of the
concentrated starch paste. However, diluted starch paste produced stronger granules,
better compressibility, and longer disintegration time (76).
Starch can also be used as a disintegrant owing to its wicking and swelling
properties. Starch paste by itself is not sufficient to introduce effective disintegration
and a disintegrant, starch or other, is usually recommended to enhance the effect of
the starch introduced as a paste. When the quantity of disintegrant starch exceeds
that of paste starch, the disintegration time decreases as the compression force
increases. When paste starch is used in higher quantity than dry, the disintegration
time increases with compression force. The effect can be related to different disintegration
mechanisms introduced by disintegrant and starch pastes (77).
When starch is dextrinized by the addition of a-amylase, the resultant paste
produces stronger granules with better flowability that produce tablets with shorter
disintegration time, lower friability, and less variability in weight and hardness.
124 Hamed et al.
 2005 by Taylor & Francis Group, LLC
However, excessive dextrinization leads to the formation of paste with lower adhesive
and tackiness properties and poor binding capabilities (78). Starch can be also
modified through pregelatinization or cross-linking. Cross-linking does not improve
the granulation ability of conventional starch. When compared to conventional
and cross-linked starches, pregelatinized starch yields coarser granules with lower
friability (79).
1. Wade A, Walker PJ, eds. Handbook of Pharmaceutical Excipients. 2nd ed. London: The
Pharmaceutical Press, 1994.
2. Lieberman HA, Lachman L, Schwartz JB, eds. Pharmaceutical Dosage Forms: Tablets.
Vol. 1. 2nd ed. New York: Marcel Dekker, Inc., 1989.
3. Kristensen HG. Binders. Encycl Pharm Technol 1993; 451465.
4. Mendes RW, Roy SB. Tableting excipients, part II. Pharm Tech 1978; 3:6174.
5. Odeku OA, Itiola OA. Evaluation of Khaya gum as a binder in a paracetamol tablet
formulation. Pharm Pharmacol Commun 1998; 4:183188.
6. Deodhar UP, Paradkar AR, Purohit, AP. Preliminary evaluation of leucaena leucocephala
seed gum as a tablet binder. Drug Dev Ind Pharm 1998; 24(6):577582.
7. Chukwunweike Onukwo G, Okoye J. Evaluation of anacardium occidentale gum as
a binder in lactose based tablet formulations. Bole Chim Farm 1997; 136:569574.
8. Antony PJ, Sanghavi NM. A new binder for pharmaceutical dosage forms. Drug Dev
Ind Pharm 1997; 23(4):417418.
9. Attama AA, Adikwu MU, Esimone CO. Binary combination of detarium gum and veegum
as a binder in sodium salicylate tablets. Bole Chim Farm 1999; 138:199202.
10. Soyeux P, Delacourte A, Delie B, Lefevre P, Boniface B. Influence of optimization of
operating parameters with a new binder in wet granulation. I: use in powder form.
Eur J Pharm Biopharm 1998; 46:95103.
11. Aiedeh KM, Taha MO. Synthesis of chitosan triethylene glycol phthalate and its evaluation
as a binder in wet granulation procedures. Die Pharmazie 1999; 54:614619.
12. Vromans H, Poels-Janssen HGM, Egermann H. Effect of high shear granulation on
granulate homogeneity. Pharm Dev Technol 1999; 4:297303.
13. Van den Dries K, Vromans H. Relationship between inhomogeneity phenomenon and
granule growth mechanisms in a high-shear mixer. Int J Pharm 2002; 247:167177.
14. Paris L, Stamm A. Optimal massing liquid volume determination by energy consumption
measurement: study of the influence of some physical properties of solvent and products
used. Drug Dev Ind Pharm 1985; 11(2,3):361386.
15. Tapper GI, Lindberg NO. Granulation of some lactose qualities with different particle
size distributions in a domestic type mixer. Acta Pharm Suec 1986; 23:4756.
16. Schaefer T, Holm P, Kristensen HG. Wet granuation in laboratory scale high shear
mixer. Pharm Ind 1990; 52(9):11471153.
17. Iveson SM, Lister JD, Ennis BJ. Fundamental studies of granule consolidation part 2:
effects of binder content and binder viscosity. Powder Technol 1996; 88:1520.
18. Iveson SM, Lister JD. Fundamental studies of granule consolidation part 2: quantifying
the effects of particles and binder properties. Powder Technol 1998; 99:243250.
19. Wells JI, Walker CV. The influence of granulating fluids upon granule and tablet properties:
the role of secondary binding. Int J Pharm 1983; 15:97.
20. Dias VH, Pinto JF. Identification of the most relevant factors that affect and reflect the
quality of granules by application of canonical analysis. J Pharm Sci 2002; 91(1):273281.
21. Schaefer T, Holm P, Kristensen HG. Comparison between granule growth in a horizontal
and a vertical high speed mixer. I. Granulation of dicalcium phosphate.. Arch Pharm
Chem Sci Ed 1984; 14:116.
Binders and Solvents 125
 2005 by Taylor & Francis Group, LLC
22. Schaefer T, Holm P, Kristensen HG. Comparison between granule growth in a horizontal
and a vertical high speed mixer. II. Granulation of lactose. Arch Pharm Chem Sci Ed
1986; 14:1729.
23. Kristensen HG, Holm P, Jaegerskou A, Schaefer T. Granulation in high speed mixers.
Part 4: effect of liquid saturation on the agglomeration. Pharm Ind 1984; 46(7):763767.
24. Holm P, Jaegerskou A, Schaefer T, Kristensen HG. Granulation in high speed mixers.
Part 1: effect of process variables during kneading. Pharm Ind 1983; 45(8):806811.
25. Rohera BD, Zahir A. Granulations in a fluidized-bed: effect of binders and their concentrations
on granule growth and modeling the relationship between granules size
and binder concentration. Drug Dev Ind Pharm 1993; 19(7):773792.
26. Holm P, Schaefer T, Larsen C. End point detection in a wet granulation process. Pharm
Dev Technol 2001; 6(2):181192.
27. Holm P. Effect of impeller and chopper design on granulation in a high speed mixer.
Drug Dev Ind Pharm 1987; 13(911):16751701.
28. Healey JNC, Rubinstein MH, Walter V. The mechanical properties of some binders used
in tableting. J Pharm Pharmac 1974; (suppl 26):41.
29. Shangraw RF, Demarest DA Jr. A survey of current industrial practices in the formulation
and manufacture of tablets and capsules. Pharm Tech 1993; 17:32.
30. Readind SJ, Spring MS. The effects of binder film characteristics on granule and tablet
properties. J Pharm Pharmac 1984; 36:421424.
31. Rowe RC. Bindersubstrate interactions in granulation: a theoretical approach based on
surface free energy and polarity. Int J Pharm 1989; 52:149154.
32. Rowe RC. Surface free energy and polarity effects in the granulation of a model system.
Int J Pharm 1989; 53:7581.
33. Rowe RC. Polar/non-polar interactions in the granulation of organic substrates with
polymer binding agents. Int J Pharm 1989; 56:117122.
34. Planinsek O, Pisek R, Trojak A, Srcic S. The utilization of surface free-energy parameters
for the selection of a suitable binder in fluidized bed granulation. Int J Pharm 2000;
35. Cutt T, Fell JT, Rue PJ, Spring MS. Granulation and compaction of a model system. I.
Granule properties. Int J Pharm 1986; 33:8187.
36. Ennis BJ, Tardos G, Pfeffer R. A microlevel-based characterization of granulation phenomena.
Powder Technol 1991; 65:257272.
37. Hoornaert F, Wauters PAL, Meesters GMH, Pratsinis SE, Scarlett B. Agglomeration
behavior of powders in a lodige mixer granulator. Powder Technol 1998; 96:116128.
38. Ritala M, Jungersen O, Holm P, Schaefer T, Kristensen HG. A comparison between binder
in the wet phase of granulation in a high shear mixer. Drug Dev Ind Pharm 1986;
39. Parker MD, York P, Row RC. Bindersubstrate interaction in wet granulation. 2: the
effect of binder molecular weight. Int J Pharm 1991; 72:243249.
40. Pepin X, Blanchon S, Couarraze G. Power consumption profiles in high-shear granulation.
I: liquid distribution in relation to powder and binder properties. J Pharm Sci 2001;
41. Wikberg M, Alderborn G. Compression characteristics of granule materials. VII. The
effect of intragranular binder distribution on the compactibilty of some lactose granulations.
Pharm Res 1993; 10(1):8894.
42. Alderborn G. Granule properties of importance to tableting. Acta Pharm Suec 1988;
43. Shah NH, Railkar AS, Phuapradit W, Zeng FW, Chen A, Infeld MH, Malick W. Effect
of processing techniques in controlling the release rate and mechanical strength of hydroxypropyl
methylcellulose based hydrogel matrices. Eur J Pharm Biopharm 1996; 42(3):
126 Hamed et al.
 2005 by Taylor & Francis Group, LLC
44. Uzunarslan K, Akbuga J. The effect of moisture on the physical characteristics of ranitidine
hydrochloride tablets prepared by different binders and techniques. Drug Dev Ind
Pharm 1991; 17(8):10671081.
45. Kiekens F, Zelko R, Remon JP. Effect of the storage conditions on the tensile strength of
tablets in relation to the enthalpy relaxation of the binder. Pharm Res 2000; 17(4):
46. Fitzpatrick S, McCabe JF, Petts CR, Booth SW. Effect of moisture on polyvinylpyrrolidone
in accelerated stability testing. Int J Pharm 2002; 246:143151.
47. Lindberg NO, Jo?nsson C. The granulation of lactose and starch in a recording high-speed
mixer, Diosna P25. Drug Dev Ind Pharm 1985; 11(2,3):387403.
48. Vojnovic D, Chicco D, El Zenary H. Doehlert experimental design applied to optimization
and quality control of a granulation process in a high shear mixer. Int J Pharm 1996;
49. Badawy SIF, Menning MM, Gorko MA, Gilbert DL. Effect of process parameters on
compressibility of granulation manufactured in high-shear mixer. Int J Pharm 2000;
50. Vojnovic D, Selenati P, Rubessa F, Moneghini M. Wet granulation in a small scale high
shear mixer. Drug Dev Ind Pharm 1992; 18(9):961972.
51. Armstrong NA, March GA. J Pharm Sci 1976; 65:200.
52. Jarosz PJ, Parrott EL. Factors influencing axial and radial strength of tablets. J Pharm
Sci 1982; 71:607610.
53. Cutt T, Fell JT, Rue PJ, Spring MS. Granulation and compaction of a model system. III.
Compaction properties of the granules. Int J Pharm 1989; 49:157161.
54. Lindberg NO, Jonsson C. Granulation of lactose in a recording high-speed mixer, Diosna
P25. Drug Dev Ind Pharm 1983; 9(6):959970.
55. Aulton M, Banks M. The factors affecting fluidized bed granulation. Manuf Chem Aeros
News December 1978.
56. Wan LSC, Heng PWS, Ling BL. Fluidized bed granulation with PVP K90 and PVP
K120. Drug Dev Ind Pharm 1995; 21(7):857862.
57. Sheskey PJ, Dasbach TP. Evaluation of various polymers as dry binders in the preparation
of an immediate-release tablet formulation by roller compaction. Pharm Tech
58. Murray M, Laohavichien A, Habib W, Sakr A. Effect of process variables on rollercompacted
ibuprofen tablets. Pharm Ind 1998; 60(3):257262.
59. Varshosaz J, Kennedy RA, Gipps EM. Effect of binder level and granulating liquid on
phenylbutazone pellets prepared by extrusionspheronization. Drug Dev Ind Pharm
1997; 23(6):611618.
60. Law MFL, Deasy PB. Use of hydrophilic polymers with microcrystalline cellulose to
improve extrusionspheronization. Eur J Pharm Biopharm 1998; 45:5765.
61. Kokubo H, Nakamura S, Sunada H. Effect of several cellulosic binders on particle size
distribution of granules prepared by a high-speed mixer. Chem Pharm Bull 1993;
62. Kokubo H, Nakamura S, Sunada H. Effect of several cellulosic binders on particle
size distribution in fluidized bed granulation. Chem Pharm Bull 1995; 43(8):
63. Xu, G, Sunada H, Zhang R. Indomethacin controlled release matrix tablet prepared by
wet granulation procedure. Chem Pharm Bull 1997; 45(1):214216.
64. Dasbach T, Sheskey P, Caswell O. Evaluation of the Use of Methocel in Fluid Bed Granulation
as a Dry Binder. Dow Product Information Publications, 2002.
65. Kokubo H, Nakamura S, Sunada H. Effect of several cellulosic binders and method of
their addition on the properties and binder distribution of granules prepared in an agitating
fluidized bed. Chem Pharm Bull 1998; 46(3):488493.
66. Nagykaldi VA, Gyarmati L. Untersuchungen zur feuchtgranulation im diosna-mscher V
25. Pharm Ind 1980; 42(5):526531.
Binders and Solvents 127
 2005 by Taylor & Francis Group, LLC
67. Solvang S, Finholt P. Effect of tablet processing and formulation factors on dissolution
rate of the active ingredient in human gastric juice. J Pharm Sci 1970; 59(1):4952.
68. El-Gindy NA, Smaha MW, El-Marandy HA. Evaluation of binder activities on the physical
properties and compression characteristics of granules prepared by two different
modes. Drug Dev Ind Pharm 1988; 14(7):9771005.
69. Schaefer T, Worts O. Control of fluidized bed granulation. IV. Effect of binder solution
and atomization on granule size and size distribution. Arch Pharm Chem Sci Ed 1978;
70. Seager H, Burt I, Ryder J, Rue PJ, Murray S, Beal N, Warrack JK. Int J Pharm Tech
Prod Manuf 1998; 1:36.
71. Seager H, Rue PJ, Burt I, Ryder J, Warrack JK. Int J Pharm Tech Prod Manuf 1981;
72. Rue PJ, Seager H, Ryder J, Burt I. Int J Pharm Tech Prod Manuf 1980; 1:2.
73. Makino T, Kitamori N. Importance of gelatinization degree of starch paste binder in
hardness and disintegration of tablets. Chem Pharm Bull 1995; 43(3):514516.
74. Makino T, Wada N, Kitamori N. Preparation temperature in fluidized-bed granulation
on resultant tablet properties. Chem Pharm Bull 1995; 43(3):473475.
75. Makino T, Kitamori N. Effect of starch paste concentration on particle size distribution
of fine granules produced by agitation granulation. Chem Pharm Bull 1995; 43(7):
76. Hill PM. Effect of compression force and corn starch on tablet disintegration time.
J Pharm Sci 1975; 65(11):16941696.
77. Hill PM. Starch paste granulation: binder dilution effects on granulations and tablets.
J Pharm Sci 1976; 65(2):313314.
78. Bandyopadhyay AK, Chaudhuri B, Bhattacharjee PK. Starch paste granulation I: effect
of dextrinized corn starch mucilage as binding agent on granulation and tablet quality.
Aust J Pharm Sci 1980; 9(3):8589.
79. Visavarungroj N, Herman J, Remon JP. Crosslinked starch as binding agent. I. Conventional
wet granulations. Int J Pharm 1990; 59:7378.
128 Hamed et al.
 2005 by Taylor & Francis Group, LLC
Spray Drying and Pharmaceutical
Metin C? elik
Pharmaceutical Technologies International, Inc., Belle Mead, New Jersey, U.S.A.
Susan C. Wendel
Elan NanoSystems, King of Prussia, Pennsylvania, U.S.A.
Spray drying is one of the oldest forms of drying and one of the few technologies
available for the conversion of a liquid, slurry, or low-viscosity paste to a dry solid
ing process schematically. The simplicity and flexibility of the spray-drying process
make it ideal for handling a wide variety of pharmaceutical products.
1.1. Background
The first detailed description of the drying of products in spray form was mentioned
in a patent of 1872 entitled Improvement of Drying and Concentration of Liquid
Substances by Atomizing (2). However, this process found its first significant applications
in the milk and detergent industries in the 1920s (3). In current times, spray
drying is utilized extensively in many aspects of our daily life from food products,
cosmetics, and pharmaceuticals to chemicals, fabrics, and electronics. Typical
pharmaceutical examples include spray-dried enzymes (such as amylase, protease,
lipase, and trypsin), antibiotics (such as sulfathiazole, streptomycin, penicillin, and
tetracycline) and many other active pharmaceutical ingredients, vitamins (such as
ascorbic acid and vitamin B12), and excipients for direct compression (such as lactose,
mannitol, and microcrystalline cellulose).
1.2. Advantages and Limitations
There are several reasons why the technology of spray drying has found many applications
in numerous industries. It is a continuous process. As long as liquid feed can
continue to be supplied to the drying system the spray-dried product will continue to
be produced. In some instances, this process has been operated for months without
interruption. The physical properties of the resulting product (such as particle size
(free-flowing powder) in one unit operation (1). Figure 1 shows a general spray-dry-
 2005 by Taylor & Francis Group, LLC
and shape, moisture content, and flow properties) can be controlled through the
selection of equipment choices and the manipulation of process variables. The actual
spray-drying process is almost instantaneous as the major portion of the evaporation
takes place in as short a time as milliseconds or a few seconds at most, depending on
the design of the equipment and process conditions. This makes spray drying well
suited for heat-sensitive products. In addition, corrosive and abrasive materials
can be readily accommodated because the contact between the mechanical parts
and materials is minimal as compared to other granulation processes. Also, spray
dryers have few moving parts. In fact, careful selection of various components can
result in a system having no moving parts in direct contact with the product. Operation
requirements of small and large dryers are the same. This makes spray drying a labor
cost-effective process, especially for high-volume products. Last, the spray-drying
process can be fully automated. Commercial-scale spray dryers are controlled by programmable
logic controllers or solid-state controllers. These control systems monitor
exhaust air temperature or humidity and provide an input signal that, by way of a
setpoint, modulates the energy supplied to the process (4).
Like all other granulation processes, spray drying also has some limitations.
For example, it is not typically well suited for producing granules with mean particle
temperatures and the exhaust air stream contains heat, which often requires sophisticated
heat exchange equipment for removal.
The spray-drying process is carried out in three fundamental stages as shown
Figure 1 A schematic representation of a general spray-drying process with primary and secondary
product separation.
130 C? elik and Wendel
size >200 mm as shown in Figure 2. It also has poor thermal efficiency at lower inlet
schematically in Figure 3. The first stage is atomization of a liquid feed into fine
 2005 by Taylor & Francis Group, LLC
droplets. In the second stage, spray droplets mix with a heated gas stream and the
dried particles are produced by the evaporation of the liquid from the droplets.
The final stage involves the separation of the dried powder from the gas stream
and collection of these powders in a chamber. The second stage (i.e., the mixing
and drying step) has also been considered as separate steps (5). The following sections
detail each of these stages, their process parameters, and related equipment details.
2.1. Atomization
Atomization is the process by which a liquid is disintegrated into many fine droplets.
The formation of a spray with high surface/mass ratio is highly critical for optimum
liquid evaporation conditions and, consequently, the desired properties of the resulting
product. Although ideally the sizes of all droplets should be the same, in practical
terms formation of droplets with a narrow size distribution would be satisfactory.
2.1.1. Atomizer Types and Designs
Formation of the atomized spray requires application of a force. The commercially
available systems employ one of the following in order to create an atomized spray:
centrifugal energy, pressure energy, kinetic energy or sonic energy, and vibrations. Centrifugal Atomizers. Centrifugal atomizers utilize either a rotating
disk or a wheel to disintegrate the liquid stream into droplets (6). Examples of rotary
wide variety of spray characteristics can be obtained for a given product through
combinations of feed rate, atomizer speed, and atomizer design. The droplet size distribution
is fairly narrow for a given method and process conditions but the mean
droplet size can be varied from as small as 15 mm to as large as 250 mm, depending
on the amount of energy transmitted to the liquid. Larger mean sizes require larger
drying chamber diameters. Wheels are well suited for producing sprays in the fine to
medium-coarse size range while disks are used to produce coarse sprays.
Rotary atomizers normally operate in the range of 500025,000 rpm with wheel
diameter of 550 cm. The mean size of the droplet produced is inversely proportional
Figure 2 Range of mean particle sizes achievable by control of atomizer operation at low to
medium feed rates.
Spray Drying and Pharmaceutical Applications 131
atomizers are shown in Figure 4. These devices form a low-pressure system and a
 2005 by Taylor & Francis Group, LLC
to the wheel speed and directly proportional to the feed rate and its viscosity. Solid
content and surface tension are the other factors having minor effects on the droplet
size. For example, an increase in feed rate may slightly increase the particle size, but
the use of a variable-speed drive on the centrifugal atomizer facilitates correction to
the specified size.
Centrifugal atomizer designs include wheels with vanes or bushings and vaneless
disks. Vaned atomizer wheels produce sprays of high homogeneity and are the
most commonly used as compared to other designs. In this type of atomizer, liquid
fed onto a wheel moves across the surface until contained by the rotating vane.
Figure 3 Schematic of spray-drying process shown in stages: stage I: atomization; stage II:
sprayair contact and evaporation; stage III: product separation. (From Ref. 5.)
132 C? elik and Wendel
 2005 by Taylor & Francis Group, LLC
The liquid flows outward under the influence of centrifugal force and spreads over
the vane, wetting the vane surface as a thin film. At very low liquid vane loadings,
the thin film can split into streams. No liquid slippage occurs on a wheel once liquid
has contacted the vanes. Whether radial or curved, the vanes prevent transverse flow
of liquid over the surface. Abrasive materials are best handled using atomizer wheels
with bushings. Since the feed material is in direct contact with rotating parts,
the bushings feature wear-resistant surfaces and require additional maintenance.
Vaneless (disk) designs are often applied when coarse powders are required at high
production rates.
Bulk pharmaceutical excipients and fine chemicals, such as antacids, are often
produced using centrifugal atomizers. The particles produced by this technique are
generally free flowing and, unless intentionally produced with very fine atomization,
dust free. The porous structure of the particles provides increased solubility and the
relatively low density and friability of these particles result in generally good compaction
properties. Also, the batch-to-batch reproducibility and dryer-to-dryer transferability
of this technique are excellent. As mentioned earlier, if larger spray-dried
particles are desired, larger production drying chambers must be employed. Kinetic Energy Nozzles. Kinetic energy is applied in the form of two-
fluid or pneumatic atomization. This is the most commonly used atomization technique
within the pharmaceutical industry. Here, atomization is accomplished by the
interaction of the liquid with a second fluid, usually compressed air. High air velocities
are generated within the nozzle for effective feed contact, which breaks up the feed
into a spray of fine droplets. Neither the liquid nor the air requires very high pressure,
with 200350 kPa being typical. A typical two-fluid nozzle is shown in Figure 5.
Particle size is controlled by varying the ratio of the compressed airflow to that
of the liquid. The main advantage of this type of atomization is that the liquid has a
relatively low velocity as it exits the nozzle; therefore, the droplets require a shorter
Figure 4 Common rotary atomizers. (Courtesy of Niro Pharma Systems.)
Figure 5 Schematic presentation of a typical two-fluid nozzle. (Courtesy of Niro Pharma
Spray Drying and Pharmaceutical Applications 133
 2005 by Taylor & Francis Group, LLC
flight path for drying. Because many pharmaceutical applications use relatively small
spray dryers, pneumatic nozzles are often used. Another advantage is the simple
design that lends itself to easy cleaning, sterile operation, and minimal contamination.
Pneumatic nozzles can be designed to meet the most stringent requirements
for sterile or aseptic applications. Special consideration must be given to supplying
a sterile source of compressed air for atomization.
Another type of kinetic energy nozzle is the three-fluid nozzle. The spray
characteristics obtained by two- and three-fluid nozzles are similar when atomizing
low-viscosity feeds at up to intermediate feed rates. Use of the second air stream with
three-fluid nozzles causes a waste of energy, except for high feed rates of low-viscosity
feeds. Pressure Nozzles. The second most common form of atomization
for pharmaceutical applications is hydraulic pressure nozzle atomization. Here,
the feed liquid is pressurized by a pump and forced through a nozzle orifice as a
high-speed film that readily disintegrates into fine droplets. The feed is made to
rotate within the nozzle, resulting in a cone-shaped spray pattern emerging from
the nozzle orifice. Rotary motion within the nozzle can be achieved by use of swirl
inserts or spiral grooved inserts (Fig. 6). The swirl inserts have comparatively large
flow passages and enable such nozzles to handle high solids feed without causing any
wear or clogging.
Because the liquid spray exits the nozzle with a relatively high velocity, a spraydrying
chamber at least 2.5m in diameter and 3.0m in cylinder height is usually
required to operate with pressure nozzles.
The differential pressure across the orifice determines the mean droplet diameter.
The distribution about the mean is similar to, but in most cases is narrower
Figure 6 Schematic presentation of pressure nozzles. (Courtesy of Niro Pharma Systems.)
134 C? elik and Wendel
 2005 by Taylor & Francis Group, LLC
than, two-fluid atomization. In contrast, sprays from pressure nozzles handling high
feed rates are generally wider in distribution and coarser than sprays from vaned
wheels. At low feed rates, spray characteristics from nozzles and wheels are comparable.
Mean size of spray is directly proportional to the feed rate and inversely
proportional to the pressure.
Pressure nozzles are generally used to form coarse spray-dried particles (120
300 mm mean particle size) with good flow properties. Antibiotics are a typical application
for such a dryer. Sonic Energy Atomizers. The use of sonic energy and vibrations
for atomization in spray drying has found growing interest in the last two decades.
However, this type of atomizer has not yet found significant commercial applications.
The advantages of sonic nozzles operating at low pressure and having wide
flow channels suggest that they may be suitable for abrasive and corrosive materials,
but it is most likely that sonic nozzles will continue to be developed as atomizers for
special applications, such as very fine sprays of mean size 20 mm, where the nature of
the spray angle and cone minimizes droplet coalescence (7).
2.1.2. Atomizer Selection
The function of any atomizer is to produce as homogenous a spray as possible. The
nature of the feed, the characteristics of the spray, and the desired properties of the
resulting dried product play very important roles in the selection of the atomizer
type. With proper design and operation, nozzles and rotary atomizers can produce
sprays having similar droplet size distribution. In all atomizer types, the size of droplets
can be altered by either increasing or decreasing the atomization energy (e.g.,
increased atomization energy results in smaller droplet size). For a given amount
of energy, the viscosity and surface tension values of the feed influence the size of
the droplet (e.g., higher values result in larger spray droplets).
In general, rotary atomizers are utilized to produce a fine to medium-coarse
product with a mean size of 20150 mm though larger spray-dried particles can also
be obtained if a very large drying chamber is used. Nozzle atomizers are used to produce
spray-dried product with a coarse mean particle size of 150300 mm (8).
For a given spray-drying application, the selection between rotary and nozzle
atomizers involves the following considerations (9):
1. The feed capacity range of the atomizer for which complete atomization is
2. Atomization efficiency
3. The droplet-size distribution at identical feed rates
4. Spray homogeneity
5. Operational flexibility
6. The suitability of dryer chamber design for atomizer operation
7. Feed properties
8. The atomizer experience available for the product in question.
2.2. SprayAir Contact and Evaporation
Once the liquid is atomized, it must be brought into intimate contact with the heated
gas for evaporation to take place equally from the surface of all droplets. This contact
step takes place within a vessel called the drying chamber. The heated gas is
Spray Drying and Pharmaceutical Applications 135
 2005 by Taylor & Francis Group, LLC
introduced into the chamber by an air dispenser, which ensures that the gas flows
equally to all parts of the chamber.
2.2.1. SprayAir Contact
The way in which spray contacts the drying air is a critical factor in spray-drying
operations. Sprayair contact is determined by the position of the atomizer in relation
to the air inlet.
Inlet air is introduced to the drying chamber via an air disperser, which uses perforated
plates, or vaned channels through which the gas is equalized in all directions.
It is critical that the air entering the disperser is well mixed and has no temperature
gradient across the duct leading into it; otherwise, the drying will not be even within
the chamber. The air disperser is normally built into the roof of the drying chamber
and the atomization device is placed in or adjacent to the air disperser. Thus, instant
and complete mixing of the heated drying gas with the atomized clouds of droplets
can be achieved.
Spray droplet movement is classified according to the dryer chamber layout
and can be designated as cocurrent, countercurrent, or mixed flow although these
designations are not complete representations of the actual conditions.
1. Cocurrent flow is the configuration in which the spray and drying air pass
through the dryer in the same direction. This arrangement is widely used
and is ideal for heat-sensitive products. Spray evaporation is rapid, the drying
air cools accordingly, and overall evaporation times are short. The particles
are not subject to heat degradation. In fact, low-temperature
conditions are achieved throughout the entire chamber in spite of very
hot air entering the chamber.
2. Countercurrent flow is the configuration in which the spray and the air
enter at the opposite ends of the dryer. This arrangement has excellent heat
utilization. Countercurrent flow is used with nozzle atomization and is well
suited for meeting the final spray-dried properties of non-heat-sensitive
3. Mixed flow is the configuration in which both cocurrent and countercurrent
flows are incorporated. The advantage of this type of arrangement
is that coarse free-flowing products can be produced in relatively small drying
chambers. In mixed flow systems the powder is subjected to higher particle
temperature. A mixed flow system can be integrated with a fluid-bed
drying chamber when lower particle temperatures are necessary.
The sprayair contact design can be selected according to the required particle
size and the temperature to which the dried particle can be subjected. For example, if
a low product temperature must be maintained at all times, a cocurrent rotary atomizer
is selected for producing fine particles while a countercurrent pressure atomizer
is preferred for obtaining coarser particles. If coarse particles with predetermined
porosity and bulk density properties are desired, a countercurrent pressure nozzle
atomizer is well suited as high product temperature can be maintained for obtaining
the desired porosity and bulk density of the resulting product. For obtaining coarse
spray-dried particles of heat-sensitive materials, a mixed flow nozzle system can be
selected. Integration with a fluid bed is recommended for agglomerated or granulated
136 C? elik and Wendel
 2005 by Taylor & Francis Group, LLC
2.2.2. Drying
The largest and most obvious part of a spray-drying system is the drying chamber.
This vessel can be tall and slender or have a large diameter with a short cylinder
height. Selecting these dimensions is based on two process criteria that must be
met. First, the vessel must be of adequate volume to provide enough contact time
between the atomized cloud and the heated gas. This volume is calculated by determining
the mass of air required for evaporation and multiplying by the gas residence
time, which testing or experience dictates.
The second criterion is that all droplets must be sufficiently dried before they
contact a surface. This is where the vessel shape comes into play. Centrifugal atomizers
require larger diameters and shorter cylinder heights. In contrast, nozzle atomizer
systems must have narrower and taller drying chambers. Most spray dryer
manufacturers can estimate, for a given powders mean particle size, the dimensions
that are needed to prevent wet deposits on the drying chamber walls.
2.2.3. Drying Gas
In pharmaceutical applications of spray drying, the feedstock can be prepared by
suspending or dissolving the product to be spray dried in water. However, the utilization
of a wide variety of organic solvents in feedstock preparations is also common.
Alcohols, such as ethanol, methanol, and isopropanol are preferred organic solvents
in spray drying of pharmaceuticals although other organic solvents such as ketones
are also used in other industries; often the synthesis process upstream from the drying
step determines the solvent selection. The drying characteristics of the solvents
are also important. For example, a solvent with a low boiling point may be the only
choice for heat-sensitive materials.
Although evaporating organic solvents by a spray drying process is very effi-
cient due to the resulting shorter residence time, as compared to the evaporation
of water, the risk of explosion makes the use of these solvents very hazardous. Therefore,
an inert gas, usually nitrogen instead of air, must be used as drying gas for the
evaporation of the solvents. Use of inert gas requires the use of a closed-cycle system
for spray drying in order to recover the solvent and to limit the gas usage. However,
for small drying tests and laboratory work, the nitrogen can be used without recirculation,
using a carbon bed on the exhaust gas to collect the solvent.
2.3. Dried Powder Separation
Powder separation from the drying air follows the drying stage. In almost every case,
spray-drying chambers have cone bottoms to facilitate the collection of the dried
Two systems are utilized to collect the dried product. In the first type of system,
when coarse powders are to be collected, they are usually discharged directly from
the bottom of the cone through a suitable airlock, such as a rotary valve. The gas
stream, now cool and containing all of the evaporated moisture, is drawn from
the center of the cone above the cone bottom and discharged through a side outlet.
In effect, the chamber bottom acts as a cyclone separator. Because of the relatively
low efficiency of collection, some fines are always carried with the gas stream. These
must be separated in a high-efficiency cyclone followed by a wet scrubber or in a
fabric filter (bag collector). Fines collected in the dry state (bag collector) are often
added to the larger powder stream or recycled. When very fine powders are being
Spray Drying and Pharmaceutical Applications 137
 2005 by Taylor & Francis Group, LLC
produced, the side outlet is often eliminated and the dried product together with the
exhaust gas is transported from the chamber through a gooseneck at the bottom of
the cone. The higher loading of entrained powder affects cyclone design but has little
or no effect on the bag collector size.
In the second type of system, total recovery of dried products takes place in the
separation equipment. This type of system does not need a product-conveying system;
therefore, the separation efficiency of the equipment becomes very critical.
Separation of dried product from the air influences powder properties by virtue of
the mechanical handling involved during the separation stage. Excessive mechanical
handling can produce powders with a high percentage of fines.
The most widely used spray-drying process layout is the open-cycle layout in which
the air is drawn from the atmosphere, passed through the drying chamber, and
exhausted back to the atmosphere. This layout is used for aqueous feedstock and
employs air as the drying gas. There are numerous variations of open-cycle layout
systems, two of which are common in pharmaceutical applications (Fig. 7) (10).
The most common and cost-effective layout utilizes a high-efficiency cyclone and
scrubber (Fig. 7A). In this layout, the loss of very fine particles to the atmosphere
cannot be prevented. If the desired particle size of the spray-dried product is too
small to be recovered by cyclone and scrubber systems then the use of a layout
employing a bag filter is recommended (Fig. 7B).
Closed-cycle layouts are mainly used for nonaqueous (i.e., organic solvents)
feedstock and generally require the use of inert gas as the drying medium. They
are also employed when flammable, explosive, or toxic products are used in the spray
flow diagram for a closed-cycle layout schematically (11). These systems require
a good control of the scrubbingcondensing stage at precise temperatures.
Figure 7 Typical layout of the open-cycle spray dryer system: (A) cyclone/scrubber and
(B) bag filter. a, air; f, feed; p, spray-dried product. 1, spray dryer chamber; 2, cyclone;
3, wet scrubber; 4, bag filter/collector. (Adapted from Ref. 10.)
138 C? elik and Wendel
drying process or atmospheric pollution is not permitted. Figure 8 illustrates the
 2005 by Taylor & Francis Group, LLC
In addition to open-cycle and closed-cycle systems, there are semiclosed-cycle
layouts which are not strict in terms of type of drying medium and are operated
under slight vacuum conditions.
4.1. Droplet Formation
4.1.1. Rotary Atomizer
During rotary atomization, bulk liquid feed is accelerated to a high centrifugal velocity.
During this acceleration, the liquid feed forms a thin film over the rotating surface.
For smooth disk atomizers, the film or liquid feed disintegrates into droplets at
the edge of the wheel by one of the three mechanisms: (a) direct droplet formation,
type of droplet formation mechanism that occurs during processing is a function of
the surface tension and viscosity of the feed as well as the wheel speed and feed rate
(12). Direct droplet formation occurs at low wheel speeds when surface tension and
viscosity dominate the atomization mechanism. The other variables that could
potentially affect direct droplet formation are inertia and air friction. However,
due to liquid slippage on the surface of the wheel, inertia is limited and the low
release velocities minimize air friction effects so that the effect of these variables
is minimized at low wheel speeds. As wheel speeds and feed rates increase, the
amount of feed in each vane increases giving rise to ligaments instead of droplets
on the periphery of the wheel. These ligaments disintegrate into droplets with larger
droplets forming from feeds with higher viscosity and higher surface tension.
While the first two atomization mechanisms are partially controlled by the physical
properties of the feed, sheet formation is a result of inertial forces becoming predominant
over these properties. At high wheel speeds and feed rates, the ligaments join
to form a liquid sheet that extends beyond the edge of the wheel. The liquid sheet
Figure 8 Typical layout of the closed-cycle spray dryer system: c, coolant (diluent); f, feed; l,
solvent recovery; p, spray-dried product. 1, dried powder; 2, cyclone; 3, liquid-phase indirect
heater; 4, heat exchanger; 5, scrubbercondenser. (Adapted from Ref. 11.)
Spray Drying and Pharmaceutical Applications 139
(b) ligament formation, or (c) sheet formation as shown in Figure 9, respectively. The
 2005 by Taylor & Francis Group, LLC
disintegrates into a broad droplet distribution as it extends from this edge. In order
to produce a narrow droplet distribution from this mechanism, high wheel speeds are
combined with low wheel loading, which is often achieved with a decreased feed rate.
In contrast, a vaned wheel directs the flow of the liquid feed across the surface
of an inner liquid distributor in which liquid slippage over the surface of the distributor
occurs until there is contact with the vane or channel. The feed then flows outward
due to centrifugal force and forms a thin film across the surface of the vane. As
the liquid film leaves the edge of the vane, droplet formation occurs as a result of the
radial and tangential velocities experienced. Atomizer wheel characteristics that
influence droplet size include speed of rotation, wheel diameter, and wheel design,
e.g., the number and geometry of the vanes.
4.1.2. Two-Fluid Nozzle
Using the two-fluid nozzle, also referred to as a pneumatic nozzle, atomization is
achieved by impacting the liquid feed with high-velocity air, which results in high
frictional forces that cause the feed to disintegrate into droplets. In order to achieve
optimal frictional conditions, this high relative velocity between liquid and air can be
accomplished by either expanding the air to sonic velocities or destabilizing the thin
liquid film by rotating it within the nozzle prior to sprayair contact.
There are several two-fluid nozzle designs available to produce the conditions
necessary for liquidair contact. A common design is one in which the liquid and
air come into contact outside the nozzle. This nozzle is often referred to as an external
mixing nozzle and its main advantage is the greater control available over the
Figure 9 Smooth disk atomizer droplet formation mechanisms: (A) direct droplet formation,
(B) ligament formation, and (C) sheet formation. (From Ref. 12.)
140 C? elik and Wendel
 2005 by Taylor & Francis Group, LLC
atomization through the independent control of both the liquid and the air streams.
Other two-fluid nozzle designs include: (a) an internal mixing design with the air and
liquid contacting within the nozzle head, (b) a combined internalexternal mixing
design created by using two airflows in the nozzle head (also called three-fluid),
and (c) a pneumatic cup design with liquidair contact occurring at the rim of
a rotating nozzle head.
In general, two-fluid nozzles are capable of producing small droplet sizes over
a wide range of feed rates. These droplets are then carried away from the nozzle by
the momentum of the spray and the expanding atomizing air. The most important
variable involved in the control of droplet size is the mass ratio of airflow to feed
rate, which is also known as the air-to-feed ratio. An increase in this ratio causes
a decrease in droplet size. This ratio generally ranges from 0.1 to 10. At ratios
approaching 0.1, atomization is difficult even for low-viscosity feeds while a ratio
of 10 approaches the limit above which atomization occurs using excess energy without
an appreciable decrease in particle size (13).
Sprays formed by two-fluid nozzles are symmetrical with respect to the nozzle
axis and have a cone-shaped pattern. The angle of this cone is called the spray angle
and, for two-fluid nozzles, it is narrow and cannot be varied greatly by adjusting the
air-to-feed ratio. The maximum spray angle available is 7080, which can be
obtained by employing the maximum feed rate and airflow in a high-throughput
nozzle. In general, an increase in air pressure will increase the spray angle if the feed
rate is maintained at a constant level as long as the maximum angle has not been
obtained. Spray angles are maintained if an increase in airflow is accompanied by
an increase in feed rate resulting in a similar air-to-feed ratio.
4.2. Droplet Drying Mechanisms
Evaporation of water from a spray is often characterized using a curve that describes
the change in drying rate as a function of time. This drying rate curve or evaporation
history is a function of temperature, humidity, and the transport properties of the
droplet formulation as well as the air surrounding the droplet. However, many characteristics
of droplet evaporation can be characterized using a general drying rate
The general drying rate curve has three main phases: an initial drying phase,
a phase in which the rate of drying is mainly constant, and a final phase during which
the rate of drying decreases (falling rate phase) (15). During the second phase, the
removal of moisture from the droplet is at a near-constant rate representing the highest
rate achieved during the evaporation history. This constant evaporation rate
results in a near-constant droplet surface temperature with the wet bulb temperature
representing the droplet temperature. During this phase, the majority of the droplet
moisture is removed. The droplet surface is maintained at saturation by moisture
migration from within the droplet to the surface. In contrast, during the falling rate
phase, the rate of moisture migration is rate limiting to the drying rate causing
a decrease in the overall rate of drying. The surface moisture content is no longer
maintained and the droplet temperature rises.
This general drying rate curve is directly applicable to the spray-drying process.
The initial drying phase begins during the sprayair contact phase immediately upon
contact of the droplet with the drying air. During this initial phase, as the drying rate
increases toward equilibrium, a slight increase in droplet surface temperature occurs.
Spray Drying and Pharmaceutical Applications 141
curve (Fig. 10).
 2005 by Taylor & Francis Group, LLC
The drying rate continues to increase until equilibrium across the dropletair interface
is established and the drying rate becomes constant. In the later phase, the solid
layer of the spray-dried particle becomes rate limiting to mass transfer and the drying
rate decreases. The evaporation rate continues to decrease until the droplet reaches
equilibrium moisture content with the surrounding air stream unless the product is
removed from the spray dryer before equilibrium moisture content is reached.
In addition, all evaporation histories, regardless of material type or spray dryer
configuration, have two main points in common: the majority of the evaporation
is completed in an extremely short time interval, usually <1.5 sec, and the temperature
of the drying air decreases rapidly during evaporation.
While the general evaporation history is representative of the processes occurring
during spray drying, the actual rate of moisture migration is affected by several
factors including the temperature of the surrounding air. If the inlet temperature is
so high that the evaporation rate is higher than the moisture migration rate needed
to maintain surface wetness, the constant rate drying phase is very short. This is
because a dried layer forms instantaneously at the droplet surface that acts as a barrier
to additional moisture transfer and retains moisture within the droplet causing
the surface temperature to be much higher. In contrast, lower inlet temperatures
actually yield a lower initial drying rate with a surface temperature equal to wet bulb
temperature for a longer period of time.
It is important to note that the drying curve is only representative. In reality,
there are no defined points during an evaporation history. Some phases may not
even occur or will be very short depending on the process conditions. One example
of this is a spray-drying process for a heat-sensitive material where the inlet temperature
is low. In this case, the initial phase may extend until a critical point where
moisture migration becomes rate limiting, effectively eliminating the constant rate
drying period. In reality, the actual evaporation rate is dependent on several factors
including the droplet shape, composition, physical structure, and solids concentration.
The actual drying time is a sum of the constant rate period and the falling rate
period until the desired moisture content is achieved.
Figure 10 General drying rate curve. (From Ref. 14.)
142 C? elik and Wendel
 2005 by Taylor & Francis Group, LLC
4.3. Effect of Formulation on Droplet-Drying Mechanisms
Droplet composition also plays a significant role in droplet evaporation history.
Typically, sprays are differentiated into three main types: pure liquids, feeds containing
undissolved solids, and feeds containing dissolved solids (16).
4.3.1. Pure Liquid Sprays
For sprays composed of pure liquids, the droplet evaporates away completely. While
this type of spray is not useful for pharmaceutical formulations, its behavior is representative
of very dilute feed materials. The evaporation of pure liquids is dependent
on the dryer configuration. For low-velocity sprays in a low-velocity air stream
(countercurrent dryer) or for low-velocity sprays in a high-velocity air stream (cocurrent
dryer), the evaporation of the pure liquid spray causes the air temperature and
the evaporation rate to decrease. Pure liquid sprays having a wide droplet distribution
evaporate more quickly than narrow distributions having the same mean droplet
size due to the smaller droplets in the wide distribution. In addition, the size
distribution of the droplet changes during evaporation. If the initial spray is a homogenous
or very narrow distribution, the mean droplet diameter decreases during evaporation.
In contrast, if the initial spray is nonhomogenous or a wide distribution,
then the mean droplet diameter initially increases prior to decreasing. In general, a
distribution is the best representation of a spray since the mean of this distribution
may not adequately describe all characteristics of the distribution. For dryer configurations
of high relative velocities such as coarse atomization in cocurrent or fountain-
type dryers, the droplets travel farther before a given fraction is evaporated.
The relative velocity between droplet and drying air affects evaporation rates more
significantly at higher velocities and higher drying temperatures.
4.3.2. Feeds Containing Insoluble Solids
For droplets containing insoluble solids, the droplet temperature is equal to the wet
bulb temperature of the pure liquid droplet during the constant rate phase since insoluble
solids have negligible vapor pressure lowering effects. The total drying is the
sum of the two drying periods. The drying time for the first period is short compared
to the falling rate period. The falling rate period depends on the nature of the solid
phase and can be estimated given the specific gravity of the feed slurry, the density of
the dried product, and the thermal conductivity of the gaseous film around the droplet,
where gaseous film temperature is the average between the exhaust temperature
and droplet surface temperature. The droplet surface temperature is equal to the
adiabatic saturation temperature of the suspension spray.
4.3.3. Feeds Containing Dissolved Solids
Droplets containing dissolved solids have lower evaporation rates than pure liquid
droplets of equal size. The dissolved solids decrease the vapor pressure of the liquid,
thus reducing the driving forces for mass transfer. Drying results in the formation of
a solid crust at the droplet surface, which does not occur from pure liquid droplets.
Vapor pressure lowering causes droplet temperature to increase over wet bulb temperature
as in the previous two examples. Formation of dried solid during evaporation
has a significant effect on the subsequent evaporation history. During evaporation,
sprayair contact and constant rate period occur but may be shorter. The main effect
of dissolved solids is seen when the first period of drying ends and droplet moisture
Spray Drying and Pharmaceutical Applications 143
 2005 by Taylor & Francis Group, LLC
content falls to critical value representing the formation of the solid phase at the surface.
During the falling rate period, the migration of moisture decreases due to resistance
to mass transfer caused by solid phase increasing. Last, the heat transfer is
greater than the mass transfer and the droplet temperature increases. Vaporization
of the moisture within the droplet during this phase may occur if the transfer is suffi-
ciently high.
The relationship between mass transfer and heat transfer for droplets containing
dissolved solids can lead to the formation of many different particle morphologies
depending on process conditions and material characteristics. Charlesworth and
Marshall have defined these morphologies as falling into two groups dependent on
the temperature of the drying air relative to the boiling point of the droplet solution
during the major part of the evaporation period (Fig. 11) (17).
If the air temperature exceeds the boiling point of the droplet solution, then a
vapor will be formed. As the solid crust forms around each droplet, vapor pressure
within the droplet is formed and the resultant effect of this pressure is dependent on
the nature of the crust. A porous crust will release the vapor, but a nonporous crust
may rupture resulting in fractured particles or fines from disintegrated particles.
Alternately, the droplet temperature may not reach boiling point levels due to
cocurrent airflow or because the residence time of droplets in the hottest regions of
the dryer is often very short. In this case, moisture migration occurs through diffusion
and capillary mechanisms.
In both cases, the porosity of the solid crust is often evident in the characteristics
of the falling rate period of the drying curve. If the film is highly nonporous, the
rate will fall sharply and the evaporation time will be prolonged. However, if a highly
Figure 11 Potential spray-dried particle morphologies in relation to process conditions and
material characteristics. (From Ref. 17.)
144 C? elik and Wendel
 2005 by Taylor & Francis Group, LLC
porous film exists, then vapor is easily removed from the dropletair interface and
the drying rate is similar to that found during the first period of drying.
These drying mechanisms result in a range of particle shapes including solid,
hollow, shriveled, and disintegrated, examples of which are shown in Figure 12.
However, it is important to note that particle morphology is also dependent on several
material characteristics including solubility, temperature of crystallization, melting
point, and thermal conductivity since they will also impact the rate of crust
formation, the porosity of the crust, and the subsequent drying rate.
It is also possible to influence particle density and size distribution through the
modification of process parameter settings such as atomizer settings, temperature
levels, and feed rates (18). For example, an increased feed rate while maintaining
a constant inlet temperature results in particles that have higher moisture content
and a resultant increased bulk density. By increasing the temperature of the feedstock,
the ability of the feed to be atomized is often improved due to the reduction
in ligament formation causing an increase in the bulk density of the dried particles.
Also, an increase in the concentration of the feed solids often increases the bulk density
of the dried particle, as does the use of a rotary atomizer, since many wheel
designs reduce air entrapment. Alternatively, bulk density may be decreased through
feed aeration or an increase in inlet temperature. Also, cocurrent spray-air contact is
often effective for reduction in bulk density because the wettest droplets encounter
the hottest air facilitating rapid evaporation and air entrapment. It is important to
note that the outlined process modifications are generally applicable, but that exceptions
to each can be found based on material characteristics.
In a similar manner, it is often possible to influence spray-dried particle
size distribution by changing process parameter settings. As mentioned earlier, the
size of the droplets formed during atomization is affected by process parameters
such as atomization type, atomizer settings, feed solids concentration, feed physical
Figure 12 Various particle forms of skim milk powder. (Courtesy of Niro Pharma Systems.)
Spray Drying and Pharmaceutical Applications 145
 2005 by Taylor & Francis Group, LLC
properties, and drying temperatures. The size of the resultant particles following evaporation
is a function of the initial droplet size as well as the material characteristics
such as solid state and film formation mechanisms.
5.1. Feasibility Assessments
Before any spray-drying application work begins, it may be advantageous to conduct
the following simple, qualitative tests at the laboratory bench using very little material
to determine the feasibility of the application (19). A rheological profile of the
solution or suspension should be evaluated or, alternatively, a small sample can be
tested to see if droplets from a stirring rod can be readily formed. In the latter test,
if the liquid strings from the surface or forms peaks, then high viscosity is indicated
and the product may not be a candidate for spray drying without formulation
changes. The behavior of non-Newtonian fluids (pseudoplastic, thixotropic, dilatant,
etc.) has been found to influence atomization and the resultant droplet size (20).
However, while it is expected that Newtonian and non-Newtonian fluids atomize differently,
this difference was not found to be as important as the more significant
effect of the wheel speed on droplet size. It is also important to note that highly
viscous materials cannot be atomized by pressure nozzles.
Once the effect of viscosity has been evaluated, it may be advisable to dry a few
drops of product on a glass slide using a heated air gun. During this bench drying
test, the air temperature is recorded and the material is observed for the presence
of stickiness, color changes, or other physical changes. If the dried powder is found
to be suitable at the air temperature applied, it can be placed on a variable temperature
hot bench to determine the temperature at which the powder becomes tacky.
For spray drying to be successful, this temperature must be higher than the outlet
temperature of the dryer.
If the initial feasibility evaluation is successful, it is reasonable to commit additional
materials for a spray-drying trial. A laboratory dryer at least 500mm in diameter
is recommended for such tests. Bench-scale spray dryers are available but are
limited in their ability to provide adequate atomization or sufficient process air flow
for the successful production of dried particles. The laboratory unit, however, combined
with very fine atomization (two-fluid or rotary) will often produce acceptable
product for further testing. A series of tests can be performed at different inletoutlet
temperature combinations using small quantities of material and these samples can
be tested for chemical stability to evaluate thermal effects from process air contact.
The relationship between outlet temperature and final product moisture can also be
established for this scale. While samples produced in a laboratory dryer are suitable
for evaluating the effect of spray drying on the product, they are not suitable for use in
downstream processing because the fine particle distribution produced as a result of
the small drying chamber dimensions may not be representative of the final spraydried
Production of coarser particles requires a larger, pilot-scale dryer, which in turn
requires larger feed volumes. This pilot-scale work is often conducted at a spraysdrying
development center since many companies have laboratory dryers but few
have the sizes and variety of process types needed to fully develop a spray-dried product
from pilot scale through 1/10th of the commercial scale and into final production.
These facilities are usually found at spray-drying manufacturers or custom
146 C? elik and Wendel
 2005 by Taylor & Francis Group, LLC
processing companies. In addition to having the equipment, manufacturers and
custom processors often have the expertise to more quickly optimize product
5.2. Spray Drying to Produce a Specific Type of Particle
Because of its inherent costs, spray drying is not always considered as a processing
option for many conventional formulations. However, when a specialized particle
type is required by the active ingredient or dosage form, spray drying can become
a feasible alternative to more conventional manufacturing processes. Such particle
types include microcapsules, controlled release particles, nanoparticles, and liposomes.
The application of spray drying to pharmaceuticals has been extensively
discussed in review articles (21,22).
5.2.1. Granulation
Spray drying is a unique process in several ways as compared to other granulation
methods. The feedstock is a homogenous liquid, which results in the uniform distribution
of all components of the spray-dried product in the same ratio in each individual
particle and eliminates the concerns of uniformly granulating dry components
with a liquid. While granule characteristics may exhibit batch-to-batch variation,
which in turn may influence the compaction behavior of the formulation or the
postcompaction properties of the tablet, granules produced using spray drying are
extremely consistent in terms of particle size, bulk density, and compaction behavior.
These features make spray drying a suitable process for the production of directly
compressible excipients such as lactose, microcrystalline cellulose, and mannitol.
Spray-dried lactose is by far the most commonly encountered spray-dried excipient
Many granulation methods utilize mechanical energy to transform very fine
particles into granules. Although shear forces are employed in nozzle and centrifugal
atomizers to create sprays, this form of energy will not destroy microencapsulated
material as can happen in high-shear granulation. In spray drying, some trial and
error is encountered in establishing the nozzle combinations and liquid pressures
to obtain equivalent particle size distribution during scale-up; however, the resultant
powder will have similar physical properties such as bulk density and compaction.
Also, it is important to note that, within the spray dryer, the product is never in contact
with moving parts, which facilitates the proper cleaning process greatly.
If the granulation size is a critical criterion for a given formulation, then the
selection of the granulation process may be determined based on the desired particle
limitations of numerous granulation techniques.
As seen in this figure, the spray-drying process results in smaller-size particles
as compared to some other granulation methods such as fluid-bed granulation or
high-shear (high-intensity) granulation. One option for producing larger agglomerates
using spray-drying technology is to employ fluidized spray drying, which combines
the features of the spray-drying process with fluid-bed granulation. The result
of this process is particles similar to those obtained from a fluid-bed granulation
operation, and yet the process is a continuous type in contrast to the batch operation
of a fluid-bed granulator. In a fluidized spray-drying system, the bottom cone of a
conventional spray dryer has been modified to include an integral fluid bed
Spray Drying and Pharmaceutical Applications 147
size and feasible operating temperatures. Figure 13 compares the general particle size
 2005 by Taylor & Francis Group, LLC
(Fig. 14). When this process is implemented, atomization and sprayair contact
occur as they do in a conventional spray dryer. However, when the partially dried
particles reach the lower portion of the dryer, instead of undergoing product separation,
the particles are fluidized by the second stream of drying gas. Controlled temperature
and humidity of this fluid bed gas stream ensure that the particles retain
enough moisture to be suitable for agglomeration. At the end of the process, each
granule is an agglomerate of spray-dried droplets. Droplets that dry completely
before agglomerating and granules that experience attrition during fluid-bed drying
create fines, which are entrained in the gas stream and carried upward to the drying
gas exhaust. As a result, these fines pass through the atomized spray, providing an
additional opportunity for agglomeration. Fines that are carried through the exhaust
Figure 14 Schematic of fluidized spray drying system (FSDTM). (Courtesy of Niro Pharma
Figure 13 Particle size range of the methods utilized in particle growth. (From Ref. 21.)
148 C? elik and Wendel
 2005 by Taylor & Francis Group, LLC
are collected in cyclones or a bag collector and can be pneumatically recycled into
the dryer.
A variation of this system is the integrated fluid-bed dryer (Fig. 15). This system
includes integrated filter bags, which are suspended from the chamber roof. This
roof is perforated and serves a dual purpose as a gas disperser. The chamber above
the roof contains clean gas that supplies the inlet process drying air and also has an
exhaust point for clean gas, since any remaining fines are entrapped in the filters.
Products produced using fluidized spray drying have a broader particle size distribution
and lower bulk density than the particles produced by conventional spray
dryers with a typical mean size particle size range of 150400 mm. This process is not
meant to replace conventional spray drying processes but instead is a feasible alternative
for spray drying applications that require larger mean particle sizes.
5.2.2. Modification of Solid-State Properties
Characterization and modification of solid-state properties of drug substances are
profoundly important in developing pharmaceutical products with the desired drug
release properties. The importance of the process understanding and improvement of
the dissolution rate for poorly water-soluble drugs has been known for decades (24).
Particle size reduction methods (such as grinding, micronization, and ball milling),
precipitation, melt quenching, freeze drying, and spray drying have been used
extensively for improving the solubility and dissolution rate of poorly water-soluble
materials (2530). These processes generally impart a polymorphic change in the
drug substance by transforming a low-energy crystalline form to a higher-energy
crystalline form or amorphous form.
Spray drying has many advantages over the other methods in achieving such
improvements. First, the spray-dried particles are generally free-flowing and spherical;
thus, no additional processing (such as dry granulation) is needed before compaction
of these particles. The hollow structure of the spray-dried particles
increases the solubility and subsequent dissolution rate of the drugs by several folds.
For example, the dissolution rate of poorly water-soluble salicylic acid was found to
be almost instantaneous and 60 times faster when spray dried as compared to that of
the original powder (25).
Figure 15 Schematic of an integrated fluid-bed dryer (IFDTM). (Courtesy of Niro Pharma
Spray Drying and Pharmaceutical Applications 149
 2005 by Taylor & Francis Group, LLC
In addition, the energy of the amorphous state depends, to some extent, on the
method of preparation (26). The rapid nature of the spray-drying process (i.e., the
short residence time of the droplets during drying) improves the stability of otherwise
unstable amorphous forms.
Last, the spray-drying process is suitable for integration of stability agents
(such as PVP and PEG) into spray-dried particles. Spray drying of the poorly soluble
drug with 50% PVP resulted in enhanced dissolution when compared to a physical
mixture of micronized drug with PVP (27). A physically stable amorphous form
of ibuprofen, which has a low melting point, was obtained when spray dried in
the presence of 5075% PVP (28). Recently, in a study on the effect of spray drying
varying lactose/PEG compositions, it was found that the most amorphous particles
were obtained when PEG was present at 10% (w/w) concentration. Conversion to
more crystalline materials occurred over time and the crystallization of lactose
appeared to be retarded at low PEG concentrations (29). In other work, the
increased amount and molecular weight of PVP was found to have the potential
to increase the physical stability of amorphous lactose (30).
5.2.3. Microencapsulation
The preparation of microcapsules involves the coating of particles or liquid droplets
with a biodegradable polymer. Applications for microspheres in the pharmaceutical
industry include controlled release, particle coating, flavor stabilization, taste masking,
and physical or chemical stabilization. Microencapsulation can be achieved through a
number of processes, but, in general, an active pharmaceutical ingredient (API) is
trapped within a reservoir or matrix. This process often begins with the preparation
of a three-phase, immiscible system containing a liquid vehicle, a core particle, and
a coating material or polymer. Several manufacturing techniques can be employed
to deposit the polymer around the particle and cause this coating to become rigid.
These methods include spray drying, Wurster fluid-bed coating, pan coating, coacervation,
and emulsion evaporation. In the spray-drying process, the encapsulation process
is achieved in one step in which desolvation and thermal cross-linking occur
concurrently and the particle is coated. A review of the main factors involved in the
application of spray drying for achieving microencapsulation references many works
which detail pharmaceutical applications, especially drug delivery systems (31).
Microencapsulation is a process that is often used for the purpose of providing
controlled release of a protein or drug. Several authors have studied microencapsulation
formulations manufactured from a spray-drying process as a means to achieve
controlled release. In one case, the effect of polymer hydrophilicity on API release
was evaluated and the most hydrophilic polymer was found to gel faster and retard
drug release the most (32). The size and cohesiveness of the resultant spray-dried
particles were found to be a function of the polymer and also affected drug release
with the smaller, more cohesive particles tending to agglomerate and delay drug
release. In another case, release of a model drug was controlled using a spray-dried,
water-activated, pH-controlled microsphere (33). Water influx into the microcapsule
caused buffer to dissolve and adjusted the inner pH causing the fraction of unionized
drug to increase resulting in the increased release of the drug.
One specific polymer type that has been employed in the spray drying of microspheres
to modify release is acrylic resin. A commercial blend of neutral methacrylic
acid esters was used for the preparation of spray-dried controlled release microcapsules
containing model drugs (34). Dissolution results of tablets compressed from the
150 C? elik and Wendel
 2005 by Taylor & Francis Group, LLC
microspheres showed successful controlled release with advantages over a matrix
system. In a similar study, sustained release and enteric tablets were prepared by
directly compressing spray-dried microspheres produced using different types of
acrylic resins (35). Complete enteric properties were observed for tablets made from
pH-dependent, anionic acrylic polymers, while a sustained release profile was
observed for tablets made from microspheres containing pH-dependent, cationic
acrylic polymers.
Two common biodegradable polymers used in microencapsulation are polylactide
(PLA) and polylactide-co-glycolide (PLGA). The efficacy of spray drying as a
method for PLA and PLGA microsphere preparation was investigated using a model
lipophilic drug (36). The spray-drying process was tailored to each polymer and the
microspheres obtained were evaluated for shape, size, drug content, and polymer
influence on these characteristics. Polymer type, polymer molecular weight, and
polymer concentration were shown to be the greatest contributing factors to these
characteristics. In-vitro dissolution testing revealed different release profiles depending
on polymer type and microsphere morphology.
5.2.4. Inhalation Dosage Forms
In order for inhalation dosage forms to be clinically effective, the drug should
deposit in the lower airways. In general, the site of drug deposition in the lungs
depends on the particle size and size distribution of the drug particles or droplets,
the inhaler device and formulation, the patients breathing patterns, and airway geometry
(37). Generally, the aerosol particles or droplets must be <5 mm aerodynamic
diameter to be deposited into the lower respiratory tract.
A formulation of mucoadhesive microspheres for nasal administration was
examined through the preparation of microspheres containing active and one of
two polymer types using a spray-drying procedure (38). The mean diameter of the
spray-dried particles was 35 mm and surface morphology was dependent on polymer
type. Microspheres containing active and either polymer were more mucoadhesive
than any of the starting materials alone and the dissolution rate decreased with
increasing polymer content.
The ability to control the particle size and density of particles for inhalation
was investigated using lactose solutions atomized with a two-fluid nozzle and dried
in a laboratory scale dryer. It was found that droplet size during atomization was
affected by nozzle orifice diameter and atomization airflow but not by feed concentration.
However, dried particle size was influenced by feed concentration and it was
suggested that the shell thickness of the hollow particles increased with increasing
feed concentration (39).
An alternative method of atomization for the formation of respirable particles
is the airblast atomizer. This type of two-fluid nozzle introduces a liquid feed
pumped at a slow rate into a high-velocity gas stream via single or multiple jets. This
atomizer type was utilized at laboratory scale to evaluate the effect of grounded vs.
electrostatically charged tower configurations on the median particle size of the
spray-dried product (40). This study found significant differences between the two
configurations with the latter producing small particles but compromising collection
Spray Drying and Pharmaceutical Applications 151
 2005 by Taylor & Francis Group, LLC
5.2.5. Nanoparticles
Attempts have been made to manufacture particles on the nanometer scale for
applications such as controlled release and intravenous delivery systems. A comparison
evaluating the processability and solid dosage performance of spray-dried
nanoparticles and microparticles was conducted (41). In this study, nanoparticle suspensions
were prepared by wet comminution in the presence of stabilizers, converted
into dried particles using a spray-drying process and subsequently compressed.
Compacts prepared from microparticles and nanoparticles were found to differ in
their internal structure and micromechanical deformations.
In another study, solid, lipid nanoparticles were produced using high-pressure
homogenization and loaded with drug using hot or cold methods for lipophilic or
hydrophilic drugs, respectively (42). Surfactant addition was investigated and stability
and entrapment efficiency were evaluated. Long-term sterile storage of these dispersions
was difficult and spray drying was investigated as a potential, feasible
The feasibility of developing nanoparticles for aerosol delivery has also been
investigated (43). The spray-dried nanoparticles produced using one carrier type
were found to be hollow while others had a continuous matrix. Particle size was measured
before spray drying and after the spray-dried powder was re-dissolved. Both
carrier types resulted in an increase in particle size after spray drying, although both
were found to remain in the nanometer range after drying and were suitable for effi-
cient lung delivery.
5.2.6. Liposomes
Another particle type capable of being produced by spray drying is liposomes. Traditional
preparation of liposomes begins with the preparation of a solution containing
the lipids to be used in a volatile organic solvent mixture. Following filtration of the
solution, the solvent mixture is removed under conditions which ensure that phase
separation does not occur. The dry lipid mixture is then hydrated by an aqueous
mixture containing the drug to be entrapped. Last, this mixture is dried. Spray drying
is one method available for accomplishing one or both of these drying steps. For
example, lipid vesicles were produced using a spray-drying process instead of the first
step of the traditional process (44). Vesicles containing phosphatidyl choline
(soybean lecithin) were produced by extruding the phospholipid through a 0.2 mm
polycarbonate membrane followed by spray drying with 10% lactose. The particle
size, vesicle size distribution, and stability of the multilamellar vesicles were measured.
The mean particle diameter after spray drying with a rotary atomizer was
7 mm and the dry particles could be reconstituted in water to liposomes without
any major change to the vesicle size distribution. In addition, the chemical stability
of the liposomes was not significantly affected by the spray-drying process. In subsequent
work, the same authors utilized spray drying for the hydration step of the
traditional process (45).
5.2.7. Peptides and Proteins
Recent advances in biotechnology have made it possible to use macromolecules such
as peptides and proteins as therapeutic agents. Spray drying has been used for decades
for processing antibiotics, vaccines, and, for the last few decades, macromolecular
152 C? elik and Wendel
 2005 by Taylor & Francis Group, LLC
The effect of spray drying process parameter settings on the activity of peptides
and proteins is often difficult to study. Consequently, enzymes are frequently used as
model protein drugs due to the ease with which their activity can be determined.
Investigations of the application of spray drying for the production of some enzymes
and proteins and the effects that processing parameters have on enzyme activity have
been discussed in a review article (21).
Many proteins and peptides are susceptible to degradation upon spray drying
due to relatively high temperatures. In a recent study, the effects of inlet and outlet
temperatures on some spray-dried peptides and proteins were reported (46). In
another study, enzyme activity was found to be susceptible to spray-drying temperature
and only half of its activity remained after spray drying without additives at outlet
temperatures <50C (47). In this study, it was found that the activity of a
formulation consisting of enzyme and mannitol was maintained at outlet temperatures
<50C and compromised at temperatures >50C. Replacing mannitol with trehalose
stabilized the spray-dried enzyme and its activity was maintained at 100% at
an outlet temperature of 100C.
5.2.8. Dry Elixirs and Emulsions
A dry elixir is a novel dosage form developed by spray-drying actives and excipients
dissolved or suspended in ethanol/water mixtures. One example is a dry elixir in
which the feed solution contained actives, dextrin, and sodium lauryl sulfate in a
mixture of ethanol/water (48). The spray-dried product was spherical in shape with
a smooth surface and a mean diameter of 13 mm. A comparison with the active in
powder form revealed a major decrease in dissolution time from >60 to 2 min.
A similar dosage form to the dry elixir is the dry emulsion. In this case, the
emulsified drug or oily drug solution with additives is spray dried to produce dry
emulsion particles. A dry emulsion of a water-insoluble nutrient was studied and
release from the spray-dried particle was found to be dependent on the type and
amount of oily carrier and surfactant used (49). Differences in release among the different
formulations were attributed to the differences in the physical state of the drug
and surfactant in the dried particle.
5.2.9. Effervescent Products
Spray-dried particles have also been incorporated into effervescent products. In one
study, spray drying was used to protect a degradation-sensitive active by coating fine
particles of the drug with a sugar alcohol solution (50). In vivo results of tablets
made using the spray-dried particles combined with coated citric acid and sodium
bicarbonate revealed that the active was rapidly absorbed from the tablet.
5.2.10. Other Process Variations
Two variations of the spray-drying process have been developed in response to product
requirements. The first variation is spray congealing. In this process, solids such
as wax or monoglycerides are melted. Other ingredients such as drugs, flavors, or fragrances
are dissolved or suspended in the molten material. This molten feed is
sprayed using the same basic spray-drying equipment except that no heat source is
required. Depending on the freezing point of the feed, ambient or chilled air may
be used during the drying process. This process has been described in more detail
and a comparison between particles produced by both the spray drying and the
Spray Drying and Pharmaceutical Applications 153
 2005 by Taylor & Francis Group, LLC
spray congealing techniques has also been drawn (51). One study compared microcapsules
produced using both methods and found that the solvent used, the lipid
type, and the chain length were variables that influenced the surface properties of
both particle types (52,53).
A second variation of the spray drying process is spray freeze drying. In this
process, the feed is sprayed into freezing air causing the droplets to freeze. The frozen
droplets are subsequently sublimed under vacuum conditions producing a dry product.
One study investigated this method further by eliminating the use of vacuum
conditions for sublimation (54). In this study, the feasibility of spraying pharmaceutical
solutions at atmospheric pressure was investigated using very low air temperatures
and desiccated air for the removal of the water from the frozen particles. The
process resulted in fine, free-flowing powder with a high surface area, good wetting,
and good solubility characteristics.
Another type of atomization employed for pharmaceuticals is supercritical
fluid nebulization. The process uses carbon dioxide as an aerosolization aid, which
permits drying at lower temperatures than is usually needed in conventional spray
drying (55). Within the atomization system, supercritical carbon dioxide is intimately
mixed with aqueous solutions containing API, often proteins or peptides. The outcome
is the formation of microbubbles, which are rapidly dried in <5 sec, resulting
in dried particles predominately <3 mm in diameter (56,57). This method is generally
applied for the production of materials for pulmonary use or to achieve increased
bioavailability (58).
Spray drying has found many applications in numerous industries. The scale-up of
the spray drying process is less troublesome as the operation requirements of small
and large dryers are the same when compared to other conventional granulation
processes such as high-shear granulation method.
Spray drying, being a continuous process, is well suited for the production of
bulk drug substances and excipients. Using this process, the physical properties of
the resulting product (such as particle size and shape, moisture content, and flow
properties) can be controlled through the selection of equipment choices and manipulation
of process variables; thus, the final spray-dried particulate matter may not
need further processing (wet or dry granulation) before compaction. In addition,
the spray-drying process matches the directives outlined in the Process Analytical
Technology Initiative which is currently being guided and championed by the
FDA (59).
Spray-drying processes offer several advantages when solid-state properties of
drug substances need to be modified. Using this process, solubility and dissolution
rates of properties of poorly soluble materials can be increased several fold and
the stability of the amorphous form of the materials can be improved significantly.
Because of its initial inherent costs, spray drying is not always considered as a
processing option for many conventional formulations, especially for small batch
size operations. However, when a specialized particle type is required by the active
ingredient or dosage form, spray drying can become a feasible alternative to more
conventional manufacturing processes. Such particle types include microcapsules,
controlled release particles, nanoparticles, and liposomes.
154 C? elik and Wendel
 2005 by Taylor & Francis Group, LLC
We gratefully acknowledge the contributions of Niro Pharma Systems, specifically
Bob Turock and Jim Schell, for providing the content for several of the figures and
Fred Shaw, whose chapter on spray drying in the first edition of this book provided
guidance and content. We also acknowledge the significant contribution of Keith
Masters whose handbook was referenced extensively as a leading text in the field.
2. Percy SR. Improvement in drying and concentrating of liquid substances by atomizing.
US Patent 125,406, April 9, 1872.
3. Masters K. Introduction. Spray Drying Handbook. 5th ed. Essex, U.K: Longman Scientific
Technical, 1991:120.
5. Masters K. Spray drying fundamentals: process stages and layouts. Spray Drying Handbook.
5th ed.. Essex, U.K: Longman Scientific Technical, 1991:24.
6. Masters K. Spray drying fundamentals: process stages and layouts. Spray Drying Handbook.
5th ed. Essex, U.K: Longman Scientific Technical, 1991:26.
7. Masters K. The process stages of spray drying. Spray Drying Handbook. 5th ed. Essex,
U.K: Longman Scientific Technical, 1991:268.
8. Masters K. Spray drying fundamentals: process stages and layouts. Spray Drying Handbook.
5th ed.. Essex, U.K: Longman Scientific Technical, 1991:31.
9. Masters K. The process stages of spray drying. Spray Drying Handbook. 5th ed. Essex,
U.K: Longman Scientific Technical, 1991:271.
10. Masters K. Spray drying fundamentals: process stages and layouts. Spray Drying Handbook.
5th ed. Essex, U.K: Longman Scientific Technical, 1991:43.
11. Masters K. Spray drying fundamentals: process stages and layouts. Spray Drying Handbook.
5th ed. Essex, U.K: Longman Scientific Technical, 1991:47.
12. Masters K. The process stages of spray drying. Spray Drying Handbook. 5th ed. Essex,
U.K: Longman Scientific Technical, 1991:199.
13. Masters K. The process stages of spray drying. Spray Drying Handbook. 5th ed. Essex,
U.K: Longman Scientific Technical, 1991:255.
15. Masters K. Drying of droplets/sprays. Spray Drying Handbook. 5th ed. Essex, U.K:
Longman Scientific Technical, 1991:309.
16. Masters K. Drying of droplets/sprays. Spray Drying Handbook. 5th ed. Essex, U.K:
Longman Scientific Technical, 1991:326.
17. Charlesworth DH, Marshall WR. Evaporation for drops containing dissolved solids.
AIChE J 1960; 6(1):923.
18. Masters K. Drying of droplets/sprays. Spray Drying Handbook. 5th ed. Essex, U.K:
Longman Scientific Technical, 1991:345.
19. Shaw F. Spray drying as a granulation technique. Parikh DM, ed. Handbook of Pharmaceutical
Granulation Technology. New York: Marcel Dekker, 1997:9396.
20. Filkova I, Weberschinke J. Apparent viscosity of non-Newtonian droplet on the outlet of
wheel atomizers. Mujumdar AS, ed. Drying 82. New York: McGraw Hill, 1982:
Spray Drying and Pharmaceutical Applications 155
Traub DA. Spray Dryers, Part 1, Process Heating, August 2001. hhttp://www.processheating.
Traub DA. Spray Dryers, Part 2, Process Heating, September 2001. hhttp://www.processheating.
com/CDA/ArticleInformation/Drying_Files_Item/0, 3274, 63175, 00.htmli.
Traub DA. The Drying Curve, Part 2, Process Heating, October 2002. hhttp://www.process-
heating.com/CDA/ArticleInformation/Drying_Files_Item/0, 3274, 84744, 00.
 2005 by Taylor & Francis Group, LLC
21. Broadhead SK, Rouan E, Rhoads CT. The spray drying of pharmaceuticals. Drug Dev
Ind Pharm 1992; 18(11,12):11691206.
22. Wendel SC, C? elik M. An overview of spray-drying applications. Pharm Technol 1997;
23. Kadam KL. Granulation Technology for Bioproducts. Boca Raton, FL: CRC Press,
24. Fincher JH. Particle size of drugs and its relationship to absorption and activity. J Pharm
Sci 1968; 57(11):18251835.
25. Kawashima Y, Satio M, Takenaka H. Improvement of solubility and dissolution rate of
poorly water-soluble salicylic acid by a spray-drying technique. J Pharm Pharmacol 1975;
26. Pikal MJ, Lukes AL, Land JE, Gaines K. Quantitative crystallinity determination for
beta-lactam antibiotics by solution calorimetry: correlations with stability. J Pharm Sci
1978; 67:767773.
27. Junginger K, Wedler M. Acta Pharm Technol 1984; 30(1):68.
28. Corrigan OI, Holohan EM, Reilly MR. Physicochemical properties of indomethacin and
related compounds co-spray dried with polyvinyl-pyrrolidone. Drug Dev Ind Pharm
1985; 11(2,3):677695.
29. Corrigan DO, Healy AM, Corrigan OI. The effect of spray drying solutions of polyethylene
glycol (PEG) and lactose/PEG on their physicochemical properties. Int J Pharm
2002; 235(1,2):193205.
30. Berggren J, Alderborn G. Effect of polymer content and molecular weight on the
morphology and heat- and moisture-induced transformations of spray-dried composite particles
of amorphous lactose and poly(vinylpyrrolidone). Pharm Res 2003; 20:10391046.
31. Re MI. Microencapsulation by spray drying. Drying Technol 1998; 16(6):11951236.
32. Wan LS, Heng PW, Chia CG. Spray drying as a process for microencapsulation and the
effect of different coating polymers. Drug Dev Ind Pharm 1992; 18(9):9971011.
33. Sutinen R, Laasanen V, Paronen P, Urtti A. pH-controlled silicone microspheres for controlled
drug delivery. J Controlled Release 1995; 33:163171.
34. Palmieri GF, Wehrle P, Stamm A. Evaluation of spray-drying as a method to prepare
microparticles for controlled drug release. Drug Dev Ind Pharm 1994; 20(18):28592879.
35. Takeuchi H, Handa T, Kawashima Y. Controlled release theophylline with acrylic polymers
prepared by spray drying technique. Drug Dev Ind Pharm 1989; 15(12):19992016.
36. Pavanetto F, Genta I, Giunchedi P, Conti B. Evaluation of spray drying as a method for
polylactide and polylactide-co-glycolide microsphere preparation. J Microencap 1993;
37. Gonda I. Targeting by deposition. Hickey AJ, ed. Pharmaceutical Inhalation Aerosol
Technology. New York: Marcel Dekker, Inc., 1992:6182.
38. Vidgren P, Vidgren M, Aronen P, Vainio P, Nuutinen J. Nasal distribution of radioactive
drug administered using two dosage forms. Eur J Drug Metab Pharmacokinet 1991;
39. Elversson J, Millqvist-Fureby A, Alderborn G, Elofsson U. Droplet and particle size
relationship and shell thickness of inhalable lactose particles during spray drying.
J Pharm Sci 2003; 92:900910.
40. Dunbar CA, Concessio NM, Hickey AJ. Evaluation of atomizer performance in production
of respirable spray-dried particles. Pharm Dev Technol 1998; 3(4):433441.
41. Lee J. Drug nano-and microparticles processed into solid dosage forms: physical properties.
J Pharm Sci 2003; 92(10):20572068.
42. Muller-Mehnert RH, Lucks JS, Schwarz C, Ruhl D, et al. Solid lipid nanoparticles (sln)-
alternative colloidal carrier system for controlled drug delivery. Eur J Pharm Biopharm
1995; 41(1):6269.
43. Sham JO, Zhang Y, Finlay WH, Roa WH, Libenberg R. Formulation and characterization
of spray-dried powders containing nanoparticles for aerosol delivery to the lung. Int
J Pharm 2004; 269:457467.
156 C? elik and Wendel
 2005 by Taylor & Francis Group, LLC
44. Goldbach P, Brochart H, Stamm A. Spray-drying of liposomes for a pulmonary administration.
Part 1. Chemical stability of phospholipids. Drug Dev Ind Pharm 1993;
45. Goldbach P, Brochart H, Stamm A. Spray-drying of liposomes for a pulmonary administration.
Part 2. Retention of encapsulated materials. Drug Dev Ind Pharm 1993;
46. Costantino HR, Andya JD, Nguyen PA, Dasovich N, Sweeney TD, Shire SJ, Hsu CC,
Maa YF. Effect of mannitol crystallization on the stability and aerosol performance of
a spray-dried pharmaceutical protein, recombinant humanized anti-ige monoclonal antibody.
J Pharm Sci 1998; 87:14061411.
47. Broadhead J, Ruan SK, Rhodes CT. The effect of process and formulation properties on
the properties of spray dried b-galactosidase. J Pharm Pharmacol 1994; 46:458467.
48. Kim CK, Soon YS. Development of digoxin dry elixir as a novel dosage form using
a spray-drying technique. J Microencap 1995; 12(5):547566.
49. Takeuchi H, Sasaki H, Niwa T, Hino T, Ozawa H, et al. Design of redispersible dry
emulsion as an advanced dosage form of oily drug (vitamin E nicotinate) by spray-drying
technique. Drug Dev Ind Pharm 1992; 18(9):919937.
50. Timmington H. Improved aspirin. Chemist Druggist 1973; 199:482483.
51. Killen MJ. Process of spray drying and spray congealing. Pharm Eng 1993; 13(JulAug),
52. Eldem T, Speiser P, Altorfer H. Polymorphic behavior of sprayed lipid micropellets and
its evaluation by differential scanning calorimetry and scanning electron microscopy.
Pharm Res 1991; 8(Feb):178184.
53. Eldem T, Speiser P, Hincal A. Optimization of spray-dried and -congealed lipid micropellets
and characterization of the surface morphology by scanning electronmicroscopy.
Pharm Res 1991; 8(1):4754.
54. Mumenthaler M, Leuenberger H. Atmospheric spray-freeze drying: suitable alternative
in freeze dry technology. Int J Pharm 1991; 72(27):97110.
55. Sievers RE, Huang ETS, Villa JA, Kawamoto JK, Evans MM, BrauerPure PR. Lowtemperature
manufacturing of fine pharmaceutical powders with supercritical fluid aerosolization
in a bubble dryer. Appl Chem 2001; 73(8):12991303.
56. Sievers RE, Karst U. Methods for fine particle formation. U.S. Patent 5,639,441, June
57. Sievers RE, Karst U. Methods and apparatus for fine particle formation. U.S. Patent
6,095,134, August 2000.
58. Sievers RE, Karst U, Milewski PD, Sellers SP, Miles BA, Schaefer JD, Stoldt CR, Xu
CY. Formation of aqueous small droplet aerosols assisted by supercritical carbon dioxide.
Aerosol Sci Tech 1999; 30:315.
59. U.S. Department of Health and Human Services, Food and Drug Administration, Center
for Drug Evaluation and Research (CDER), Center for Veterinary Medicine (CVM),
Office of Regulatory Affairs (ORA). Guidance for Industry: PATA Framework
for Innovative Pharmaceutical Development, Manufacturing, and Quality Assurance,
September 2004.
Spray Drying and Pharmaceutical Applications 157
 2005 by Taylor & Francis Group, LLC
Roller Compaction Technology
Ronald W. Miller
Bristol-Myers Squibb Company, New Brunswick, New Jersey, U.S.A.
This chapter is different from the original; new works, findings, and best practices
from various investigators since 1997 have been added. The interested reader will
have to wait for a comprehensive book (in preparation now) to link numerous journal
articles, chapters, and investigative works for technology completeness.
Powder granulation is a process of powder size enlargement that incorporates small
particles into larger ones. The definition of granulation comprises a range of different
size enlargement methods that can be classified as either dry or wet. In wet methods, a
suitable liquid is used to agglomerate the small powder particles into a mass. The wet
mass is subsequently dried and sized for further downstream processing needs. Wet
granulation methods have been the most widely used powder granulation technology
in the production of pharmaceutical products, particularly in modern pharmaceutical
The chief reasons to granulate powders for the manufacture of pharmaceutical
dosage forms are described by Kristensen and Schaefer (1):
 To improve powder flow properties for dosage filling and compression
 To eliminate wet granulation induced degradants and to improve product
 To prevent active product ingredient from segregating
 To reduce bulk volume, thereby minimizing storage and enhancing transport
 To reduce potential environmental and safety hazards.
Kristensen and Schaefer provide ample literature references in their chapter in
the Encyclopedia of Pharmaceutical Technology about granulation size enlargement
methods (1): Capes (2,3), Pietsch (4,5), Sherrington and Oliver (6), Kapur (7). Others
referenced about wet granulation technologies are Kristensen and Schaefer (8),
Lindberg (9), Fonner et al. (10), Anderson et al. (11), and Ghebre-Sellassie (12).
 2005 by Taylor & Francis Group, LLC
Granulation methods are used in the pharmaceutical industry to enlarge and densify
small powder particles into larger ones typically to improve powder flow so that the
material can be processed effectively and efficiently further into solid dosage forms.
There are two methods of granulation, wet and dry. Wet granulation methods are
widely described in the literature: Encyclopedia of Pharmaceutical Technology (Granulations)
is solely devoted to this technology (1). Dry granulation operations do not
use moisture or heat to process powders into densified granules. The pharmaceutical
industry employs two methods of dry granulation: slugging and roller compaction.
Little has been written about pharmaceutical dry granulation technology. Its
contemporary use in the industry is 5060 years, beginning in the late 1940s. However,
its popularity has risen in the last 15 years in parallel with the increased
research on new efficacious active pharmaceutical ingredients (API) in the pharmaceutical
industry. A number of these new API cannot be processed so easily using
wet granulation and drying processing steps, because of their chemical fragility
and sensitivity. Therefore, this pushes the need for the use of dry granulation processing
techniques to advance new API in the 21st century.
Briefly, in dry pharmaceutical granulation processing, the powder particles are
aggregated under high pressure, typically a pressure of 3070 bar. Particulate matter
can be aggregated when compressed at high pressure because of bonding forces developed
by the direct contact between the solid surfaces. The high pressure serves to
improve the contact area between the surfaces and thus the overall bonding strength.
Sometimes a binding agent is needed to provide additional bonding strength.
In the pharmaceutical industry, dry granulation processing in the 1950s1970s
favored a process called slugging. This process design consisted of feeding powder
into a large compression machine, such as a Stokes D3 type compression machine,
where the powder was compressed into large tablets or slugs, typically in the
order of 1 in. diameter with a tablet gauge of about 0.25 in. The tablet slugs were
subsequently milled by a separate sizing machine to an appropriate particle size
distribution, and further processed into pharmaceutical capsules, powder for oral
suspensions, sachets, or tablet dosage forms. The slugging process is still used today
by only a few manufacturing firms that have old pharmaceutical formulation processes.
Today, modern pharmaceutical formulation processes introduced into the
Americas, Western Europe, Australia, and parts of Asia do not use this kind of
dry granulation equipment in newly developed formulations. The slugging process
is a relic of the past in modern pharmaceutical technology; roller compaction is
the key technology to future dry granulation processing.
Some characteristics are described briefly about the slugging process to complete
the technology information for the reader. The slugging process is externally
influenced by raw material feed properties such as powder cohesiveness, density, flow
characteristics, and powder particle size distribution. The slugging machines design
characteristics such as machine type, feed hopper, feed frame, die diameter, tooling
features, compression speed, and slugging pressure also influence the slugging process
and the final product properties. In general, the key processing operational
aspect of slugging is to maintain a uniform powder fill weight into the dies during
the dynamics of the slugging process. This assures the best chance to manufacture
uniform powder slugs and, ultimately, uniform densified granules. The compression
slugging setup is a key essential to maximizing the slugging throughput and
minimizing the hopper feed-frame and die powder flow problems associated with
160 Miller
 2005 by Taylor & Francis Group, LLC
the process. Slugging compression is normally performed at 46 tons hydraulic pressure,
at a rate of 1030 turret revolutions per minute. The specific machine tonnage,
turret speed, and roll dwell time required for the process are dependent on the
powder blends physical properties, the tooling configuration, machine parts, and
ultimately the slug specifications. Typical slugging machine output ranges from 30
to 50 kg/hr and the machines are not instrumented with modern devices to control
their performance. There are many disadvantages with the slugging technology in the
pharmaceutical industry (Table 1).
This chapter briefly identifies key aspects of roller compaction technology. Unlike
the slugging process technology, roller compaction technology is well suited for
dry granulation agglomeration in the era of modern development of active pharmaceutical
ingredients in pharmaceutical plants.
The increasing scale of manufacturing pharmaceutical products worldwide, the
need for high processing rates, together with increased levels of good manufacturing
practices, necessitate controlled dry granulation processes with as few processing
steps as possible. This has been accomplished by instrumenting roller compactors
to automate and control the mechanical process. Roller compaction technology
plays a very important role in providing competitive cost control, safety, and quality
products in the pharmaceutical industry. Key roller compaction benefits are identi-
fied in Table 2.
Table 1 Disadvantages of Slugging
Single batch processing Excessive air and sound pollution
Frequent maintenance changeover Increased use of storage containers
Poor process control Increased needs of manufacturing space
Poor economies of scale Increase of logistics
Low manufacturing throughput per hour More energy and time required to produce
1 kg of slugs than 1 kg of roller compact
Source: From Ref. 13.
Table 2 Advantages of Roller Compaction
Simplifies processing Uses less raw materials
Facilitates powder flow Eliminates water-induced degradants
Uses minimal energy to operate Improves process cycle time
Requires less man-hours to operate Prevents particle segregation
Improves drug dosage weight control Facilitates continuous manufacturing
Reproduces consistent particle density Improves content uniformity
Produces good tablet and capsule
Does not require explosion proof room/
Eliminates aqueous and solvent granulating Produces a dry product that is process
Source: From Ref. 13.
Roller Compaction Technology 161
 2005 by Taylor & Francis Group, LLC
The bonding forces in a dry aggregate are important to granulation properties such
as granule integrity, flowability, friability, density, compressibility, and size for
downstream manufacturing process steps (1). Rumpf and coworkers described the
bonding mechanisms occurring during dry granulation as a mixture of van der
Waals forces, mechanical interlocking, and a recombination of bonds established
between freshly created surfaces, and solid bridges, created because of partial melting
and solidification during compression (14). A general theory describes particle bonding
related to roller compaction in the Handbook of Pharmaceutical Granulation
Technology (13). The process of dry granulation relies on interparticulate bond formation.
Granule bond formation is characterized in different stages, which usually
occur in the following order:
1. Particle rearrangement
2. Particle deformation
3. Particle fragmentation
4. Particle bonding.
Particle rearrangement occurs initially as powder particles begin filling void
spaces. Air begins to leave the powder blends interstitial spaces, and particles begin
to move closer together. This action increases the powder blends density. Particle
shape and size are key factors in the rearrangement process. Spherical particles will
tend to move less than particles of other shapes because of their close initial packing
to one another. Particle deformation occurs as compression forces are increased.
This deformation increases the points of contact between particles where bonding
occurs and is described as plastic deformation (13).
Particle fragmentation follows as the next bonding stage. This occurs at
increased compression force levels. At this stage, particle fracturing creates multiple
new surface sites, additional contact points, and potential bonding sites. Particle
bonding occurs when plastic deformation and fragmentation occur. It is generally
accepted that bonding takes place at the molecular level, and that this is due to
the effect of van der Waals forces (13).
When powder granules undergo an applied force or stress, a stress force is
released from the granules. The granules attempt to return to their original shape
or form; this is described as elastic deformation. A deformation that does not totally
recover after the stress is released is a plastic deformation. Elastic and plastic deformations
can occur simultaneously, but one effect usually predominates.
Parrott identified three theories of compression bonding: mechanical, intermolecular,
and liquid-surface film. Mechanical bonding purports that individual particles
undergo elastic, plastic, and brittle deformation. Bonding of this nature occurs
because particle surfaces intertwine, forming mechanical bonds. Intermolecular theory
identifies that there are some unsatisfied surface ions that have a potential need to
bond to one another. Under pressure, intermolecular forces become pushed together
close enough so that van der Waals forces can act to consolidate particles. The liquidsurface
film theory identifies that bonding occurs because of the existence of a thin
liquid film. The thin liquid film is generated from pressure induced by the energy
of compression. This mechanism acts as a bonding agent promoting mechanical
strength and an enlarged particle (15). Very little information is written about this last
162 Miller
 2005 by Taylor & Francis Group, LLC
Dehont et al. provided a simplified approach to roller compaction theory (16).
They described that powder granules move through stages in the feed area. The
material is drawn into the gap by rubbing against the roll surfaces. The densification
that occurs in this area is particle rearrangement. At this stage, the speed of the
powder is slower than the peripheral speed of the rollers. Figure 1 represents compactor
rolls in the horizontal plain; powder is pushed vertically downward into
the compaction area.
Note in Figure 1, a ? nip angle, b ? material in volume space. The material is
located in the compaction area between a and the horizontal axis (Fig. 1). At this
stage, the material undergoes additional compaction forces. The particles undergo
plastic deformation and are bonded. Dehonts team noted that nip angle varies
according to the material characteristics of particle size and density and the angle
is about 12 (16). They defined the neutral angle, g, which corresponds to the point
where the pressure applied by the rollers is the greatest on the material. They also
defined elastic deformation, d, and that occurs after the compact begins leaving
the compression roll area. Compacted flakes may increase in size due to material
elastic deformation and actually may have a larger thickness than the roll gap, e (16).
Dehont et al. developed Eq. 5.1 for the linear variation of flake thickness at a
specific roll diameter (16):
e1 ? D?d0=d1  d0??1  cos a? ?5:1?
where e1 is flake thickness, D is roll diameter, d0 is material density at angle a, and d1
is flake density. Dehont et al. assumed that the material in the compaction
area remains horizontal and moves at the peripheral speed of the rollers. They
also considered that the angle a is independent of the roller diameter size and noted
that the flake thickness e1 depends on the roller speed, the roller surface, and the
compaction pressure. All these parameters influence the density of the flake d1.
Dehont et al. concluded that if the same flake thickness were obtained with different
roller diameters, the flake density would be greater with larger-diameter rollers (16).
This is due to the greater nip angle formed with the larger rolls allowing more material
to be compacted.
R. W. Heckel considered the compaction of powders analogous to that of a
first-order chemical reaction. The pores were the reactant and the densification of
the material the product. The proportionality between the change of the density with
the pressure and the pore fraction was the process kinetics (17). Heckel explained
Figure 1 Front view of compactor rolls in horizontal plane. (From Ref. 16.)
Roller Compaction Technology 163
 2005 by Taylor & Francis Group, LLC
mathematical constants that described the compaction behavior of a given powder
and developed a mathematical relationship, Eq. 5.2 (17):
ln?1=1  D? ? KP ? ln?1=1  D0? ?5:2?
where D is the relative powder density, D0 is the relative loose powder density at zero
pressure, P is the pressure applied, 1  D is the pore fraction, and K is the proportionality
The expression of densitypressure relationship permitted the determination of
density values in the range of the pressures investigated. Heckel described mathematically
that the curved region when plotting ln(1/1  D) vs. P is associated with powder
densification (17). This occurred by a mechanism of individual particle movement
in the absence of interparticle bonding. Heckel concluded that the densification represented
by the linear region of the plot, ln (1/1  D) vs. P, occurred by plastic
deformation of the compact after an appreciable amount of interparticle bonding
had taken place. For quantitative reasons, Heckel, redefined the mathematical
expression denoting a constant, A ? ln(1/1  D0), which is quantitatively valid
except at low pressures. He postulated that the constant A, which is somewhat larger
than ln(1/1  D0), represents the degree of packing achieved at low pressures as
a result of rearrangement processes before appreciable amounts of interparticle
bonding take place (17). Heckel postulated that K, the slope of the linear region,
is a measure of compact densification by plastic deformation.
Heckel concluded that densitypressure data indicate that the rate of the
change of density with pressure, any pressure, is proportional to the pore fraction
in the compact at that pressure. Additionally, densitypressure curves may be
described by two parameters. He theorized that one was related to low-pressure densification
by interparticle motion. The second measured the ability of the compact to
densify by plastic deformation after appreciable interparticle bonding.
J. R. Johanson identified, through very comprehensive mathematical models
and relationships, material properties, press dimensions, and operating conditions
for roll compactors. For more information, the interested person should read the
entire reference (18). In part, he explained that roller compaction involves a continuous
shear deformation of the granules into a solid mass.
To satisfy the theorys assumption, it was postulated that the material be isothat
undergoes shear deformation into a solid mass (18). In Figure 2, Po ? horizontal
pressure between rolls, y ? angular position of roll bite, a ? nip angle, 2d ? roll
diameter D, h ? height above the roll centerline at which feed pressure Po is applied,
Pm ? horizontal pressure at y ? 0, and S ? roll gap.
Johanson pointed out that no roller compactor theories at that time determined
the angle of the nip and the bulk density at y ? a, except by actually rolling
the granular solid in a roll press. He also provided a method to calculate the nip
angle and the pressure distribution between the rolls. His calculations determined
the pressure distribution above and in the nip area (18).
He provided the technical rationale to calculate the nip pressures in the nip
region. He showed that material trapped in a volume Va between arc-length segments
4L, must be compressed to volume Vy between the same arc-length segments.
The relationship requires that the bulk densities ga, gy in volumes Va, Vy be related
164 Miller
tropic, frictional, cohesive, and compressible. Figure 2 depicts material in a press
by Eq. 5.3 (Fig. 3) (18):
 2005 by Taylor & Francis Group, LLC
ga=gy ? VyVa ?5:3?
In Figure 3, P0 ? horizontal pressure between rolls, y ? angular position of roll
bite, a ? nip angle, 2d ? roll diameter D, h ? height above the roll center line at
which feed pressure P0 is applied, Pm ? horizontal pressure at y ? 0, S?roll gap,
DL?arc-length segments, Va?material trapped in volume space described by
arc-lengths, Vy?compressed volume space described by arc-lengths, ga and gy?
respective powder bulk densities in volume spaces Va and Vy, and K?a material
property constant for a given moisture content, temperature and time of compaction.
Johanson stated that the pressure sy at any y < a can be determined as a
Figure 3 Front view of compactor rolls, depicting nip angle. (From Ref. 18.)
Figure 2 Front view of compactor rolls in horizontal plane, depicting powder regions at
different compaction forces. (From Ref. 18.)
Roller Compaction Technology 165
 2005 by Taylor & Francis Group, LLC
function of the pressure sa, at y?a, by the pressuredensity relationship. It was
understood that, for increasing pressures, log density was a linear function of log
pressure (19).
Johanson found that the nip angle does not depend on the magnitude of the
roll force or the roll diameter. He demonstrated that the nip angle was affected very
little by the geometry of the press or the cut grooves on the roll surface. It was mostly
influenced by the nature of the materials that were compressed. The very compressible
materials, with small K-values, had very large nip angles. On the other hand,
incompressible materials, with large K-values, had very small nip angles. Ultimately,
Johansons results showed that material properties determine the maximum pressure
that a roller press applies to material.
Parrott evaluated compacting of several pharmaceutical powders into compact
sheets that were sized by an oscillating granulator into granules. He studied flake
thickness, bulk and tap densities, angle of repose, and flow rate after sizing. His work
demonstrated that one could improve most granule densities but not necessarily
improve granule flow properties (20). Nearly all particulate matter can be aggregated
when compressed at high pressure. In some cases, pressure alone cannot achieve suf-
ficient bonding strength; therefore, a binding agent is required to be added to the
powder blend (1). Polymeric binders added to mixtures at certain percentages form
highly viscous bridges between particle-to-particle powders. When the polymeric
mixtures are compressed, they increase the overall compact strength and enhance
particle-to-particle bonding and can improve powder flowability.
Pietsch described a compaction capacity throughput equation for a roller compactor
using flat rolls. The equipment capacity equation is theoretical in nature; it
assumes that all compact is 100% usable for downstream processing needs. Compact
capacity throughput, Cc, can be determined by Eq. 5.4 (4):
Cc ? pDlhAn60g ?5:4?
where Cc is the roller compactor throughput (kg/hr), D is the roller diameter (cm),
l is the roller length, working width (cm), hA is the gap width between the rollers, sheet
thickness (cm), n is the number of revolutions per minute (1/min), and g is the apparent
sheet density (kg/cm3).
Certainly, a key enhancement that highlights todays pharmaceutical industry stateof-
the-art roller compactors is programmable logic controllers (PLCs). They are
used to control and monitor mechanical parts that regulate screw feed rate, roll
speed, pressure and gap, vacuum deaeration, and mill speed. This section will not
discuss PLC designs.
Briefly, a key machine innovation, vacuum deaeration, was a new important
feature design added by some roller compactor vendors in the early mid-1990s.
The design feature has been shown to help premodify raw material density prior
to compacting and increase throughput (21). Other equipment features such as multiple
horizontal or angled feed screws have assisted in manufacturing a uniform raw
material feed across the rolls (21,22). Newly designed roll machine blocks, featuring
cantilever roll systems, offer more efficient ways to clean, handle, and facilitate
product and equipment changeovers. New storage hoppers and various screw feeder
designs have improved delivery of poor flow powder to the rolls.
166 Miller
 2005 by Taylor & Francis Group, LLC
A history of hopper and feed screw designs showed that each design evolved to
facilitate and improve powder flow to the compactor feed screw conveyance system.
Feed hopper designs shown in Roller Compaction Technology for the Pharmaceutical
Industry depict designs used in older compactor models (23). Dehont et al. described
the powder compaction feeding systems that were in use up to 1989 (24).
Specific hopper designs were used when material had good rheological characteristics.
Materials that had good but unique flows, and needed to be evenly distributed
across the roll-surface, such as granular salts, used hoppers that had guiding
flaps. When the powder material had poor rheological characteristics, a vertical flap
distribution box was fitted under the hopper. It primarily was used when compacting
ores. Force feeder designs are more commonly used in the pharmaceutical industry.
They are employed with various configured designs and shapes to supply poorly
flowing powder to the nip roll region (23). Pietsch described modern feed screw system
designs; depicting a vertical force-feeding screw with a slightly tapered end, an
inclined feed screw, a vertical, tapered, and blade-angled feed screw, and a horizontal
(single or dual) straight feeder screw(s). These designs are described and pictured
(25). Innovatively designed feed screw systems are used commercially in roller compactors
worldwide. An example is shown in Figure 4.
Sizing devices are now trimly fitted to the compactor body and controlled by
variable speed drives. Most compacters no longer require a separate milling machine
in tandem to size compacts as required by slugging technology. Roller compactors
have clean-in-place systems that offer environmental and safety features. These systems
minimize human exposure to chemicals. See Roller Compaction Technology for
Compactor design features have evolved over the years. By the mid-1970s, research
revealed a number of roll design improvements that increased compacting efficiency.
Three key conditions were identified, at that time, which optimized the roll compact
throughput and minimized leakage of noncompacted powder (26):
 Adequate powder supply must enter the gripping zone,
 Powder must be conveyed fully into the narrowest part of the roller gap,
 Compaction pressure must be distributed as uniformly as possible across
the whole roller-gripped powder mass.
Figure 4 Twin horizontal feed screw system with vacuum deaeration system. (Courtesy of
Alexanderwek Inc., USA.)
Roller Compaction Technology 167
the Pharmaceutical Industry for additional information (23).
 2005 by Taylor & Francis Group, LLC
Equipment engineers and researchers worked on improving feeding equipment
systems and roll designs to satisfy and maximize the above conditions. Some of the
key advances are identified in Ref. 21 and are re-emphasized in this chapter.
Because of powder feed variability at the nip and in the roll gap regions, powder
leakage is produced during the compaction process. This situation produces
excessive fines and possible undesirable processed material. Usually, this problem
is caused by uneven powder flow and compact formed when the powder is fed
toward the middle of the roll width. Granules produced under these conditions
are typically not optimal for further pharmaceutical processing.
Trials performed by Funakoshi et al. using rectangular aperture chutes, fitted
at the end of a feed screw, aided in preventing uneven powder flow across the roll
width. Funakoshis team demonstrated the positive effects of having a concavoconvex
roller pair fitted with inner ring walls. This feature counteracted side seal
effects (fractured or incomplete compacted edges). They also designed and tested
other rolls (concavo-convex with rim), which allowed the powder to distribute more
uniformly across the roll width. This design reportedly minimized powder leakage
during compaction (26). Funakoshis team also showed that when the roll rim inner
angle is zero the powder is not adequately and uniformly delivered to the gripping
and compacting zones. This occurred because the stationary side seals acted as resistors
to the powder flow (26). Their work demonstrated that the formed compact and
the compaction pressure (using standard rolls) produced an uneven compact across
the roller ends because of side seal friction. When they employed the newly designed
concavo-convex rimmed rolls, it protected the compact from the adverse side seal
fracturing effect. Funakoshi et al. also determined a proper selection of rimmed rollers,
which delivered adequate powder to the compaction zone and also conveyed
powder fully across the roller gap region. They concluded that the height and slope
of the inner walls rims optimally influenced the side seal effect. A 65 inner wall
slope produced 2.53.0% fines leakage during the compaction of lactose. When
the rolls had no inner wall rims, the fines leakage was 15%. In summary, the team
optimized the compact and pressure distribution along the rolls by using the concavo-
convex rimmed shaped rolls and found the best roll design was still dependent
on raw material properties (26).
Parrott further substantiated the usefulness of the rimmed concavo-convex
roller pair to increase the density of several pharmaceutical powders. His work
resulted in optimizing a process with an uncompacted leakage rate of 5%. The optimization
depended on the physical properties of the powder and the machine operating
conditions such as the roller gap, feed screw speed, and roll speed (27). Some of
their key design feature findings are summarized:
 Installing a concavo-convex roll surface rather than a flat roller pair
 Installing rectangular feed chute and flaps
 Designing cylindrical, conical/cylindrical, or tapered variable speed auger
feed screws
 Optimizing the roll rim angle to 65
 Installing digital and analog variable feed screw controllers
 Developing horizontal and variable screw feed systems.
Jerome et al. in the 1980s studied the effects of compactor adjustments on powder
blend properties. His teams research showed that pressure applied by the movable
roller is not a predominant factor. They found that the most important variable was
the compactor feed screw in relation to the roll speed. More will be said about this
168 Miller
 2005 by Taylor & Francis Group, LLC
point later in the chapter. Jerome correlated improved tablet compression hardness to
the feed screw speed in relation to the speed of the rolls (28). Roll configurations are
well documented (23). Most compactors now have one floating roll and one fixed roll
each on separate bearing blocks and this is illustrated in Figure 5. Currently, one pharmaceutical
manufacturer of roller compactor equipment has a fixed roll-pair system.
This roll design system relies on powder gravity feed and a specially designed compression
feed screw to continuously deliver powder to the fixed roll pair (23).
Pietsch described two pairs of rollers with different diameters D1 and D2, with
identical roll gaps hA. He theorized that if the peripheral speed of both pairs of rollers
is the same, roller compaction takes place more gradually in the case of the larger
roll pair. At the same time, the larger-diameter roll pair pulls a larger powder volume
into the nip region, resulting in a higher-density compacted product. The largerdiameter
roll pair also minimizes air entrainment more efficiently than the smallerdiameter
roll pair when both are operating at the same peripheral speed (25). Pietsch
noted that the peripheral roll speed and particulate powder speed are not equivalent
in the entire compaction zone. Throughput does not increase proportionally with
roll speed.
There are two effects that hinder throughput: starved conditions in the feed
zone and too much squeezed air from the particle mass that flows upward and
against the powder flow, reducing the supply of material to the nip area (25).
What is the optimum roll speed for a compaction process?What factors does the
formulating scientist or process engineer need to consider to maximize compact quality
and compaction throughput? Johanson (29) attempted to answer these questions
by predicting roll-limiting speeds for briqueting presses. He developed mathematical
expressions considering even the gas and liquid effects as they can theoretically be
Figure 5 Fixed and floating roll pair. (From Ref. 13.)
Roller Compaction Technology 169
 2005 by Taylor & Francis Group, LLC
squeezed from a solid mass. Solid properties, press dimensions, and operating conditions
were evaluated to predict optimumroll speeds. The results necessary for a quality
briquette are most critical for low-density fine particles. Johansons work showed the
relation between feed pressure and roll speed to essentially be proportional to the
materials permeability (porosity). For example, when the initial powder was compacted,
sized, and recompacted, the bulk powder permeability increased; however,
when the powder was compressed, the compression force decreased, and compactor
feed pressure force requirement decreased significantly. Johanson (29) also demonstrated
that if the compactor feed pressure is kept constant and the press speed is
increased, the maximum pressure applied to the briquette decreases as the roll speed
increases. As a result, the briquette density and strength decrease likewise. For
additional information, the interested reader should peruse the reference list.
Sheskey and Hendren in 1999 studied the effect of roll surface configuration on
the drug release and physical properties of a hydroxylpropylmethyl cellulose
(HPMC) matrix controlled-release dosage form (30). Smooth and axial-grooved
roll surface designs were studied using a Vector Model TF-Mini roller compactor
(Vector Corporation, Marion, IA). Their hypothesis was that the greater the depth
of the concavity on the roll surface, the pressure exerted on the powder being compacted
would not be as evenly displaced as with a smooth roll surface. Therefore, as
with tablet tooling design, the top of the crown area of the compacted ribbon would
theoretically be softer than the rest of the ribbon. However, the results of a particle
size distribution test performed on milled ribbons generated using both smooth and
axial-grooved roll pairs showed similarity between tested samples. In addition,
results showed little difference in tablet crushing strength values between samples
manufactured using either type (smooth, axial-grooved) of roll surface design. Drug
release profiles of tablets prepared from roller compacted granulations using both
roll surface configurations were also similar. It was noted that the authors did not
measure product throughput rates to address the possibility of improved efficiency
when using an axial-grooved roll pair vs. a smooth roll pair.
The consistency and evenness of the powder feed into a roll pair determines to
a large extent, how complete a compact is made and ultimately the success of a compaction
process. Most roller compacting systems suffer the disadvantage of leakage,
i.e., 2030% powder particles (depending on the formulation) are not compacted.
This primarily occurs because of uneven powder feed and powder slippage between
individual loose particles and the roll surfaces. Under these conditions, it is usually
necessary to recycle the uncompacted powder or fines. Recycling a compaction process
is a significant drawback because of additional capital expenditure, labor costs,
and increased throughput time.
The chief objective of roller compaction is to consistently make an agglomerate
of sufficient strength that meets the required density, granulometry, and powder flow
specifications. The key operational goal in compacting is to maintain a pressure
range on the feedstock, independent of the fluctuating powder granulometry and
flow into the rolls, to maintain a consistent compact. The compaction process is
managed by controlling the input material, the quantity per unit of time, the roll
speed, and the roll gap. Allowing the roll gap to float unchecked could influence
the production rate and the compact quality. Therefore, it is important to control
170 Miller
 2005 by Taylor & Francis Group, LLC
the compaction process by setting a constant powder feed rate during the compaction
Design innovations on the powder feed input side of the roller compactor are
complex. The complexity is double-edged: powder materials, which flow and move
well, are more easily handled on the feed side of the compactor. These types of materials
generally do not necessarily need significant densification to marginally improve
their flow handling characteristics on a large scale. On the other hand, poor flowing
and low-density powders require special equipment and considerations to feed a
compactor. Additionally, it is necessary to maintain a constant powder flow and
quantity of material to the rolls during the compaction process. Therefore, the delivery
feed system plays a very important role in delivering poor flowing low-density
bulk powder materials to the roll compactor.
Johanson, in 1995, described an arch-breaking hopper design that effectively
delivered poor flowing powder from the hopper to a horizontal feed screw. The
unique hopper design eliminated the rat-holing effect of poor flowing powders
during hopper voiding and is described in Figure 6 (31).
Miller noted that the total compaction power requirement is the sum of the
power required for the feed screw(s) and the roll drive (13). Feed screws not only
convey the powder material from the compactor storage hopper but they also help
deaerate the powder in the process. The deaeration of the powder acts as a minicompactor
by precompacting the material just prior to roll compaction. Optimum compaction
pressures and feed screw designs vary widely for different powder material
properties. Changes in bulk powder density and feed screw speeds will affect the roll
gap, the compaction pressure, the throughput, and the quality of the compact (13).
Weggel indicated that feed screw torque varies directly with the precompaction
pressure (32). He suggested that by maintaining a constant feed screw pressure, the
compactor operator can control the compact quality. Variations in precompaction
pressure and in the compacting pressures are directly related to feed screw amperages
and the roll drive motor. The compaction pressures are dependent on a continuous
flow of powder into the feed screw area. If the powder feed flow is intermittent or
Figure 6 Arch-breaking hopper design. (From Ref. 31.)
Roller Compaction Technology 171
 2005 by Taylor & Francis Group, LLC
interrupted, this will affect the feed screw amperage readings and eventually the roll
drive readings. Weggel noted that forcefeeding the material deaerated the feedstock
and reduced the roll pressure loads. He found that it was more efficient to achieve
maximum densification during several stages of the roller compaction process (32).
Miller noted, in 1996, that the design of vertical cylindrical feed screws appears
to feed powder uniformly to the circumference of the feed screw (13). However, he
concluded that the powder feeding to the rolls did not coincide with the general rectangular
shape of the compactor throat. The observed drawback was that the powder
did not get delivered evenly across the rolls; the middle of the rolls received more
powder than the rolls edge. This potentially created a poor quality compact;
a strong middle compact that is weak at the ends because of frayed edges. Frayed
compact edges give rise to uncompacted material and excess fines (13). An example
of the phenomena was depicted when using a light yellow-green color feedstock,
which after roller compaction using a twin feed screw system produced a homogenous
colored compact. The same feedstock passed through a compactor designed with
a single vertical feed screw, exhibited a sinusoidal colored compact (Fig. 7). Miller indicated
that a multiple horizontal feed screw design system provided a more uniform
powder distribution across the rolls in area and in volume than a single vertical feed
screw system. This applied to either vertical or horizontal single feed screw systems
A number of these attributes, best technology practices, and features were rated for
their industrial and pharmaceutical importance and reported by Miller and Sheskey
in 2001 (33). Survey findings evaluated future industry usage of roller compaction
Worldwide pharmaceutical manufacturing technical operations predict signifi-
cantly increased roller compaction usage in the future. Research and development
scientists indicated worldwide that they are currently involved with roller compac-
Figure 7 Roller compaction: compacts. (Courtesy of R.W. Miller.)
172 Miller
technology in technical operations and in research and development (Fig. 8).
 2005 by Taylor & Francis Group, LLC
tion technology to develop new formulations. Some key reasons why this is the case
is that most API discovered now are very sensitive to water and heat and have low
bulk densities. The authors concluded that these combined factors push roller
compaction technology to the forefront as the preferred processing technology for
new pharmaceutical chemicals for solid-dose formulations in the pharmaceutical
industry (33). It also appears that pharmaceutical companies are chiefly content to
have the right technology in place to deliver the granulation form needed in the
easiest, quickest, and most efficacious manner.
A significant survey response indicated the importance of both design and service
features to advance roller compaction technology (33). There seems to be an
assorted blend of needs that are due in part to vendor past performances, such as
reliability, technical support programs, and service history. There is also a component
of innovation, new technology advances, such as cantilevered rolls and operator
ceutical Review (33)].
P. Guigon and O. Simon, Compiegne Universite de Technologie, Compiegne Cedex,
France, recently presented experimental results using a laboratory Komarek
out vacuum deaeration system), to feed a roll press, the stress exerted on the compact
was neither homogenous on the roll width nor constant with time.
The heterogeneity was characterized by measuring light transmission through
a sodium chloride compact. The team showed that the powder packing that took
place in the last flight of the feed screw caused heterogenous feeding pressure. The
heterogenous feeding pressure caused the heterogenous compact properties (34).
The team developed a mathematical model to link feed screw geometry to the stress
distribution in the compact. The author notes that this finding is contrary to
Figure 8 (A) Roller compaction current usage in company technical operations. (From
Ref. 33.) (B) Roller compaction current usage in R&D projects. (From Ref. 33.)
Roller Compaction Technology 173
B100QC press (Fig. 9). They demonstrated that when using a single feed screw (withtouch
screens with parametric feedback [Table 3, courtesy of the American Pharma-
 2005 by Taylor & Francis Group, LLC
the Johanson model noted earlier in this chapter and quite revealing. The laboratory
press was specially instrumented with two flush mounted piezoelectric transducers
fitted on the smooth upper roll (130mm diameter by 50mm width). The transducers
measured stress (0200 MPa) exerted on the roll surface at 15mm from both sides of
the roll. The position of the piezoelectric transducers was defined once per revolution;
the roll speed, screw feeder speed, and hydraulic pressure were also measured
The team measured the compact density by the distribution of light transmitted
through the sodium chloride compact. Guigon indicated that the milled sodium
chloride crystals were oriented by the applied stress and therefore the incident light
was not diffused similarly in all directions (34). The stress applied on the compact is
Table 3 Preferences for Compactor Design and Service Features
Compactor design and
service features Number of responses Average rating
Machine performance and reliability 56 4.5
Vendor spare parts availability 55 4.2
Variety of model sizes 61 4.1
Vendor IQ and OQ program 51 3.9
Vendor service history 48 3.9
Cantilevered vs. double-bearing rolls 49 3.8
Vendor delivery time 54 3.8
Ease of cleaning 54 3.7
Variety of roll patterns 51 3.7
Vendor training support 54 3.5
Touch screens with parameter feedback 60 3.2
Other 11 3.2
Note: 5 ? high, 1 ? low preferences.
Source: From Ref. 33.
Figure 9 Side and front view of Komarek B100QC press used to determine geometry feed
system and stress distribution and applied on compact. (From Ref. 34.)
174 Miller
and recorded (Fig. 10).
 2005 by Taylor & Francis Group, LLC
neither homogenous on the roll width nor constant with time. Guigon points out
that the stress maximum for the same sensor varies from one rotation to the next
and comparing the left and right sensors, one has a low signal while the other has
a high signal during each turn. He noted that the maximum stress measured for each
roll revolution was a cycle of the time equal to the time period of the feed screw
stress, illustrating that the maximum stresses are not applied to the center of the roll
but more toward the roll edges. The scientists showed that the applied stress on the
compact in the roller gap is due to the feed pressure distribution from the feed screw.
The nature of the feed pressure was due to the pushing of the powder in the last
spiral of the feed screw. The team took many stress profile readings to measure
the heterogeneity of the applied stress, as it was not constant or homogenous along
the roller width (Fig. 12).
Graphically, the light transmitted through the salt compact and the maximum
measured stresses (from the two locations on the roll) enabled the stress measurement
throughout the compact in the gap. The compact zones that had endured less
stress transmitted less light and therefore appeared darker (34).
indicated that the feed screw exit pressure was a function of the geometry and surface
Figure 10 (A) Applied stress on the compact during one rotation period of feed screw. (B)
Corresponding gray level on the numerical image of light transmitted through the compact.
Sodium chloride d50 ? 74 mm, Vr ? 10 rpm, Vs ? 25.3 rpm. (From Ref. 34.)
Roller Compaction Technology 175
(Fig. 11).
Figure 12 shows a good representation of the heterogeneity of the applied
Guigon sketched the feeding zone just prior to roll compaction (Fig. 13). He
 2005 by Taylor & Francis Group, LLC
properties of the feed screw and of the compressibility of the powder. Summarizing
his findings, the pressure at a location (y, z planes, perpendicular to the x-direction of
motion) of the feeding plane will vary due to the motion of the last flight of the
screw. When the wall of the screw flight is far from the feeding plane (top part of
and the localized pressure at the feeding plane is low. When the flight at this location
moves forward, the powder first compacts without moving significantly, increasing
progressively its density and local pressure at the feeding plane.
The team proved the observation with two dynamic motion picture experiments.
The first study showed the motion of particles at the check plate wall and
confirmed by video analysis that the particles do not move continuously in the feed
zone (34). Guigon noted that the powder particles are almost stationary. Then, the
powder moves forward before stopping again. The velocity fluctuations have
the same period as the screw. The second experiment consisted of using sodium
chloride and pushing the material to the rolls by means of a piston-feeding device.
the densest part of the compact is through the center, where most of the light is
The work by Guigon and Simon provided new meaning to the importance of
the feed screw pushing powder into the rolls and the feed screw designs influence
on compact formation. The author of this chapter raises some questions on how a
multiple feed screw system under vacuum deaeration environment would affect compact
formation. No information has been published on investigating these aspects.
Figure 11 Stress profiles recorded simultaneously on the (A) left and (B) right transducer vs.
number of roller rotations during consecutive turns. (Screw speed, Vs; roller speed, Vr; sodium
chloride (Vr ? 8.4 rpm, Vs ? 22.3 rpm). (From Ref. 34.)
176 Miller
Fig. 13), the powder between the flight and the feeding plane is not compacted
The compact is shown in Figure 14 with light passing through it. Most notably,
 2005 by Taylor & Francis Group, LLC
Figure 12 Distribution of stresses applied to a compact during the dynamic state of roller
compaction. (From Ref. 34.)
Figure 13 Sketch of the feeding zone. (From Ref. 34.)
Roller Compaction Technology 177
 2005 by Taylor & Francis Group, LLC
A key factor limiting compaction throughput and quality is air entrapment in powder
materials. During compression, air-occupying voids between particles are compressed
and squeezed. The gas pushes through the powder causing powder
fluidization and a nonuniform level of powder at the roll gap. It is best described
in Figure 15. This situation limits compact throughput and creates a nonuniform
compact density. Air entrapment creates excess fines prior to sizing because of spidering
compact edges.
The spidering condition occurs when gases rush across the inside of a compact
to thinly and weakly formed flaked edges. The flake edges break apart perpendicular
to the compaction direction. The compact edge breakage appears saw tooth in
Figure 14 Distribution of stresses applied to a compact, by piston pushed material into rolls,
not feed-screw pushed. (From Ref. 34.)
Figure 15 Pattern of gas escape from nip region. (From Ref. 36.)
178 Miller
 2005 by Taylor & Francis Group, LLC
structure and varies in length depending on the nature of the powder binding properties,
the amount of air entrainment, and the roll dwell time.
Johanson predicted theoretical compactor operating conditions to handle air
entrapment effects in materials. In general, he concluded that the critical compacting
pressure level is dependent on a number of factors: roll diameter and speed, powder
permeability, compressibility, and compact strength. Johanson indicated, when
applying these principles in commercial application, that a compactor operator
would have to operate the press at slightly less than maximum pressures to allow
for material inconsistencies and variable in-feed flow rates (35).
Johanson and Pietsch described the forces at work in a roll press when powders
are compacted (25,35,36). Johanson illustrated the three typical work phases during
compaction Figure 16. The first region is the initial solids contact pressure with the
feed screw. Pressure at this point is nil, P0. Later, as the solids begin to get pushed
and gripped by the rolls, pressure continues to build. Ultimately, the solids are
moved into the narrowest point of the nip angle where the maximum pressure occurs
at pmax. The second region describes the solids density increase through compaction
and densification occurs rapidly, in less than a second. Some materials have properties
that after discharge from the compactor undergo elastic deformation, which can
reduce the compacts density, expressed by the dotted curve. The third region of
force at work is air pressure. Models such as the one described take into account that
it is dependent on the porosity of the particulate solids compacted. If the porosity of
the compacted material remains high enough during the compaction process, air
pressure can escape and vent during and after compaction. On the other hand, if
the material has low porosity, air pressure builds up to high levels because it cannot
escape easily. This can be seen in several ways, one of which is through expansion of
the compact bursting. Bursting can be associated with a popping sound when operating
the compactor. The second way occurs when the compacted sheet breaks into
slivers or spidering at the compact edges.
Both Johanson and Pietsch reported that expanding gas in a compact is detrimental
to the compaction process: by reducing the compaction throughput and
increasing the amount of fine particles. Johanson illustrates the effects of roll speed
and powder porosity on air pressure in a compacted sheet (43). He shows a relative
large roll speed operating range when compacting a permeable (porous) powder. Air
entrainment does not limit roller speed for coarse granular powders. On the other
Figure 16 Compaction work phases. (From Ref. 43.)
Roller Compaction Technology 179
 2005 by Taylor & Francis Group, LLC
hand, when compacting very fine powders, the operating roll speed range is signifi-
cantly reduced because of air entrainment.
Miller indicated that the evenness of the powder feed into the rolls determines,
to a large extent, the success of compaction. Roller compactor systems suffer from
two disadvantages: as the powder feed bulk density approaches 0.3 g/cm3 or less,
the compaction throughput efficiency decreases. Secondly and concurrently, the
uncompacted powder leakage generally increases around the rolls (21). Miller, in
1994, described a new machine design improvement that used vacuum deaeration
to remove air entrainment from the powder just prior to the nip angle during roller
compaction. The multiple benefits of such action are significant when compacting
low-density raw materials (21):
 More uniform powder feed to the rollers
 Less voltage and amperage variability for the roll pair
 More uniform and strong compact
 Less powder leakage
 Greater yield
 Less powder adhering to the compact prior to sizing
 Higher compact throughput
 Less airborne particles.
The newly designed equipment involved a compactor fitted with two horizontal
feed screws, which featured vacuum deaeration. Specifically, the roller compactor
was equipped with a conical storage hopper containing a variable speed agitator.
Bulk powder was fed directly from the top of the hopper to the top of twin horizontal
auger feed screws, which directly transported the powder to the nip roll area (13).
Top and side views of the design features are shown in Figure 17 .
A novel stainless steel encasing that leads to the compactor rolls encloses the
variable-speed auger screws. Just before the nip area, a pair of sintered stainless steel
segments are assembled within the horizontal auger feed system, which can operate
under partial vacuum. A small, self-contained vacuum pump draws negative pressure
through a dry filter and a stainless steel line connected to the sintered assembly
plates. The partial vacuum is adjustable from 0.1 to 0.8 bars. The compaction rolls
Figure 17 Compactor side view of auger feed screw system and sintered plate segment.
(From Ref. 13.)
180 Miller
 2005 by Taylor & Francis Group, LLC
operate at different speeds and are supported on heavy-duty bearings in such a way
that the lower roll is fixed and the upper roll is slightly movable in the vertical plane.
The deaerated material passes through the roll pair, which is under infinitely variable
hydraulic pressure. The deaeration, auger feed screws design and speed, roll speed,
and hydraulic roll pressures are the main factors in producing a compact with
specified properties.
In several experiments,Miller studied the effects of using horizontal twin auger feed
screws under partial vacuum (Alexanderwerk Inc. Horsham, PA, USA, model 50/75
compactor). The feed powder was vacuum deaerated just before roller compaction.
The experimental design showed that the compactors deaeration feed system signifi-
cantly increased compaction output and minimized powder leakage when compacting
very low-density blends (<0.35 g/cm3); remarkable results were observed (22).
To evaluate the effectiveness of the compactors deaeration system, a test was
designed to process a low-density active drug blend with and without the activated
deaeration system. The test showed how much material was compacted with the
deaeration system engaged and how much material was compacted when it was
not engaged. The test also determined how much material was not compacted (powder
roll leakage) in each case. Details of the experimental plan are found in Ref. 38.
Results of this experiment showed that under the influence of vacuum deaeration,
the compactor produced throughputs of 100 kg/hr and the resultant noncompacted
material leakage rate was <2%. The vacuum deaeration was so effective that there
was no need to recirculate the noncompacted powder back to the rolls to meet processing
specifications. Under a second set of compacting conditions, vacuum deaeration
was not activated and the process typically produced 7080 kg/hr of densified
compact. The powder leakage rate increased to 2030%. During this set of conditions,
the powder flow to the roll pair was uneven and the ribbon compact was
not uniform. Processing this formulation for a long period of time under these conditions
would have required recirculation equipment to return the uncompacted
meters (trials 1 and 2).
In several other trials, Miller also studied the effects of using two different comviously
described, were carried out using differently designed compactors. The
objective of these trials was to increase the active bulk density of the feedstock from
0.22 to 0.61 g/cm3 (13). Miller reported that the compactor, machine 1, containing
3 density at 80 kg/hr yield. When the
vacuum deaeration was not activated the desired granule density specification could
drug granules during the first compaction run when employing vacuum deaeration.
After a second compaction pass with vacuum deaeration, trial 2, the powder density
was increased slightly to 0.58 g/cm3. It appeared that the level of vacuum being
applied to the active drug substance, in machine 2, was insufficient, ineffective, or
Millers experiments evaluated the effects of powder density; screw feed speed,
roll speed, roll pressure, vacuum deaeration pressure, compaction rate, and the compaction
leakage rate. Test results demonstrated that the first compactors deaeration
and feed system designs significantly increased compaction output. The new equip-
Roller Compaction Technology 181
the horizontal twin screws and the vacuum deaeration design as described in Figure.
17, produced granules with 0.630.64 g/cm
A second compactor, machine design 2 (Fig. 18), did not densify the active
powder to the rolls for further processing (21,37,38). Table 4 for compactor parapactors
vacuum deaeration systems (Table 5). The tests, similar to the ones prenot
be achieved. Table 6 (trials 1 and 2) for operating parameters and results.
not optimally positioned, even at its maximum vacuum deaeration level (13). Table
7 (trials 13) for machine 2 operating parameters and results.
 2005 by Taylor & Francis Group, LLC
ment design and process provided high compact yields and virtually eliminated powder
leakage, obviating the need for expensive powder recirculation equipment. The
of approximately 0.2 g/cm3. In summary, a new critical condition, vacuum deaeration,
had been identified in optimizing roller compacting effectiveness and efficiency
(13). Miller concluded that four key processing conditions must exist to optimize
roller compaction throughput and minimize powder leakage around the rolls (13):
 Adequate powder supply must enter the gripping zone.
 Powder must be fully conveyed into the narrowest part of the roller gap.
 Compaction pressure must be distributed as uniformly as possible over the
whole of the roller-gripped powder mass.
 Sufficient vacuum deaeration must be effectively distributed prior to the nip
roll region, particularly for low bulk density powder feed stock.
Work by this author in the mid-1990s investigated roller compaction and near-infrared
spectroscopy (NIRS) technology, by examining three different active ingredient
Table 4 Compactor Operating Conditions and Yields for Poor
Following Low-Density Active Drug Blend, Formulation 1, Vacuum
Deaeration and Non Vacuum Deaeration Trials
Formulation 1
Conditions Trial 1 Trial 2
Powder density (g/cm3) 0.250.35 0.250.35
Screw feed (rpm) 52 52
Roll speed (rpm) 8 8
Vacuum (bar) (0.780.78) 0
Roll pressure (bar) 6065 6065
Compact rate (kg/hr) 100 7080
Compact leakage rate (kg/hr) 2 1520
Source: From Ref. 13.
Table 5 Compactor Operating Conditions and Yields for Poor
Flowing Low-Density Active Drug Blend, Formulation 2, Vacuum
Deaeration and Non-vacuum Deaeration Trials
Formulation 2
Conditions Trial 1 Trial 2
Powder density (g/cm3) 0.250.35 0.250.35
Screw feed (rpm) 52 52
Roll speed (rpm) 8 8
Vacuum (bar) (0.780.78) 0
Roll pressure (bar) 6065 6065
Compact rate (kg/hr) 150 100110
Compact leakage rate (kg/hr) 1.3 2030
Source: From Ref. 13.
182 Miller
first compactors vacuum deaeration design (Fig. 17) proved to be superior to that
of the second machine (Fig. 18) when compacting an active bulk drug with a density
 2005 by Taylor & Francis Group, LLC
blends subjected to different compaction conditions and mapped specific in-process
roller compaction processing steps and relationships. Miller illustrated the use of
NIRS to map, in the static mode, roller compaction unit operations. This was the
first published work about using any spectroscopy to map roller compaction processing
steps (39).
Blends were compacted using an Alexanderwerk roller compactor, model
WP50/75N (Alexanderwerk Inc., Horsham, PA) fitted with vacuum deaeration.
All compacts were sized through a continuous series double rotary granulator fitted
with 3.15 and 1.25mm mesh screens, respectively. Blends were finally mixed for
2 min and compressed into tablets using a Kilian E 150 press (Kilian & Co Inc.,
Horsham, PA), which was equipped with standard concave 11mm diameter tooling.
NIR spectral analysis was conducted using a Rapid Content Analyzer, model
6500 NIR spectrophotometer (Foss NIRSystems, Silver Spring, MD). NIRS
libraries were developed for each active drug concentration and the specific
Table 6 Operating Parameters and Results of CompactorMachine
Design 1When Compacting Poor Flowing Low-Density Active Drug
Substance Blends
Conditions Trial 1 Trial 2
Initial density (g/cm3) 0.22 0.22
Powder density (g/cm3) 0.6364 0.45
Screw feed (rpm) 52 52
Roll speed (rpm) 8 8
Vacuum (bar) 0.65 0
Roll pressure (bar) 50 50
Compact rate (kg/hr) 80 42
Source: From Ref. 13.
Figure 18 Compactor front view of vertical feed screw system with vacuum deaeration.
(From Ref. 13.)
Roller Compaction Technology 183
 2005 by Taylor & Francis Group, LLC
compactors process critical control points (PCCPs) operating conditions, details of
the experimental plan are referenced (39).
The NIRS absorbance shows (optical density) differences between 10 and
20 bar roll pressure compacts and also differences between compacts manufactured
with and without vacuum deaeration (Fig. 19). The associated NIR ibuprofen/
starch compact spectra also show an absorbance (optical density) hierarchy. The
higher the roll pressure with vacuum deaeration the greater the NIR optical density
(10 bar, no vacuum<10 bar, vacuum<20 bar, no vacuum<20 bar, vacuum).
tion. Figure 20 compares the two levels of roll pressure with and without vacuum
deaeration to a third parameter, compaction roll power consumption (CRPC).
CRPC is a power force measured from the roll drive that is continuously monitored
throughout the compaction operating process. An ibuprofen/starch compact trend
analysis shows that the greater the roll pressure with vacuum deaeration the greater
the CRPC (10 bar, no vacuum<10 bar, vacuum<20 bar, no vacuum<20 bar,
vacuum). The analysis demonstrates that employing vacuum deaeration increased
the CRPC at both low and high roll pressure levels. As the vacuum deaeration
increased, the width and gauge of the ibuprofen/starch powder compacts also
increased. Additional characterizations for aspirin/starch and acetaminophen/
starch compacts were completed, Ref. 39.
Table 7 Operating Parameters and Results of Compactor Machine Design 2 When
Compacting a Poor Flowing LowDensity Active Drug
Conditions Trial 1 Trial 2 Trial 3
Initial density (g/cm3) 0.22 0.45 0.22
Powder density (g/cm3) 0.45 0.58 0.46
Screw feed (horizontal) (rpm) 30 30 30
Screw feed (vertical) (rpm) 260 260 260
Roll speed (rpm) 6 6 6
Vacuum (bar) 0.34 0.34 0
Roll pressure (bar) 50 50 50
Compact rate (kg/hr) 40 48 36
Source: From Ref. 13.
Figure 19 NIR absorbance processing 50/50 Ibuprofen/Starch compacts. (From Ref. 39.)
184 Miller
Figure 20 further illustrates the NIRS ibuprofen/starch 50/50-compaction evalua-
 2005 by Taylor & Francis Group, LLC
Figure 21 depicts granule NIRS sensitivity to vacuum deaeration and roll pressure
sized compacts. In general, granules produced from sized compacts manufactured
without vacuum deaeration and at lower roll pressure levels exhibited higher
made from the same corresponding conditions. The results demonstrated that successive
ibuprofen/starch processing steps, compacting, and sizing produced uniquely
different NIR absorbance amplitudes.
A new FDA guideline initiative (20032004), Process Analytical Technologies (PAT)
encourages manufacturers to use real time in-line, on-line, or at-line nondestructive
sensory technologies to monitor process critical control points. The PAT framework
addresses every aspect from incoming raw materials to optimization to continuous
improvement. It starts with processability of the incoming raw materials; their attributes
would be used to adjust the process parameters. The incoming material attributes
could be used to predict or adjust optimal processing parameters. The link
between PAT and the new 21st Century Good Manufacturing Practices initiative,
from the FDAs perspective, is quality depends on knowledge and PAT brings
Figure 20 50/50 Ibuprofen/Starch compacts CRPC compactor and NIR corresponding
values. (From Ref. 39.)
Figure 21 90/10 Aspirin/Starch granules NIR absorbance with compactor settings. (From
Ref. 39.)
Roller Compaction Technology 185
optical density. NIR spectra of 90/10 ibuprofen-processing effects are shown in Figure
22: final blend, compact 20 bar roll pressure (vacuum deaeration) and granules
 2005 by Taylor & Francis Group, LLC
more knowledge and understanding of processes. This is the point where science and
risk-based decisions can be made in terms of manufacturing per Dr. A. Hussain
In 20022003 Gupta, Morris, and Peck from Purdue University Industrial
Physical Pharmacy School began a series of noninvasive real-time investigations.
Evaluating roller compaction by NIR in the dynamic mode, the team monitored
the variation of compacted ribbon strength that could adversely affect particle size
and distribution of granules obtained after milling a compact (41). Their thinking
was that both the ribbon strength and the particle-size distribution can be determined
off-line but it is time consuming. The ability to use NIR spectra to quantify
the compact strength, the particle size, and to estimate postmilling particle-size
distribution could be done in real time. The most obvious NIR spectral change that
occurs in compacts or tablets prepared under increasing pressure is an upward shift
in the NIR spectral baseline (39). Kirsch and Drennen used the slope of the best-fit
line through spectra to quantify this upward shift and found a linear correlation
between this slope and the tablet hardness determined by the diametral compression
test (42). Gupta et al. applied this treatment to correlate the slope of the best-fit line
through the NIR spectra with the strength of compacts was determined using the
three-point beam bending technique (41). Additionally, slope values were correlated
Figure 22 NIR absorbance 90/10 Ibuprofen/Starch processing effects. (From Ref. 39.)
Figure 23 MCC compacts: effect of roll speed. (From Ref. 41.)
186 Miller
 2005 by Taylor & Francis Group, LLC
to the particle-size distribution of granules produced from milled compacts. Realtime
on-line techniques were used to monitor the roller compaction unit operation
for microcrystalline cellulose and 10% active blend. The spectra collected on the
different compact segments prepared from the same material under the same roller
compactor settings showed little variation, suggesting good reproducibility in the
data collection by the NIR sensor. A significant shift in the NIR spectra was
observed for compacts prepared at different roll speeds.
At constant feed screw speeds, the spectra shifted toward lower absorbance
prepared from both the placebo and the 10% tolmetin blends. The shift in the
NIR spectra may be quantified, using the slope of the spectral baseline or using
the slope of the best-fit line, by regression analysis through the spectrum (41). Use
of the slope of the best-fit line was advantageous over the slope of the baseline since
the best-fit line uses information from the entire spectrum. The normalized force
values for the compacts showed strong linear correlation with the slopes of the spectral
best-fit lines (Fig. 24). It was observed that the particle-size distribution of the
milled compacts exhibits a steady decrease in the d90, d50, and d10 values with increasing
roller-compactor roll speed at constant rate (Fig. 25).
Figure 24 Relationship between slope of NIR spectrum and compact strength. (From
Ref. 41.)
Figure 25 Relationship of milled 100% compacts at various roll speeds. (From Ref. 41.)
Roller Compaction Technology 187
values with increasing roll speed (Fig. 23). This shift was observed with compacts
 2005 by Taylor & Francis Group, LLC
The values for the particle-size distribution for the milled 10% tolmetin compacts
correlated well with the slopes of the spectral best-fit line (Fig. 26). This clearly
shows the potential of the technique in predicting the particle-size distribution of
milled compacts under a given set of milling conditions. Multivariate analysis of
the spectral data using PLS regression analysis confirmed the ability of the slope
of the spectral best fit to estimate the compact strength and the postmilling particle-
size distribution. For greater detail in this expanding new technology area the
interested reader should examine the references.
The author gratefully acknowledges and thanks Carol G. Miller for her contributions
and assistance in editing and data manuscript control. The author thanks
and appreciates Dr. Pierre Guigon, Professor and Director of Chemistry at the University
of Technology of Compiegne, France, for his collaborative contributions to
the text. The author thanks and appreciates Mr. Abhay Gupta, Ph. D. former graduate
student at the Industrial Physical Pharmacy School of Purdue University, West
Lafayette, IN, USA. The author acknowledges and thanks Mr. Leonard Minervini,
Alexanderwerk, Inc., Hatboro, PA, USA, for electronic files of photographs and
1. Kristensen HG, Schaefer T. Granulations. Swarbrick J, Boyland J, eds. Encyclopedia of
Pharmaceutical Technology. Vol. 7. New York: Marcel Dekker, 1993:121126.
2. Capes CE. Particle size enlargement. Williams J, Allen T, eds. Handbook of Powder
Technology. Vol. 1. Amsterdam: Elsevier, 1980.
3. Capes CE. Particle size enlargement. Grayson M, Eckroth D, eds. Kirk-Othmer Encyclopedia
of Chemical Technology. Vol. 21. 3rd ed. New York: Wiley, 1978:77105.
4. Pietsch WB. Size enlargement by agglomeration. In: Fayed M, Otten L, eds. Handbook of
Powder Science and Technology. New York: van Nostrand Reinhold Co, 1984:Chap. 7.
5. Pietsch WB. In: Roth D, ed. Agglomeration Techniques for the Manufacturing of
Granular Materials with Specific Product Characteristics. The Institute for Briquetting
and Agglomeration Proceedings. Vol. 25. Erie, PA: Gannon University, 1997:49164.
6. Sherrington PJ, Oliver R. In: Granulation. London: Heyden & Son, 1981.
Figure 26 Milled tolmetin 10% compacts. (From Ref. 41.)
188 Miller
 2005 by Taylor & Francis Group, LLC
7. Kapur PC. Adv Chem Eng 1978; 10:55.
8. Kristensen HG, Schafter T. Drug Dev Ind Pharm 1987; 13:803.
9. Lindberg N. Industrial wet granulation. Acta Pharm Suec 1988; 25(special issue):
10. Fonner DE, Anderson NR, Banker GS. Granulation and tablet characteristics. In:
Lieberman HA, Lachman L, eds. Pharmaceutical Dosage Forms. Tablets. Vol. 2. New
York: Marcel Dekker, Inc., 1982:Chap. 5.
11. Anderson NR, Banker GS, Peck GE. Principles of improved tablet production system
design. In: Lieberman HA, Lachman L, eds. Pharmaceutical Dosage Forms. Tablets.
Vol. 3. New York: Marcel Dekker, Inc., 1982:Chap.1.
12. Ghebre-Sellassie I. Pharmaceutical Pelletization Technology. New York: Marcel Dekker
Inc, 1989.
13. Miller RW. Roller Compaction Technology. Parikh DM, ed. Handbook of Pharmaceutical
Granulation Technology. Vol. 81. New York: Marcel Dekker Inc, 1997:99150.
14. Pietsch WB. Roll Pressing. London: Heyden, 1976.
15. Parrot EL. Pharmaceutical Dosage Forms. Vol. 2. New York: Marcel Dekker Inc,
16. Dehont FR, Hervieu PM, Jerome E, Delacourte A, Guyot JC. Briquetting and granulation
by compaction: a new granulator-compactor. Wells J, Rubinstein M, eds.
Pharmaceutical Technology, Tableting Technology. Vol. 2. London: Ellis Horwood,
17. Heckel RW. Density-pressure relationships in powder compaction. Trans Metall Soc
AIME 1961; 221:671675.
18. Johanson JR. Rolling Theory for Granular Solids. Trans Am Soc Mech Eng 1965:
19. Jenike AW, Shield RT. Plastic flow of Coulomb solids beyond original failure. J Appl
Mech 1959; 26:599602.
20. Parrott EL. Densification of powders by concavo-convex roller compactor. J Pharm Sci
1981; 70(30):288291.
21. Miller RW. Advances in pharmaceutical roller compactor feed system designs. Pharm.
Technol 1994:154162.
22. Shileout G, Lammens RL, Kleinebudde P. Dry granulation with a roller compactor Part
1: the function units and operational modes. Pharm Technol Eur 2000:2435.
23. Miller RW. Roller Compaction Technology for the Pharmaceutical Industry. Encyclopedia
of Pharmaceutical Technology. New York: Marcel Dekker, Inc., 2003 (on-line).
24. Dehont FR, Hervieu PM. Briquetting and granulation by compaction new granulator
compactor for the industry. Drug Dev Ind Pharm 1989; 15(1416):22452263.
25. Pietsch W. Size enlargement by agglomeration. Fayed M, Otten L, eds. Handbook of
Powder Science, and Technology. 2nd ed. New York: Chapman and Hall, 1997:
26. Funakoshi Y, Asogawa T, Satake E. Use of a novel roller compactor with a concavoconvex
roller pair to obtain uniform compacting pressure. Drug Dev Ind Pharm 1977;
27. Parrott EL. Densification of powders by concavo-convex roller compactor. J Pharm Sci
1981; 70(3):288289.
28. Jerome E. Measurement of resulting forces on a roller compactor. Drug Dev Ind Pharm
1991; 17(12):15711591.
29. Johanson JR. Predicting limiting roll speed for briquetting presses. Proceedings of the
13th Institute for Briquetting and Agglomeration. Vol. 13. 1975:8999.
30. Sheskey PJ, Hendren J. The effects of roll compaction equipment variables, granulation
technique, and HPMC polymer level on a controlled-release matrix model drug formulation.
Pharm Technol 1999; 23(3):90106.
31. Johanson JR. Roll press feed systems. Proceedings of the 24th Institute for Briquetting
and Agglomeration, October. Vol. 24. 1995:149163.
Roller Compaction Technology 189
 2005 by Taylor & Francis Group, LLC
32. Weggel RW. The basics of force-feed roll briquetting. Manuf Eng 1982:107109.
33. Miller RW, Sheskey PJ. Survey of current industrial practices and preferences of roller
compaction technology and excipients year 2000. Am Pharm Rev Spring 2001;
34. Guigon P, Simon O. Correlation between the geometry of feeding system and the stress
distribution applied on the compact. In: Roth D, ed. Proceedings of the 27th Institute for
Briquetting and Agglomeration, Providence, RI, Nov 2001. Vol. 27. Erie, PA: Gannon
University, 2001:3141.
35. Johanson JR, Cox BD. Fluid entrainment effects in roll press compaction. Proceedings of
the 20th Institute for Briquetting and Agglomeration. Vol. 20. 1987:251263.
36. Dec RT. Problems with processing of fine powders in roll press. Proceedings of the 24th
Institute for Briquetting and Agglomeration. 1995; 24:199210.
37. Miller RW. Using vacuumdeaeration feed system to minimize powder leakage during
roll compaction. Powder Bulk Eng 1997; 11(2):7175.
38. Miller RW. Vacuum deaeration advances in pharmaceutical roller compaction technology.
Proceedings of the 24th Institute for Briquetting and Agglomeration 1995;
Vol. 24:165173.
39. Miller RW. Roller compaction optimizationNIR in-process mapping. J Pharm Technol
1999; yearbook supplement:3039.
40. Miller RW. Process analytical technologies PAT part 2. Am Pharm 2003; 6(1):5261.
41. Gupta A, Peck GE, Morris KR. Near infrared monitoring of roller compaction. In:
Roth, ed. Proceedings of the Institute of Briquetting and Agglomeration. Vol. 28. Erie,
PA: Gannon University, 2003:2034.
42. Kirsch JD, Drennen JK. Nondestructive tablet hardness testing by near-infrared spectroscopy:
a new and robust spectral best-fit algorithm. J Pharm Biomed Anal 1999; 19(34):
43. Johanson JR. Reducing Air Entrainment Problems in Your Roll Press. Powder and Bulk
Engineering 1989; 3(2):4346.
190 Miller
 2005 by Taylor & Francis Group, LLC
High-Shear Granulation
Rajeev Gokhale*
Incyte Corporation, Wilmington, Delaware, U.S.A.
Yichun Sun and Atul J. Shukla
College of Pharmacy, University of Tennessee, Memphis, Tennessee, U.S.A.
Granulation is the process of agglomeration of a powder mixture, which results in the
enlargement of the particles. This is often necessary for manufacturing of solid dosage
forms such as tablets. The materials, which are compressed into tablets, must possess adequate
flowability, density, and compressibility. This is because the requisite amount of
powder mixture required to compress each tablet is filled into the die cavity by volume
and not by weight. This requirement of adequate flowability, density, and compressibility
is particularly important during a high-speed tablet production where the dwell time is
often short. For example, active pharmaceutical ingredients such as ibuprofen and acetaminophen,
which have inadequate flow and compression properties, and a relative high
dose, are often granulated prior to compression into tablets. Thus, the overall purpose of
granulation is to improve the flowability and compressibility of the powder mixture.
Besides improving the flowability and compressibility, the granulation process can also
 Densify the powder mixture and reduce dust
 Narrow the particle size distribution of the powder mixture
 Ensure uniform distribution of the drug in the powder mixture
 Improve the dissolution characteristics of the finished tablets.
The three commonly used granulation methods include wet granulation, dry
granulation, and hot-melt granulation. These methods are categorized based on
the type of binder and the process employed during granulation. The equipment that
is used during the granulation processes is classified into the following three major
categories, based on the shearing strength it generates on the powder bed:
1. Low-shear granulatorstwin shell (Peterson Kelly, PK) with an agitator
bar, dough mixer or planetary mixer, ribbon blenders, and fluid bed
granulator without the rotogranulator
* Present Address: Merck Research Laboratories, West Point, Pennsylvania, U.S.A.
 2005 by Taylor & Francis Group, LLC
2. Medium-shear granulatorsfluid bed granulators with a rotogranulator
3. High-shear granulators.
The granulation techniques can affect the physical properties of the resulting
granules, which then subsequently affect the tableting process, the quality, and the
performance of the finished tablets, as illustrated in the following examples.
An extended release of matrix tablet formulation for metoprolol tartrate (100mg)
was prepared with three different granulation processes: direct compression, fluid-bed,
or high-shear granulation (1). Different grades of hydroxypropylmethylcellulose
(HPMC) (Methocel K4M, K15M, K100M, and K100LV) were used as fillers and
binders. Direct compression formulations exhibited poor flow, picking and sticking
problems during tableting. High-shear granulation produced hard granules, which were
difficult to mill. However, they yielded good tablets. Granules produced by the fluid-bed
granulation process appeared to have satisfactory flow and tableting performance.
A hydrophilic matrix tablet formulation containing 6070% drug, and lowor
medium-viscosity grades of HPMC for extended oral delivery of zileuton, was
prepared using the wet-granulation techniques (2). The granules were prepared with
planetary (low shear) and high-shear granulators. In vitro drug release was evaluated
using USP apparatus 1. Slower drug release was achieved from the granules prepared
with the high-shear granulation process, because of the slower water penetration into
the denser and less porous granules.
A high-shear mixer was utilized for the evaluation of pregelatinized,
cross-linked, waxy corn starches as binding agents in the wet-granulation process (3).
Lactose granules prepared in a high-shear mixer showed a larger average size than
granules prepared in a low-shear mixer (planetary mixer), using the same amount of
binding solution. The granules produced with high-shear mixer had lower friability
than those prepared with the planetary mixer.
High-shear granulation has been one of the most commonly used methods
to produce granules since the early 1980s (4). Hence, this chapter discusses in detail,
the equipment, process variables, formulation requirements, granulation end-point
determination, and scale-up considerations of the high-shear granulation process.
Most of the high-shear granulators consist of a mixing bowl, a three-bladed impeller,
and an auxiliary chopper. The shape of the mixing bowl could be cylindrical or conical.
The mixing bowl can be jacketed for heating or cooling the contents in the bowl,
by circulating hot or cool liquid or steam through the jacket. An impeller is
employed to mix the dry powder and spread the granulating fluid. The impeller of
the high-shear mixer granulator normally rotates at a speed ranging from 100 to
500 rpm. The function of the chopper is to break down the wet mass to produce
granules. The rotation speed of the chopper ranges from 1000 to 3000 rpm. The
high-shear granulator could be termed as either vertical or horizontal, based on
the orientation and the position of the impeller. The vertical high shear granulator
view of a top-driven vertical high-shear granulator, ULTIMAGRALTM/ULTIMAPRO
vertical high-shear granulators: ULTIMATM, GMXTM, and GMATM
shows photographs of top-driven, vertical, high-shear granulators: UltimaGral 150,
192 Gokhale et al.
could be either a top-driven or a bottom-driven unit. Figure 1 shows the schematic
 2005 by Taylor & Francis Group, LLC
. Figure 2(ac) shows photographs of choppers and impellers of top-driven
. Figure 3(ad)
Figure 1 Schematic view of top-driven vertical high shear granulator. (ULTIMAGRALTM/
ULTIMAPROTM courtesy of Niro Pharma Systems.)
Figure 2 (A) Photograph of an ULTIMATM top-driven impeller and chopper. (Courtesy
of Niro Pharma Systems.) (B) Photograph of a GMXTM top-driven impeller and chopper
(Courtesy of Vector Corporation.) (C) Photograph of GMATM top-driven impeller and
chopper (Courtesy of L. B. Bohle.)
High-Shear Granulation 193
 2005 by Taylor & Francis Group, LLC
GMX 10, GMA 1200, and Glatt. shows the schematic view of a
bottom-driven vertical high-shear granulator, with a horizontal chopper shaft
Pharma MatrixTM
photograph of a plow-shaped mixing blade and a chopper inside a horizontal
Figure 3 (A) Photograph of a top-driven vertical high shear granulator (UltimaGral 150).
(Courtesy of Niro Pharma Systems.) (B) Photograph of a top-driven vertical high shear granulator.
(GMX 10 courtesy of Vector Corporation.) (C) Photograph of a top-driven vertical
high shear granulator. (GMA 1200 courtesy of L. B. Bohle.) (D) Photograph of a top-driven
vertical high shear granulator. (Courtesy of Glatt Air Techniques.)
194 Gokhale et al.
Figure 4
(PMA through-the-wall design). Figure 5(A) and 5(B) shows photographs of top
views of bottom-driven impellers with horizontal choppers (Diosna and Glatt). Figure
PMA600 Podium, Diosna P 150, and Glatt. Figure 7 shows the
schematic view of a horizontal high-shear granulator (Loedige). Figure 8 shows a
 2005 by Taylor & Francis Group, LLC
6(AC) shows the photographs of bottom-driven vertical high-shear granulators:
high-shear mixer granulator (FKM-1200Loedige).
The size of the high-shear granulators varies considerably from the small
laboratory-scale models to large production-scale models, depending on the type of
the model and end use. High-shear granulators are often equipped with granulation
end-point control devices, which are used to detect the end point of the granulation
process. Some of the recently introduced features in high-shear mixer granulators are
clean in place (CIP) or wash in place (WIP), and one-pot processing. High-shear
granulators with CIP or WIP features can be cleaned easily, without removing the
bowl, the impeller, and the chopper. The advantage of the one-pot processing system
is that the wet granules can be dried in the same bowl without additional equipment
for drying. Some of the suppliers of the commercially available high-shear granula-
3.1. Wet Granulation
The composition of a powder mixture for granulation generally consists of an active
pharmaceutical ingredient (API), a filler, a disintegrant (usually not included in controlled
release tablets), and a binder.
A high-shear wet-granulation process includes the following steps:
1. Loading all the ingredients into the mixing bowl, which can be accomplished
by either of the following methods: gravity feeding with manual
or pneumatic valve, and vacuum feeding.
2. Mixing of dry ingredients such as API, filler, and disintegrant, at high
impeller and chopper speeds for a short period of time (25 min).
3. Addition of a liquid binder (either binder solution or solvent) into the powder
mixture, while both the impeller and the chopper are running at a low
Figure 4 Schematic view of a bottom-driven vertical high shear granulator with horizontal
chopper shaft. (PMA through-the-wall courtesy of Niro Pharma Systems.)
High-Shear Granulation 195
high-shear mixer granulator (Loedige). Figure 9 shows a photograph of a horizontal
tors are listed in Table 1.
 2005 by Taylor & Francis Group, LLC
4. Wet massing with both the impeller and the chopper running at a high
5. Removal of the resulting wet granules from the granulator bowl, and drying
them using an appropriate drying technique such as fluid-bed or tray
6. Sieving the dried granules.
The high-shear wet-granulation process offers several advantages over the
other granulation processes. These include:
 Short processing time
 Use of less binder solution
 Granulation of highly cohesive materials containing hydrophilic polymers,
which is not achievable with low-shear granulation processes
 Greater densification and production of less friable granules
 Production of reproducible granules with a uniformparticle size distribution
Figure 5 (A) The top view of a bottom-driven impeller with a horizontal chopper. (Courtesy
of Diosna.) (B) The top view of a bottom-driven impeller with a horizontal chopper inside a
conical shaped mixing bowl. (Courtesy of Glatt Air Techniques.)
196 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
 Reduction of process dust, thus minimizing exposure to workers
 Obtaining predictable granulation end-point determination.
Despite the above advantages, the process is not immune to challenges such as:
 Production of less compressible granules, compared to low-shear granulation
 Narrow range of operating conditions
A commonly employed high-shear granulation process requires multiple unit
operations, such as drying and sieving after wet granulation. The preferred drying
method is fluid-bed drying, because of the distinct advantages over tray drying.
However, a primary problem inherent in the two-step process of high-shear mixer
granulation and fluid-bed drying is the possibility of exposure of the workers and
Figure 6 (A) Photograph of a bottom driven Pharma MatrixTM PMA600 Podium/
Through-The-Wall. (Courtesy of Niro Pharma Systems.) (B) Photograph of a bottom driven
MixerGranulator. (P 150 courtesy of Diosna Dierks & Soehne GmbH.) (C) Photograph of a
bottom driven high shear granulator. (Courtesy of Glatt Air Techniques.)
High-Shear Granulation 197
 2005 by Taylor & Francis Group, LLC
the environment to potentially toxic materials during the transfer of the wet granules
from the high-shear granulator to the fluid bed. In order to remedy the contamination
problem, the processing suites need to be operated under a negative pressure using
powerful air filtration systems. Moreover, dehumidifying and heating of a large volume
of air is also necessary during the fluid-bed drying process. Regulating the airflow and
filtering the exhaust air during the drying process could also be challenging.
However, the above problems can be remedied by a one-pot approach in which
the high-shear wet-granulation and drying processes are combined into a single step.
This approach saves time, since cleaning and validating a single bowl is inherently
easier than tackling two or three units. For example, a Mi-Mi-Pro high-shear granulator,
combined with microwave, was employed to study the effect of process variables
on the properties of the granules obtained, for the mixture of alpha-lactose
monohydrate and microcrystalline cellulose (MCC) (5). The one-pot high-shear
granulators are commercially available from several manufacturers. Options such
Figure 7 Schematic view of a horizontal high shear mixer granulator. (Courtesy of Loedige.)
Figure 8 Photograph of a plough-shaped mixing blade and a chopper inside a horizontal
high shear mixer granulator. (Courtesy of Loedige.)
198 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
as oscillating the bowl during drying, vacuum drying, gas-assisted vacuum drying,
single-pot processor ULTIMAPROTM, based on the ULTIMAGRALTM with a
cessor at a swinging position (Niro Inc., Pharma Systems Division).
Another approach is to modify the wet-granulation process. This process is
called moisture-activated dry-granulation process. The moisture-activated drygranulation
process consists of two steps, wet agglomeration of the powder mixture
followed by moisture absorption stages. A small amount of water (14%) is added
first to agglomerate the mixture of the API, a binder, and excipients. Moisture
absorbing material such as MCC and potato starch is then added to absorb any
excessive moisture. After mixing with a lubricant, the resulting mixture can then
be compressed directly into tablets. Hence, this process offers the advantages of
wet granulation, but eliminates the need for a drying step.
The applicability of a 25 L high-shear mixer for moisture-activated dry granulation
of phenobarbital was investigated by Christensen et al. (6). MCC, potato starch,
or a mixture of 50% of each was used as moisture absorbing material. The results of
the study showed that the physical properties of the tablets were primarily affected by
the water content, the moisture absorbing material, and the compression force.
3.2. Hot-Melt Granulation
Hot-melt granulation in a high-shear granulator is a versatile approach. Hot-melt
granulation utilizes a binder, which is a solid or semisolid at room temperature
and melts at a temperature below the melting point of API. Generally, the melting
point of such binders is between 30C and 100C. The binder, when heated near
its melting point, liquifies or becomes tacky. This tacky and liquified form of binder
agglomerates the powder mixture, which upon cooling forms a solid granulated
mass. The energy to melt the binder may be derived from heat dissipated from the
circulating hot liquid, such as steam, water, or oil through jacketed bowl, or heat
generated by the friction of high-shear mixing. A solvent such as water or an organic
Figure 9 Photograph of a horizontal high shear mixer granulator. (FKM-1200 courtesy of
High-Shear Granulation 199
and microwave drying are also available. Figure 10 shows the schematic view of a
vacuum drying system. Figure 11 shows the photo of UltimaPro 600 single-pot pro-
 2005 by Taylor & Francis Group, LLC
compound is not necessary to initiate particle binding in this method. The resultant
granules are compressed into tablets.
A series of studies have been carried out by Schaefer and his colleagues to
investigate the effects of formulation and process variables on the hot-melt granulation
process, and the properties of the obtained granules (713).
In spite of numerous granulation options available to a formulator, high-shear
wet granulation remains the method of choice.
The high-shear wet-granulation process can be divided into five stages as shown
below, namely, mixing, adding binder solution, wetting and nucleation, consolidation
and growth, and granule attrition and breakage.
The binder solution or the granulating liquid is distributed through the powder
bed by mechanical agitation created by the impellers. At this stage, the powder mixture
becomes wetted and initiates agglomeration via nucleation. Nucleation of particles
occurs by the formation of liquid bridges between primary particles, which
adhere together to form agglomerates. At this stage, the concentration of liquid
phase in the powder mixture is relatively low, but high enough to establish liquid
bridges. The addition and distribution of binder solution can have a major impact
on the particle size distribution and quality of the granules. Poor liquid distribution
produces granules with a wide particle size distribution. These granules could include
large, overwetted particles, as well as small dry particles.
Granule growth is dominated by one of the two mechanisms: coalescence
or layering (14). Coalescence is agglomeration which occurs by collision and consolidation
of deformable nuclei/granules, provided the agglomerates could withstand
the shear forces applied by the impellers. Layering occurs when fine particles stick
to larger particles or granules coated with binder. Free liquid on the surface of
agglomerates, which renders the necessary bonding strength and plasticity to the
agglomerate, is required for coalescence or layering. The free liquid on the surface
of the agglomerates can be from the addition of more binder solution or from
200 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
the expulsion of liquid inside the agglomerates due to the consolidation of the
Granule attrition and breakage reduce the granule size. Granule breakage
is determined by the dynamic granule strength and the shear forces within the
Table 1 List of High-Shear Granulator Suppliers
Supplier Granulator Contact information
DIOSNA Dierks &
Soehne GmbH
Diosna P
270 Route 46 Rockaway, NJ 07866,
Tel: (973) 5863708,
Fax: (973) 5863731,
e-mail: marc_kaufman@servo-lift.com
Glatt Air Techniques VG high-shear
20 Spear Rd,
Ramsey, NJ 07446,
Tel: (201) 8258700,
Fax: (201) 8250389,
e-mail: info@glattair.com
Key International,
480 Route 9,
Englishtown, NJ 07726,
Tel: (732) 5369700,
Fax: (732) 9722630,
e-mail: info@keyinternational.com
L.B. Bohle Inc. GMA, VMA, and
BMG high-shear
L.B. Bohle, LLC,
700 Veterans Circle,
Suite 100,
Warminster, PA 18974,
Tel: (215) 9571240,
Fax: (215) 9571237,
e-mail: info@lbbohle.com
Littleford Lodige KFM and MGT
Lodige Process Technology, Inc.,
203 Forrester Creek Way,
Greenville, SC 29607,
Tel: (864) 2549194,
Niro Inc. Pharma
Systems Division
and PMA highshear
9165 Rumsey Road,
Columbia, MD 21045,
Tel: (410) 9977010,
Fax: (410) 9975021,
e-mail: info@niroinc.com
Vector Corporation GMX
Vector Corporation,
675 44th St.
Marion, IA 523023800,
Tel: (319) 3778263,
Fax: (319) 3775574,
(A division of
Romaco USA,
242 West Parkway,
Pompton Plains, NJ 07444,
Tel: (973) 6160440,
Fax: (973) 6169985,
e-mail: usa@romaco.com
High-Shear Granulation 201
 2005 by Taylor & Francis Group, LLC
granulator. If the impact forces are larger than the granule strength, continuous
breakage and immediate coalescence of the granules takes place (15). However, when
the granule strength exceeds the impact forces, granules will not break. In that case,
granule growth is more static, i.e., the exchange of primary particles between the
Figure 10 Schematic view of a single-pot processor ULTIMAPROTM based on the ULTIMAGRAL
TM with a vacuum drying system. (Courtesy of Niro Pharma Systems.)
Figure 11 Photograph of an UltimaPro 600 single pot processor (granulator is in a swinging
position with an operator). (Courtesy of Niro Pharma Systems.)
202 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
granules is minimal. Thus, there is a balance between agglomerate growth and degradation
of the granules.
The liquid level in the powder bed in a high-shear granulator plays a key role in
the growth of granules. Granule growth and consolidation (densification) occur at
the same time in the high-shear granulation process. Since deposition of free liquid
on the surface of agglomerates is required for coalescence or layering to occur, maintenance
of the free liquid on the surface of the agglomerates is critical for the growth
of granules in a high-shear wet-granulation process (16). The agglomeration of fine
particles is controlled mainly by the liquid saturation of the moist agglomerates,
which is the percentage of intragranular voids filled with the liquid. Increasing the
amount of granulating liquid reduces the yield stress necessary to deform the
agglomerates in collision, thus allowing growth by coalescence with other granules.
This deformation also makes it possible to consolidate the granules, which further
increases the liquid saturation level and squeezes additional fluid to the granule surface
for continued growth.
The moist agglomerates can exist in the following three states, based on the
amount of the liquid they contain: the pendular state, the funicular state, and the
capillary state. The three states can be distinguished by the relative liquid saturation
of the pores S, which is the ratio of pore volume occupied by the liquid to the total
volume of the pores available in the agglomerate.
Agglomerates are in the pendular state when the liquid saturation of the pores
S<25%; in the funicular state when S is between 25% and 80%; and in the capillary
state when S>80%. The liquid saturation S is determined by the amount of binder
solution and the intragranular porosity, which can be calculated based on the
following equation (17):
S ?
H?1  e?
r ?4:1?
where H is the ratio of the mass of liquid binder to the mass of solid particles, e is the
intragranular porosity, and r is the particle density.
Therefore, the formulation, process, and granulator factors that affect the
degree of liquid saturation would play a critical role in defining the robustness of
the granulation process and the properties of the granules obtained. The effects of
the starting material properties, the type of granulator, and its operation on the
granule growth can be evaluated by assessing the changes in the intragranular
porosity, which occur during the process, since intragranular porosity affects the
degree of liquid saturation as indicated by Eq. (4.1). The properties of the
granules, such as particle size distribution, porosity, and content uniformity would
affect the properties and performance of the dosage forms prepared from these
4.1. Liquid Requirement
The optimal amount of liquid binder used for high-shear granulation is within a narrow
range. If the optimal amount of binder solution is not used, there could be
neither growth nor lump formation during the wet-granulation process. In theory,
the amount of liquid used during the wet-granulation process should be equal to
or just exceed the liquid content corresponding to 100% liquid saturation. However,
some grades of lactose and dicalcium phosphate (DCP) could be granulated by coalescence
at liquid saturation far below 100% (18). It would be desirable to predict the
High-Shear Granulation 203
 2005 by Taylor & Francis Group, LLC
required amount of liquid binder based on the formulation composition. However,
this is a challenging task, since the liquid binder required for high-shear granulation
depends on a number of formulation and process factors, such as physicochemical
properties of the starting materials (API and excipients), type of the binder used,
impeller speed, wet-massing time, etc.
Performance and physical properties of the compressed tablets are driven by properties
of the granules. Yajima et al. (19) examined the relationship between particle size
distribution and physical properties of the obtained granules (e.g., angle of repose,
angle of spatula, loose bulk density, tapped bulk density, compressibility) and the
resulting tablets (hardness and weight variation of tablets). The particle size distribution
of the granules used for tablet compression was expressed as a function of median
particle size, and standard deviation with a logarithmic normal distribution.
Physical properties of granules and the compressed tablets were significantly affected
by this factor. Tablet hardness increased as the median particle size decreased, when
the standard deviation was <1.0.
In another example, Wikberg et al. (20) produced 11 granulations of a common
filler (lactose) and three granulations of a high-dose drug (dipentum) by wet
granulation with polyvinylpyrrolidone (PVP) as binder, in a high-shear mixer granulator.
The agglomeration process was varied to produce granules with varying
granule porosity. The granule fragmentation during compression was evaluated by
measurements of air permeability of the tablets. The results showed that the degree
of granule fragmentation during compression was related to the granule porosity
before compaction. A granulation with a higher porosity had a higher fragmentation
propensity during compression. The tablet strength correlated well with the degree of
fragmentation, i.e., a granulation with a higher degree of fragmentation yielded
tablets of a higher hardness. Variations in compactibility can be explained by variations
in granule porosity for the same formulation that was wet granulated under
different process conditions. Therefore, the formulation variables and granulation
process should be controlled to produce granules with the desired properties. Since
granule formation in high-shear wet granulation is dependent on the degree of liquid
saturation, the factors that influence the degree of liquid saturation should be optimized
to obtain an optimal granulation formulation. The factors that affect the
degree of liquid saturation (which depends on the amount of liquid added as well
as on the densification of the granules) are formulation, process, and apparatus variables.
Hence, the granulation process and the properties of the granules are affected
by the aforementioned variables.
5.1. Formulation Variables
Besides the API, excipients such as a filler or fillers, a disintegrant (not applicable to
controlled release), and a binder are also included in the powder mixture of a tablet
formulation. The fillers used in tablet formulations can be classified into two categories,
based on their water solubility: soluble fillers such as lactose, sucrose, mannitol,
etc., and insoluble fillers such as MCC, starch, calcium carbonate, calcium
phosphate, etc. The binders used in the wet granulation process are water-soluble
204 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
polymers such as gelatin, PVP, HPMC and sugars such as glucose, sucrose, and sorbitol.
Some of the commonly used disintegrants are sodium starch glycolate, crosslinked
PVP, and cross-linked sodium carboxymethylcellulose. These excipients are
available in various grades from vendors. Hence, the physical properties such as particle
size distribution, particle shape, surface morphology, surface area, and solubility
of API and excipients in the binder solution could vary considerably. The rate
and the final degree of densification are controlled by the physical properties such
as particle size distribution, particle shape, surface morphology, surface area, solubility
in the binder solution, etc., of the starting materials. Therefore, the physical
characteristics of drug and drug loading, the types and amount of excipients, and
the type and amount of binder used could affect the rate and the final degree of
densification of the resulting granules. Moreover, these aspects of the powder mixture
can also affect the amount of liquid required for granulation and the degree
of liquid saturation in the agglomerates, during the wet-granulation process, which
in turn affect the physical properties of the obtained granules.
5.2. Effects of Starting Materials
The physical properties of the granules are influenced by the physical properties of
the starting materials. For example, the shape, size, and size distribution of the particles
of the starting materials can affect the packing pattern of a powder mixture,
and ultimately the final degree of densification and the strength of the resulting granules.
The extent of granule densification affects the degree of liquid saturation of the
powder mixture, and the amount of liquid binder required for granulation. The total
surface area of the starting materials, related to the particle size and porosity of the
starting materials, will affect the amount of liquid binder required for granulation.
The liquid adsorption of the starting material could also affect the amount of liquid
binder required for wet granulation. Therefore, the amount of binder solution required
for granulation depends on the formulation composition, and the physical
characteristics of the starting materials. In a study by Schaefer et al. (21), when
the mean particle size of lactose was decreased, the amount of liquid required for
granulation was increased to obtain a similar particle size of granules in high-shear
The influence of the primary particle size of DCP on granule growth in a
Lo?dige M5GR high-shear mixer granulator was investigated by Schaefer et al.
(22). Two batches of DCP, one with a narrower particle distribution and smaller
geometric mean diameter, and the other with a wider particle size distribution and
a larger geometric mean diameter were used. The batch with the narrower particle
size distribution and smaller geometric mean diameter was more easily densified than
the one with the wider particle size distribution and a larger geometric mean diameter,
because it was overwetted at lower moisture content. The granules with wider
particle size distribution were produced from the DCP with a wider particle size distribution
and larger geometric mean diameter.
The strength of wet agglomerates can be dominantly affected by the degree of
liquid saturation and the primary particle size of the starting materials. Consequently,
the growth mechanism and drug content uniformity of the resulting granules
can be affected.
The influence of the primary particle size of lactose on the breakage of erythrosine
granules was investigated by van den Dries et al. (4). The results showed that
a decrease in particle size of the starting material, lactose, led to a decrease in
High-Shear Granulation 205
 2005 by Taylor & Francis Group, LLC
breakage of granules. When granule breakage was absent, the granules remained
intact and the preferential layering of the smallest particles yielded inhomogenous
The effect of particle size of the lactose on the homogeneity of low-dose steroid
hormone distribution in granules produced in high-speed mixers was studied by
Vromans et al. (23). When a micronized low-dose steroid hormone was granulated
with unmicronized lactose at 250 rpm, inhomogeneity of active was observed. The
coarse particle size fractions of the granules were found to be superpotent, up to
150% of the mean drug content, whereas the fine size fraction showed a corresponding
subpotency of 50%. However, when the particle size of lactose was reduced,
a better drug distribution was seen. This was attributed to an increase in the tensile
strength of the nuclei due to the smaller particle size of lactose.
The effect of particle size of lactose on the homogeneity of estradiol granules
prepared in a 10 L high-shear mixer was studied by van den Dries et al. (15). Three
different particle sizes of lactose were granulated with 0.1% micronized estradiol
(5 mm) using an aqueous solution of hydroxypropyl cellulose (HPC) as a binder.
Granules prepared with the largest lactose particles (141 mm) yielded a homogenous
granulate. However, after a prolonged processing time, demixing was observed.
Contrary to the largest particles, granulation with the smaller lactose particles (23
and 50 mm) led to demixing in the first minute, although to a lesser extent. It was
concluded that granulation with the largest particles resulted in breakage of the granules,
thereby preventing demixing. However, once the granules were strong enough
(smaller particle size and prolonged process time) to survive the shear forces, demixing
was observed. The extent of demixing depended on the particle size difference.
The size distribution of granules prepared from three different tablet excipients,
lactose, glucose, and mannitol, in a high-shear granulator, were studied by both
sieve analysis and laser light diffractometry (24). Both the shapes and the size distribution
of the granules produced from the three excipients differed.
The crystallinity of the starting materials affected the granulation process and
the properties of the granules. The effect of crystallinity of MCC on wet granulation,
in binary mixtures of MCC and corn starch (CS), was investigated by Suzuki et al.
(25). The crystallinity of MCC was modified, producing amorphous material by
milling in a jet mill (J-MCC) and a vibrational rod mill (R-MCC). Crystallinity of
J-MCC was nearly equal to that of intact MCC (I-MCC), but that of R-MCC
was remarkably low due to the mechano-chemical effect. The growth rate of granules
with R-MCC was greater than that with I-MCC and J-MCC. The formulations containing
I-MCC and J-MCC produced granules with a core (MCC and CS) around
which CS was layered. R-MCC provided granules in which CS and abraded MCC
were homogenously mixed. In addition, R-MCC remarkably decreased the difference
in drug content among the size fractions of granules compared with I-MCC.
The decrease in crystallinity of MCC increased the abrasion of the material by the
impeller. This resulted in a greater growth rate of granulation.
The physical properties of the API also affect the wet granulation process and
the properties of the resulting granules. For example, when the particle shape of the
API was changed from spherical to plate like due to a change in the crystallization
solvent, the compressibility of the resulting granules dramatically decreased (26).
The effect of the particle size of DPC 963 (API) on the characteristics of granules
prepared by high-shear wet granulation was evaluated by Badawy et al. (27).
Particle agglomeration was affected by the particle size of the drug substance.
Granule geometric mean diameter and fraction with particle size greater than
206 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
250 mm was inversely proportional to the particle size of the drug substance. The
granules prepared with the smaller particle size had higher porosity, thus suggesting
lower tendency for granule densification than those manufactured with the larger
particle size. The compressibility of the granules was increased with decreasing particle
size of the drug substance. The effect of particle size on granulation growth is
the result of increased densification propensity from increased drug particle size.
In another study, a high-shear mixer granulator was used for pelletizing a drug.
High-dose pseudoephedrine-HCl pellets were produced with MCC in a high-shear
mixer granulator (28). The drug loading and drug particle size influenced the bulk
and true densities of the resulting pellets.
5.3. Effects of Type and Amount of Binder Solution Used
A binder is normally required for the wet-granulation process. The binder can be
added into the powder mixture as a dry powder followed by the addition of water,
or an appropriate solvent to activate binding. Alternatively, the binder can also be
added into the powder mixture as a binder solution. Because of the higher densification
of the wet agglomerates, the high-shear granulation process normally only needs
two-thirds to three-quarters of the liquid required during the low-shear granulation
process (18). The bonding characteristics of the binders used for wet granulation
could vary due to the differences in their physicochemical properties. Therefore,
the type of binder used for granulation can influence the granulation process, the
amount of binder required for granulation, and the physical properties of the
obtained granules.
For example, the effect of binder solutions on the size of DCP granules was
investigated by Ritala et al. (29). Five different binders, PVP (Kollidone 90), polyvinylpyrrolidone
polyvinylactate copolymer (PVPPVA copolymer), hydrolyzed
gelatin and HPMC (Methocel E5 and Methocel E15) were studied. The granules prepared
with PVP and hydrolyzed gelatin possessed a larger mean particle size and
lower intragranular porosity when the same volume of binder solutions was used.
The other three binder solutions behaved similarly. This could be due to the higher
densification of DCP granule caused by PVP and hydrolyzed gelatin. Thus, the granules
had higher liquid saturation with lower volumes of the binder solutions. It was
apparent that increasing the concentration of the binder solution in the powder
mixture increased the mean granule size.
In another study, Ritala et al. (30) investigated the differences between five
binders during wet granulation using a high-shear granulator. The binders used
for the study were PVP K90, PVP K25, PVPPVA copolymer, Methocel E5, and
Methocel E15. The results demonstrated that the binder solutions with high surface
tension, such as PVP solutions, produced denser granules with larger mean particle
size than those with low surface tension, such as PVPPVA copolymer,Methocel E5,
and Methocel E15. The difference of granule growth by different binders in highshear
wet granulation could be due to the variation of surface tension of the binder
The amount of binder and granulating liquid used during granulation affect
the granulation process and the physical characteristics of the prepared granules.
In general, the granules are finer when the amount of granulating liquid used for
granulation is decreased. For example, the size of the granules prepared from three
tablet excipients, lactose, glucose, and mannitol, increased when the amount of
granulation liquid was increased (24). Similar results were observed for granulating
High-Shear Granulation 207
 2005 by Taylor & Francis Group, LLC
mixtures of mannitol and MCC (31). When the mixtures were granulated in a highshear
mixer using HPC as the binder, the mean particle size of the granules increased
with increasing amounts of water and concentration of the binder.
Generally, an increase in binder concentration could result in the production of
larger granules. However, this is not always the case with starch paste since it could
behave differently from the other binders. For example, a high-shear mixer was utilized
for evaluation of pregelatinized and pregelatinized, cross-linked, waxy corn
starches as binding agents in the wet-granulation process (3). Increasing the concentration
of the binders decreased the average granule size; however, no difference in
friability of the granules was observed. The use of pregelatinized or pregelatinized,
cross-linked starches as a paste provided larger and less friable granules than when
the binder was added in the dry form.
The effect of starch paste concentrations on the particle size distribution of fine
granules, produced by high-speed mixer, was investigated by Makino (32). When a
low concentration (5%) of starch paste was used, the growth of granules occurred
mainly during the massing time, since a relatively large amount of free water was
available from the paste. Only the mean particle size of the granules became larger
without changing the particle size distribution of the granules during granulation.
However, when a high concentration (15%) of starch paste was used, rapid deaggregation
of granules occurred due to the applied shear force, following the formation
of large granules, immediately after the addition of the starch paste. Deaggregation
of these large agglomerates became predominant during the wet-massing stage, and
hence, the size distribution of the granules became narrower. Longer massing time
was also required when a high concentration (15%) of starch paste was used, since
only a small amount of water exuded from the more concentrated starch paste during
the agitation.
Besides the effect on the particle size of the granules, the amount of binder
solution used in the wet-granulation process can also affect the content uniformity
of the granules. The effects of the amount of binder solution on the content uniformity
of the oily drug D-alpha-tocopheryl acetate (VE) in granules obtained by highshear
wet granulation was investigated by Kato et al. (33). When the amount of
binder solution was below the amount of water required to reach the plastic limit
(minimum amount of water required for the powder mixture to form a large agglomerate
when kneaded), the content of VE was <50% in the fractionated fine granules,
but was >200% in the fractionated large granules. Large variations were seen in the
contents of VE even if the granulation time was extended up to 30 min. When the
amount of binder solution was at or above the amount of water required to reach
the plastic limit, less variation was observed in the content of VE throughout the
granules, and the content of VE was fairly uniform.
In another study, the effect of the amount of water as the granulating liquid
and HPC as a binder on the pharmaceutical properties of granules of two model
drugs, ascorbic acid and ethenzamide, was investigated by Miyamoto et al. (34).
The dependence of drug content uniformity of these two model drugs on the granule
size was also investigated in this study. For a water-soluble drug, ascorbic acid, an
appreciable dependence of drug content on granule size was not observed in model
formulations. However, for a poorly water-soluble drug, ethenzamide, more drug
was found in small-size granules (<75 mm). The drug content of ethenzamide in
small-size granules decreased with increasing amounts of HPC and granulation
liquid. These observations suggested that drug content uniformity is influenced
208 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
not only by drug solubility in the binder solution, but also by the amounts of granulation
liquid and the type of the binder used.
Granule friability is also influenced by the binder concentration and the amount
of granulating liquid added to the formulation during granulation. For a new API
granulated with lactose and Avicel PH 101 using low-viscosity HPMC (Pharmcoat
603), increasing the binder concentration and the amount of water tended to produce
higher wet-mass consistencies. Granule friability decreased with an increase in the
binder level. An inverse relationship was observed between granule friability and the
amount of water added to the formulation, especially at lower drug concentrations (35).
The binder can influence not only the physical properties of the granules, but
also the performance of the finished tablets. For example, within a certain range of
binder concentrations and granulating liquid, increasing the amount of binder or
granulating liquid actually decreases the compressibility of the granules (26). In
another study, the effect of different types of binders such as PVP K30, HPMC
(Cellulose HP-M 603), maltodextrin (Lycatab DSH), pregelatinized starch (Lycatab
PGS), and low substituted hydroxypropyl cellulose (L-HPC) (type LH 11) on the
hardness of placebo and paracetamol tablets using high-shear granulation technique
was compared by Becker et al. (36). The placebo granules comprised only lactose and
MCC (Avicel PH 102). The binders were added at 2%, 6%, and 10% in the dry form
and water was used as the granulating liquid. The median particle size increased with
increasing binder concentrations from 2% to 6% except when L-HPC was used as the
binder. The median particle size of the granules decreased with increasing binder
concentrations from 2% to 6% when L-HPC was used as the binder. The granule
strength, and the hardness (crushing strength and friability) of the resulting tablets
increased with increasing binder concentrations. In the preparation of model tablets
containing paracetamol, PVP K30 (6%) and Cellulose HP-M 603 (6%) turned out to
be the binders of choice with respect to crushing strength of the finished tablets.
Lycatab PGS, Lycatab DSH, and L-HPCC-LH 11 could not be used to produce
paracetamol tablets that met the requirements.
The amount of granulating water can also affect the physical properties and
compressibility of MCC, which is one of the most commonly used fillers. The
mechanism of forming hard granules with MCC, using a high-shear mixer granulator,
was investigated by Suzuki et al. (37). The hardness of the MCC granules
increased with granulation time and the amount of water added. The specific surface
area decreased during the granulation process. Crystallite size of cellulose decreased
with granulation time and with increasing amounts of water added. MCC granules
were strengthened with longer granulation time and greater amounts of water, thus
resulting in a more intricate network due to the strong shear force of the impeller.
The effect of granulating water level on the physicalmechanical properties of
MCC and silicified microcrystalline cellulose (SMCC) was investigated by Habib
et al. (38). Granules containing either MCC or SMCC were prepared at different
water levels using a high-shear mixer. The resulting granules were tray dried. The
water level ranged from 0% to 100%. Increasing the water level affected the granule
particle size, increased the granular density and flow properties of the granules, and
decreased the porosity and compactibility. The compactibilities of both materials
were similar and acceptable at each granulating water level up to 40%. However,
both the materials showed poor compactibility at higher water levels. The effect of
amount of water used for granulation did not have a statistically significant difference
on the compressibility of MCC and SMCC. SMCC did not offer practical
advantages over MCC, other than better flow in the powder form, which could be
High-Shear Granulation 209
 2005 by Taylor & Francis Group, LLC
attributed to slightly larger particle size and the presence of silicon dioxide in its
5.4. Effect of Process Variables
Process variables play a critical role in the granulation process since they influence
how the binder liquid is distributed in the powder bed and the degree of densification
for the powder mixture. The degree of powder densification affects the level of liquid
saturation of the moist agglomerates. Therefore, process variables could influence
properties such as particle size distribution and the drug content uniformity of the
obtained granules. Specifically, the process variables affecting the granulation process
and the physical properties of the obtained granules are:
 Load of the granulator bowl
 Impeller speed
 Granulating solution addition method
 Granulating solution addition rate
 Chopper speed
 Wet-massing time.
The following examples illustrate how varying the aforementioned process
variables influences the properties of the granules obtained using a high-shear wetgranulation
Processing variables such as impeller speed, granulating solution addition rate,
total amount of water added in the granulation step, wet-massing time, etc., were
evaluated using a PlackettBurman experimental design in a study by Badawy
et al. (39). The results showed that granule compressibility of the lactose-based formulation
is extremely sensitive to the processing conditions. The granule particle size
was increased by increasing the amount of water added, high impeller speed, and
short wet-massing time. The wet-massing time and impeller speed also had a signifi-
cant impact on granulation compressibility. Increasing the impeller speed and wetmassing
time decreased granule porosity and fragmentation propensity, which led
to the decreased hardness of the finished tablets.
The shear effects of the impeller on the properties of granules prepared from
surface treated sericite were investigated by Oulahna et al. (40). Surface treated
sericite was granulated with an alcoholic solution of polyethylene glycol 20,000 using
a high-shear granulator at three different impeller speeds (100, 500, and 1000 rpm).
The properties of the granules produced under different impeller speeds were examined
in terms of porosity, friability, and binder content. The results indicated that
heterogenous granules (in terms of binder concentration) with finer particles and a
wider particle size distribution were obtained at a low impeller speed (100 rpm).
However, at higher impeller speeds, homogenous granules with less fine particles
and a narrower size distribution were observed. Moreover, the porosity of the granules
decreased with increasing impeller speeds. Therefore, mechanical energy
brought to the powder bed by the impeller is as important as the physicochemical
characteristics of the powderbinder pair in affecting the granule properties.
The effect of granulation processing parameters, fill ratios, impeller speed,
chopper speed, and wet-massing time on granule size distribution of placebo formulations
was studied by Bock et al. (41). High fill ratios in the bowl resulted in an
increased proportion of fines in the obtained granules. Increasing the impeller speed
210 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
and the massing time increased the granule size. The speed of the chopper did not
affect granule size distribution for the formulations tested.
The mode of addition of the liquid binder can affect the characteristics of the
granules (42). When water, used as a binder liquid, was added to the powder mixture
by atomization, granules with a slightly narrower particle size distribution were
The effects of massing time on the properties of the granules of hydrophilic
polymer-based controlled-release formulations were studied by Timmins et al.
(43). The formulations consisted of 30% sodium alginate, 10% HPMC, and 50%
diltiazem hydrochloride or verapamil hydrochloride. The increase in massing time
resulted in an increase in the mean granule size of the formulations. This could be
true for all the matrix controlled release formulations, which contain a high concentration
of the hydrophilic polymers.
The effect of reducing drug loading and mixing time on the content uniformity
of a low-dose drug was investigated by Kornchankul et al. (44). Buspirone hydrochloride
was used as a model drug and was mixed with other ingredients in two different
concentrations (0.5% and 5%, w/w) in a T. K. Fielder high-shear mixer at a
high impeller speed (522 rpm) and a high chopper speed (3600 rpm) for up to
32 min. Samples were withdrawn from nine locations in the mixer at specific time
points using a side-sampling thief probe. The optimum time to mix the 0.5% w/w
formulation was 8 min, while it was only 1 min for the 5% w/w formulation.
MCC was granulated with water in a high-shear mixer (37). The hardness of
the MCC granules increased with granulation time and the amount of water added.
The specific surface area was reduced during the process. These findings suggested
that the long-chain structures in MCC were disrupted, resulting in smaller units with
shorter chain lengths due to the strong shear force of the impeller. These smaller
units then formed a network within the granules. Thus, MCC granules were strengthened
with longer granulation time and greater amounts of water, resulting in a
more intricate network. The change in MCC chain length and physical structure
can be experimentally detected using the small-angle x-ray scattering and wide-angle
powder x-ray diffraction methods.
MCC granules were prepared by wet granulation with water in a high-shear
mixer granulator (45). Samples of wet granules were taken at different time points
after addition of water to examine the physical characteristics of the granules using
near-IR spectrometry, thermogravimetry, and isothermal water vapor adsorption.
The results indicated that the degree of interaction between MCC and water increased
with granulation time due to the change of physical structure of MCC during
the granulation process.
5.5. Effect of Granulator Design
A variety of high-shear granulators with varying processing designs are commercially
available. The orientation of the mixing chamber in the granulators is either
vertical or horizontal. The vertical granulators could be bottom or top driven.
Besides the configuration difference, a significant difference exists among the mixers
in terms of the shapes of the mixing bowl and impeller blades. For example, the vertical
top-driven mixer granulators utilize a curved blade design, thus allowing the
tips of the mixing blade to reach the side wall of the mixing bowl, which is also
curved in a shape similar to the blade. On the other hand, the vertical bottom-driven
high-shear granulators (such as Diosna) have a conical mixing bowl with a flat blade
High-Shear Granulation 211
 2005 by Taylor & Francis Group, LLC
design. The blade is parallel to the lid and the bottom of the bowl, and only a small
portion of the impeller blade directly sweeps off the side wall of the conical mixing
bowl. During mixing and granulation, the conical mixing bowl pushes the powder
toward the center of the bowl for efficient mixing. The differences in impeller blades
and mixing bowl designs of the different commercially available high-shear granulators
are expected to result in different flow patterns and powder flow dynamics in the
During the wet-granulation process, the impellers impact the powder mixture
and keep the mixture moving. During this process, binder solution is distributed
and the powder mixture is compacted and densified. Relative swept volume, the
volume swept out per second by impellor and chopper divided by the volume of
the mixer, is the indicator of the work input on the powder material inside the bowl
of the granulator. Changes in the granulator design could affect the relative swept
volume of the granulation process, which in turn could change the degree of densi-
fication of the wet agglomerates. As a result, the optimal amount of granulating
liquid and the physical properties of the resulting granules may be altered. Therefore,
the design of high-shear granulators must be taken into consideration during the wet
granulation process.
Schaefer et al. found that there was a difference between the horizontal (Lo?dige
M5GR) and vertical (Fielder PMAT 25 VG) granulators in terms of granule growth
for DCP (22). The degree of liquid saturation required was higher for the vertical
(Fielder PMAT 25 VG) granulator in order to obtain the same granule size. Agglomeration
of DCP began at lower degrees of liquid saturation in the horizontal (Lo?dige
M5GR) granulator. This could be because the rolling motion of granules in the horizontal
(Lo?dige M5GR) granulator facilitated agglomeration, whereas stronger
mechanical forces in the vertical (Fielder PMAT 25 VG) granulator promoted the
breakage of granules. A similar trend was observed for lactose granulation (46) as
The effects of impeller and chopper design upon granule growth of DCP in
a laboratory high-speed mixer (Fielder PMAT 25) were investigated by Holm (47).
Three different changeable impeller blades with the same surface area were constructed.
The angles of inclination were kept at 30, 40, and 50. The granules with
low porosity were obtained at high impeller speed for the inclination angles at 40
and 50. The effects of the impeller design with respect to the blade inclination
and impeller rotation speed can be explained in terms of the volume of powder mixture
swept out by the impeller. The relative swept volumes for the impellers at three
different inclination angles were 2.2, 2.82, and 3.36 at 400 rpm. A high-swept volume
causes high densification of the agglomerates and narrow granule size distributions.
Chopper size and rotation speed had no effect on the granule size distribution
because the primary function of the chopper is to disturb the uniform flow pattern
of the mass.
The granule properties for a given formulation are a function of the process parameters
such as the impeller speed, amount of granulating fluid, and granulation time
(wet-massing time). Therefore, the time to end the granulation process during a highshear
granulation process becomes critical. The properties of the granules produced
determine, in part, the ultimate quality and performance of the finished dosage
212 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
forms. Thus, the determination of the granulation end point becomes necessary. The
ultimate goal of end-point determination in a granulation process is to obtain an
indication of formation of granules with the desired physical properties, such as
acceptable mean particle size range and porosity. The measurements of changes during
the granulation process are employed for end-point determination. The advantages
of using an appropriate method to determine the granulation end point are
listed below:
 Process optimization
 Evaluate raw material
 Determine optimal end point
 Batch reproducibility
 Use end point to achieve batch to batch consistency
 Document adherence to batch protocol
 Process trouble shooting
 Detect mechanical problems
 Identify mixing irregularities
Several different approaches have been explored for granulation end-point
determination. These approaches can be classified into two major categories: indirect
measurements and direct measurements. In the indirect measurements, the electrical
and mechanical parameters of the granulator motor are monitored since the changes
of these parameters are related to the changes in the consistency of the powder mixture
in the wet granulation process. Faures study (48) has confirmed that there is a
close relationship between the wet-mass consistency/viscosity of samples prepared in
a mixer granulator and physical properties of the dry granules produced from the
wet mass. The physical properties of the dry granules include granule size distribution,
bulk density, friability, and flow characteristics. Variations in the formulation
and process affect the relationships between the wet-mass consistency and drygranule
properties, and the net power consumption of the mixer granulator.
In the direct measurements, the physicochemical properties of the powder mixture
are monitored during the wet-granulation process. These properties could be
mass conductivity and granule size.
6.1. Indirect Measurements
In the indirect measurement, both electrical and mechanical characteristics of the
motor are monitored to control the granulation end point. The electrical characteristics
of the motor are motor current and power consumption. The mechanical characteristics
of the motor are torque and tachometry.
The power consumption of a granulator motor is related to the resistance of
the mass mixture to the granulator blades, which varies with the consistency of
the powder mixture. Power consumption used as granulation end point control
has been related to the level of liquid saturation of the moist agglomerates (49), densification
of wet mass (30,50,51), and granule growth (52). Leuenberger proposed
that the liquid amount required for granulation corresponds to the plateau in the
power consumption record profile (53).
High-Shear Granulation 213
 2005 by Taylor & Francis Group, LLC
High-shear granulators capable of monitoring the granulation end point using
power consumption are commercially available. Figure 12 shows a typical power
consumption profile obtained from a commercially available high-shear granulator
with an end-point determination (53). It is evident that the profile consists of five
phases during the wet granulation process. In Phase I, the powder mixture is moistened
by adding the granulation liquid. Since the formation of liquid bridges is not
observed between primary particles, power consumption does not increase during
this granulation phase. In Phase II, the liquid bridges begin to form among the primary
particles, thus resulting in granule formation. The power consumption increases
dramatically in this phase. In Phase III, as more granulation liquid is added, the
interparticular voids are filled by the granulation liquid and coarser granules are
formed. The power consumption remains relatively constant in this phase. In Phase
IV, liquid saturation of the powder bed is reached, and more coarse granules are
formed. The power consumption increases in this phase. In Phase V, as more granulation
liquid is added, a suspension is formed. The power consumption starts
declining in this phase. It has been noticed that only the granules formed during
Phase III are usable for further dosage form development. Thus, the power consumption
profile can be used to monitor the granulation process and to determine
the granulation end point.
The drawback of monitoring power consumption is that the signal is affected
by a number of factors, such as product (formulation), equipment, or process variables.
For example, the densification properties of the starting materials, type and
amount of binder used, addition rate of binder solution, impeller speed, etc., influence
the power consumption profile. Wear and tear of the granulator could also
affect the power consumption signal. Thus, the power consumption profile for a
granulation process is formulation and process specific.
The end-point control of granulation by power consumption measurement in a
25 L high-shear mixer was investigated by Holm et al. (42). The effect of impeller
design, impeller speed, liquid addition rate, type of binder, and mixing ratios between
lactose and starch on the correlation between power consumption and granule
Figure 12 A typical power consumption profile obtained from a commercially available high
shear granulator. (From Ref. 53.)
214 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
growth was investigated in the liquid addition phase of the process using a fractional
factorial experimental design. A linear correlation between power consumption and
mean granule size was observed. The correlation was dependent on the impeller
design, the impeller speed, and the type of binder. However, a granulation end-point
control, based on power consumption was not found to be sensitive to variations in
the lactose:starch ratio. An end-point control based on the peak detection method
was not generally applicable, because a peak in the differentiated power consumption
signal could not be identified in all the experiments.
Monitoring of power consumption was also explored for automating the wetgranulation
process (54). A vertical high-shear mixer (300 L Collette Gral) was
instrumented with a control unit, which controls the continuous addition of granulating
liquid in relation to the consistency of the powder mass by power consumption
measurement. The instrumentation comprised an important approach for automated
process control, and on-line documentation of critical process parameters with
respect to validation. Similarly, a signal analysis system with memory-programmable
automatic process control for determining power consumption in a high-shear mixer
was introduced by Laicher et al. (55). The signal analysis system made it possible to
perform gliding mean value calculations of the measuring signal and to evaluate the
signal fluctuations. The granulation end point was determined by curve analysis.
Using this system, granules with a reproducible particle size distribution were obtained
from active ingredient paracetamol or ibuprofen/excipient mixtures, which
either showed a tendency to form lumps, in the case of overwetting (paracetamol),
or required a short granulation time due to low melting point (ibuprofen).
Since power consumption is related to both current and voltage of the motor, it
would be logical to use motor current for granulation end-point control. However,
the motors used in high-shear granulators are induction motors, which use alternating
current that lags behind the voltage. Therefore, it is inappropriate to determine
granulation end point by monitoring motor current or the simple combined product
of current and voltage (51). The power consumption of induction motors can be
expressed by the following equation:
Power ? ffiffiffi 3 p IV cos y ?6:1?
where I is the current, V is the voltage, and y is the phase difference between current
and voltage.
An alternative to power consumption is torque measurement. A high-shear
mixer was instrumented with a new capacitive sensor, a watt meter, and a straingauge
torque sensor (50). Placebo formulations containing DCP, MCC 101, MCC
102, and lactose were granulated in the high-shear granulator. A similar map (profile
of torque measurement or power consumption) of the granulation process was
obtained for power consumption and torque measurement.
In another study (51), a high-shear vertical mixer/granulator was instrumented
to monitor power consumption, direct torque (signal from impeller shaft), and reactive
torque (signal from the motor). Of the three methods used, direct torque generated
the most descriptive profile for the granulation process. Granules produced for
the test formulations, based on the end points determined using direct torque measurement,
resulted in tablets with acceptable hardness.Measurement of direct torque
could also detail the granulation process of an overwetted formulation.
Torque profiles of granulation of MCC or lactose in a high-shear granulator
with various operating conditions were studied by Kornchankul et al. (56). The
High-Shear Granulation 215
 2005 by Taylor & Francis Group, LLC
torque measurement for each batch of granules obtained was correlated with the corresponding
tablet characteristics. The scatter plots of tablet hardness indicated that
the hardness of tablets prepared with MCC granules was >7 Kp, as long as the end
torque was <100 lb in., granulation rate was <0.4 lb in./sec, and granulation extent
was <13,000 lb in. sec.
The relevance of end-point determination based on torque measurement in
controlling performance characteristics of compressed tablets such as hardness, friability,
disintegration time, and dissolution time was determined by Achanta et al.
(57). The measurement of torque was used to determine the end point of the granulation
process in a high-shear granulator. With increasing end-point levels, tablets
compressed with the corresponding granules were found to be harder.
6.2. Direct Measurements
During the high-shear wet-granulation process, the binder liquid is dispersed in the
powder mixture. At this point, the wetted powder becomes agglomerated and densi-
fied. Thus, the physical properties (particle size, density, consistency) of the powder
mixture change with time during the granulation process. This change in the properties
of the powder mixture can be used as an indication of the end point of the granulation
6.3. Consistency Measurement Using a Probe
One approach to determine the granulation end point is based on monitoring the
change of the consistency or strength of the wet mass during the granulation process.
This approach was first developed by the Boots Company for Diosna high-shear
mixers (18). The Disona-Boots probe is designed to detect changes in the momentum
of granules, in a constant velocity region of the mass, with the use of strain gauges
mounted on the probe. To avoid bias due to random events (e.g., production of large
lumps due to inhomogenous solution distribution), signal pulse heights, sampling
times, and pulse density are all considered to attain the final signal. The signal is normalized
over a calibrated range of forces from 0% to 100%. The criterion for endpoint
indication is that the signal obtained during the granulation process must be
greater than the preset value. When properly developed and calibrated for a given
high-shear granulator, the probe can be successfully used to determine a repeatable
granulation end point, i.e., granules with a similar particle size and density are
obtained. However, the signal obtained from the probe is granulator and formulation
6.4. Acoustic Emission
Acoustic emission monitoring detects and analyzes the sound produced by a process
or system. During the wet-granulation process, particle size, flow properties of the
powder mixture, and degree of powder densification change with time because of
the addition of binder and formation of agglomerates. The change of physical properties
of the powder mixture could affect the acoustic emission signal. Thus, this
technique can be used to map the granulation process and determine its end point.
The technique is noninvasive, sensitive, and relatively inexpensive.
The acoustic emissions during the wet granulation of a model formulation containing
lactose and MCC were monitored using acoustic emission sensors (58). The
216 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
sensors were attached to the outside of the high-shear mixer bowl. Undergranulated
to overgranulated granules were prepared by varying the quantity of the binder solution.
Average signal level increased with increasing amount of the added binder solution
and wet-massing time. A strong correlation between the acoustic emissions and
the physical properties of the granules at the end of the granulation process was
demonstrated. This technique was capable of monitoring changes in physical properties
such as particle size, flow properties, and compression properties of powder
material during granulation.
6.5. Image Analysis
In the direct measurement techniques, one of the successful methods is monitoring
the granulation processing based on the particle size of the granules.
A novel system to continuously monitor granule growth in a high-shear
granulation has been developed by Watano et al. (59). The system consists of an
image processing system and a particle image probe comprising a CCD camera,
lighting unit, and air purge system. High-shear granulation was conducted using
pharmaceutical powders, and granule size and products yield of various size ranges
were continuously measured by the developed system. Sieve analysis of the granules
sampled out during the granulation was simultaneously conducted, and the data
were compared with that of the on-line image processing system. A close relationship
was found between both the data, thus proving that the system could monitor the
granule growth accurately and continuously throughout.
Such an image processing system was further developed to automatically control
the wet-granulation process by Watano et al. (60). Besides the image probe, a
fuzzy control system using a linguistic algorithm employing if-then rules with a
process lag element was developed to control granule growth accurately. The system
consisted of an image processing and a fuzzy control system. An image probe continuously
monitored granule images through the sidewall of the vessel. The images
were digitized by the image processing system, and granule median diameter and
shape factor were calculated automatically. The difference between the desired and
measured granule diameters was used as an input for fuzzy reasoning. The result of
fuzzy reasoning was used to control the output power of the liquid feed pump. The
system could control granule growth with high accuracy, regardless of changes in the
operating conditions.
The pharmaceutical properties of the granules, which can significantly affect the performance
and properties of the finished dosage forms are controlled by multiple
interrelated variables such as formulation, process, and granulator parameters. In
most cases, the nature and characteristics of the excipients and API used in the formulation
are fixed. Thus, only the amount of granulating liquid and granulation process
variables need to be optimized. Design of experiment (DOE) is often used to
optimize the granulation process. In experimental designs, a number of cause factors
which comprise volume of binder solution, impeller speed, massing time, etc., and
response variables (dependent variables), which comprise mean particle size, tablet
hardness, time for 50% of drug released, etc., are selected. The relationship between
High-Shear Granulation 217
 2005 by Taylor & Francis Group, LLC
cause variables and response variables is established mathematically. The optimal
response variables can then be obtained by selecting the optimal cause variables.
A Box-Behnken design was used to optimize the composition and process factors
for formulations containing mannitol and MCC, which were granulated in a
high-shear mixer and compressed into tablets (31). Simultaneous optimization of
crushing strength, disintegration time, and ejection force of the tablets was carried
out to find optimal regions in the design space for these tablet properties. The composition
of the tablet mixture and the process of tablet manufacturing were optimized
using such statistical techniques.
A central composite experimental design was used to optimize a granulation
process in a high-shear mixer for a hydrophilic matrix tablet formulation (61).
The parameters tested were the amount of water in the hydroalcoholic granulation
liquid, the amount of granulation liquid, and the massing time. The results indicated
that the amount of granulation liquid was the most important parameter, followed
by the amount of water in the granulation liquid. The influence of the massing time
was negligible. A formulation with granule friability below 20% was obtained under
optimal processing conditions.
The wet-granulation process by a high-speed mixer granulator for two model
drugs, ascorbic acid and ethenzamide, was optimized through experimental design
(spherical central composite experimental design) and simultaneous optimization
methodology (34). The amount of water (granulation liquid) and HPC (binder) were
simultaneously optimized with regard to the four pharmaceutical properties, including
yield, drug content uniformity, geometrical mean diameter of granules, and
uniformity of granule size. A simultaneous optimal point incorporating four pharmaceutical
properties was obtained using the generalized distance function. The
experimental values of the four response variables obtained in newly prepared granules
were found to correspond well with the predicted values of granules containing
both ascorbic acid and ethenzamide.
Various experimental designs have been explored by Vojnovic et al. to optimize
wet-granulation process variables. Some of these examples are described below.
Wet granulation of lactose and corn starch in a 10 L high-shear mixer was optimized
using a response surface design (62). The effect of the process factors, impeller
speed, amount of water added, and granulation time on the properties of the granules
were investigated. Moisture level had a major effect on geometric mean diameter
and flow rate of the granules. The impeller speed markedly influenced the geometric
mean diameter, geometric standard deviation, compactibility index, and percentage
of granules smaller than 1250 mm. The granulation time affected the compactibility
index of the granules. Theoretical optimum conditions were obtained for the following
five response variables: geometric mean diameter, geometric standard deviation,
flow rate of the granules, percentage of granules smaller than 1250 mm, and compactibility
index. These results were comparable to the experimental results.
A central composite response surface design was employed to optimize a wetgranulation
process for lactose and corn starch in a 10 L high-shear mixer (63). The
effect of the optimal combination of three independent variables (moisture level,
impeller speed, and granulation time) on four properties (geometric mean diameter,
percentage of particles smaller than 1250 mm, geometric standard deviation, and flow
rate) of the granules was investigated by the simultaneous optimization method. The
optimum zone determined in 10 L high-shear mixer was analyzed for scale-up. The
same product was manufactured at a 50 L scale and optimum theoretical results were
218 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
obtained for the four response variables. These results were in agreement with the
experimental data.
Optimization of wet granulation of a placebo formulation in a 10 L high-shear
mixer was carried out using an experimental design, which combined the Scheffe
simplex-centroid design and Doehlert experimental matrix (64). Lactose, corn starch,
and MCC were used as excipients. Mixing ratios of these excipients were selected as
formulation factors. The impeller speed and granulation time were chosen as independent
variables. A two-phase experimental strategy was employed. The optimal
composition of the three excipients was determined in the first phase and the optimal
process conditions to obtain a geometric mean diameter >150 mm was determined in
the second phase.
The optimization of wet granulation in a 10 L high-shear mixer was investigated
using a nonclassical mixture experimental design, which combines mixture
experimental design with the ponderation function (65). HPMC, lactose, corn starch,
and MCC were used as excipients, while PVP was used as a binder. The introduction
of such a modification in the experimental design allowed selection of an experimental
matrix, which could provide enough information with a relatively low number of
experiments and at a minimal cost.
The formulation and process for a new API are normally optimized using small-scale
equipment, owing to a limited amount of API. The granulators used during the
development stage are laboratory scale models. The formulation is generally fixed
during the clinical trial studies. However, the production batches are manufactured
with large-scale equipment. For a solid dosage form prepared using the wet-granulation
process in a high-shear granulator, product scale-up could be challenging. This
is because of significant differences between laboratory and production models of
high-shear granulators in terms of the design, shape, size, and geometry of the impellers.
Even though the composition of the formulation is similar, the physical properties
of the obtained granules could be dramatically different because of the
differences between laboratory scale and production scale granulators. These differences
in the physical properties of the granules could significantly affect the properties
and performance of the final tablets.
The ultimate goal of product scale-up is to maintain similar processing conditions
for the production batches as those used at the laboratory scale so that the
properties of the granules produced after scale-up would not change significantly.
Therefore, it is imperative to identify and monitor key process parameters during
the product development stage, so that the tablets produced from the large-scale
granulation batch perform similarly to those produced from the small-scale batch.
The liquid saturation level of the agglomerates during a high-shear wetgranulation
process plays an essential role in terms of granule growth, as discussed
previously. It is desirable to maintain a constant level of liquid saturation during the
scale-up. From Eq. 4.1, liquid saturation level is controlled by both moisture
content and intragranular porosity. The moisture content is determined by the
amount of applied granulation liquid. The intragranular porosity is affected by
the impaction caused by the impeller and wet-massing time. Due to the different geometric
shapes and sizes of the granulator and impeller, and fill ratio in the mixing
bowl, maintaining constant liquid saturation level could be problematic.
High-Shear Granulation 219
 2005 by Taylor & Francis Group, LLC
8.1. Linear Scale-up
As described earlier, the process and formulation factors that affect the granulation
process and physical properties of the obtained granules are: rate of addition of the
binder solution, impeller speed, wet-massing time, and the amount of binder or granulating
liquid used. Ideally, the amount of binder and granulation liquid will be
scaled up by linearly increasing the amount of water with the batch size. If the time
of binder solution addition is kept constant by increasing the rate of binder addition
during scale-up, then the variables that need to be controlled narrow down to impeller
speed and wet-massing time. Different approaches have been proposed to control
impeller speed. These approaches include maintaining a constant impeller tip speed,
maintaining a constant relative swept volume, or maintaining a constant Froude
number (dimensionless number) during the scale-up process. The wet-massing time
can be determined by the granulation end-point determination such as power consumption.
However, due to the difference in the size and shape of the mixing bowl
and impeller between laboratory and production scale granulators, linear scale-up
may not work.
A constant relative swept volume and a constant impeller tip speed during the
granulation of lactose in three Gral high-shear mixers (Gral 10, 75, and 300 L), did
not result in a comparable process for different sizes of granulators (66). Thus, even
within the same type of high-shear granulators, compensatory changes in the volume
of binder fluid, impeller speed, and wet-massing time to produce equivalent granules
during scale-up may be required.
The ease of scale up could be formulation dependent, and some formulations
are not affected by the batch size. If the formulation is easy to densify, then the speed
of the impeller may not affect the properties of the obtained granules. In one study
(67), alpha-lactose monohydrate was granulated using a PVP K 30 (2.5% on dry
material) binder solution. The wet granulation process was optimized in a Gral
10 L granulator using a two-factor (water content and impeller tip speed), three-level
experimental design. The granulation process was then scaled down from a Collette
Gral 10 L (8 L batch size) to a serial minigranulator (Pro-C-epT Mi-Pro) with different
bowl volumes of 5 L, and 1900, 900, and 250 mL. In all mixer volumes, the impeller
tip speed range used did not influence the granule or tablet properties. In all bowl
volumes, the influence of water concentration on actual yield, particle size distribution,
and granule friability was similar.
In another study, the formulation containing 5% API, 47.35 % of MCC, and
47.35% of pregel starch was granulated using different sizes of high-shear granulator
ranging from 2 to 25 L (68). A 23 factorial design was used to investigate the effects
of scale-up variables including amount of binder solution, wet-massing time, and
impeller tip speed on the physical properties of the obtained granules. It was shown
that the characteristics of the granulations made under different conditions were
highly reproducible. The excipient system of MCC and pregel starch produced a
robust formulation that was resistant to changes during the scaling-up process in
the high-shear mixers.
Unfortunately, most formulations are not scalable. For example, a less homogenous
liquid distribution, a wider granule size distribution, and a higher intragranular
porosity of the obtained granules were found for the scale-up batches when
DCP was granulated using different sizes of high-shear granulator. This could be
due to the decreased swept volume in the larger granulator during scale-up (18).
220 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
In another study, a model formulation was developed in a small-scale highshear
granulator (Diosna P 1-6) (41). It was easy to scale up to the P 10 granulator
without changing the formulation or the granulation conditions. However, further
scale-up to Diosna P 25 and P 100 granulators, which are large-scale machines,
resulted in granules which were smaller than those prepared in the laboratory scale
To achieve linear scale-up for formulations that are sensitive to the changes
during the scale up, the parameters used during the laboratory scale granulation
process have to be similar to those used during the production scale process. Two
processes may be considered similar if there is a geometrical, kinematic, and dynamic
similarity (69). Two systems are considered to be geometrically similar if they have
the same ratio of linear dimensions. Two geometrically similar systems are kinematically
similar if they have the same ratio of velocities between corresponding points.
Two kinematically similar systems are dynamically similar when they have the same
ratio of forces between corresponding points. For any two dynamically similar systems,
all the dimensionless numbers necessary to describe the process are the same
(70). Dimensionless number of the process is scale independent, and thus can be
employed for scale-up of granulation.
Dimensional analytical procedure was first introduced by Lord Rayleigh in
1915 (71). It is a process whereby the variables pertinent to a physical problem are
systematically organized into dimensionless groups/numbers. Assuming the absence
of chemical reaction and heat transfer, the following influencing and independent
variables are pertinent to the wet-granulation process (72):
DP: net impeller power consumption (motor power consumption minus the
dry blending baseline level)
R: impeller radius
N: impeller speed
h: height of granule bed in the bowl
r: granule bulk or specific density
Z: granule dynamic viscosity
The only three fundamental quantities are mass (M), length (L), and time (T).
Buckingham theorem (p-theorem) (73) states that when there are i physical
variables and dimensional constants and j fundamental qualities (such as mass,
length, and time), the number of dimensionless groups/numbers in a complete set
is equal i  j.
Therefore, according to the Buckingham theorem, there are four dimensionless
numbers for a high-shear granulator.
Newton (power) number (Np) relates the drag force acting on a unit area of the
impeller to the inertial stress:
Np ?
rN3R5 ?8:1?
Reynolds number (Re) relates the inertial force to the viscous force:
Re ?
Z ?8:2?
Froude number (Fr) is the ratio of the centrifugal acceleration to the gravitation
constant (g). It can be used as a criterion for dynamic similarity of granulators:
High-Shear Granulation 221
 2005 by Taylor & Francis Group, LLC
Fr ?
g ?8:3?
Geometric number (ratio of characteristic lengths) h/R is related to fill level in
the bowl.
An ideal scenario is that the granulation processes in laboratory scale and production
scale granulators are geometrically, kinematically, and dynamically similar.
However, the shape and size of granulators currently available vary from manufacturer
to manufacturer. Therefore, it is difficult for the granulation processes to be
geometrically, kinematically, and dynamically similar. For example, Collette Gral
10, 75, and 300 are not geometrically similar (66).
One practical approach during scale-up is to keep the processes dynamically
similar, which is defined by the same dimensionless numbers for the two different
sizes of granulators. Therefore, dimensionless numbers such as Froude number
can be used for scale-up.
Froude number was suggested as a criterion for dynamic similarity and a
scale-up parameter during wet granulation (66). Lactose granules prepared in Gral
high-shear mixers of three different capacities (Gral 10, 75, and 300 L) were
compared. The granulation process regarding temperature increase and particle size
distribution could be scaled up by keeping the Froude number constant. It therefore
seems appropriate to characterize and compare different mixers by the range of the
Froude numbers. A matching range of the Froude numbers would indicate the
possibility of scaling up the batch even for the mixers such as Collette Gral 10,
75, and 300 that are not geometrically similar. Hence, systematic consideration is
needed to select laboratory and high-production scale granulators to develop a solid
dosage form, which requires high-shear granulation in the manufacturing process.
The Froude numbers of each granulator can be calculated based on Eq. 8.3. If
the Froude numbers for the laboratory scale granulator cannot be matched, linear
scale-up is unlikely.
Dimensionless power relationship between power number and the product of
Froude number, Reynolds number, and fill ratio could be another approach for
scale-up (74). A model formulation granulated in different conditions (batch size,
blade speed) with different sizes of high-shear granulators, but presenting similar
wet-mass characteristics (bulk density and consistency) led to dry granules of similar
properties such as granule size distribution, density, friability, and flow. The wet
masses were characterized by the bulk density and consistency, as measured by mixer
torque rheometry. The power consumption of the granulator was monitored with a
power meter. The power number, Reynolds number, Froude number, and bowl fill
ratio were calculated for each running condition. The relationship between the
power number and the product of Reynolds number, Froude number, and bowl fill
ratio can be established by plotting the decimal logarithms of power number vs. the
decimal logarithms of the product of Reynolds number, Froude number, and bowl
fill ratio. The results showed that under certain conditions, a common scale-up master
curve could be drawn from the data gathered for each bowl, and such a curve
could be used for the determination of power consumption of a mixer granulator
at a defined granulation end point for scale-up.
Dimensionless numbers such as Froude number can also be used to design and
select high-shear granulators of different scales. It is therefore necessary to characterize
and compare high-shear granulators of different types and sizes by the range of
222 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
the Froude numbers. A matching range of the Froude numbers would indicate the
possibility of a scale-up, even for the mixers that are not geometrically similar
(66). Froude number of a granulator is determined by both impeller speed and the
diameter of the impeller as shown in Eq. 8.3. When the diameter of the impeller is
fixed, the Froude number of a granulator can still be varied by adjusting the impeller
speed. Therefore, it would be desirable for the production scale granulator to be
equipped with variable speeds. Hence, the range of Froude numbers for each size
of granulator needs to be considered in the design of the high-shear granulators.
A systematic approach should be followed to select laboratory scale and production
scale granulators.
The matching Froude numbers for different size granulators provide the possibility
of linear scale-up. However, some precautions are still needed for the geometric
dissimilarity of mixing bowl vessels, and the shape of the impellers in the
granulators of different types and scales.
During scale-up, after selecting the granulator with matching Froude number
and similarly shaped impeller blades, the granulation end point can be determined by
any one of the techniques described earlier. A new approach for scale-up based on
the use of power consumption, as the granulation end-point determination, is shown
Integration of the impeller power (watts) vs. time (seconds) profile for highshear
granulator was also explored for scale-up of high-shear granulation (75).
The energy parameter impeller work during the wet-granulation process of a model
MCC formulation in 5 or 10 L was recorded and normalized, based on the weight of
the dry powder mixture. The relationship between work of granulation and cohesion
indexes (slopes of the tablet breaking strength vs. compression force profiles), or
granulation size distributions was evaluated. Independent of granulator make or
model, the impeller work measured during granulation correlated quantitatively with
changes in the granulation bulk and tapped densities, average particle size of the finished
powders, and cohesion index. Granulation end points were accurately predicted
for 25, 65, and 150 L high-shear granulators on the basis of development
work on 5.0 and 10 L equipment using impeller work values normalized for the
weight of dry powders in the granulator (W sec/g).
8.2. Nonlinear Scale-up
If linear scale-up cannot be achieved, the amount of binder solution, the impeller
speed, and the wet-massing time should be optimized. For example, Kristensen
attempted to compensate the less efficient mixing from large-scale granulation by
increasing the relative amount (ratio between granulation liquid and starting materials)
of granulation liquid used (76). Rekhi et al. recommended maintaining a constant
tip speed, scaling up the granulation liquid linearly, and adjusting the
granulation time based on the ratio of impeller speeds between the granulators used
in both the scales (77).
8.3. Utilization of Optimization Techniques
Since scale-up of high-shear granulation is still a trial and error process due to the
complexity of the granulation process and lacks standards for granulators of different
types and sizes, the DOE approach can be useful during production. The process
variables for the large-scale production can be optimized by using the DOE
High-Shear Granulation 223
 2005 by Taylor & Francis Group, LLC
technique with relatively less number of trials. Alternatively, the granulation process
can be optimized in such a way that the process is scale independent.
A spherical central composite experimental design was used to design experiments
(78). Granules were prepared in the two batches (2 and 5 kg batches) using a
high-speed mixer granulator. Except for the rotation speed of the impeller, the
operation conditions such as chopper speed, mixing time, and granulation time were
kept constant during the granulation processes. The concentration of ethanol in the
aqueous binder solution and the total volume of binder solution used were selected
as independent variables. The yield of granules, geometric mean particle size, and
geometric standard deviation were employed as dependent variables. A computerized
optimizing technique based on a response surface methodology was developed.
A universal optimal formulation unaffected by manufacturing scale could be
obtained by minimizing the integrated optimization function. The optimized characteristics
measured at the production scale coincided well with those obtained at
laboratory scale, suggesting that this approach could be useful in minimizing the
scale-up problems.
High-shear granulation is one of the most important unit operations in the production
of tablets. Various innovative approaches have been explored to simplify and
control the granulation process, and improve the quality of the produced granules.
Some of these novel approaches include using a one-pot granulation system, moisture-
activated dry granulation and hot-melt granulation techniques, and different
granulation end-point techniques. However, both the process and formulation parameters
for each formulation still need to be individually optimized due to the complexity
of the number of variables involved during the granulation process and the
uniqueness of the formulation. Moreover, the relationship between the properties
of the finished tablets and unit operations for tableting, such as wet granulation, drying,
milling, and tablet compression, is inconclusive. Therefore, a systematic evaluation
of all formulation and process variables involved in the manufacturing of tablets
is required for the optimization of production of tablets. Future advancement in the
equipment and granulation techniques could further improve the granulation process,
thus resulting in better quality of the granules.
1. Nellore RV, Rekhi GS, Hussain AS, Tillman LG, Augsburger LL. Development of
metoprolol tartrate extended-release matrix tablet formulations for regulatory policy
consideration. J Control Release 1998; 50:24756.
2. Qiu Y, Hui HW, Cheskin H. Formulation development of sustained-release hydrophilic
matrix tablets of zileuton. Pharm Dev Technol 1997; 2:197204.
3. Visavarungroj N, Remon JP. Crosslinked starch as a binding agent. II. Granulation in
a high shear mixer. Int J Pharm 1990; 65:4348.
4. van den Dries K, de Vegt OM, Girard V, Vromans H. Granule breakage phenomena in
a high shear mixer; influence of process and formulation variables and consequences on
granule homogeneity. Powder Technol 2003; 133:228236.
224 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
5. Kiekens F, Cordoba-Diaz M, Remon JP. Influence of chopper and mixer speeds and
microwave power level during the high-shear granulation process on the final granule
characteristics. Drug Dev Ind Pharm 1999; 25:12891293.
6. Christensen LH, Johansen HE, Schaefer T. Moisture-activated dry granulation in a high
shear mixer. Drug Dev Ind Pharm 1994; 20:21952213.
7. Schaefer T, Holm P, Kristensen HG. Melt granulation in a laboratory scale high shear
mixer. Drug Dev Ind Pharm 1990; 16:12491277.
8. Schaefer T, Holm P, Kristensen HG. Melt pelletization in a high shear mixer. II. Power
consumption and granule growth. Acta Pharm Nordica 1992; 4:141148.
9. Schaefer T, Holm P, Kristensen HG. Melt pelletization in a high shear mixer. I. Effects of
process variables and binder. Acta Pharm Nordica 1992; 4:133140.
10. Schaeferand T, Mathiesen C. Melt pelletization in a high shear mixer. IX. Effects of binder
particle size. Int J Pharm 1996; 139:139148.
11. Schaeferand T, Mathiesen C. Melt pelletization in a high shear mixer. VIII. Effects of
binder viscosity. Int J Pharm 1996; 139:125138.
12. Schaefer T, Taagegaard B, Thomsen LJ, Kristensen HG. Melt pelletization in a high
shear mixer. V. Effects of apparatus variables. Eur J Pharm Sci 1993; 1:133141.
13. Schaefer T, Taagegaard B, Thomsen LJ, Kristensen HG. Melt pelletization in a high
shear mixer. IV. Effects of process variables in a laboratory scale mixer. Eur J Pharm
Sci 1993; 1:125131.
14. Plank R, Diehl B, Grinstead H, Zega J. Quantifying liquid coverage and powder flux in
high-shear granulators. Powder Technol 2003; 134:223234.
15. van den Dries K, Vromans H. Relationship between inhomogeneity phenomena and
granule growth mechanisms in a high-shear mixer. Int J Pharm 2002; 247:167177.
16. Kristensen HG. Agglomeration of powders. Acta Pharm Suecica 1988; 25:187204.
17. Kristensen HG, Holm P, Jaegerskou A, Schaefer T. Granulation in high speed mixers.
Part 4: effect of liquid saturation on the agglomeration. Pharm Ind 1984; 46:763767.
18. Kristensenand HG, Schaefer T. Granulation. A review on pharmaceutical wet-granulation.
Drug Dev Ind Pharm 1987; 13:803872.
19. Yajima T, Itai S, Hayashi H, Takayama K, Nagi T. Optimization of size distribution of
granules for tablet compression. Chem Pharm Bull (Tokyo) 1996; 44:10561060.
20. Wikbergand M, Alderborn G. Compression characteristics of granulated materials. IV.
The effect of granule porosity on the fragmentation propensity and the compactibility
of some granulations. Int J Pharm 1991; 69:239253.
21. Schaefer T, Holm P, Kristensen HG. Wet granulation in a laboratory scale high-shear
mixer. Pharm Ind 1990; 52:11471153.
22. Schaefer T, Holm P, Kristensen HG. Comparison between granule growth in a horizontal
and a vertical high speed mixer. I. Granulation of dicalcium phosphate. Arch Pharm
Chem, Scientific Edition 1986; 14:116.
23. Vromans H, Poels-Janssen HG, Egermann H. Effects of high-shear granulation on granulate
homogeneity. Pharm Dev Technol 1999; 4:297303.
24. Juppo AM, Yliruusi J, Kervinen L, Strom P. Determination of size distribution of
lactose, glucose, and mannitol granules by sieve analysis and laser diffractometry. Int
J Pharm 1992; 88:141149.
25. Suzuki T, Watanabe K, Kikkawa S, Nakagami H. Effect of crystallinity of microcrystalline
cellulose on granulation in high-shear mixer. Chem Pharm Bull (Tokyo) 1994;
26. Chowhan ZT. Aspects of granulation scale-up in high shear mixers. Pharm Technol 1998;
27. Badawy SIF, Lee TJ, Menning MM. Effect of drug substance particle size on the characteristics
of granulation manufactured in a high-shear mixer. AAPS Pharm Sci Tech 2000:1.
28. Vertommen J, Michoel A, Rombaut P, Kinget R. Production of pseudoephedrine hydrochloride
pellets in a high-shear mixer granulator. Eur J Pharm Biopharm 1994; 40:3235.
High-Shear Granulation 225
 2005 by Taylor & Francis Group, LLC
29. Ritala M, Jungersen O, Holm P, Schaefer T, Kristensen HG. A comparison between binders
in the wet phase of granulation in a high shear mixer. Drug Dev Ind Pharm 1986;
30. Ritala M, Holm P, Schaefer T, Kristensen HG. Influence of liquid bonding strength on
power consumption during granulation in a high shear mixer. Drug Dev Ind Pharm 1988;
31. Westerhuis JA, de Haan P, Zwinkels J, Jansen WT, Coenegracht PJM, Lerk CF. Optimization
of the composition and production of mannitol/microcrystalline cellulose
tablets. Int J Pharm 1996; 143:151162.
32. Makinoand T, Kitamori N. Effect of starch paste concentration on particle size distribution
of fine granules produced by agitation granulation. Chem Pharm Bull (Tokyo) 1995;
33. Kato Y, Moroshima K, Hashizume M, Ando H, Furukawa M. Further observation of
content uniformity of d-alpha-tocopheryl acetate as an oily drug in granules obtained
by wet granulation with a high-shear mixer. Drug Dev Ind Pharm 2001; 27:781787.
34. Miyamoto Y, Ryu A, Sugawara S, Miyajima M, Ogawa S, Matsui M, Takayama K,
Nagai T. Simultaneous optimization of wet granulation process involving factor of drug
content dependency on granule size. Drug Dev Ind Pharm 1998; 24:10551065.
35. Hariharanand M, Mehdizadeh E. The use of mixer torque rheometry to study the effect
of formulation variables on the properties of wet granulations. Drug Dev Ind Pharm
2002; 28:253263.
36. Becker D, Rigassi T, Bauer-Brandl A. Effectiveness of binders in wet granulation: a comparison
using model formulations of different tabletability. Drug Dev Ind Pharm 1997;
37. Suzuki T, Kikuchi H, Yamamura S, Terada K, Yamamoto K. The change in characteristics
of microcrystalline cellulose during wet granulation using a high-shear mixer.
J Pharm Pharmacol 2001; 53:609616.
38. Habib YS, Abramowitz R, Jerzewski RL, Jain NB, Agharkar SN. Is silicified wet-granulated
microcrystalline cellulose better than original wet-granulated microcrystalline
cellulose? Pharm Dev Technol 1999; 4:431437.
39. Badawy SI, Menning MM, Gorko MA, Gilbert DL. Effect of process parameters on
compressibility of granulation manufactured in a high-shear mixer. Int J Pharm 2000;
40. Oulahna D, Cordier F, Galet L, Dodds JA. Wet granulation: the effect of shear on granule
properties. Powder Technol 2003; 130:238246.
41. Bockand TK, Kraas U. Experience with the Diosna mini-granulator and assessment of
process scalability. Eur J Pharm Biopharm 2001; 52:297303.
42. Holm P, Schaefer T, Larsen C. End point detection in a wet granulation process. Pharm
Dev Technol 2001; 6:181192.
43. Timmins P, Delargy AM, Minchom CM, Howard JR. Influence of some process variables
on product properties for a hydrophilic matrix controlled-release tablet. Eur
J Pharm Biopharm 1992; 38:113118.
44. Kornchankul W, Hamed E, Parikh NH, Sakr A. Effect of drug proportion and mixing
time on the content uniformity of a low-dose drug in a high shear mixer. Pharmazie
2002; 57:4953.
45. Suzuki T, Kikuchi H, Yonemochi E, Terada K, Yamamoto K. Interaction of microcrystalline
cellulose and water in granules prepared by a high-shear mixer. Chem Pharm Bull
(Tokyo) 2001; 49:373378.
46. Schaefer T, Holm P, Kristensen HG. Comparison between granule growth in a horizontal
and a vertical high speed mixer. II. Granulation of lactose. Arch Pharm Chem, Scientific
Edition 1986; 14:1729.
47. Holm P. Effect of impeller and chopper design on granulation in a high speed mixer.
Drug Dev Ind Pharm 1987; 13:16751701.
226 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
48. Faure A, Grimsey IM, Rowe RC, York P, Cliff MJ. Process control in a high shear
mixer-granulator using wet mass consistency: the effect of formulation variables. J Pharm
Sci 1999; 88:191195.
49. Holm P, Schaefer T, Kristensen HG. Granulation in high-speed mixers. Part VI. Effects
of process conditions on power consumption and granule growth. Powder Technol 1985;
50. Corvari V, Fry WC, Seibert WL, Augsburger L. Instrumentation of a high-shear mixer:
evaluation and comparison of a new capacitive sensor, a watt meter, and a strain-gage
torque sensor for wet granulation monitoring. Pharm Res 1992; 9:15251533.
51. Kopcha M, Roland E, Bubb G, Vadino WA. Monitoring the granulation process in
a high shear mixer/granulator: an evaluation of three approaches to instrumentation.
Drug Dev Ind Pharm 1992; 18:19451968.
52. Holm P, Schaefer T, Kristensen HG. Granulation in high-speed mixers. Part V. Power
consumption and temperature changes during granulation. Powder Technol 1985;
53. Leuenberger H. Granulation, new techniques. Pharm Acta Helv 1982; 57:7282.
54. Werani J. Production experience with end point control. Acta Pharm Suecica 1988;
55. Laicher A, Profitlich T, Schwitzer K, Ahlert D. A modified signal analysis system for
end-point control during granulation. Eur J Pharm Sci 1997; 5:714.
56. Kornchankul W, Parikh NH, Sakr A. Correlation between wet granulation kinetic parameters
and tablet characteristics. Drugs Made Germany 2001; 44:7887.
57. Achanta AS, Adusumilli PS, James KW. Endpoint determination and its relevance to
physicochemical characteristics of solid dosage forms. Drug Dev Ind Pharm 1997; 23:
58. Whitaker M, Baker GR, Westrup J, Goulding PA, Rudd DR, Belchamber RM, Collins
MP. Application of acoustic emission to the monitoring and end point determination of
a high shear granulation process. Int J Pharm 2000; 205:7992.
59. Watano S, Numa E, Miyanami K, Osako Y. On-line monitoring of granule growth in
high shear granulation by an image processing system. Chem Pharm Bull (Tokyo) 2000;
60. Watano S, Numa T, Miyanami K, Osako Y. A fuzzy control system of high shear granulation
using image processing. Powder Technol 2001; 115:124130.
61. Bouckaert S, Massart DL, Massart B, Remon JP. Optimization of a granulation procedure
for a hydrophilic matrix tablet using experimental design. Drug Dev Ind Pharm
1996; 22:321327.
62. Vojnovic D, Selenati P, Rubessa F, Moneghini M, Zanchetta A. Wet granulation in a
small scale high shear mixer. Drug Dev Ind Pharm 1992; 18:961972.
63. Vojnovic D, Moneghini M, Rubessa F, Zanchetta A. Simultaneous optimization of several
response variables in a granulation process. Drug Dev Ind Pharm 1993; 19:1479
64. Vojnovic D, Moneghini M, Rubessa F. Optimization of granulates in a high shear mixer
by mixture design. Drug Dev Ind Pharm 1994; 20:10351047.
65. Vojnovic D, Moneghini M, Chicco D. Nonclassical experimental design applied in the
optimization of a placebo granulate formulation in high-shear mixer. Drug Dev Ind
Pharm 1996; 22:9971004.
66. Horsthuis GJB, van Laarhoven JAH, van Rooij RCBM, Vromans H. Studies on upscaling
parameters of the Gral high-shear granulation process. Int J Pharm 1993; 92:143150.
67. Ameye D, Keleb E, Vervaet C, Remon Jean P, Adams E, Massart Desire L. Scaling-up of
a lactose wet granulation process in Mi-Pro high shear mixers. Eur J Pharm Sci 2002;
68. Iskandarani B, Shiromani PK, Clair JH. Scale-up feasibility in high-shear mixers: determination
through statistical procedures. Drug Dev Ind Pharm 2001; 27:651657.
High-Shear Granulation 227
 2005 by Taylor & Francis Group, LLC
69. Leuenberger H. Scale-up of granulation processes with reference to process monitoring.
Acta Pharm Technol 1983; 29:274280.
70. Zlokarnik M. Problems in the application of dimensional analysis and scale-up of mixing
operations. Chem Eng Sci 1998; 53:30233030.
71. Lord R. The principle of similitude. Nature 1915; 95:6668.
72. Faure A, Grimsey IM, Rowe RC, York P, Cliff MJ. A methodology for the optimization
of wet granulation in a model planetary mixer. Pharm Dev Technol 1998; 3:413422.
73. Buckingham E. On physically similar systems; Illustrations of the use of dimensional
equations. Phys Rev NY 1914; 4:345376.
74. Faure A, Grimsey IM, Rowe RC, York P, Cliff MJ. Applicability of a scale-up methodology
for wet granulation processes in Collette Gral high shear mixer-granulators. Eur
J Pharm Sci 1999; 8:8593.
75. Siroisand PJ, Craig GD. Scale-up of a high-shear granulation process using a normalized
impeller work parameter. Pharm Dev Technol 2000; 5:365374.
76. Kristensen HG. Particle agglomeration in high shear mixers. Powder Technol 1996;
77. Rekhi GS, Caricofe RB, Parikh DM, Augsburger LL. A new approach to scale-up of
high shear granulation process. Pharm Technol 1996; 20:5867.
78. Ogawa S, Kamijima T, Miyamoto Y, Miyajima M, Sato H, Takayama K, Nagai T. A
new attempt to solve the scale-up problem for granulation using response surface methodology.
J Pharm Sci 1994; 83:439443.
228 Gokhale et al.
 2005 by Taylor & Francis Group, LLC
Low-Shear Granulation
Tom Chirkot
Patterson-Kelley Co., East Stroudsburg, Pennsylvania, U.S.A.
Cecil Propst
SPI Pharma, Norton Shores, Michigan, U.S.A.
Shearing of the powder bed occurs in many granulators. The low-shear granulators
included in this chapter are granulators that, for reasons of agitator speed, sweep
volume, or bed pressures, generate less shear than granulators discussed in other
chapters such as extruders or the high-shear mechanical granulators. Also considered
to be low shear, but covered in other chapters, are fluid bed granulators.
1.1. Comparison of Granulators
Comparing final granule characteristics in fluid bed, low-shear, and high-shear applications
is problematic because it is often difficult for the same formulation to be successfully
processed in each piece of equipment. However, some broad conclusions
may be drawn. Two of the characteristics most frequently reported are bulk density
and particle size; therefore, a comparison will be shown for these characteristics.
These granulators may accomplish the same unit operation but the final product outcome
may be quite different. The differences are a result of varied process requirements
in each granulator. The process of wet granulation involves several steps
including blending, liquid binder addition, and wet massing or distribution of the
liquid. After charging the powder to the mixer, a blending step is required to achieve
a homogenous blend. The time required to achieve the blend depends on the amount
of movement characteristic of the unit and the size of the unit. Also, the degree of
homogeneity differs from one type of mixer class to the next. In some classes of
mixers, one can easily overblend and segregate components of the mix.
The binder solution addition step follows the blending step. The selection of
the type of binder and quantity depends on the type of mixer selected for the wet
granulation. Nouh (1) studied a sulfadiazine formulation using several different binders.
He worked in both a fluid bed unit and a conventional wet massingscreening
method. When using 5% gelatin as a binder, he noted a 0.968mm average granule
size in wet massing and 0.574mm in the fluid bed. For acacia as binder at the 5%
 2005 by Taylor & Francis Group, LLC
level, the values were 0.90mm for wet massing and 0.605mm for the fluid bed. For
a polyvinylpyrrolidone (PVP) binder at the 5% level, the wet massing produced
0.962mm granule size vs. 0.247mm for the fluid bed. Surprisingly, the difference
between the methods did not transfer to bulk density results. Both the acacia and
the PVP formulations had similar bulk densities in both methods. The gelatin did
show a difference, yielding 0.476 g/mL for the wet massing and 0.294 g/mL for
the fluid bed. Gore et al. (2) made a comparison among a fluid bed granulator, a
planetary mixer, and a high-shear granulator. They found that the granule yield
between 20 and 100 mesh was 97.7% in the fluid bed, 53.3% in the planetary mixer,
and 71.8% in the high-shear granulator. The respective bulk densities were 0.39, 0.66,
and 0.71 g/mL. For comparison, in a low-shear tumble granulator, Scarpone et al.
(3) found 84% yield between 20 and 100 mesh in a 0.05 m3 (2 ft3) vessel and
50.7% yield in a 2.83 m3 (100 ft3) vessel. Density measurements averaged 0.486 g/
mL for the two trials of a cardiovascular drug formulation in the smaller vessel.
Sheskey and Williams (4) studied a niacinamide formulation in both low- and
high-shear granulators and concluded that the resulting apparent densities were
nearly identical in both cases. In general, the bulk density values produced in
low-shear tumbling granulators are intermediate in value between those of a fluid
bed and of a high-shear granulator.
Similar conclusions may be drawn about granule morphology as lower-shear
granulators produce fluffier, more porous granules than do high-shear granulators.
A final note related to morphology: If the low-shear tumbling granulator has a rotating
shell, some rounding of granules may be expected as the material flows through
the angle of repose.
1.2. Binder Issues
Because of improved liquid distribution early in the granulation process, some highshear
granulators require 6080% of the liquid needed in a low-shear mixer. Liquid
requirement for an antacid granulation made in a high-shear horizontal mixer was
4.5% of the total batch, whereas 7% liquid was required for a low-shear planetary
mixer. Some of the low-shear granulators such as the sigma mixer densify granules.
The liquid requirement for the sigma mixer is slightly reduced at 6% for the same
antacid. Some studies indicate that dissolving a solid binder and adding it to the
dry mix vs. adding the solid binder to the dry mix and then wetting increases dried
granule hardness (5).
The final stage of the wet granulation process is the liquid distribution or the
wet massing. This step can be compared with a kneading step during which the voids
between granules are compressed and the granules are thereby densified. The final
density of the granules, therefore, is dependent on the amount of shearing available
in the unit. Because the shear is introduced in the mixing process by mechanical
means by moving impellers or blades, these low-shear mixers cannot compress the
voids between the granules solely by mechanical means and thus require more binder
solution to form granules of some integrity.
The combination of less mechanical shear and lower bed pressure allows for
a thicker binder deposit, especially when the binder is poured into the granulation
unit instead of sprayed. This thicker film often dries to a more effectively spreadable
film during the compression of these granules to form tablets. This increased plasti-
230 Chirkot and Propst
city is especially true at the lower compression forces (Fig. 1). The tablets also have
 2005 by Taylor & Francis Group, LLC
much lower friability (Fig. 2). The improved friability can be over a much broader
range of compression forces.
The machine classes to be considered under mechanical agitator granulators are;
(1) ribbon or paddle blender; (2) planetary mixers; (3) orbiting screw mixers; and
(4) sigma blade mixers.
Figure 2 Friability vs. compression forces.
Figure 1 Hardness vs. compression forces.
Low-Shear Granulation 231
 2005 by Taylor & Francis Group, LLC
2.1. Ribbon or Paddle Blenders
The ribbon blender-type mixer (Fig. 3) is very popular as a dry mixer. However, if
small amounts of liquid are added or if a dry paste is formulated for the machine,
the ribbon blender can serve as a very reliable granulator (5). Most ribbon blenders
are not made to withstand the shaft torque required for manufacturing granules.
One should ensure that the granulator shaft is strong enough before granulating
in a ribbon blender designed to dry mix.
Zoglio et al. (5) studied several kneading times in a chopper ribbon blender and
discovered an optimum range of kneading times for providing the best granule
mechanical properties. A kneading time that is too long densifies the granules. This
increases moisture exposure and particle size and reduces porosity. The tendency of
the material to stick to the side wall is pronounced in the ribbon blender.
Paddles instead of ribbons decrease sticking problems and the torque required.
Paddle blenders as batch granulators can handle wetter paste. These paddle blenders
are occasionally used as continuous granulators and have both lower torque and
more applications than a continuous ribbon blender. The movement of the paste
helps remove material from the paddles. Formulation of a nontacky paste is aided
by using a substance such as microcrystalline cellulose that helps absorb the excess
moisture from the mix, yet is plastic and nonsticky.
Two very popular ribbon or paddle blenders used as granulators are the topogranulator
and the turbulizer.
either compress or mechanically fluidize the granulation. Compression while slightly
wet increases the overall influence of the liquid on the particle size of the granulation.
At the other extreme, fluidization reduces density if used during the granule growth
phase or speeds drying. The topogranulator is also a vacuum dryer.
The topogranulator is used extensively to make effervescent products by liquid
addition under vacuum or by the Murry fusion method (6). Murrys method uses
liberated moisture from the acid in the mix (i.e., hydrous citric acid) to start the
acidbase reaction, which generates more water. Thus, granulating of the sodium
bicarbonatecitric acid mixture can be accomplished. The water produced must be
removed quickly to reproducibly stop the reaction. The topogranultor, because of
its ability to compress the particles into the binding moisture, makes a larger, denser
Figure 3 Ribbon blender.
232 Chirkot and Propst
The topogranulator (Fig. 4) is a batch-style ribbon blender with the ability to
 2005 by Taylor & Francis Group, LLC
granulation with lower moisture content. Also, mechanical fluidization and vacuum
drying removes the water from the reaction more quickly, making the process more
Granulation in vacuum removes entrapped air from the particle surface.
Liquid addition into the vacuum provides immediate wetting and also begins the
drying process. This granulation in vacuum allows the manufacturing of a calcium
carbonate-based effervescence with rapid reactivity (7).
The turbulizer is a continuous paddle granulator. By using continuous powder
feeders and liquid metering pumps, the unit produces large quantities of product per
hour in a very small space. The unit provides adjustable mixing, shear, and impact
action based on the revolutions per minute (rpm) of the shaft and the angle of the
impact blade. The paddle adjustments also vary the retention time. This machine
has a very low silhouette. Paddles are easily accessible for cleaning and inspection
either through a clam shell- or drop door shell-opening side wall. The unit is jacketed
for material temperature control.
Another continuous paddle granulator made by Teledyne Readco has been
studied by Ghali et al. (8). The unit is adjustable to produce either low- or highshearing
action. The amount of energy induced by the rotating shaft can be selected
by using various types of pins. A rounded granule results from the action of the pin
tip speed and some rolling of the granulation against the fixed vessel wall.
2.2. Planetary Mixers
The planetary motion of these granulators is created by rotating the agitator off an
assembly in a direction opposite that of the rotation of the agitator assembly as it
mercial names including Hobart, Kitchen Aide, Pony, and AMF Glen granulators.
All of these mixers have the same basic makeup, which includes: (a) planetary
motion, (b) removable bowl, and (c) top-drive agitators.
These mixers tend to be better at mixing dry powders in a horizontal plane than
the vertical. Lack of vertical mixing may require the materials to be dumped and
returned to the bowl to obtain an acceptable dry mix. Reduced vertical mixing
Figure 4 Topogranulator.
Low-Shear Granulation 233
moves around the bowl. The planetary mixers (Fig. 5) are represented by many com-
 2005 by Taylor & Francis Group, LLC
is much less a problem in the wet-mix phase, because the materials adhere to the
agitator and move in groups without vertical stratification.
Remon and Schwartz, working with microcrystalline cellulose and lactose mixtures
in a planetary mixer, saw decreased friability of the granule with increasing
massing time (9). This increased massing time improves binder distribution and
mechanical strength. Ghanta et al. demonstrated the mechanism of granule growth
in a Hobart mixer (10).
2.3. Orbiting Screw Granulators
The orbiting screw granulator (Fig. 6) is also used mainly for dry mixing. However,
the unit has been fitted with the nozzle through the center agitator to add liquids to
dry powders (11). Also, a jacket can provide both heating and cooling. A sintered
metal plate can allow entry of compressed air through the skin of the mixer or
can exhaust moisture to obtain drying. All of these added features, along with a
Figure 5 Planetary mixer.
Figure 6 Orbiting screw mixer.
234 Chirkot and Propst
 2005 by Taylor & Francis Group, LLC
chopper in the side wall, allow this blender to be an effective granulator for powders,
slurries, suspensions, and paste. The unit is a very gentle granulator mixer.
2.4. Sigma Blade Mixers
The sigma blade mixer (Fig. 7) is a compressive granulator. Generally, it develops
material into a paste- or dough-type consistency. Often the sigma blade mixer is preceded
by a dry mixer. The unit generates liquid distribution with pressure during
granulating. This creates a uniform granulation with binder matrix and good binder
These vessel shapes are usually some derivation of a cylinder, with a double cone and
counterparts, the machine shells rotate around an axis parallel to the ground. The
rotation speed is moderate, and generally falls into a range of 72.2106.7 m/min
(250350 ft/min) peripheral speed. The rpm value changes as vessel size grows.
laboratory model may rotate at 2530 rpm, whereas a larger production model
may rotate at 48 rpm. The peripheral speed remains constant in order to maintain
scale-up relationship among vessel sizes.
Figure 7 Sigma blade mixer.
Low-Shear Granulation 235
V-shape (Fig. 8) being common examples. Unlike their fluid bed and high-shear
 2005 by Taylor & Francis Group, LLC
A second rotating device is located on the same axis of rotation as the shell.
This bar or arm may be supported at both ends or may be cantilevered, with support
only on a single end. The bar is used to impart substantially more energy into the
system than that delivered by the rotating shell. The bar is driven by an additional,
larger motor that runs independently of the shell motor. The manufacturers descriptive
names for these bars, such as agitator or intensifier bars, confirm their use.
Unlike the gentle rolling action induced by the shell movement, the bar movement
has a high-speed nature. This should not be confused with high shear because the
bar movement produces mainly a convective motion within the material. Indeed,
these machines were originally designed as mixers, with elements of diffusive and
convective mixing. Only later in their evolution did they become granulator-dryers.
The increased peripheral speed of the bar is substantial. As a general rule, they
operate at 10 times the speed of the shell or about 914.4 mpm (3000 fpm). The bars
may also serve as liquid addition devices. The vessels may be jacketed for heating
and cooling if there is a necessity for these options during granulation. They are
vacuum capable, which makes them ideal candidates as a single-pot processor for
mixing, granulating, and drying in the same vessel. Ultimate vessel size for these
granulators is constrained by the weight of the agitator bar. The bars have many
parts and must be disassembled often for cleaning. Large bars become increasingly
unwieldy and difficult to remove from the interior of the vessel.
Ample opportunity exists for fine-tuning a granulation process in these machines.
Apparatus variables that can be adjusted are shell speed, bar speed, bar size, and bar
design. Process changes can be made in mix time, liquid addition time, bed temperature,
internal vessel pressure, and spray droplet size.
Overwhelmingly, these granulator types are used in batch operations. However,
a few manufacturers offer continuous granulating systems that emulate the
combination of forced and rolling agglomeration found in the batch machines.
Besides the issue of validating a continuous process, throughput requirement is
usually the influencing criterion when considering a change from batch to continuous
Figure 8 Schematic of the twin-shell blender that allows blending and granulating in a single
vessel. (Courtesy of Patterson-Kelley Co., Division of Harsco Corp., East Stroudsburg, PA.)
236 Chirkot and Propst
 2005 by Taylor & Francis Group, LLC
Much of the granulation done in the pharmaceutical industry is by wet massing
and, more recently, by fluidized bed (12). Consequently, the literature is rich with
articles describing the many variables studied in these machines. On the other hand,
the rotating shape granulators originated as mixing devices and a paucity of research
articles exist on their usage as granulators. This makes them a fruitful area for future
3.1. Bar Speed and Energy Input
The combination of agitator bar speed and its accompanying energy input wields
a wide influence on the physical properties of the granulated material. Watano
et al. (13), using a fluidized bed granulator with an agitator blade, varied the rotation
speed of the agitator and observed the effect it had on product density. They were
able to derive a predictive equation that yielded the density value when the agitator
speed was known. The same group measured shape indexthe mean ratio of the
short/long diametersand concluded that as granulation times increased, the shape
tended more toward sphericity. The rotating shape granulators vary energy input to
the powder bed by changing the agitator speed or extending the running time of the
bar after liquid addition has stopped. A common term for the extended bar period is
postmix time.
A study of a low-shear granulator, with a formulation consisting of lactose,
PVP, and water, that varied the peripheral speed of the bar showed no statistically
significant change in granule density despite the substantial change in bar tip speed
from 636 to 846 to 1056 mpm (14). The tip speed was statistically influential on the
amount of granules found in the yield fraction. The lowest speed produced the best
yield indicating that the higher speeds can degrade the formed granules during
a postmix phase and thereby cause a reduction in yield. The ultimate lower limit
on tip speed is the point where adequate distribution of the binder solution is
3.2. Disk Size and Bar Design
Addition of binder liquid with spray heads is available with rotating shape granulators.
The double-cone shape is often used with a dry intensifier bar and spray head
combination for granulating. A wide range of bar designs exists for the more typical
situation of adding liquid through the bar. Blade or knife design is one variable to
consider. The blades extend perpendicularly from the disk circumference and then
bend 90
The blade shape is critical for the liquid addition method. As the blades rotate,
they carve out a toroidal void volume in the flowing powder bed. The liquid exiting
the disk coats the interior of the torus. Proper vessel loading ensures that no liquid
impinges on the wall of the vessel.
The action of the blades is quite vigorous and may cause a problem with very
friable materials. A modification to the bar that removes the blades and recesses all
nuts and bolts can be used to granulate these fragile materials. Further fine-tuning of
the spray pattern can be accomplished with changes in the positioning of liquid evacuation
spacing on the disk circumference. A straight pattern creates droplet flow
orthogonal to the axis of rotation, which allows only a small area of the interior
torus to be coated. An angled pattern creates a substantially wider spray effect, with
more efficient droplet distribution.
Low-Shear Granulation 237
in a shape often called a dog-eared design (Fig. 9).
 2005 by Taylor & Francis Group, LLC
Disk diameter is critical. A large-diameter disk imparts more energy as torque
is a function of diameter. The larger diameter offers more circumference for liquid to
be evacuated and carves out additional void volume in the material. The sweep
volume of the disks is considered a major factor in properly scaling up a granulator
3.3. Liquid Addition Rate
Particle motion is so complex when granulating that it becomes difficult to determine
the optimum rate of liquid addition. Stamm and Paris (16) studied rates of 5, 10, and
20 mL/min in a fixed shell, helicoidal mixer with vertical bar, and determined that
the best results occurred at the slowest rate.
Cliff (17), working in a high-shear granulator, found that a long binder addition
time was needed to prevent overdensification. Lipps and Sakr (18), using a topspray,
fluidized bed granulator, related geometric mean granule size, specific surface
area, and granule flow properties to the binder flow rate.
In rotating shape granulators, the main energy-imparting function and the
liquid distribution function are encompassed within the agitator bar, yielding
a highly efficient system. Liquid is literally ripped into droplets by the peripheral
speed of the bar. Tumbling action ensures that fresh material replaces the wet material
during each shell revolution (19). A wide range of liquid addition rates is permissible.
The lowest rate is undefined and may involve a slow dropwise addition to the
bar. The highest rate is limited to the point when the liquid rate reduces the droplet
making efficiency of the bar and results in liquid being forced out in a sheet-like
pattern to the powder bed.
When one extends the liquid addition period, the bar continues to abrade the
granules. Favorably, any particles that have been fractured can be rewetted and the
fracture surface can easily bind to another particle. This mechanism tends to hold
down the oversized fraction without the expected increase in the undersized fraction.
Extremely rapid liquid addition overwhelms the blending capability of the system.
Many particles simply do not have enough time to become associated with the
droplets, whereas others become overwet due to the concentrated onrush of liquid.
An additional consequence of high liquid addition rate is that liquid may have diffi-
culty releasing from the bar. A back pressure situation develops that may impede
liquid flow in the feed tube and ultimately cause it to reverse its flow. High enough
back pressure can induce a spring-loaded bar to jump from its moorings and damage
Figure 9 Intensifier bar for the twin-shell blender/granulator showing a dog-eared design.
(Courtesy of Patterson-Kelley Co., Division of Harsco Corp., East Stroudsburg, PA.)
238 Chirkot and Propst
 2005 by Taylor & Francis Group, LLC
the interior of the machine. Because the rotating shape granulators are often heatjacketed,
the potential exists for them to provide a partially or fully coated, hard
Use of the dissolved solids binder addition method through an agitator bar
could be problematic. Any undissolved solids may clog the liquid evacuation spacing
on the disk. Also, binder solutions with a viscosity much beyond 300500 cP may not
atomize readily when released from the disk.
3.4. Droplet Size
Droplet size is a very important variable in fluidized bed granulation. Droplet size is
unimportant in high-shear granulators because the energy exists to readily distribute
and redistribute the liquid. Liquid addition in the rotating shape granulators is
unique, for the droplets are released from the fluidizing device or agitator bar. An
analogous situation in a high-shear application would mean that liquid was being
added through the impeller or chopper.
The liquid in a rotating shape granulator is fed through a tube and exits in the
interior of the bar through a distribution slot. A pump can be used for metering the
flow, or gravity flow may be sufficient, if desired. The speed of rotation of the bar
actually induces an area of lower pressure within the bar to draw in the liquid.
The liquid exits the agitator bar through openings on the circumference of the rotating
disk. Openings can be adjusted by placement of spacers that vary in thickness.
A droplet diameter cannot exceed this thickness value, although smaller droplets
may be present. Typical spacer openings are 0.025 cm (0.010 in.) with common adjustment
to 0.0125 cm (0.005 in.) and 0.05 cm (0.020 in.), depending on the viscosity of the
Many granulation methods deliver the liquid at some pressure through a spray
head. With liquid addition through the bar, only enough pressure is needed to overcome
line loss. Undue pressure is actually counterproductive. The atomization is
influenced by pressure and the liquid exiting the disks under high pressure tends
to be sheet-like rather than easily distributed droplets.
Droplet size in a fluid bed has been measured by capturing droplets on a slide
covered with viscous oil. The size has been determined to be in the 20100 mm
range (20). Less delicate techniques in the rotating shape granulators have shown
droplet sizes of about 250 mm. Agland and Iverson provide an experimental study
showing the relation of liquid droplet size to granule characteristics such as size
and penetration of liquid (21).
3.5. Vessel Loading
A negative factor for rotating tumble granulators with high-speed internal bars is the
tight constraint for fill level. The fill level is usually about 5065% of the total
volume, and the powder level must have some contact with the bar. Overloading
the vessel impedes the mixing action.
Underloading causes material to flow beneath the bar during vessel rotation,
resulting in liquid spraying on the interior wall. This spray-through can be a factor
in material sticking during a drying step.
Even careful loading of the machine with the dry powder may not be sufficient
to preclude spraying through to the walls. If substantial densification occurs during
early liquid addition, the load may drop enough during the final stages of liquid
Low-Shear Granulation 239
 2005 by Taylor & Francis Group, LLC
addition for liquid to find its way to the shell wall. The effective working volume fill
level can be increased if the tip-to-tip blade diameter is extended.
3.6. Low Shear Single Pot Processing
Their original design as mixers ensures even minor ingredients are well-distributed in
the dry mix phase prior to granulation.
After granulation, the heated shell and vacuum capability are used for gentle
tumble drying and collection of condensed vapors. The bar can then be used to provide
a measure of dry sizing, followed by lubricant addition, and tumble blending.
3.7. Continuous Granulation
Although batch granulation is commonly used, there are some examples of lowshear
continuous granulation. A true adaptation of the batch, rotating shape
sists of a high-speed mixing chamber connected to a series of three V-shaped tumble
blenders. The principles of liquid addition and disk speeds covered earlier are equally
applicable to the Zig Zag.
The typical hold-up volume is half of the total volume of the vessel. The holdup
volume can be adjusted somewhat by raising or lowering the discharge end a few
degrees from the horizontal. Residence times are in the 35 min range for granulation
applications. Two types of granulation mechanisms occur in this machine.
When the material is in the high-speed mixing chamber, the liquid droplets contact
the powder in a manner similar to that of batch units. As the wet material travels
into the tumbling section, rolling agglomeration occurs. The tumbling section splits
its load in half on each revolution, allowing half the material to proceed forward
while the other half recycles to the rear. This action tends to smooth out any inconsistencies.
Scale-up to larger Zig-Zag units is based on constant residence times. Therefore,
even in the very largest models, a residence time of 35 min can be expected.
The largest models are capable of producing 30,000 kg/hr.
Often, the most difficult part of implementing a granulation process is the scale-up
step. The many variables involved and the poorly understood relations among the
variables often places scale-up in the realm of trial and error. Hancock et al. (22),
using a fixed bowl mixer with meshing blades, linked a torque arm to a dynamometer
to monitor the process. The amplitude of oscillation (torque range) and the mean
torque increase from baseline were the items recorded. Torque showed a relation
for the force required to shear the wet mass and provided an output for determining
the wet mass strength. Integrating the torque-time curve yields the total energy into
the system for a particular time period. Hancocks group defined the term, cumulative
energy of mixing, and postulated its use as a means of scaling up.
Vojnovic et al. (23) determined that peripheral speed of the impeller in a vertical
high-shear mixer was an important factor and adjusted this speed as a function
240 Chirkot and Propst
The rotating shape granulators offer fine potential for single-pot processing (Fig. 10).
granulators is the Patterson-Kelley Zig Zag mixer (Figs. 11 and 12). This vessel con-
 2005 by Taylor & Francis Group, LLC
Figure 10 Schematic diagram of Patterson-Kelley Solids Processor. (Courtesy of Patterson-Kelley Co., Division of Harsco Corp., East Stroudsburg, PA.)
Low-Shear Granulation 241
 2005 by Taylor & Francis Group, LLC
of the diameter of the larger mixer. They found this to be an adequate scale-up
method despite a geometric dissimilarity in the two mixers studied.
Further work on the rheologic character of the wet mass and the effect on
scale-up was performed by York in a high-shear mixergranulator (24). It was theorized
that the power numberReynolds number relationship used in the scale-up of
fluid mixing impeller systems could be applicable to the granulation mechanism in
high-shear mixer. To analyze the relationship, appropriate machine information
needed to be measured, as did the density and viscosity of the wet mass. Mixer
torque rheometry was used to determine wet mass viscosity, and yielded a pseudo-
Reynolds number with the units of torque. It is interesting to note that some unit
operations textbooks show similar logic in developing scale-up equations for dry
powder mixing (25).
Figure 12 Zig-Zag mixeragglomerator. (Courtesy of Patterson-Kelley Co., Division of
Harsco Corp., East Stroudsburg, PA.)
Figure 11 Schematic of a Zig-Zag, which is an example of a continuous granulating machine.
(Courtesy of Patterson-Kelley Co., Division of Harsco Corp., East Stroudsburg, PA.)
242 Chirkot and Propst
 2005 by Taylor & Francis Group, LLC
Scaling up the process in a low-shear granulator is a daunting task. Scale-up
equations are usually proprietary to the mixer manufacturers and little is written
about them in the literature. The equations have a sound basis in science but are
ameliorated by constants and factors that take into consideration empirical feedback
and intuition concerning the machine function.
Most commonly, an energy/volume ratio is established in the smaller vessel.
This ratio is emulated in the larger vessel, with appropriate consideration for the
added mass and the larger motor, to yield an estimated scale-up time.
Scarpone et al. (3) studied a change in granulation technology from a highshear
mixing bowl to a V-shaped granulator. Developmental work was done in a
0.056 m3 (2 ft3) vessel and the ultimate production in a 2.83 m3 (100 ft3) vessel.
Scale-up based on the manufacturers recommendation was not exact and required
some fine-tuning before the process was established in the larger vessel.
This study provides a detailed look at some of the hazards encountered in scaling
up. The initial trial in the 2.83m3 (100 ft3) production vessel failed to provide an
adequate granulation. The main reason for the failure was that the scale-up trial
lacked the active ingredient and contained only excipient. This highlights the
necessity of maintaining as much constancy as possible, both with formulation
and process variables. Additional tests with the active ingredient were much closer
to the results gained in the smaller vessel. By the third trial, a successful scale-up
had been achieved. Further studies were conducted to evaluate the influence of
changes in the physical characteristics of the active ingredient. For example, density
changes had an influence on the binder liquid needed. The researchers were able to
use their knowledge from the scale-up to achieve another scale-up on a different
M. Uliveri, M. Eli, and R. Bianchini (unpublished data) also attempted a scaleup
in a V-shaped granulator. The laboratory vessel was a 0.056m3 (2 ft3) model and
the scale-up vessel was a 0.283m3 (10 ft3). They found it was necessary to increase the
batch size 1.2 times over the direct 10:2 ratio, that the total liquid was 82% of direct
scale-up, and the granules were slightly larger in the scale-up vessel.
Another scale-up attempt in a V-shaped granulator involved a 2 and 57 L vessel
(15). The author used a cumulative torque to mass ratio as a means of relating the
process in both vessels. A separate, individual method is also provided for scaling-up
the liquid addition time.
Suitably, an article by Ennis et al. (26) highlights the fact that understanding
granulation scale-up is a needed growth area for further research.
The rapidity by which granulation proceeds makes end-point determination a diffi-
cult problem. Growth behavior is nonlinear, preventing easy solutions to the differential
equations describing the process (27). Watano et al. (13) used an infrared
moisture sensor to continuously monitor moisture through a PID feedback loop.
The current was used to control the liquid addition pump until a suitable size was
achieved. Leuenberger et al. (28) plotted the power consumption curve in a planetary
mixer and observed five distinct phases that were dependent on the amount of liquid
added. Zoglio et al. (5), working in a ribbon blender, postulated a potential
end-point determination using specific pore area. Ghanta et al. (10) installed a slip
ring torque sensor between the Hobart mixer agitator and motor and observed five
Low-Shear Granulation 243
 2005 by Taylor & Francis Group, LLC
phases in the torque profile. This process was repeatable and could be considered for
granulation end point in this type of mixer. Another method uses effusivity to determine
the amount of moisture in the granule for end-point determination (29).
Low-shear granulators offer a middle ground solution for many of the problems formulators
may encounter. Much denser granules may be produced in these vessels,
compared with a fluid bed device, yet the energy expended is not as great as that
found in a high-shear machine. Many of the low-shear granulators are extremely
adaptable devices capable of mixing the formulation constituents before granulating,
and some even dry the materials after granulation is complete. Ample opportunity is
available in these granulators for adjustments: energy input may be altered with variable
speed drives; droplet size may be changed through judicious selection of spray
heads; and various agitating bar designs may be selected. As with many other granulators,
scale-up and end-point detection remain poorly defined. A trial and error
procedure is often the only method to determine end point and scale-up.
1. Nouh ATI. The effect of variations in concentration and type of binder on the physical
characteristics of sulfadiazine tablets and granulations prepared by wet and fluidized-bed
granulation method. Pharm Ind 1986; 48(6):670.
2. Gore AY, McFarland DW, Batuyios NH. Fluid-bed granulation: factors affecting the
process in a laboratory development and production scale-up. Pharm Technol
3. Scarpone AJ, Dalvi UG, Delorimier AE. Preparing tablet granulations in a 100 cu ft
solids processor. Pharm Technol 1986; 10:44.
4. Sheskey PJ, Williams DM. Comparison of low-shear and high-shear wet granulation
techniques and the influence of percent water addition in the preparation of a controlled-
release matrix tablet containing HPMC and a high-dose, highly water-soluble
drug. Pharm Technol 1996:8090.
5. Zoglio MA, Huber HE, Koehne G, Chan PL, Carstensen JT. Physical aspects of wet
granulation II: factors involved in prolonged and excessive mixing. J Pharm Sci 1976;
6. Murry R. J Pharm Sci 1968; 57:17761779.
7. Gergely et al. Effervescent system for effervescent tablets and effervescent granules. US
Patent 5888544, March 30,1999.
8. Ghali S, Contractor AM, Mankad AD, OConnor RE, Schwartz JB, Auslander DE,
Grim WM. A high-speed mixer for continuous wet granulation. Pharm Technol 1990;
9. Remon JP, Schwartz JB. Effect of raw materials and processing on the quality of granules
prepared from microcrystalline cellulose-lactose mixtures. Drug Dev Ind Pharm
1987; 13:1.
10. Ghanta SR, Srinivas R, Rhodes CT. Use of mixer torque measurements as an aid to optimizing
wet granulation process. Drug Dev Ind Pharm 1984; 10:305311.
11. Day JH & Co. News bulletin: monitoring particle density, moisture content, particle size,
and process weight. Cincinnati, OH, 1983.
12. Schaefer T. Equipment for wet granulation. Acta Pharm Suecica 1988; 25:205.
244 Chirkot and Propst
 2005 by Taylor & Francis Group, LLC
13. Watano S, Terashita K, Miyanami K. Determination of end-point with a complex granulation
applying infrared moisture sensor. Chem Pharm Bull 1991; 329:1013.
14. Chirkot TS. Characterization of a pharmaceutical wet granulation process in a V-type
granulator. Pharm Eng 1999; 19(4).
15. Chirkot TS. Scale-up and endpoint issues of pharmaceutical wet granulation in a V-type
low shear granulator. Drug Dev Ind Pharm 2002; 28:871.
16. Stamm A, Paris L. Influence of technological factors on the optimal granulation liquid
requirement measured by powder consumption. Drug Dev Ind 1985; 11:333.
17. Cliff MJ. Granulation end point and automated process control of mixer-granulators:
part I. Pharm Technol 1990; 14:112.
18. Lipps D, Sakr AM. Characterization of wet granulation process parameters using
response surface methodology. 1. Top-spray fluidozed bed. J Pharm Sci 1994; 83:937.
19. Fischer JJ. Liquid-solids blending. Chem Eng 1962; 3(Feb 5).
20. Lindberg NO, Jonsson C. The granulation of lactose and starch in a recording high-speed
mixer, Diosna P25. Drug Dev Ind Pharm 1985; 1:387.
21. Agland S, Iverson S. The impact of liquid droplets on powder bed surfaces. CHEMECA
99, Newcastle, Australia, Sep 2629, 1999:25.
22. Hancock C, York P, Rowe RC, Parker MD. Characterization of wet masses using
a mixer torque rheometer: 1. Effect of instrument geometry. Int J Pharm 1991; 76:239.
23. Vojnovic D, Moneghini M, Rubessa F, Zanchetta A. Simultaneous optimization of several
response variables in a granulation process. Drug Dev Ind Pharm 1993; 19:1479.
24. York P. Granulation using mechanical agitation. Proceedings of IFPRI Annual Meeting,
Urbana, IL, 1995.
25. McCabe WL, Smith JC, Harriot P. Unit Operations of Chemical Engineering. 4th ed.
New York: McGraw-Hill, 1985:846.
26. Ennis BJ, Green J, Davies R. The legacy of neglect in the US. Chem Eng Prog 1994:32.
27. Oliver R, Ford LJ. The role of agglomeration in chemical and process technology. Proceedings
of the Institute for Briquetting and Agglomeration, Orlando, FL, 1987:132.
28. Leuenberger H, Bier HP, Sucker HB. The theory of the granulating liquid requirement in
the conventional granulation process. Pharm Technol 1986; 11:76.
29. Mathis Instruments. Technical Literature. Fredericton, NB, Canada: Mathis Instruments,
Low-Shear Granulation 245
 2005 by Taylor & Francis Group, LLC
Batch Fluid Bed Granulation
Dilip M. Parikh
Synthon Pharmaceuticals Inc., Research Triangle Park, North Carolina, U.S.A.
Martin Mogavero
Niro Pharma Systems, Columbia, Maryland, U.S.A.
Fluidization is the unit operation by which fine solids are transformed into a fluidlike
state through contact with a gas. At certain gas velocities, the fluid will support
the particles, giving them freedom of mobility without entrainment. Such a fluidized
bed resembles a vigorously boiling fluid, with solid particles undergoing extremely
turbulent motion, which increases with gas velocity. Fluidized bed granulation is a
process by which granules are produced in a single piece of equipment by spraying
a binder solution on to a fluidized powder bed. This process is sometimes classified as
the one-pot system. The fluid bed granulation process has received considerable attention
within the pharmaceutical industry; however, other process industries, such as
food, agro-chemical, dyestuffs, and other chemical industries, have adopted the fluid
bed granulation process to address particle agglomeration, dust containment, and
material handling. The fluidization technique, as it is known today, began in 1942,
with the work of the Standard Oil Company (now known as Exxon, in the United
States) and M.W. Kellogg Company, in an effort to produce the first catalytic cracking
plant on a commercial scale (1).
Fluid bed processing of pharmaceuticals was first reported by Wurster, when
he used the air suspension technique to coat tablets (2,3). In 1960, he reported on
granulating and drying of a pharmaceutical granulation, suitable for the preparation
of compressed tablets, using the air suspension technique. In 1964, Scott et al. (4)
and Rankell et al. (5) reported on the theory and design considerations of the process,
using a fundamental engineering approach and employing mass and thermal
energy balances. They expanded this application to the 30 kg-capacity pilot plant
model designed for both batch and continuous operation. Process variables, such
as airflow rate, process air temperature, and liquid flow rate were studied. Contini
and Atasoy (6) later reported the processing details and advantages of the fluid
bed process in one continuous step.
Wolf (7) discussed the essential construction features of the various fluid bed
components, and Liske and Mobus (8) compared the fluidized bed and traditional
 2005 by Taylor & Francis Group, LLC
granulation process. The overall results indicated that the material processed by the
fluid bed granulator was finer, more free-flowing, and had homogenous granules,
which, after compression, produced stronger and faster disintegration of tablets
than the materials processed by conventional wet granulation. Reviews by Sherrington
and Oliver (9), Pietch (10), and a series published on the topic of Fluidization
in the Pharmaceutical Industry (1117) provide an in-depth background on the
fundamental aspects of the fluidized bed and other granulation technologies. The
fluidized bed was used only for drying the pharmaceutical granulation efficiently
in early days, but now it is employed routinely for drying, agglomeration, pelletization,
and production of modified-release dosage forms using air suspension coating.
Because of this, these units are normally classified as multiprocessor fluid bed
The size enlargement of primary particles has been carried out in the pharmaceutical
industry in a variety of ways. One of the most common unit operations used
in the pharmaceutical industry is fluid bed processing. The batch size increase using
fluid bed granulation requires a good understanding of the equipment functionality,
theoretical aspect of fluidization, excipient interactions, and most of all identifying
the critical variables that affect the process of agglomeration.
This chapter will provide the essential understanding of the fluidization theory,
system description that make up the fluid bed processor, and will discuss the critical
variables associated with the equipment, the product, and the process.
A fluidized bed is a bed of solid particles with a stream of air or gas passing upward
through the particles at a rate great enough to set them in motion; this velocity,
according to Ku? lling and Simon (18), is higher than the incipient fluidizing velocity,
but lower than the entrainment velocity. When the rate of flow of gas increases, the
pressure drop across the bed also increases until, at a certain rate of flow, the frictional
drag on the particles equals the effective weight of the bed. These conditions,
and the velocity of gas corresponding to it, are termed incipient fluidization and incipient
velocity, respectively. The relationship between the air velocity and the pressure
At low gas velocities the bed of particles is practically a packed bed, and the
pressure drop is proportional to the superficial velocity. As the gas velocity is increased,
a point is reached at which the bed behavior changes from fixed particles to suspended
particles. The superficial velocity required to first suspend the bed particles is known as
minimum fluidization velocity (umf). The minimum fluidization velocity sets the lower
limit of possible operating velocities, and the approximate pressure drop can be used
to approximate pumping energy requirements. For the agglomeration process in the
fluid bed processor, the air velocity required is normally five to six times the minimum
fluidization velocity.
At the incipient point of fluidization, the pressure drop of the bed will be very
close to the weight of the particles divided by the cross-sectional area of the bed
(W/A). For the normal gas fluidized bed, the density of the gas is much less than
the density of the solids and the balance of forces can be shown as
DPmf ? W=A where W ? ?1  emf ?rp?g=gc?
248 Parikh and Mogavero
drop is as shown in Figure 1 (19).
 2005 by Taylor & Francis Group, LLC
where DP?pressure drop, emf?minimum fluidization void fraction, A?crosssectional
area, W?weight of the particles, rp?density of particles, g/gc?ratio of
gravitational acceleration and gravitational conversion factor.
The fundamental phenomenon of fluidization was recently studied by researchers
(20) using a small-scale fluid bed unit. The purpose of the study was to compare
experimental and computational minimum fluidizing velocities (umf) of pharmaceutical
materials using miniaturized fluid bed device. Using various materials, researchers
found that the experimental method was more capable of describing the fluidizing
behavior of pharmaceutical materials than the computational approach. Computational
models of fluidization are based on the behavior of various model particles.
Computational models do not take into account particle size and shape distributions,
cohesion and adhesion of pharmaceutical materials.
As the velocity of the gas is increased further, the bed continues to expand and
its height increases with only a slight increase in the pressure drop. As the velocity of
the gas is further increased, the bed continues to expand and its height increases,
whereas the concentration of particles per unit volume of the bed decreases. At a certain
velocity of the fluidizing medium, known as entrainment velocity, particles are
carried over by the gas. This phenomenon is called entrainment. When the volumetric
concentration of solid particles is uniform throughout the bed all the time,
the fluidization is termed as particular. When concentration of solids is not uniform
throughout the bed, and if the concentration keeps fluctuating with time, the fluidization
is called aggregative fluidization.
A slugging bed is a fluid bed in which the gas bubbles occupy the entire crosssection
of the product container and divide the bed into layers.
A boiling bed is a fluid bed, in which the gas bubbles are approximately of the
same size as the solid particles.
A channeling bed is a fluid bed, in which the gas forms channels in the bed
through which most of the air passes.
A spouting bed is a fluid bed in which the gas forms a single opening through
which some particles flow and fall on the outside.
Figure 1 Relation between the air velocity and pressure drop. (From Ref. 19.)
Batch Fluid Bed Granulation 249
 2005 by Taylor & Francis Group, LLC
Figure 2 shows various types of fluid beds (21).
The mechanisms by which air affects fluidization have been discussed by various
researchers (13,2226). When the fluidizing velocity is greater than the incipient
velocity, bubbles of air rise through the bed causing mixing of particles. Mixing does
not generally occur when the bed is fluidized at very low or zero excess gas velocities,
because insufficient bubbles are formed to cause bulk displacement of particles. It is
the gas passing through the bed in the form of bubbles that determines the degree of
mixing. The extent of mixing appears to vary with the particle size. Mixing of particles
having a mean particle size of less than approximately 150 mm decreases as their
mean size approaches zero. Different types of beds, described above, are formed
depending on the movement of bubbles through the bed. The pattern of movement
of the gas phase in and out of bubbles depends on several factors, including minimum
fluidization velocity and particle size. These movements affect heat transfer
between air bubbles and particles. The air distributor at the bottom of the container
has a controlling influence on the uniform distribution of gas, minimization of dead
areas, and maximization of particle movement. The most common reason for mixing
problems, such as segregation in the fluid bed, is the particle density differences. The
main characteristic of the fluid bed is the relative velocity imparted to the particles,
U0, which is a strong function of the size of the particles and the gas velocity in the
bed, and was shown to be given by (27)
U0  aY0 ? 18Uba=Dbd2
Figure 2 Various types of fluid beds. (From Ref. 21.)
250 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
where a is the average particle size, Ub is the bubble velocity, Db is the bubble diameter,
and d is the dimensionless bubble spacing. The first expression on the righthand
side of equation applies to fluidized beds with no rotating parts, where shear
is induced by the motion of bubbles only.
The extent of segregation can be controlled in part by maintaining high fluidizing
velocities and high bowl heightbowl diameter ratio. There are standard air velocities
for various processes that can be used as guidelines. The standard velocities are
based on the cross-sectional area at the bottom of the product container.
This is calculated by using the following formula for calculating the air velocity.
Velocity ?m=sec? ? Airflow ?m3=hr?=Area ?m2?  3600
where airflow in cubic meters per hour (CMH)?airflow (CFM)1.696.
Standard air velocities are based on the application. Low air velocities such as
0.81.4 m/sec are required for drying. The velocities are higher during the early
stages of drying because of the wet mass present in the bowl, but will be reduced
when the product loses its moisture. The objective is to have good particle movement,
but to keep the material out of filters. Particle movement and quick drying
are important during the agglomeration process. Airflow velocities are normally
1.02.0 m/sec.
An indication of good fluidization is a free downward flow of the granulation
at the sight glass of the drying container. However, improper fluidization can also be
detected by monitoring the outlet air temperature. Every product has a unique constant
rate of drying in which the bed temperature remains relatively constant for a
significant length of time. Therefore, if the outlet temperature rises more rapidly than
anticipated, it will indicate an improper fluidization and the process may have to be
stopped, and manual or mechanical intervention may be required to assist the
A fluid bed processor is a system of unit operations involving conditioning of process
air, a system to direct it through the material to be processed, and have the same air
cal fluid bed processor with all the components. These components and their utility
in granulation will be reviewed.
An exhaust blower or fan is situated at the downstream end of the fluid bed
processor to draw the air through the entire unit. This arrangement provides negative
pressure in the fluid bed, which is necessary to facilitate material loading, maintain
safe operation, prevent material escape, and carry out the process under good
manufacturing practice guidelines, all of which will be discussed later in the chapter.
3.1. Air Handling Unit
A typical air preparation system includes sections for prefiltering air, air heating, air
dehumidification, rehumidification, and final high-efficiency particulate air (HEPA)
filtering. Generally, outside air is used as the fluidizing medium in a fluid bed processor.
For the air to be used for pharmaceutical products, it must be free from dust and
contaminants. This is achieved by placing coarse dust filters (3085%) in the air
Batch Fluid Bed Granulation 251
(usually laden with moisture) exit the unit void of the product. Figure 3 shows a typihandling
unit (AHU). Figure 4 shows a typical AHU.
 2005 by Taylor & Francis Group, LLC
Figure 3 Fluid bed processor installation with all components.
Figure 4 Typical air handling unit.
252 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
Through years of experience and dealing with various types of materials and
various climatic conditions, it is known that incoming air must be controlled very
closely. As an example, it has been found that the humidity of the incoming air
can greatly affect the quality of spray granulation, drying, or coating. Therefore,
air preparation systems are now designed to better control the conditions of the
incoming air. After the installation of the filters, distinct heating or cooling sections
are installed in the air handler, depending on the geographical location of the plant.
In an extremely cold climate, where cooling coils (needed in summer months for
maintaining uniform dew point) can freeze in winter, a preheating section is placed
ahead of the cooling coils. A typical range for the air after pretreatment that one
should aim at achieving is 1530C dry bulb and 35C wet bulb. If the unit is
located in a tropical or humid climate, the humidity removal section is employed
first. The dehumidification of the air is extremely important where the outside air
moisture varies over a wide range. In summer, when the outside humidity is high,
dehumidification of the process air is required to maintain a specific dew point of
the incoming process air. Rehumidification may be necessary during the winter
months in some regions. A steam injector is used for rehumidifying the dry air. Generally,
lower the process air dew point, higher the affinity to entrain moisture and
shorter the process time. When granulating extremely fine powders, inlet air dew
point of 15C is beneficial to reduce static charges and facilitate uniform fluidization.
In many processes, when preheating is required, a bypass loop can be used for preconditioning
the air. This loop allows the required process temperature and humidity
to be attained within the system ducts before the product is subjected to fluidization.
After the conditioned air leaves the humidification/dehumidification section of the
AHU, it is finally heated to the desired process air temperature and then passed
through an HEPA filter of about 99.9099.99% capacity. As the process air is treated
and filtered, it is transported by the inlet duct. The air is thus brought into the
process vessel in the lower plenum.
3.2. Product Container and Air Distributor
With the air at the desired humidity and temperature, it is ready to be passed through
The air must be introduced evenly at the bottom of the product container,
through an inlet air plenum. Proper airflow in the inlet air plenum is critical to ensure
that equal airflow velocities occur at every point on the air distributor plate. If the air is
not properly distributed before it reaches the bottom of the container, uneven fluidization
can occur. To facilitate the even flow of powder in the product container, conditioned
air is brought in the plenum at various locations by certain manufacturers.
To properly fluidize and mix the material in the container, correct choice of the
container and air distributor must be made. The container volume should be chosen,
such that the container is filled to at least 3540% of its total volume and not more
than 90% of its total volume. Correct choice of the air distributor is important. These
distributors are made of stainless steel and are available with a 230% open area.
Typically, the distributor should be chosen such that the pressure drop across the
product bed and air distributor is 200300mm of water column. A fine screen of
60325 mesh normally covers the air distributor and retains the product in the
container. This type of sandwiched construction has been used for the last 30 years
in the fluid bed processors. The classic air distributor with the fine product retaining
screen is shown in Figure 5.
Batch Fluid Bed Granulation 253
the bed of solids. Figure 5 shows a typical product container with the air distributor.
 2005 by Taylor & Francis Group, LLC
Keeping the screen and air distributors clean has been challenging. Partially to
address the cleaning problems and partially to provide an efficient processing, a overlap
gill air distributors eliminate the need for a fine screen and perform dual functions as
efficient air distributors and product retainers. Other advantages claimed by the manufacturer
are validatable clean in place (CIP), controlled fluidization, and directional
flow of air to discharge the processed product from the container. Because there is no
fine screen, these types of air distributors sometimes sift very fine particles through
the plate, thus losing part of the batch in the plenum. This sifting of fine powder through
these types of air distributors is of concern when a container containing product is
moved around on the production floor, losing some product due to movement of the
container. Before selecting these types of single-plate distributors, product particle size
and a sifting test should be performed. But these types of air distributors offer advantages
when the airflow can be directed to discharge the granulated product from the
3.3. Spray Nozzle
A spray is a zone of liquid drops in a gas, and spraying is the act of breaking up a
liquid into a multitude of these droplets. The general purpose of spraying is to
increase the surface area of a given mass of liquid, in order to disperse it over the
product area. The primary concern is with the increase of surface area per unit mass
achieved by spraying. The nozzle is an orifice through which liquid is forced,
normally by compressed air. This is done by one of three general methods: (a) liquid
may be sucked up by a pressure drop created over the nozzle cap, after which
compressed air atomizes the liquid stream by disintegrating it with air jets (b) the
compressed air operates a piston arrangement that pushes the liquid through the
orifice and then lets surface tension create droplets or (c) two pressure streams of
liquid impinge upon each other, and so form a highly dispersed, uniform spray.
Figure 5 Product container with air distributor.
254 Parikh and Mogavero
gill plate, shown in Figure 6(A) and (B), was introduced in 1990 (28). These new overlap
 2005 by Taylor & Francis Group, LLC
container in a contained manner. (Please see the material handling options section.)
The type of spray system is usually characterized by one of four nozzle designs
1. Pressure nozzle: The fluid under pressure is broken up by its inherent instability
and its impact on the atmosphere, on another jet, or on a fixed plate.
2. Rotating nozzle (rotary atomizer): Fluid is fed at a low pressure to the center
of a rapidly rotating disk, and the centrifugal force breaks up the fluid.
These types of nozzles are used mainly in a spray drying application.
3. Airless spray nozzle: The fluid is separated into two streams that are
brought back together at the nozzle orifice, where upon impingement, they
form drops.
4. Gas atomizing nozzle (two-fluid nozzle): The two-fluid (binary) nozzle
where the binder solution (one fluid) is atomized by compressed air (second
C). These nozzles are available as single-port or multiport designs. Generally,
the single-port nozzles are adequate up to the 100 kg batch, but for
the spray undergoes three distinct phases. In the first, the compressed air
Figure 6 (A) Schematic of the overlap gill plategill arrangements. (B) Container with the
overlap gill distributor. (Courtesy of Niro Pharma Systems.)
Batch Fluid Bed Granulation 255
(29). Figure 7 shows four types of nozzles.
fluid) is the most commonly used nozzle for fluid bed granulation (Fig. 8A
larger-size batches multiport nozzles, such as either three-port (Fig. 9) or
six-port (Fig. 10) nozzles are required. When these nozzles are air atomized,
 2005 by Taylor & Francis Group, LLC
(gas) expands, essentially adiabatically, from the high pressure at the nozzle
to that at the fluid bed chamber. The gas undergoes a JouleThomson
effect, and its temperature falls. In the second, the liquid forms into discrete
drops. During this atomization, the liquids specific surface area usually
increases 1000 times. In the third, the drops travel after being formed, until
they become completely dry or impinge on the product particles. During
Figure 7 Types of nozzle. (From Ref. 29.)
Figure 8 (A) Schematic nozzle showing different parts. (B) Schematic of two-fluid nozzle.
(C) Typical single-port nozzle. (Courtesy of Niro Pharma Systems.)
256 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
this phase, the solvent evaporates and the diameter of the drops decreases.
The energy required to form a drop is the product of the surface tension and
the new surface area. About 0.1 cal/g energy is needed to subdivide 1 g of
water into 1 mm droplets. The air pressure required to atomize the binder
liquid is set by means of a pressure regulator. The spray pattern and spray
angle are adjusted by adjusting the air cap.
Optimum atomization is achieved by fine adjustment of the air cap and atomization
air pressure measured at the nozzle. The binder solution is delivered to the
nozzle port through a spray lance and tubing. The peristaltic or positive displacement
pump is commonly used to pump the binder solution. The pneumatically controlled
nozzle needle prevents the binder liquid from dripping when the liquid flow is
stopped. Nozzle port openings of between 0.8 and 2.8mm in diameter are most common
and are interchangeable.
The two-fluid nozzle in its simplified model is based on energy transmission as
shown below:
Energy ? Liquid!Two fluid nozzle ! Droplets ? Heat
The ratio of energy dissipation by heat and by the droplet-making process is
difficult to measure. Masters (30) suggested that <0.5% of applied energy is utilized
in liquid breakup. Virtually, the whole amount is imparted to the liquid and air as
kinetic energy.
Figure 9 Three-port nozzle. (Courtesy of the Vector Corporation.)
Figure 10 Six-port nozzle. (Courtesy of the Glatt Group.)
Batch Fluid Bed Granulation 257
 2005 by Taylor & Francis Group, LLC
3.4. Disengagement Area and Process Filters
Once the air leaves the product bed, fine particles need to be separated from the air
stream. Two zones are used in the fluid bed to separate the particles from the air
streamthe disengagement area and the exhaust filter. In the disengagement area,
larger particles lose momentum and fall back into the bed. The velocity of the
process air is the highest at the center of the processor and approaches zero at
the sidewalls. A process-air filter system removes the particles from the exhaust
air. The process air is filtered by using bags or cartridges. The bag filters are widely
used and are available with single-bag or with double-bag configuration, where one
bag is mechanically shaking the particles while the other bag remains functional,
thus facilitating uninterrupted fluidization. This alternate shaking of dual bags
allows the process to be consistent from batch to batch. These filter bags can be constructed
out of nylon, polyester, polypropylene, and polytetrafluroethylene (PTFE)
lined materials (Figs. To dissipate the potential static charges from the
product particles, conductive fabrics are also available and are recommended.
Figure 11 Conventional bag filters. (Courtesy of the Glatt Group.)
Figure 12 Conventional bag filters with hanging arrangements. (Courtesy of the Vector
258 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
Cartridge filters lined with PTFE were introduced to the industry in the 1980s (31).
The standard filtration system normally contains a multiple-cartridge filter system
with an alternating blowback pulse arrangement allowing continuous product fluidization.
A cleanable polyester 2 mm material is utilized for processing water-soluble
and -insoluble materials, which has an electrical conductivity for static-free operation.
Recently, cartridges made of stainless steel suitable for CIP have been introduced
(32). Various suppliers of the process equipment have filter arrangements.
The vertical filter cartridge claimed to provide better cleaning, however, requires
mechanical means to bring the filters down to replace them. Cartridge filters located
at an angle do provide better access to take them out from and place them in the
unit. They are equally effective. and (B) shows the different cartridge-
filter arrangements in the fluid bed processor. The stainless steel cartridge filpossibility
of cleaning, using the automated CIP system. For a potent compound
processing these cartridge filters with a CIP capability is normally recommended.
During the granulation or drying process, cloth filters are mechanically shaken
to dislodge any product adhered, while cartridge filters use low-pressure compressed
air blowback system to do the same. Figure 14(A) and (B) shows various PTFE lined
cartridge filters used in the fluid bed processors.
3.5. Exhaust Blower or Fan
Once the air leaves the exhaust filters, it travels to the fan. The fan is on the outlet
side of the system, which keeps the system at a lower pressure than the surrounding
atmosphere. The airflow is controlled by a valve or damper installed just ahead or
after the fan. Manufacturers of the fluid bed normally make the selection of the fan,
Figure 13 Production size unit with filter bags. (Courtesy of the L.B. Bohle Group.)
Batch Fluid Bed Granulation 259
Figure 14(A)
ters (Fig. 15) are an expensive alternative to the cloth filter bags, but provide the
 2005 by Taylor & Francis Group, LLC
based on the layout and the complexity of the system. Fan size is determined by calculating
the pressure drop (DP), created by all the components that make up the fluid
bed processor, including the product at the highest design airflow volume.
3.6. Control System
A fluid bed granulation process can be controlled by pneumatic analog control
devices, or state of the art, programmable logic controllers (PLCs) or computers.
The electronic based control system offers not only reproducible batches according
to the recipe but also a complete record and printout of all the process conditions.
Process control technology has changed very rapidly and it will continue to change
as advances in computer technology take place and as the cost of control systems
fall. The CFR Part 11 requirements (33) mandated by the US FDA has created a
number of approaches to ensure that these control systems are complying with the
Figure 14 (A) Cartridge filter during processing mode. (B) Cartridge filters during cleaning
mode. (Courtesy of the Vector Corporation.)
260 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
current regulation. The concept of process analytical technology (PAT) is covered in
another chapter of this book, as it pertains to the process control, and is discussed
further in this chapter, as it pertains to fluid bed processing in particular.
3.7. Solution Delivery System
The liquid delivery systems operate at a low pressure. A peristaltic pump capable of
delivering fluid at a controlled rate is desirable. The liquid is transported from the
solution vessel through the tubing, and atomized using a two-fluid (binary) nozzle
in the fluid bed processor.
3.8. Laboratory Units
During feasibility of process development, the supply of active ingredient is scarce.
A smaller unit with all the functionalities of a larger unit is desirable. The process
parameters developed at this stage may not be fully scalable; however, one can proused
for process feasibility and preliminary development work.
Agglomeration can be defined as the size enlargement process, in which the starting
material is fine particles and the final product is an aggregate in which primary particles
can still be identified. The granules are held together with bonds formed by the
binder used to agglomerate. Various mechanisms of granule formation have been
described in the literature (3436). The chapter on the theory of granulation in this
book discusses the theory of granule growth. To summarize, the researchers have
suggested three mechanisms for granule formation. These are
1. Bridges due to immobile liquids form adhesional and cohesional bridging
bonds. Thin adsorption layers are immobile and can contribute to the
bonding of fine particles under certain circumstances.
2. Mobile liquids where interfacial and capillary forces are present.
3. Solid bridges formed due to crystallization of dissolved substances during
Figure 15 Stainless steel filters. (Courtesy of the Glatt Group.)
Batch Fluid Bed Granulation 261
cess 20100 g of material. Figures 1618 show various smaller-size units that can be
 2005 by Taylor & Francis Group, LLC
The type of bonds formed approaches through four transition states, described
by Newitt and Conway-Jones (34) as (1) pendular, (2) funicular, (3) capillary, and (4)
droplet, which normally happens during spray drying.
Tardos et al. (37) investigated comprehensive model of granulation. They
developed a pendular bridge apparatus that can be used to test the bridge-forming
characteristics of the binder, and to determine binder penetration and spreading
rates and the critical time of binder strengthening. Iveson (38) reported a mathema-
Figure 16 Tabletop unit. (Courtesy of Pro-cept.)
Figure 17 Tabletop fluid bed processor. (Courtesy of the Glatt Group.)
262 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
tical model for granule coalescence during granulation. He found that current models
had one of two limitations; either they only consider whether a bond formed on impact
is strong enough to survive subsequent impacts or they fail to consider the possibility
of bond rupture after formation at all. He developed a new model that takes into
account both the effects of bond strengthening with time and the distribution of
impact forces. He suggested that his model be combined with existing models that
predict whether or not two granules stick initially on impact, to be able to predict
the probability of permanent coalescence.
Most of the fluid bed granulated products require much less wetting than the
high-shear granulation or spray dryer processed product. In the fluid bed granulation
process, the particles are suspended in the hot air stream and the atomized liquid
is sprayed on it. The degree of bonding between these primary particles to form an
agglomerated granule depends on the binder used, the physicochemical characteristics
of the primary particles being agglomerated, and the process parameters.
Schaefer et al. (39) and Smith and Nienow (40) have reported a description of
the growth mechanisms in the fluid bed, where the bed particles are wetted by liquid
droplets in the spray zone. Atomized liquid from the nozzle tends to spread over the
particle surface, as long as there is an adequate wettability of the particle by the fluid
(41) Wet particles, on impact, form a liquid bridge and solidify as the agglomerate
circulates throughout the remainder of the bed. Solid bridges then hold particles
together. The strength of the binder determines whether these particles stay as
agglomerates. These binding forces should be larger than the breakup forces and
in turn depend on the size of the solid bridge. The breakup forces arise from the
movement of the randomized particles colliding with each other and are related to
the excess gas velocity and particle size.
If the binding forces are in excess of the breakup forces, either in the wet state or
in the dry state, uncontrolled growth will proceed to an overwetted bed or production
of excessive fines, respectively. If a more reasonable balance of forces is present, con-
Figure 18 Lab unit. (Courtesy of Heinen, Germany.)
Batch Fluid Bed Granulation 263
 2005 by Taylor & Francis Group, LLC
trolled agglomeration will occur, the growth of which can be controlled. Maroglou
and Nienow presented a granule growth mechanism in the fluid bed by the use of
model materials and scanning electron microscope (42). Figure 19 shows the various
paths a liquid droplet can take and their consequences on the particle growth.
The mechanism of formation of a granule and subsequent growth primarily
progresses through three stages:
1. Nucleation
2. Transition
3. Ball growth
beginning of the spraying stage, primary particles form nuclei and are held together
by liquid bridges in a pendular state. The size of these nuclei depends on the droplet
size of the binder solution. As the liquid addition continues, more and more nuclei
Figure 19 Mechanism of granulation in fluid bed.
264 Parikh and Mogavero
Figure 20 shows the growth of the granule relative to the liquid added. In the
 2005 by Taylor & Francis Group, LLC
agglomerate and continue the transition from the pendular state to the capillary
The uniqueness of the fluid bed agglomeration process is in, how the liquid
addition and drying (evaporation) steps are concurrently carried out. When the granulation
liquid is sprayed into a fluidized bed, the primary particles are wetted and
form together with the binder relatively loose and very porous agglomerates. Densi-
fication of these agglomerates is brought about solely by the capillary forces present
in the liquid bridges. It is therefore important that the quantity of liquid sprayed into
the bed should be relatively large compared with that used in high-shear granulation.
Drying a wet product in a fluid bed is a separate topic, but during the granulation
process it becomes an integral part of the process; hence, understanding fluid
bed drying is important before we review the agglomeration process.
Drying is usually understood to be the removal of moisture or solvent. Drying
involves heat transfer and mass transfer. Heat is transferred to the product to evaporate
liquid, and mass is transferred as vapor in the surrounding gas; hence, these
two phenomena are interdependent. The drying rate is determined by the factors
affecting the heat and mass transfer. The transfer of heat in the fluid bed takes place
by convection. Convection is the transfer of heat from one point to another within a
fluid (gas, solid, liquid), by the mixing of one portion of the fluid with another.
The removal of moisture from a product granulated in the fluid bed granulator or
in other equipment essentially removes the added water or solvent. This free moisture
content is the amount of moisture that can be removed from the material by drying at
a specified temperature and humidity. The amount of moisture that remains
associated with the material under the drying conditions specified is called the
equilibrium moisture content (EMC).
Figure 20 States of liquid saturation. Liquid bridging state of agglomerates undergoing
(A) binding liquid addition and (B) densification.
Batch Fluid Bed Granulation 265
 2005 by Taylor & Francis Group, LLC
The evaporation rate of liquid film surrounding the granule being dried is
related to the rate of heat transfer by the equation:
dt ? h
where dw/dt is the mass transfer rate (drying rate), h is the heat transfer coefficient,
A is the surface area, H is the latent heat of evaporation, and dT is the temperature
difference between the air and the material surface.
Because fluid bed processing involves drying of a product in suspended hot air,
the heat transfer is extremely rapid. In a properly fluidized processor, the product
temperature and the exhaust air temperatures should reach equilibrium. Improper
air distribution and hence poor heat transfer in a fluidized bed causes numerous problems
such as caking, channeling, or sticking. The capacity of the air (gas) stream to
absorb and carry away moisture determines the drying rate and establishes the
duration of the drying cycle. Controlling this capacity is the key to controlling the
drying process. The two elements essential for this control are inlet air temperature
and airflow. The higher the temperature of the drying air, the greater its vapor
holding capacity. Since the temperature of the wet granules in a hot gas depends
on the rate of evaporation, the key to analyzing the drying process is psychrometry
Psychrometry is defined as the study of the relationships between the material
plify the crucial calculations of how much heat must be added and how much moisture
can be added to the air. The process of drying involves both heat and mass
transfer. For drying to occur, there must be a concentration gradient, which must
exist between the moist granule and the surrounding environment. As in heat transfer,
the maximum rate of mass transfer that occurs during drying is proportional to
the surface area, the turbulence of the drying air, the driving force between the solid
and the air, and the drying rate. Because the heat of vaporization must be supplied to
evaporate the moisture, the driving force for mass transfer is the same driving force
required for heat transfer, which is the temperature difference between the air and
the solid.
Scha?fer and Worts (47) have shown that the higher the temperature difference
between the incoming air and the product, the faster the drying rate. Therefore, product
temperature should be monitored closely to control the fluidized bed drying
During fluid bed drying, the product passes through three distinct temperature
the ambient temperature to approximately the wet-bulb temperature of the air in the
dryer. This temperature is maintained until the granule moisture content is reduced
to the critical level. At this point, the material holds no free surface water, and the
temperature starts to rise further.
The drying capacity of the air depends on the relative humidity (RH) of the
incoming air. At 100% RH, the air is holding the maximum amount of water possible
at a given temperature, but if the temperature of the air is raised, the RH drops, and
the air can hold more moisture. If air is saturated with water vapor at a given temperature,
a drop in the temperature will force the air mass to relinquish some of its
moisture through condensation. The temperature at which moisture condenses is the
dew point temperature. Thus, the drying capacity of the air varies significantly during
processing. By dehumidifying the air to a preset dew point, incoming air can be
266 Parikh and Mogavero
and energy balances of water vaporair mixture. Psychometric charts (Fig. 21) simphases
(Fig. 22). At the beginning of the drying process, the material heats up from
 2005 by Taylor & Francis Group, LLC
maintained at a constant drying capacity (dew point) and hence provide reproducible
process times.
Julia Gao et al. (48) studied the importance of inlet air velocity to dry the
product granulated in a high-shear granulator and dried in a fluid bed dryer. The
manufacturing process involved granulating the dry components containing 63%
water-insoluble, low-density drug in a high-shear granulator, milling the wet mass,
and drying in a fluid bed dryer. The granules were dried at an inlet air temperature
of 60C. Two different air velocities were examined for their effect on drying uniformity
of the product. The authors observed that the excessive velocity, indicated by
Figure 21 Psychrometry chart.
Figure 22 Product temperature changes during drying. (From Ref. 21.).
Batch Fluid Bed Granulation 267
 2005 by Taylor & Francis Group, LLC
the rapid rise in the exhaust air temperature resulted in nonuniform drying of the
product besides resulting in an inefficient process.
6.1. Process
As with any granulating system, in fluid bed granulation processing, the goal is to
form agglomerated particles through the use of binder bridges between the particles.
To achieve a good granulation, the particles must be uniformly mixed, and the liquid
bridges between the particles must be strong and easy to dry. Therefore, this system
is sensitive to the particle movement of the product in the unit, the addition of the
liquid binder, and the drying capacity of the air. The granulation process in the fluid
bed requires a binary nozzle, a solution delivery system, and compressed air to atomize
the liquid binder. Figure 23 shows the equipment setup for granulation using
the fluid bed processor.
Thurn (49), in a 1970 thesis, investigated the details of the mixing, agglomerating,
and the drying operations, which take place in the fluid bed process. Results
indicated that the mixing stage was particularly influenced by the airflow rate and the
air volume. It was suggested that the physical properties of the raw materials, such as
hydrophobicity, might exert a strong influence on the mixing stage. At the granulation
stage, particular attention was paid to the nozzle, and it was concluded that a
Figure 23 A typical fluid bed processor setup for fluid bed granulation.
268 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
binary design (two-fluid) nozzle gave a wide droplet size distribution, yielding a
homogenous granule. The need for strong binders was recommended to aid granule
formation and it was suggested that the wettability of the raw materials required
particular attention. Several research papers have been published, on the influence
of raw material (4765), binder type (5,8,47,49,59,61,6676), binder concentration,
and binder quantity (8,52,57,61,63,66,6870,7375,7793). Recently, researchers
granulated various products using foam instead of solution and found that fluid
bed granulation using foams of aqueous solutions of low molecular weight Methocel
hypromellose polymers (E3PLV and E6PLV) or conventional solution did not have
an effect on the physical properties of granules or tablets compressed from these
granules. However, they found that due to foam, the granule formation is achieved
more efficiently. It was further claimed that variables associated with nozzles were
eliminated by using foam, and water requirement was reduced along with shorter
production time (94).
Each phase of the granulation process must be controlled carefully to achieve
process reproducibility. When the binder liquid is sprayed into a fluidized bed, the
primary particles are wetted and form together with the binder, relatively loose and
very porous agglomerates. Densification of these agglomerates is brought about,
almost solely by the capillary forces present in the liquid bridges. It is therefore
important that the liquid binder sprayed into the bed should be relatively large in
quantity, compared with that used in high- or low-shear granulation processes. During
spraying, a portion of the liquid is immediately lost by evaporation, so the system
has little tendency to pass beyond the liquid bridge phase. The particle size of the
resulting granule can be controlled to some extent by adjusting the quantity of binder
liquid and the rate at which it is fed, i.e., the droplet size. The mechanical
strength of the particles depends principally on the composition of the primary product
being granulated and the type of the binder used. Aulton et al. (83) found that
lower fluidizing air temperature, a dilute solution of binder fluid, and a greater spray
rate produced better granulation for tableting.
6.2. Variables
Factors affecting the fluid bed granulation process can be divided into three broad
1. Formulation related variables
2. Equipment related variables
3. Process related variables
6.2.1. Formulation Related Variables Properties of Primary Material. Ideally, the particle properties
desired in the starting material include a low particle density, a small particle size,
a narrow particle size range, the particle shape approaching spherical, a lack of particle
cohesiveness, and a lack of stickiness during the processing. Properties such as
cohesiveness, static charge, particle size distribution, crystalline or amorphous nature,
and wettability are some of the properties which have an impact on the properties
of granules formed. The cohesiveness and static charges on particles present
fluidization difficulty. The same difficulties were observed when the formulation contained
hydrophobic material or a mixture of hydrophilic and hydrophobic materials.
The influence of hydrophobicity of primary particles has been shown by Aulton and
Batch Fluid Bed Granulation 269
 2005 by Taylor & Francis Group, LLC
Banks (17) where they demonstrated that the mean particle size of the product was
directly related to wettability of the primary particles expressed as cos y (where y is
the contact angle of the particles). It was also reported that as the hydrophobicity of
the mix is increased, a decrease in granule growth is observed. Aulton, Banks, and
Smith, in a later publication, showed that addition of a surface-active agent, such
as sodium laurel sulfate, improves the fluidized bed granulation (57). In a mixture
containing hydrophobic and hydrophilic primary particles, granule growth of hydrophilic
materials takes place selectively, creating content uniformity problems. Formulating
a controlled release granulation can be accomplished by using fluid bed
granulation. A controlled release matrix formulation of Naproxen was successfully
developed using fluid bed granulation (95). Low-Dose Drug Content. Wan et al. (96) studied various methods of
incorporating a low-dose drug, such as chlorphenirmine maleate, in lactose formulation
with Povidone (PVP) as the granulating binder. They concluded that the randomized
movement of particles in the fluid bed might cause segregation of the drug and
that uniform drug distribution was best achieved by dissolving the drug in the granulating
solution. The mixing efficiency of drug particles with the bulk material was
found to increase, in the proportion of the granulating liquid used to dissolve the
drug. The optimum nozzle atomizing pressure was deemed to be important to avoid
spray drying the drug particles or overwetting, which creates uneven drug distribution.
Higashide et al. (97) studied the fluidized bed granulation using 5-fluorouracil
in a concentration of 0.3% in 1:1 mixture of starch and lactose. Hydroxy propyl cellulose
(HPC) was used as the binder. The ratios of starch and lactose contained in
the granules were measured gravimetrically. The researchers found that a larger
amount of the drug and starch was found in larger granules than in smaller granules.
The results were attributed to the hydrophobicity of the 5-fluorouracil, starch, and
the hydrophilicity of lactose. Binder. A more general discussion on the types of binders used in
the pharmaceutical granulations and their influence on the final granule properties
was presented in a previous chapter of this book. Different binders have different
binding properties, and the concentration of the individual binder may have to be
changed to obtain similar binding of primary particles. Thus, the type of binder, binder
content in the formulation, and concentration of the binder have a major influence
on granule properties. These properties include friability, flow, bulk density,
porosity, and size distribution.
Davies and Gloor (98,99) reported that the types of binder such as povidone,
acacia, gelatin, and HPC, all have different binding properties that affect the final
granule properties mentioned above. Hontz (91) investigated microcrystalline cellulose
concentration, inlet air temperature, binder (PVP) concentration, and binder
solution concentration effects on tablet properties. Binder and microcrystalline cellulose
concentration were found to have a significant effect on tablet properties.
Alkan and Ulusoy (76) studied binder (PVP) addition in solution and as a dry powder
in the powder mix. They found a larger mean granule size when the dry binder
was granulated with ethanol. However, when the binder was in solution, the granules
produced were less friable and more free-flowing. A similar finding was confirmed by
other researchers (92,93). Binder temperature affects the viscosity of the solution,
and in turn affects the droplet size. Increased temperature of the binder solution
reduces the viscosity of the solution, reducing the droplet size, and hence producing
smaller mean granule size. Binder solution viscosity and concentration affect the
droplet size of the binder. Polymers, starches, and high molecular weight PVP cause
270 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
increased viscosity, which in turn creates a larger droplet size and subsequently a larger
mean granule particle size (68).
Diluted binders are preferred because they facilitate finer atomization of the
binder solution, provide control of the particle size, reduce friability, and increase the
bulk density, even though the tackiness or binding strength may suffer (8,69,79,83,99). Binder Solvent. In most instances, water is used as the solvent. The
selection of solvent, such as aqueous or organic, depends on the solubility of the binder
and the compatibility of the product being granulated. Generally, organic solvents,
due to their rapid vaporization from the process, produce smaller granules
than the aqueous solution. Different solvents have different heats of vaporization
as shown in Table 1. Incorporating the binder or a mixture of binders of low melting
point with the drug substance in the dry form can eliminate the requirement of a solvent
for the binder. The temperature of the incoming air is sufficient to melt the binder
and form the granules. Seo et al. (100) studied fluid bed granulation using
meltable polymers such as polyethylene glycol (PEG) 3000, or esters of polyethylene
glycol and glycerol (Gelucire 50/13). They showed that melt agglomeration by atomization
of a melted binder in a fluid bed occurs by initial nucleation followed by
coalescence between nuclei. The nuclei are formed by immersion of the solid particles
in the binder droplets, provided that the droplet size is larger than the size of the
solid particles. The agglomerate growth rate is supposed to be practically independent
of the droplet size, if the binder viscosity is so low that the droplets are able
to spread over the agglomerate surface. If the droplets are unable to spread, because
of high viscosity, the growth rate is supposed to be inversely proportional to the droplet
size. These effects of droplet size are different from those seen in aqueous fluid
bed granulation, probably because the aqueous process is affected by the evaporation
of the binder liquid.
6.2.2. Equipment Related Variables Design. To fluidize, and thus granulate and dry the product, a certain
quantity of process air is required. The volume of the air required will vary
based on the amount of material that needs to be processed. The ratio of drying
capacity of the process air to the quantity of the product needs to be maintained constant
throughout the scaling-up process. However, some suppliers of the equipment
provide higher drying capacity for their laboratory unit, but cannot maintain the
same ratio for the production units. This lack of proportionality reduces the drying
capacity per unit volume of the process air, resulting in a longer process time in the
production units. The current design of the fluid bed is a modular one, where multiple
Table 1 Heats of Vaporization for Commonly Used Solvents
Solvent boiling point
Heat of vaporization
Methylene chloride 40.0 1.327 77
Acetone 56.2 0.790 123.5
Methanol 65.0 0.791 262.8
Ethanol 78.5 0.789 204.3
Isopropanol 82.4 0.786 175.0
Water 100.0 1.000 540.0
Batch Fluid Bed Granulation 271
 2005 by Taylor & Francis Group, LLC
processes such as drying, granulating, Wurster coating, rotary fluid bed granulating
or coating, etc., can be carried out by changing the container specially designed for
the individual process. Air Distributor Plate. The process of agglomeration and attrition
due to random fluidization requires control of the particle during the granulation
process. This is a complex phenomenon due to the prevailing fluidizing conditions
and a particle size distribution, which undergoes changes during the process. As
the conditioned air is introduced through the lower plenum of the batch fluid bed,
the fluidizing velocity of a given volume of air determines how fluidization will be
Perforated air-distributor plates covered with the 60325-mesh fine stainless
steel screen, described previously, provide an appropriate means of supplying air
to the product. These plates are identified by their percentage of open area. Air distributor
plates that have 430% open area are normally available. These interchangeable
plates provide a range of loading capacities, so that batches of various sizes can
be produced efficiently and with uniform quality. To prevent channeling, an operator
can select a plate with optimum lift properties. For example, a product with low bulk
density requires low fluidizing velocity. A distributor plate having a small open area
to give a large enough pressure drop may provide uniform fluidization of such a product
without reaching entraining velocity and impinging the process filters. Alternatively,
a product with higher bulk density can be fluidized and processed using a plate
with a larger open area. The air distributor plate consists of a perforated plate and a
fine mesh screen. This arrangement sometimes causes problems like product leakage,
due to a torn screen, and difficulty in cleaning without separating the perforated plate
and the fine mesh screen. To overcome these deficiencies, various air distributor
design have been recently introduced. These distributor designs were discussed earlier
in the chapter. Pressure Drop. The blower creates flow of air through the fluid bed
processor, or a fan located downstream from the process chamber. This fan imparts
motion and pressure to air using a paddle-wheel action. The moving air acquires a
force or pressure component in its direction of motion, because of its weight and inertia.
This force is called velocity pressure and is measured in inches or millimeters of
water column. In operating duct systems, a second pressure that is independent of
air velocity or movement is always present. Known as static pressure, it acts equally
in all directions. In exhaust systems such as fluid bed processors, a negative static pressure
exists on the inlet side of the fan. The total pressure is thus a combination of static
and velocity pressures. Blower size is determined by calculating the pressure drop (DP)
created by all the components of the fluid bed processing system. Proper selection of
the blower is essential in fluid bed design. A blower with appropriate DP will fluidize
the process material adequately. However, a blower without enough DP will not allow
proper fluidization of the product, resulting in longer process time and improper granulation.
A similar effect can be seen when a product with unusually high bulk density
is processed in place of normal pharmaceutical materials, or an air distributor offers
high resistance due to its construction. This creates a pressure drop that the blower
was not designed to handle. A properly sized blower or fan should develop sufficient
DP so that the exhaust damper can be used in the 3060% open position. Any additional
components such as scrubbers, exhaust HEPA, police filters, or additional components
in the air handling unit would require a larger blower/static pressure which
can be recommended by the supplier of the fluid bed processor.
272 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC Shaker/Blow Back Cycle Mechanism. To retain entrained particles
of a process material, process filters are used. To maintain these filters from building
up layers of fine process material, and causing a higher pressure drop and thus
improper fluidization, these filters are cleaned during the granulation process. When
bag filters are used, mechanical means are used to clean them. This mechanical cleaning
of the bag filters requires a cessation of airflow and thus the fluidization during
the filter cleaning process. In units with a single-bag house, this results in a momentary
dead bed, where no fluidization takes place. This interruption in the process
extends the process time. To avoid process interruptions, a multishaking filter bag
arrangement is desired, where the granulation process is continuous. Using bag filters
with a blowback or using cartridge filters where air under pressure is pulsed
through the filters also achieves the continuous process. Generally, filters should
be cleaned frequently during the granulation step, to incorporate the fines back in
the granulation. This is possible if the cleaning frequency is high and the period
between the filter cleaning is short. Rawley (101) reported the effect of bag-shake/
interval cycle. He discussed the possibility of improving particle size distribution
by optimizing the shake time and the corresponding interval between bag shakes.
The following general guidelines for filter cleaning frequency and duration are
Single-bag shaker unit: Frequency, 210 min between filter cleaning, 510 sec
for shaking. This may vary as the fine powders form granules, and the
frequency between the shakes or duration of shaking interval, can be
extended. In any case, the occurrence of a collapsed bed should be kept
at a minimum in a single-shaker unit.
Multiple bags shaker unit: Since this is a continuous process, frequency of shaking
for each section is approximately 1530 sec between filter cleanings,
and about 5 sec for shaking the filters. If a low-pressure blow back system
is used for the bags, the frequency of cleaning is about 1030 sec.
Cartridge filters: These offer continuous processing and require cleaning frequency
of 1030 sec.
The cleaning frequency and cleaning duration are now offered as an automated
system, where instead of having to base the cleaning frequency on time, the trigger
point for filter cleaning is the buildup of a pressure drop across the filters. This automates
the process and eliminates operator input. Other Miscellaneous Equipment Factors. Granulator bowl geometry
is considered to be a factor that may have an impact on the agglomeration process.
The fluidization velocity must drop from the bottom to the top rim of the bowl by
more than half to prevent smaller, lighter particles from being impinged into the filter,
creating segregation from heavier product components in the bowl. Generally,
the conical shape of the container and expansion chamber is preferred, where the
ratio of cross-sectional diameter of the distributor plate to the top of the vessel is 1:2.
Most of the suppliers of this equipment offer units with a multiprocessor concept
where a single unit can be used for drying, agglomerating, air suspension coating,
or rotary fluid bed processing by changing the processing container, while the rest
of the unit is common. This approach does eliminate the concerns about the geometry
of the processor, because of the way these units are constructed.
Batch Fluid Bed Granulation 273
 2005 by Taylor & Francis Group, LLC
6.2.3. Process Related Variables
The agglomeration process is a dynamic process where a droplet is created by a two-
fluid nozzle, and deposited on the randomly fluidized particle. The binder solvent
evaporates, leaving behind the binder. Before all of the solvent is evaporated, other
randomized particles form bonds on the wet site. This process is repeated numerous
times to produce the desired agglomerated product. There are a number of process
variables that control the agglomeration. Process variables, most important to consider
are listed as follows:
1. Process inlet air temperature
2. Atomization air pressure
3. Fluidization air velocity and volume
4. Liquid spray rate
5. Nozzle position and number of spray heads
6. Product and exhaust air temperature
7. Filter porosity and cleaning frequency
8. Bowl capacity
These process parameters are interdependent and can produce the desired product
if this interdependency is understood. Inlet process air temperature is determined
by the choice of binder vehicle, whether aqueous or organic, and the heat
sensitivity of the product being agglomerated. Generally, aqueous vehicles will
enable the use of temperatures between 60C and 100C. On the other hand organic
vehicles will require the use of temperatures from 50C to below room temperature.
Higher temperatures will produce rapid evaporation of the binder solution and will
produce smaller, friable granules. On the other hand, lower temperatures will produce
larger, fluffy, and denser granules.
let air humidity during the granulation process.
The process of drying while applying spraying solution is a critical unit operation.
This mass transfer step was previously discussed. The temperature, humidity,
and volume of the process air determine the drying capacity. If the drying capacity
of the air is fixed from one batch to the next, then the spray rate can also be fixed. If
the drying capacity of the air is too high, the binder solution will have a tendency to
spray dry before it can effectively form bridges between the primary particles. If on
the other hand, the drying capacity of the air is too low, the bed moisture level will
become too high and particle growth may become uncontrollable. This will result in
unacceptable movement of the product bed.
As previously discussed, the appropriate process air volume, inlet air temperature,
and binder spray rate are critical for achieving proper and consistent particle
size distribution and granule characteristics. There are many ways to arrive at the
proper operating parameters. The following procedure was found by the authors
to be one of the ways one can set the operating parameters when granulating with
fluid bed processors.
1. Determine the proper volume of air, to achieve adequate mixing and particle
movement in the bowl. Avoid excessive volumetric airflow so as to
entrain the particles into the filters.
2. Choose an inlet air temperature that is high enough to negate weather
effects (outside air humidity or inside room conditions). The air temperature
should not be detrimental to the product being granulated. (To
274 Parikh and Mogavero
Figure 24 shows the relationship of inlet and product air temperature, and out-
 2005 by Taylor & Francis Group, LLC
achieve consistent process year round, a dehumidification/humidification
system is necessary, which provides the process air with constant dew
point, and hence constant drying capacity).
3. Achieve a binder solution spray rate that will not dry while spraying (spray
drying) and will not overwet the bed. This rate should also allow the nozzle
to atomize the binder solution to the required droplet size.
4. As stated earlier, a typical air velocity used for spray granulation is from
a first guess at determining the proper spray rate for a spray granulation
process in a fluid bed processor.
The presence of variables in the fluid bed granulation process, and their impact
on the final granulation, was summarized by Davies and Gloor (102), where they
state that the physical properties of granulation are dependent on both the individual
formulations and the various operational variables associated with the process. The
solution spray rate increase and subsequent increase in average granule size, resulted
in a less friable granulation, a higher bulk density, and a better flow property for a
lactose/corn starch granulation. Similar results were obtained by an enhanced binder
solution, decreasing nozzle air pressure, or lowering the inlet air temperature
during the granulation cycle. The position of the binary nozzle with respect to the
fluidized powders was also studied. It was concluded that by lowering the nozzle,
binder efficiency is enhanced, resulting in average granule size and a corresponding
decrease in granule friability.
Figure 24 Temperature and humidity changes during the granulation process.
Batch Fluid Bed Granulation 275
1.0 to 2.0 m/sec. Table 2, which is based on the psychometric chart, gives
 2005 by Taylor & Francis Group, LLC
The significant process parameters and their effect on the granule properties
Maroglou (43) listed various parameters affecting the type and rate of growth
meters and the material parameters on the product.
The agglomeration process is a batch process, and accurate repeatable control of all
critical process parameters is necessary for a robust system. Earlier designs of the
fluid bed processor used pneumatic control, which provided safe operation in a
hazardous area but relied heavily on human actions to achieve repeatable product
quality and accurate data acquisition. Current designs use PLCs and personal computers
(PCs) to achieve sophisticated control and data acquisition. The operating
conditions are controlled to satisfy parameters of multiple user-configured recipes,
and critical data are collected at selected time intervals for inclusion in an end-ofbatch
report. Security levels protect access to all user-configured data with passwords
permitting access only to selected functions. With the appropriate security
level, not only are operating conditions configured, but also identification of each
Table 2 Calculation of Fluid Bed Spray Rate
Given process data
Air volume range:
Minimum (1.2 m/sec) _________________ m3/hr
Maximum (1.8 m/sec) _________________ m3/hr
Inlet air temperature and humidity to be used: _______ C _______ %RH
% Solids in sprayed solution: _________________ % solids
From psychrometry chart
Air density at point where air volume is measured: ___________ m3/kg air
Inlet air absolute humidity (H): ___________ g H2O/kg air
Maximum outlet air absolute humidity (H): ___________ g H2O/kg air
(Follow line of constant adiabatic conditions)
Use 100% outlet RH for spray granulator or 3060% RH (as required for column coating)
Calculations for spray rate
Step 1 Convert air volumetric rate to air mass rate
Minimum ________ m3/hr(60 ________ m3/kg air)? ________ kg air/min
Maximum _________ m3/hr(60 ________ m3/kg air)? ________ kg air/min
Step 2 Subtract inlet air humidity from outlet air humidity
___________________ (g H2O/kg air) Hout___________________ (g H2O/kg air) Hin
? ____________________________________________________ g H2O removed/kg air
Step 3 Calculate (minimum and maximum) spray rate of solution
This will provide range of generally acceptable spray rates based on the airflow used
in the unit
Step 1 (minimum) _________ step 2 ________ [1( ________ % solids100)]
? _______________________ spray rate (g/min) at minimum airflow
Step 2 (maximum) ________ step 2 ________ [1( ________ % solids100)]
? _______________________ spray rate (g/min) at minimum airflow
276 Parikh and Mogavero
are summarized in Table 3.
in batch fluidized granulation (Table 4) and showed the influence of process para-
 2005 by Taylor & Francis Group, LLC
valid recipe and operator is entered. The identification is verified before any operator
actions are permitted and is included with the end-of-run report. The use of computer
related hardware requires some additional validation, but with coordination
between the control system provider and the end user, the validation of software
(B) shows a PLC-based control panel with a typical operator screen.
The most important sensors for control of the drying process are product, inlet
and exhaust air temperature, and sensor for airflow measurement, located in the air
transport system. Other sensors for the spray agglomeration process are atomization
air pressure and volume, pressure drops (across the inlet filter; the product container
with the product being processed, and outlet process air filter), inlet air humidity or
dew point, process filter cleaning frequency and duration, spray rate for the binder
solution, and total process time.
All of these sensors provide constant feedback information to the computer.
These electronic signals may then be stored in the computers memory and then
recalled as a batch report. With this ability to recall data analysis, a greater insight
can be gained into the process.
Table 3 Significant Variables and Their Impact on the Fluid Bed Granulation Process
Process parameter Impact on process References
Inlet air temperature Higher inlet temperature produces finer
granules and lower temperature produces
larger stronger granules
Humidity Increase in air humidity causes larger
granule size, longer drying times
Fluidizing airflow Proper airflow should fluidize the bed
without clogging the filters. Higher airflow
will cause attrition and rapid evaporation,
generating smaller granules and fines
Nozzle and position A binary nozzle produces the finest droplets
and is preferred. The size of the orifice has
an insignificant effect, except when binder
suspensions are to be sprayed. Optimum
nozzle height should cover the bed surface.
Too close to the bed will wet the bed faster
producing larger granules, while too high
a position will spray dry the binder, create
finer granules, and increase granulation time.
Atomization air
volume and
Liquid is atomized by the compressed air. This
mass-to-liquid ratio must be kept constant to
control the droplet size and hence the granule size.
Higher liquid flow rate will produce larger droplet
and larger granule and the reverse will produce
smaller granules. At a given pressure an increase
in orifice size will increase droplet size
Binder spray rate Droplet size is affected by liquid flow rate, and
binder viscosity and atomizing air pressure and
volume. The finer the droplet, the smaller the
resulting average granules
Batch Fluid Bed Granulation 277
can be managed. Figure 25 shows the pneumatic control panel and Fig. 26(A) and
 2005 by Taylor & Francis Group, LLC
7.1. Advances in Process Control and Automation
The degree of the instrumentation of pharmaceutical unit operations has increased.
This instrumentation provides information on the state of the process and can be
used for both process control and research. A central part of optimizing production
is increasing the level of automation. Besides monitoring the process parameters,
a number of approaches are being developed for measuring the moisture of the
product to determine the end point of the process, and consequently the in-process
particle size analysis. A number of publications discuss the on-line moisture
measurement and process end point determination using near-infrared (NIR).
7.1.1. Near-Infrared
The nondestructive character of vibrational spectroscopic techniques, such as NIR,
makes them novel tools for in-line quality assurance (103). NIR has been widely used
for the measurement of water in various applications (28). NIR can be applied both
for quantitative analysis of water and for determining the state of water in solid
material. This gives a tool for understanding the physicochemical phenomena during
manufacture of pharmaceutical granulation.
Table 4 Influence of Operating and Material Parameters on the Granulated Product
Operating parameters
Droplet size NARa
Atomization air velocity
Surface tension
Nozzle position
Nozzle type
Bed moisture content Solution type and feed rate
Bed temperature
Fluidization velocity
Aspect ratio
Nozzle position and atomization velocity
Air distributor design
Jet grinding
Binder solution/suspension Concentration Bridge strength and size
Material parameters
Binder solution/suspension Concentration Bridge strength and size
Type of binder Molecular length and weight
Wettability Particle-solvent interaction
Surface tension
Material to be granulated Average particle size
Size distributiona
Shape and porosity
Drying characteristics
Density and density differencesb
aNAR is the ratio of air to liquid flow rates through the nozzle of a twin fluid atomizer expressed either in
mass units or in volume units (air at STP).
bEspecially important relative to elutriation and segregation.
278 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
The history of NIR dates back to the studies, by Herschel, in 1800. The modern
NIR analysis was developed in 1950 by the works of a group at United States
Department of Agriculture (USDA), headed by Karl Norris (28).
Other branches of chemical industry have also applied for various applications.
One feature of NIR is that the applications have been ahead of theoretical aspects.
This has hindered the general approval of NIR in the pharmaceutical industry. However,
the pharmacopoeias have defined some characteristics of analysis with NIR
(104,105). The NIR spectrum is just above the visible region of electromagnetic spectrum
(EMS). The NIR region covers the interval between 4000 and 12,500 cm1
(0.82.5 mm). Molecules that absorb NIR energy vibrate in two fundamental modes:
stretching and bending. Stretching is a continuous change in the interatomic distance,
along the axis between two atoms, and it occurs at lower wavelengths than
bending vibration. A bending vibration is a change in the bond angle between diatomic
molecules. A band observed at 1940nm is known to be caused by OH stretching
and bending vibrations, and is the most used analytically (106,107). It has
recently been reported, from measurements on silica gel layers, that water content
has an effect on NIR absorption at all wavelengths, even where water is absorbed
minimally (108). Usually, for a solid sample, the reflected light is the parameter measured
in NIR spectroscopy, known as diffuse reflectance. The reflected light consists
of two components: specular and diffuse. The specular, or mirror-like, component in
the boundary between two media occurs at the sample surface, and it contains little
information about the chemical composition of the substance. The NIR spectroscopy
is particularly based on the diffused component of the reflected light, and it can be
affected by particle size and shape distribution, bulk density surface characteristics,
and temperature (109,110). This portion of the Electro Magnetic Spectrum (EMS)
has, for the last 30 years, been studied and investigated in great detail as an analytical
tool for the analysis of many natural and man-made materials (109,111114).
Developing a functional automation system requires new measuring techniques;
new in-line measuring devices are needed (115119). Solidwater interactions
are one of the fundamental issues in the pharmaceutical technology. The state of
Figure 25 Pneumatic control panel. (Courtesy of the Glatt Group.)
Batch Fluid Bed Granulation 279
 2005 by Taylor & Francis Group, LLC
water in the solid material may be characterized using x-ray diffraction, microscopic
methods, thermal analysis, vibrational spectroscopy, and nuclear magnetic resonance
spectroscopy (120). Traditionally, the control of fluidized bed granulation has been
based on indirect measurements. These control methods applied utilize the properties
of process air by Schaefer and Worts (47). Frake et al. (121) demonstrated the
use of NIR for in-line analysis of the moisture content in 0.050.07mm pellets during
spray granulation in fluid bed processor. Rantanen et al. (122,123) described a
similar approach for moisture content measurement using a rationing of three to
four selected wavelengths. He and his coworkers reported that the critical part
of in-line process was the sight glass for probe positioning, which was continuously
Figure 26 (A) PLC based system. (Courtesy of Niro Pharma Systems.) (B) Production unit
with a PLC control panel. (Courtesy of the L.B. Bohle Group.)
280 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
blown with heated air. They also reported spectra baselines caused by particle size and
refractive properties of the in-line samples; they analyzed several data pretreatments
to eliminate these effects on their fixed wavelength setup. Solvents other than water
have also been evaluated for real-time quantification.
On-line measurement has also been possible enabling the monitoring of film
coating on pharmaceutical pellets, in an industrial manufacturing process. Andersson
et al. (124) conducted measurements on solid coated tablets using fiber-optic
probe positioned in the fluid bed processor. In this case, they secured a representative
sampling during processing, by using a sample collector that was emptied with compressed
air inside the processor. Va?zquez recently provided comprehensive review of
FT-NIR application in measuring fluid bed drying end (125). Rantanen et al. (126)
used NIR to monitor the moisture as well as airflow. Using in-line multichannel NIR
the multivariate process data collected were analyzed using principal component
analysis (PCA). The authors showed that robust process control and measurement
system combined with reliable historical data storage can be used for analyzing
the fluid bed granulation process. PCA modeling proved a promising tool to handle
multidimensional data that were collected, and for reduction of the dimensionality of
process data. FT-NIR spectra gave useful information for understanding the phenomenon
during granulation. Rantanen et al. (127) further studied the application
of NIR for fluid bed process analysis. The authors used NIR to study moisture measurement
combined with temperature and humidity measurements. By controlling
the water during the fluid bed granulation, the granulation process was controlled.
They concluded that the varying behavior of formulations during processing can
be identified in a real-time mode. Thus, they found that NIR spectroscopy offered
unique information of granule moisture content during all phases of granulation.
7.1.2. Other Approaches for Process Control Self-Organizing Maps. On-line process data are usually multidimensional,
and it is difficult to study with traditional trends and scatter plots. Rantanen
et al. (128) has suggested a new tool called self-organizing maps (SOM), for
dimension reduction and process state monitoring. As a batch process, granulation
traversed through a number of process states, which was visualized by SOM as a
two-dimensional map. In addition, they demonstrated how the differences between
granulation batches can be studied. At-Line Measurement. Laitinen et al. presented a paper at a recent
conference (129) proposing at-line optical techniques to study particle size. Using
a CCD camera with optics and illumination units, with stabilized collimated light
beams, authors took two images of 36 granule samples by illuminating the samples
alternatively. Two digital images, with matrices of their gray scale values, were
obtained and the differences between the two matrices were calculated. This method
provided very rapid (1 min per sample) measurement of particle size, with a very
small sample size (<0.5 g). Focused Beam Reflectance Measurement. This device uses a focused
beam of laser light that scans, in a circular path, across a particle or particle structure
passing in front of the window. Upon hitting the particle, light is scattered in all
directions. The light scattered back toward the probe is used to measure the chord
length or the length between any two points on a particle. Such devices are supplied
commercially and claim to be useful for monitoring on-line measurement of particle
size in the fluid bed granulation process.
Batch Fluid Bed Granulation 281
 2005 by Taylor & Francis Group, LLC Artificial Neural Network. Neural networks have been used by scientists
for optimizing formulations as an alternative to statistical analysis because of its
simplicity for use and potential to provide detailed information. The neural network
builds a model of the data space, which can be consulted to ask what if kinds of
questions. Recently, there has been interest in using artificial neural network (ANN)
for process control. Similar to the human brain, an ANN predicts events or information
based on learned pattern recognition. ANNs are computer systems developed to
mimic the operations of the human brain, by mathematically modeling its neurophysiological
structure (i.e., its nerve cells and the network of interconnections between
them). In ANN, the nerve cells are replaced by computational units called neurons,
and the strengths of the interconnections are represented by weights (130). This
unique arrangement can simulate some of the neurolocal processing ability of the
brain, such as learning and drawing conclusions from experience (131). Using the
process control system, quality assurance results, or energy usage data, an ANN
develops supervisory set points for the system. When ANN and process control systems
are used together, they form a product control system. Product control occurs
when a system measures defined product attributes in real time, and uses the knowledge
to adjust the control system. While process control system runs the process (i.e.,
fans, motors, heaters, etc.), the ANN controls the product (moisture level, consistency).
The fluidized bed processors process control system includes an operator interface,
sensing elements, and final control elements. The inputs in that case are inlet air
temperature, outlet air temperature, airflow rate, and energy consumption. Additional
contributing factors are the fouling coefficient of the dryer bags, the quantity of product
in the processor, and the type of product with its unique characteristics.
Watano et al. (132) described a practical method for moisture control in fluid bed
granulation by means of neural network. Wet granulation of pharmaceutical powder
was conducted using an agitation fluidized bed, and the moisture content was continuously
measured by IR moisture sensor. A neural network system for moisture control
was developed using moisture content and its changing rate as input variables, and the
moisture control characteristics were investigated by the neural network system with a
back propagation learning. Good response and stability without overshoot were
achieved by adopting the developed systems. This system also maintained favorable
stability under various operating conditions. Several researchers have published
papers, detailing the use of ANN for different applications (133135).
8.1. Regulatory
Scale-up is normally identified with an incremental increase in batch size, until a
desired level of production is obtained. In 1991, American Association of Pharmaceutical
Scientists (AAPS) with US FDA held a workshop on scale-up (136), where
several speakers presented scale-up issues from an industrial and regulatory perspective.
For example, Shangraw divided scale-up problems into two general categories:
those related to raw materials or formulation and those related to processing
equipment. He also indicated that it is essential to ascertain whether or not changes
in raw materials have occurred, before one looks at processing/equipment changes
as a source of any problem. The workshop report, as it pertains to the process
and equipment, is reproduced below:
282 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
It is generally recognized that many NDAs and ANDAs contain provision for
multiple manufacturers of the drug substance(s), and that not all drug substance
suppliers, a priori, produce equivalent material. There is then a need for material
quality control to assure the performance and reproducibility of the finished
product. Particle size and distribution, morphology, and intrinsic dissolution of
the drug substance are important considerations. Polymorphism, hygroscopicity,
surface area, wettability, density (bulk and tapped), compressibility (for dry
blending), and powder flow effects should be controlled.
Additionally, the process should be controlled by employment of a validation
protocol, which defines the critical parameters and also establishes the acceptance
criteria for the granulation or blend; which may include sieve analysis, flow,
density, uniformity, and compressibility, moisture content, etc.
In the milling, blending, granulating and or drying processes, the operating
principles of the equipment employed should be defined, and the variables determined.
The impact and mechanism of measurement on in-process variables should
be defined. Time, temperature, work input of equipment, blend/granulation
volume, and granulating rate should be determined. . . . The parameters selected
should be appropriate for the process, . . . In those cases where the manufacturing
process has been controlled and validated as specified in the foregoing discussion;
batch scale-up, changes in site of manufacture, allowance for equipment change
(where the operating principle is the same), minor formulation changes, etc.,
should be determined on the basis of the comparability of both the blend/granulation
and the final product; as assured by: a) appropriate tests; b) specifications; c)
process validation; and d) comparative accelerated stability.
Regulatory guidelines for scale-up and postapproval changes (SUPAC) IR
were released in 1995 (137) and are discussed in another chapter of this book.
8.2. Scale-Up and Equipment Design
The scale-up from the laboratory equipment to production size units is dependent on
equipment design, which may or may not have been scalable as far as its dimensional
feature or components selection is concerned. The importance of scalability is well
understood and accepted by the manufacturers of fluid bed processors. Various sizes
in their product line are logically designated and manufactured. Airflow in the fluid
bed process is a critical parameter. The design and selection of the processor is very
important for the laboratory and production unit. Because airflow is one of the components
of the drying capacity of a fluid bed system, the ratio of air volume per kilogram
or liter of the product is very critical to achieve a scale-up that is linear. The
other design feature is the cross-sectional area of the product container, and how
it has been designed throughout the various sizes that a manufacturer supplies.
The relationship between various sizes of the process containers can be utilized to
calculate the scale-up of binder spray rate, and if the cross-sectional area is designed
linearly, then the spray rate scale-up can be linear.
8.3. Scale-Up and Process Factors
The fluid bed agglomeration process is a combination of three steps, namely, dry
mixing, spray agglomeration, and drying to a desired moisture level. These process
steps are equally important. But the quality of the granules is really determined
during the spraying stage, when constant building of granules and evaporation of
Batch Fluid Bed Granulation 283
 2005 by Taylor & Francis Group, LLC
binder solvent are taking place. Granule size is directly proportional to the
bed humidity during granulation (47) and hence, control of this humidity during
scale-up is essential.
Gore et al. (138) studied the factors affecting the fluid bed process during scale
up. The authors found that processing factors that affected granule characteristics
the most, were process air temperature, height of the spray nozzle from the bed, rate
of binder addition, and the degree of atomization of the binder liquid.
The atomizing air pressure and the wetness of the bed are two of the most
important elements of fluid bed granulation. A higher atomizing air pressure yields
a finer droplet of binder solution. Therefore, granule growth (as described earlier in
this section) is affected by the atomizing air pressure. A major factor which must be
considered during the scale up of fluid bed granulation process is maintaining the
same droplet size of the binder for assuring successful scale-up. A recent study
(139) confirmed the influence of spray nozzle setup parameters on the drying capacity
of the air. The study concluded that more attention should be paid to the easily
overlooked nozzle atomizing air pressure and volume. When considering the atomizing
air pressure, attention must be paid to ensure that enough air is delivered to the nozzle
tip. This can be ensured by placing air pressure and volume measurement devices at the
nozzle. The data also show that the drying capacity of the process air influences the
final granulated particle size.
Jones (140) has suggested various process related factors that should be considered
during the scale-up of a fluid bed processing. The suggestions are listed in the
following paragraphs.
Due to the higher degree of attrition in the larger unit compared to the smaller
unit, the bulk density of the granulation from the larger fluid bed is 20% higher
than the smaller unit. He also re-emphasized the importance of keeping the bed
moisture level below the critical moisture level, to prevent the formation of larger
agglomerates. Since the higher airflow along with the temperature (drying capacity)
in a larger unit provide higher evaporation rate, one must maintain the drying capacity
in the larger unit, such that the bed temperature is similar to the smaller-unit bed
temperature. This can be accomplished, either by increased spray rate, increased air
temperature, and increased airflow, or by the combination of these variables to
obtain suitable results. Since the ratio of bed depth to the air distributor increases
with the size of the equipment, the fluidization air velocity is kept constant by
increasing the air volume. In the past, the scale-up was carried out by selecting best
guess process parameters. The recent trend is to employ the factorial and modified
factorial designs and search methods. These statistically designed experimental plans
can generate mathematical relationships between the independent variables, such as
process factors, and dependent variables, such as product properties. This approach
still requires an effective laboratory/pilot scale development program and an understanding
of the variables that affect the product properties.
In summary, when scaling up, the following processing conditions should be
similar to the pilot scale studies:
1. Fluidization velocity of the process air through the system
2. The ratio of granulation spray rate to drying capacity of fluidization air
3. Droplet size of the binder spray liquid.
284 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
Each of these values must be calculated based on the results of the operation of
the pilot size unit. Pilot size equipment studies should also be conducted in a wide
range to determine the allowable operating range for the process.
Another chapter of this book provides fluidized bed scale-up and should be
8.4. Case Study
The following case study illustrates how a product is scaled up from 15 to 150 kg in
the equipment supplied by Aeromatic, when one understands the critical process
parameters used when scaling up.
A spray granulation process was developed for a common pharmaceutical
compound. The granulation process involved spraying of a 5% w/w binder solution
onto the fluidized powder. Table 5 shows the data from the 15 kg run and the resulting
successful 150 kg run condition for a spray agglomeration process.
8.4.1. Airflow Calculations
To maintain the same fluidization velocity, the air volume in a larger unit must be
increased, based on the cross-sectional area of the product bowl. In this case, the
cross-sectional area of the base of the larger container was 0.77m2 and that of the
smaller was 0.06m2. The correct airflow should be calculated as 300(0.77/
0.06)?3850CMH. This number was further modified after considering the increase
in bed depth in a larger unit to 4000CMH.
8.4.2. Spray Rate Calculations
To maintain the same particle size, the triple-headed nozzle could spray three times
the pilot unit spray rate at a 2.5 atomization air pressure. However, this could result
in a longer process time. Another approach to maintain a similar droplet size is to
maintain the mass balance of spray rate and the atomization pressure. Thus, by
increasing the atomization pressure to 5 bar, the spray rate was increased to 800 g/
min, keeping the same droplet size, and hence obtaining granulation with the desired
8.4.3. Temperature Calculations
Finally, the required inlet temperature was recalculated based on the change in the
ratio of the air volume to the spray rate. Because the air volume was increased over
13 times but the spray rate was only increased eight times, the inlet temperature was
reduced to 50C. This adjustment in drying capacity was necessary to avoid spray
drying of the spray solution. (A three-headed nozzle used in this scale-up can be
Table 5 Scale-Up Process Parameters from a 15 to a 150 kg Batch
Process parameters 15 kg 150 kg
Airflow (m3/hr) 300 4000
Inlet air temperature (C) 55 50
Spray rate (g/min) 100 800
Nozzle air pressure (bar) 2.5 5
Container cross-sectional area of the base (m2) 0.06 0.77
Number of nozzles 1 3
Batch Fluid Bed Granulation 285
 2005 by Taylor & Francis Group, LLC
replaced by a six-headed nozzle. This would have resulted in the ability to increase
the spraying rate 13 times above the pilot size unit to match the airflow. The maintenance
of droplet size and temperature could have been achieved with the
six-headed nozzle. The end result would be reduced process time.) Figure 27 shows
the particle size distribution produced, using 15 and 150 kg units.
Recently, Matharu and Patel (141) presented a scale-up case study, where a
low-dose multiple-strength product (0.55% w/w/ active) was spray granulated
and scaled up from a pilot scale fluid bed processor up to production size equipment.
Their approach was based on matching air velocity between the two scales of operation.
The impact of droplet size was determined by varying the independent parameters.
Based on their study, authors have suggested an equation, which takes into
account, material and equipment parameters. Rambali (142) scaled up granulation
process from small (5 kg) to medium (30 kg) to large (120 kg) with an aim to obtain
the target geometric mean granule size of 400 mm. The scaling up was based on the
relative droplet size and the powder bed moisture content at the end of the spraying
cycle. Authors found that the effect of the change in relative droplet size on the granule
size was different for each fluid bed. They applied the experimental design on the
small- and medium-scale unit, and regression models for the granule size were proposed
in order to scale up the granulation process on the small to medium scale.
Using only the relative droplet size, authors were able to scale up the process to
the larger unit.
For an explosion to occur, three conditions must exist: an ignition source, a fuel, and
oxygen. With an explosion, oxygen reacts with the fuel, releasing heat and gases. If a
dust explosion occurs in free space, a fireball of considerable extent arises. If the dust
explosion occurs in a closed container, there is a sudden pressure rise that is mainly
decided by the following factors: type of dust, size of the dust, dust/oxygen ratio,
turbulence, precompression, temperature, shape of the container, and ignition
source. In a container without precompression and with an organic dust of sufficient
fineness, the pressure inside the container can rise to over 10 bar overpressure.
Figure 27 Scale-up case study and resultant particle size distribution.
286 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
The fluid bed process handles a large amount of air. This air, in the presence of
fine product dust, poses potential for an explosion. This hazard can be enhanced
when using flammable solvents. If sufficient ignition energy (static charge) is introduced,
an explosion within the processor can take place. In order to contain these
dust or flammable solvent induced explosions, fluid bed processors are normally constructed
to withstand an overpressure of 2.0 bar. In this case, the fluid beds are provided
with explosion relief flaps, to release the pressure as soon as it starts to build up
inside the processor. The explosion flaps, mounted either horizontally or vertically
(Fig. 28), are designed to vent the pressure buildup as low as 0.06 bar. The explosion
isolates the overpressure within the processor. The 2 bar vented design shows the
of the building. These panels are gasketed, and sealed so that normal fluid bed operation
is not affected. It was an accepted practice to have the production unit with
2 bar pressure shock integrity; however, the cleaning of the gasket area around the
flaps is always difficult. To avoid having the product exposed to the outside during
such an event, a suppression system is used to contain the possible overpressure front
sors located within the processor. These sensors are designed to trigger a series of fire
extinguishers (containing ammonium phosphate), as soon as a preset level (generally,
0.1 bar) of pressure is set within the processor.
Figure 28 Processor with horizontal and vertical relief.
Batch Fluid Bed Granulation 287
protection valve, shown in Figure 29, acts to cut off the airflow to the blower and
propagation of the overpressure (Fig. 30). The explosion flaps open up to the outside
from leaving the unit (Fig. 31). The suppression system consists of low-pressure sen-
 2005 by Taylor & Francis Group, LLC
With the introduction of potent and costly drug substances, the 2 bar design is
being replaced with a 10 bar design and higher, based on the specific mixtures of the
product(s) being processed. These units can withstand explosions up to 10 bar. Most
of the pharmaceutical dust explosions studied (143) show the overpressure reaching
Figure 29 Explosion protection valve.
Figure 30 Vented design (2 bar unit). (Courtesy of the Glatt Group.)
288 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
9 bar, with a Kst value (constant of explosion speed) of 200. An explosion in a 10 bar
unit is contained within the unit. A 10 bar designed unit does not require any explosion
relief panels or gaskets. This eliminates the concerns about cleaning of the gaskets
and panels. Another advantage of a 10 bar unit is that, in case of explosion, the
processor containing the potent drug substance is contained inside the unit, and exploand
(B) shows the Ventex-ESI valve for passive control of the explosion. The Ventexsion
force (pressure wave), moving ahead of the flame front, hurls the poppet forward to
the valve seat providing an airtight seal. The poppet, once seated, is locked in by a
mechanical shut-off device, which retains the seal until manually reset. The three basic
versions of the standard mechanical Ventex valve are available with a set pressure of
1.5 psiandamaximumpressure of 150 psi.TheVentex-ESI valve closes bythe explosion
shows the various types of Ventex valves installed in a fluid bed system.
In case of granulation requiring flammable solvents, process air and nozzle atomization
air are replaced by an inert gas such as nitrogen and the system is designed
as a closed cycle with a solvent recovery capability (144). A number of approaches
ods for solvent emission control systems.
The inert gas used for fluidization circulates continuously. An adjustable volume of
gas is diverted through the bypassed duct, where solvent vapors are condensed and
solvent collected. The circulating gas passes through the heat exchanger to maintain
the temperature necessary for evaporation of the solvent from the product bed. During
Figure 31 Explosion suppression system.
Batch Fluid Bed Granulation 289
sion does not pose an environmental problem as with the 2.0 bar unit. Figure 32(A)
ESI valve requires less maintenance than the active valve shown in Figure 29.Anexplopressure
wave, without external power for horizontal or vertical operation. Figure 33
Kulling and Simon (18) reported the closed loop system shown in Figure 34.
can be taken to handle solvent from the process. Table 6 summarizes various meth-
 2005 by Taylor & Francis Group, LLC
the agglomeration and subsequent drying process, the solvent load in the gas stream
does vary. The bypass valve controls the flow of the gas to the heat exchanger and
the condenser. By controlling the gas stream in this manner, the drying action is continued
until the desired level of drying is reached. Even though the cost of fluid bed
processor with the solvent recovery is generally double the cost of a regular single-pass
fluid bed processor, such a system offers effective measures for both explosion hazard
reduction and air pollution control.
In 1994 European parliament issued the ATEXdirective (145) on the approximation
of the laws of the member states, concerning equipment and protective systems
intended for use in potentially explosive atmospheres. The scope, these regulations specifically
require protective systems, intended to halt incipient explosions immediately
Figure 32 (A) Schematic of the Ventex valve. (B) Ventex valve. (Courtesy of Rico-Sicherheitechnik
AG, Switzerland.)
290 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
and to limit the effective range of explosion flames and explosion pressures. All manufacturers
of fluid bed processors in Europe must comply with this directive.
The transfer of materials to and from the fluid bed processor is an important consideration.
The loading and unloading of the processing bowl can be accomplished
either by manual mode or by automated methods.
Figure 33 Setup of fluid bed with different Ventex valves. (Courtesy of Rico-Sicherheitechnik
AG, Switzerland.)
Table 6 Comparison of Different Solvent Emission Control System
Water scrubbing Catalytic burning Carbon absorption Condensation
Open cycle Open cycle Open cycle Closed cycle with
High capital cost Low capital cost Moderate capital cost Low capital cost
High energy
Low energy Low energy Low energy
External installation External installation External installation Internal installation
Medium space
High space
Moderate space
Small space
Medium flexibility Medium flexibility Low flexibility Good flexibility
Waste treatment
CO2/H2O emission
waste treatment
Waste treatment Concentrated waste
Batch Fluid Bed Granulation 291
 2005 by Taylor & Francis Group, LLC
10.1. Loading
The contemporary method for loading the unit is by removing the product bowl
from the unit, charging the material into the bowl, and then placing the bowl back
into the unit. This loading is simple and cost-effective. Unfortunately, it has the
potential of exposing the operators to the product and contaminating the working
area. To avoid the product from having a dust and cleaning hazard, a dust collection
system should be installed to collect the dust before it spreads. A manual process also
depends on the batch size, the operators physical ability to handle the material, and
the container full of product. Furthermore, this can be time consuming, since the
material must be added to the product container, one material at a time.
The loading process can be automated and isolated to avoid worker exposure,
minimize dust generation, and reduce loading time. There are two main types of
loading systems. These systems are similar because both use the fluid beds capability
to create a vacuum inside the unit. Here, the product enters the fluid bed through a
product in-feed port, on the side of the unit. This is done by having the fan running
and the inlet air control flap set, so that minimum airflow may pass through the product
container and the outlet flap is almost fully open. Typically, when the highshear
granulated material needs to be charged into the fluid bed, this approach helps
valve is closed and the granulating process started. This transfer method uses some
amount of air to help the material move through the tube. Loading can be done
either vertically from an overhead bin, or from the ground. Less air is required
through the transfer pipe when the material is transferred vertically because gravity
is working to help the process. Vertical transfer methods do require greater available
height in the process area. Loading by this method has the advantages of limited
operator exposure to the product, allows the product to be fluidized as it enters
the processor, and reduces the loading time. The disadvantage of this type of system
is the cleaning required between different products.
Figure 34 Schematic of closed loop solvent recovery system.
292 Parikh and Mogavero
(Fig. 35). Once the material has been charged to the fluid bed, the product in-feed
 2005 by Taylor & Francis Group, LLC
10.2. Unloading
As with loading, the standard method for unloading is by removing the product
bowl from the unit. Once the bowl is removed, the operator may scoop the material
from the bowl, which is the most time consuming and impractical method, because
of its potential of exposure to the product. Alternatively, the product can be vacuum
transferred to a secondary container or unloaded by placing the product bowl into
This hydraulic device is installed in the processing area. The mobile product
container of the fluid bed processor is pushed under the cone of the bowl dumper
and coupled together by engaging the toggle locks. Subsequently, the container is
lifted hydraulically, pivoted around the lifting column, and rotated 180 for discharging.
Use of the bowl dumping device or vacuum unloading device still requires that
the product bowl be removed from the unit. There are contained and automated
methods for unloading the product, while the product bowl is still in the fluid bed
processor. The product may be unloaded either from the bottom of the product
container or from the side. Until recently, the most common contained method
is to unload the material from the bottom of the unit. This requires a ceiling
height high enough to accommodate it, or the installation becomes multistoried
There are two types of bottom discharge options, gravity or pneumatic.
tainer, which is located below the lower plenum. If the overall ceiling height limitation
prevents discharge by gravity, the gravity/pneumatic transfer combination can
be considered. The gravity discharge poses cleaning problems, since the process air
and the product discharge follow the same path; assurance of cleanliness is always of
prime concern.
The desire to limit the processing area and the development of the overlap gill
air distributor mentioned earlier in the chapter has prompted the consideration of
the side discharge as an option. The product bowl is fitted with the discharge gate,
Figure 35 Loading of fluid bed from high-shear mixer. (Courtesy of the Glatt Group.)
Batch Fluid Bed Granulation 293
bowl dumping discharge device, as shown in Figure 36(A) and (B).
Gravity discharge (Figs. 37 and 38) allows for collection of the product into conas
shown in (Fig. 39A and B).
 2005 by Taylor & Francis Group, LLC
Most of the product, being free-flowing granules, flows through the side
discharge into a container. The remainder of the product is then discharged by
manipulation of the airflow through the overlap gill air distributor. The discharged
product can be pneumatically transported to an overhead bin if the dry milling of the
granulation is desired.
The contained system for unloading the product helps to isolate the operator
from the product. The isolation feature also prevents the product from being
contaminated due to exposure to the working environment. Material handling consideration
must be thought of early in the equipment procurement process. Fluid bed
processing, whether used as an integral part of high-shear mixer/fluid bed dryer or as
a granulating equipment option, production efficiency, and eventual automation can
be enhanced by considering these loading and unloading options.
The fluid bed processor was used as an efficient way to dry a product due to suspension
of particles in the hot air stream. However, over the last 35 years, the development
in the pharmaceutical industry and the proliferation of the batch fluid bed
processing technology in other industries as food, polymer, detergent, etc., has
provided the opportunity to use the batch fluid bed processor for granulation,
coating of particles, and pelletization. The advances in the fluid bed can be attributed
to several factors. The needs of formulators, the requirements of the regulators,
and technological innovations from the manufacturers of these equipments are
Figure 36 (A) Product discharge system. (B) Bowl inverter. (Courtesy of Niro Pharma
294 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
responsible for these advances. The result of these changes provided units that are
paint free, modular, safer, in compliance with the cGMPs, and are capable of performing
various processes that were not thought of before.
11.1. Developments in Equipment and Process Applications
Parikh (146) has presented a review of all of the equipment developments. Among
various advances, the development of production units that can withstand more than
10 bar pressure shock resistance is a very significant development. These units do not
require pressure relief duct and associated cleaning problems. Units are now equipped
with the air handler that can provide designated humidity and dew point air, throughout
the year and at any geographical location. The fluid bed CIP became a reality with
the introduction of the overlap gill air distributors and the stainless steel cartridge filters
described earlier in this chapter. The coating of the particles is carried out most
Figure 37 Loading and unloading setup with bottom discharge. (Courtesy of the Vector
Batch Fluid Bed Granulation 295
frequently using Wurster column (Fig. 40A and B). The Wurster process is the most
 2005 by Taylor & Francis Group, LLC
popular method for coating particles. The Wurster based coating process does not
contain any fluid bed regions in the traditional sense, as it is a circulating fluid bed
process. Four different regions within the equipment can be identified: the upbed
region, the expansion chamber, the downbed region, and the horizontal transport
region. The coating process consists of three phases: the start-up phase, the coating
phase, and the drying and cooling phase. During the coating phase, several processes
take place simultaneously. They are atomization of the coating solution or suspension,
transport of the atomized droplets of the coating solution to the substrate, and drying
of the film. In 1993, the design was modified as in Wurster HS (147).
However, the technology has certain disadvantages, such as nozzle accessibility,
prolonged process time, minimum volume requirement, and difficulty of loading
and unloading. In 1995, to address these shortcomings of the Wurster system, the
Precision Coater incorporating modified air suspension technique was introduced
(148). It was designed to allow removal of the nozzles for cleaning, faster process
time because of the patented particle accelerator, good utilization of thermal and
kinetic energies, and scalability from a single-column to a multicolumn setup.
Recently, Glatt organization introduced the STRATOS2 system for applying sustained
release tablet coating, using bottom spray. The modification of the bottom
spray setup offers better efficiency than Wurster HS for coating tablets (149). The
Wurster setup for granulating was presented by Walter et al. (150). Authors claim
that the quality of granulation is superior to the top spray granulation.
Researchers have discussed the incorporation of microwave in the laboratory
fluid bed processor (151,152). Fluid bed process using organic solvent requires an
inert gas such as nitrogen to replace the air used for fluidization, as discussed earlier
in the chapter. It is accompanied by the solvent recovery system. In 1989, a vacuum
fluid bed system was presented by Luy et al. (153). The main feature was the generation
and sustaining of a fluidized bed under vacuum, thereby eliminating the use of
inert gas. Several advantages are claimed by the authors, such as emission reduction,
increased recovery rate of solvent, and an application for oxygen sensitive materials.
Figure 38 Bottom discharge. (Courtesy of the Glatt Group.)
296 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
11.2. Rotary Fluid Bed
The other advance of significance is the development of a rotary fluid bed for producing
denser granulation. Modules were introduced by various manufacturers, and
the technology is discussed below. The 1972 patent (154) for the rotor technology
was awarded for the equipment and coating of the granular material. The subsequent
patents (155,156) were awarded for the coating of the spherical granules.
An advantage of rotary fluid bed processing to produce granules over the conventional
top-spray granulation technique was reported by Jager and Bauer (155). In
this unit, the conventional air distributor is replaced by the rotating disk. The material
to be granulated is loaded on the rotating disk. The binder solution is added
through the atomization nozzle, located tangentially to the wall of the bowl. The centrifugal
force creates a dense, helical doughnut-shaped pattern. This type of motion
is caused by the three directional forces.
The vertical movement is caused by the gap or slit air around the rotating disk,
the gravitational force folds back the material to the center, and the centrifugal force
caused by the rotating disk pushes the material away from the center. The granulation
produced in the rotary fluid bed processor shows less porosity, compared to the
Figure 39 (A) Side discharge. (Courtesy of the Glatt Group.) (B) Side discharge. (Courtesy
of Niro Pharma Systems.)
Batch Fluid Bed Granulation 297
 2005 by Taylor & Francis Group, LLC
conventionally agglomerated product in the fluid bed processor. There are essentially
two designs of this rotary fluid bed module which are available: the single-chamber
conventional air distributor is replaced by the rotating disk, which has variable size
slit opening between the bowl wall and the rotor disk. The fluidizing air enters the
mixing zone in the bowl through the slit. The single-chamber design is manufactured
by Glatt (W.Glatt GMBH, Binzen, Germany), as Glatt Rotor Granulator, and by
Figure 40 (A) Wurster Coater schematic. (B) Wurster Coater. (Courtesy of the Vector
298 Parikh and Mogavero
design and the double-chamber design (Fig. 41). In the single-chamber design, the
 2005 by Taylor & Francis Group, LLC
Freund (Freund Industrial Co. Ltd., Tokyo, Japan), as a Spir-a-flow granulator.
The double-chamber design was patented and manufactured by Aeromatic-Fielder
AG (Aeromatic-Fielder AGnow Niro Pharma Systems, Bubendorf, Switzerland)
(158). An inner stainless steel wall, encompassing the area called the forming zone
is surrounded by an annular drying zone. The forming zone has a rotating disk.
The gap air around the rotating disk allows the free movement of the disk. The stainless
steel wall separating the forming zone and drying zone can be raised, so that the
rotating product transfers in the drying zone through the gap created during the drying
mode between the stationary and movable walls. The process air fluidizes the product
through the annular drying zone. As the partially dried product reaches the upper
boundary of the forming zone, the particle velocity slows down due to the design of
the unit and the particles fall back in the forming zone. These partially dried particles,
eventually, go through the same path in a cyclic pattern, until the desired moisture content
is reached. The rotary fluid bed, with its different designs listed above, do provide
granules which have less porosity, and higher bulk density compared to the granulation
produced by the typical top spray granulated fluid bed process.
Tu?rkoglu et al. produced Theophylline granulation using a rotary fluid bed
(159). The formulation contained lactose, starch, and microcrystalline cellulose,
Figure 41 Various rotary fluid bed inserts.
Batch Fluid Bed Granulation 299
 2005 by Taylor & Francis Group, LLC
along with Theophylline. They reported that the granules produced were spherical
and dense. Three different drug level formulations were evaluated. The authors concluded
that the rotary fluid bed, as a wet granulator, has the potential to obtain a
better drug content uniformity for tablets even at low active levels such as 1% in
comparison with conventional fluidized beds. The use of rotary fluid bed to produce
spherical granules for modified release application is reported by a number of
authors. Rotary fluid bed technology was reviewed by Li et al. (160), and its usefulness
was described to produce the pellets. The comparison of the rotary fluid bed
processing with the multiple-step extrusion and spheronization was reported by
Robinson et al. (161). The authors manufactured acceptable immediate release acetaminophen
pellets, using both of these techniques. The quality of the pellet produced
improved as the minimum quantity of product was increased in the rotary fluid bed
processor. The advantage of using a single unit, such as the rotary fluid bed, over the
multiple-unit process involving several pieces of equipment was described.
The rotary fluid bed is used for producing a pellet by layering the active drug
suspension or solution on to nonpareil cores, and subsequently coating them with
polymers to impart modified release properties (162) Hileman et al. (163) reported
the manufacture of immediate spheres of a poorly water-soluble drug in a rotary
fluid bed by layering the active drug suspension on to nonpareil cores. These immediate
release spheres were then overcoated with an ethylcellulose/HPMC hydroalcoholic
solution in the same unit, eliminating the need for additional process and
handling steps. Iyer et al. evaluated layering of aqueous solution of phenylpropanolamine
hydrochloride with different binders (164). The layered beads were coated in
the rotoprocessor and the Wurster Coater to compare the utility of the rotoprocessor
as an equipment, not only to produce pellets, but to coat them as well. Various
equipment manufacturers have promoted powder layering on the pellets, in a rotary
fluid bed. In 1992, Jones et al. received a patent for such a process (165). The process
claims to have advantages of layering a drug substance with a relatively small
amount of liquid, thus making this layering process more efficient. The commercial
application of this process has not been reported in the literature. Korakianiti et al.
(166) studied the preparation of pellets using rotary fluid bed granulator. The
authors concluded that the rotor speed and the amount of water significantly
affected the geometric mean diameter of the pellets, and they proposed an equation
to show that correlation. Pis?ek et al. (167) studied the influence of rotational speed
and surface of rotating disk on pellets produced by using rotary fluid bed. They used
a mixture of pentoxifylline and microcrystalline cellulose to produce pellets, using a
suspension of Eudragit NE 30 D as a binder. The results showed that both the surface
and the rotational speed of the disk have an influence on the shape, surface, and
size of the pellets, while there was less effect on the density, humidity content, and
yield. They found that the textured surface of the disk produced pellets with a
rougher surface when the rotational speed was increased compared to the smooth
surface, where increased rotational speed produced more spherical pellets with a
larger diameter. Jan Vertommen (168) has summarized the use of rotary fluid bed
processor to produce pellets.
11.3. Integrated Systems
The fluid bed technology is used for drying, agglomerating, coating, and pelletization.
The trend in the industry is toward integrating various steps currently used
to produce the solid dosage products. In an attempt to minimize the number of steps
300 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
involved, fluid bed granulation and drying is increasingly planned in the following
three ways:
1. Fluid bed processor for granulation and drying
2. Integration of high-shear mixer and a fluid bed dryer
3. Integration of high-shear mixer with fluid bed dryer for containment of
potent compounds.
11.3.1. The Granulation and Drying Carried Out in a Single Unit
To minimize material handling steps, the fluid bed units are loaded by gravity or by
vacuum, as described previously. Discharge of the product from the unit is accomplished
by either side or bottom discharge, by employing pneumatic transport
system. The CIP of such a unit can be accomplished by using stainless steel cartridge
filters and an overlap gill air distributor, if desired.
11.3.2. Integration of a High-Shear Mixer and a Fluid Bed Dryer
for controlling dust and cross-contamination. When these two-unit operations are
integrated as a single unit, a number of points must be considered. The following is a
list of some of the questions that the reader may want to consider:
1. Engineering layout and the footprint, ceiling height requirements.
2. How will the high-shear mixer be loadedby gravity, vacuum, or
Figure 42 Integrated system. (Courtesy of the Glatt Group.)
Batch Fluid Bed Granulation 301
Figures 42 and 43 show a typical integrated system, where containment is considered
 2005 by Taylor & Francis Group, LLC
3. How will the binder solution be prepared and delivered to the mixer?
4. How will the granulation end point be determined and reproduced?
5. How will the discharge from the high-shear mixer be accomplished?
6. Are the process parameters for granulation and fluid bed drying established
and reproducible, indicating a robust process?
7. How will the product, discharged from the fluid bed dryer be handled?
Does it require sizing, blending with the lubricants?
8. Is this system dedicated for a single product or multiple products?
9. How will this system be cleaned?
10. Will the control of a process be done individually for each unit or by an
integrated control system?
For the potent compound processing, requiring high-shear granulation and
fluid bed drying, all the questions mentioned previously must be considered. In
addition, the operator safety, special sampling valves, isolated room air handling
systems, and CIP must be considered. Such an integrated system can be dedicated
for a single product.
materials are dispensed, operating personnel do not have to handle the product during
the granulation drying steps. Such a system is costly and does require an enormous
amount of time for process and cleaning validation.
The fluid bed process, like other granulation techniques requires understanding
of the importance of characterization of the raw materials, especially of a drug substance,
the process equipment, limitations of the selected process, establishment of
in-process control specifications, characterization of the finished product, and clean-
Figure 43 Integrated system. (Courtesy of the Glatt Group.)
302 Parikh and Mogavero
Figure 44 shows a system where a potent compound is processed. After the raw
 2005 by Taylor & Francis Group, LLC
ing and process validation. It is equally important that the formulation and development
scientists do not lose sight of the fact that the process being developed will be
going on the production floor and should be robust enough.
My sincere thanks to Ms. Christine Budny, my assistant at Synthon Pharmaceuticals
Inc., for her help in preparing this chapter.
Figure 44 Integrated system for potent compound. (Courtesy of Niro Pharma Systems.)
Batch Fluid Bed Granulation 303
 2005 by Taylor & Francis Group, LLC
1. Kunii D, Lvenspiel O. Fluidization Engineering. New York: John Wiley & Sons, Inc.,
2. Wurster DE. J Am Pharm Assoc Sci Ed 1960; 49:82.
3. Wurster DE. J Am Pharm Assoc Sci Ed 1959; 48(8):451454.
4. Scott MW, Liberman HA, Rankell AS, Battista JV. J Pharm Sci 1964; 53(3):314319.
5. Rankell AS, Scott MW, Liberman HA, Chow FS, Battista JV. J Pharma Sci 1964;
6. Contini S, Atasoy K. Pharm Ind 1966; 28:144146.
7. Wolf G. Drugs Made Germany 1968; 11:172180.
8. Liske T, Mobus W. The manufacture and comparative aspects of fluidized layer spray
granulation. Drugs Made Germany 1968; XI:182189.
9. Sherrington PJ, Oliver R. Granulation monograph (chapter 3). In: Goldberg AS, ed.
Powder Science and Technology. London: Heyden, 1981.
10. Pietch WB. In: Fayed ME, Otten L, eds. Handbook of Powder Science and Technology.
New York: Van Nostrand Reinhold, 1984.
11. Hersey JA. Fluidized bed technologyan overview. Int J Pharm Tech Prod Manuf
1981; 2(3):34.
12. Thiel WJ. The theory of fluidization and application to the industrial processing of
pharmaceutical products. Int J Pharm Tech Prod Manuf 1981; 2(5):58.
13. Thiel WJ. Solids mixing in gas fluidized beds. Int J Pharm Tech Prod Manuf 1981;
14. Littman H. An overview of flow in fluidized beds. Pharm Technol 1985; 9(3):48.
15. Whitehead AB. Behavior of fluidized bed systems. Pharm Technol 1981; 2(13).
16. Story MJ. Granulation and film coating in the fluidized bed. Pharm Technol 1981;
17. Aulton MJ, Banks M. Fluidized bed granulation; factors influencing the quality of the
product. Pharm Technol 1981; 2(24):2427.
18. Kulling W, Simon EJ. Fluid bed technology applied to pharmaceuticals. Pharm Technol
1980; 4(1):7983.
19. Gomezplata A, KugelmanAM.Processing systems. In: MarchelloJMand Gomezplata A,
eds. GasSolid Handling in the Process Industries, Chemical Processing and Engineering
Vol. 8. New York: Marcel Dekker, Inc., 1976:chap 4.
20. E. Ra?sa?nen et al. AAPS PharmSci, 3(3) Abstract, AAPS Annual Meeting 2001.
21. Parikh DM. Airflow in batch fluid-bed processing. Pharm Technol 1991; 15(3):100110.
22. Row PN, et al. Trans Inst Chem Eng 1965; 32T:271.
23. Davis L, et al. Trans Inst Chem Eng 1966; 44T:293.
24. Gorrodnichev VI, et al. Pharm Chem J (USSR) 1974; 8:298.
25. Thiel WJ. The theory of fluidization and application to the industrial processing of
pharmaceutical products. Int J pharm Tech Prod Manuf 1981; 2(5).
26. Zenz FA. In: Kirk-Othmer, ed. Fluidization: Encyclopedia of Chemical Technology,
Vol. 10, 3rd ed. New York: Wiley-Interscience Publication, 548581.
27. Ennis BJ, Tardos GI, Pfeffer R. A micro-levelbased characterization of granulation
phenomena. Powder Technol 1991; 65:257272.
28. Osborne BG, Fearn T, Hindle PH. In: Practical NIR Spectroscopy with Applications in
Food and Beverage Industry Analysis. 2nd ed. Harlow, U.K.: Longman, 1993:227.
29. Long GE. Spraying theory and practice. Chem Eng 1978:7377.
30. Masters K. Spray Drying, an Introduction to Principles, Operational Practice and
Application. 2nd ed. New York: John Wiley & Sons Inc., 1976.
31. Japanese Patent Application 61259696, 1986.
32. Swiss Patent 0176/93, 1993, U.S. Patent 5,444,892, 1995, European Patent 0572356A1.
33. Guidance for Industry, Part 11, Electronic Records; Electronic SignaturesScope and
Application, February 2003.
304 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
34. Newitt DM, Conway-Jones JM. A contribution to the theory and practice of granulation.
Int J Pharm Tech Prod Manuf 1958; 36:422.
35. Record PC. A review of pharmaceutical granulation technology. Int J Pharm Tech Prod
Manuf 1980; 1:32.
36. Rumpf H. In: Krepper W, ed. Agglomeration. New York: Interscience, 1962:374.
37. Tardos GI, Khan MI, Mort PR. Critical parameters and limiting conditions in binder
granulation of fine powders. Powder Technol 1997; 94:245258.
38. Iveson SM. Granule coalescence modeling: including the effects of bond strengthening
and distributed impact separation forces. Chem Eng Sci 2001; 56:22152220.
39. Schaefer T, Worts O. Control of fluidized bed granulation I. Effects of spray angle, nozzle
height and starting materials on granule size and size distribution. Arch Pharm
Chem Sci Ed 1977; 5:5160.
40. Smith PJ, Nienow AW. Chem Eng Sci 1983; 38(8):12231231, 13231240.
41. Aulton ME, Banks M. Int J Pharm Tech Prod Manuf 1981; 2(4):2429.
42. Maroglou A, Nienow AW. 4th Symposium on Agglomeration, Toronto, Canada, Iron
& Steel Society Inc., June 1985:465470.
43. Maroglou A, Nienow AW. Fluidized bed granulation technology and its application to
tungsten carbide. Powder Metall 1986; 29(4):195291.
44. McCabe WL, Smith JC. In: Lachman L, Liberman HA, Kanig JL, eds. Unit Operations
of Chemical Engineering. New York: McGraw-Hill, 1956:Inc.
45. Green Don W, ed. In: Perrys Chemical Engineers Handbook. New York : McGraw-
Hill, Inc., 1984:section 20.
46. Rankell RS, Liberman HA, Schiffman RF. Drying. In: The Theory and Practice of
Industrial Pharmacy. 3rd ed. Philadelphia, PA: Lea & Feabiger, 1986.
47. Schaefer T, Worts O. Control of fluidized granulation III, effects of the inlet air temperature,
and liquid flow rate on granule size and size distribution. Control of moisture
content of granules in the drying phase. Arch Pharm Chem Sci Ed 1978; 6(1):113.
48. Gao JZH, et al. AAPS PharmSci, 1(4) Abstract (Tech Note), AAPS Annual Meeting
49. Thurn U. Dissertation, Eidgenossischen Technischen, Hochschule, Zurich, 1970.
50. Liske T, Mobus W. Drugs Made Germany 1968; 11(4):182189.
51. Schaefer T, Worts O. Control of fluidized bed granulation, I. Effects of spray angle,
nozzle height and starting materials on granule size and size distribution. Arch Pharm
Chem Sci Ed 1977; 5:5160.
52. Schaefer T, Worts O. Control of fluidized bed granulation V, factors affecting granule
growth. Arch Pharm Chem Sci Ed 1978; 6:6982.
53. Ormos Z, Pataki K. Hung J Ind Chem 1979; 7:89103.
54. Ormos Z, Pataki K. Hung J Ind Chem 1979; 7:105117.
55. Banks M. Ph.D. Thesis, C.N.A.A., Leicester Polytechnic, 1981.
56. Galmen MJ, GreerW. Paper 2. Fluid Technology and Pharmaceutical Manufacturing
International Conference, 1982.
57. Aulton ME. Paper 3. Fluid Technology and Pharmaceutical Manufacturing International
Conference, Powder Advisory Center, London, 1982.
58. Veillard M, et al. Int J Pharm Tech Prod Manuf 1982; 3(4):100107.
59. Georgakopoulos PP, et al. Pharmazie 1983; 38(4):240243.
60. Jinot JC, et al. STP Pharm 1986; 2(13):126131.
61. Shinoda A, et al. Yakuzaigaku 1976; 36(2):8388.
62. Aulton ME, et al. J Pharm Pharmacol 1977; 29(suppl):59.
63. Schepky G. Acta Pharm Technol 1978; 24(3):185212.
64. Aulton ME, et al. Proceedings of the International Conference on Powder Technology
in Pharmacy, Powder Advisory Center, Basel, Switzerland, June 1979.
65. Kocova El-Arini S, et al. Drugs Made Germany 1983; 26(4):205211.
66. Davies WL, Gloor WT. J Pharm Sci 1972; 61(4):618622.
67. Rouiller M, et al. Acta Pharm Technol 1975; 21(2):129138.
Batch Fluid Bed Granulation 305
 2005 by Taylor & Francis Group, LLC
68. Schaefer T, Worts O. Control of fluidized bed granulation II, estimation of droplet size
of atomized binder solution. Arch Pharm Chem Sci Ed 1977; 5:178193.
69. Ormos Z, et al. Hung J Ind Chem 1979; 7:131140.
70. Ormos Z, et al. Hung J Ind Chem 1979; 7:141151.
71. Ormos Z, et al. Hung J Ind Chem 1979; 7:153163.
72. Ormos Z, et al. Hung J Ind Chem 1979; 7:221235.
73. Kocova El-Arini S. Pharm Ind 1981; 43(7):674679.
74. Jager KF, Bauer KH. Acta Pharm Technol 1984; 30(1):8592.
75. Nouh ATI. Pharm Ind 1986; 48(6):670673.
76. Alkan H, et al. Doga Tu J Med Pharm 1987; 11(1):17.
77. Bank A, et al. Proc 2nd Conf Appl Phys Chem 1971; 2:687692.
78. Davies WL, Gloor WT. J Pharm Sci 1973; 62(1):170172.
79. Ormos Z, et al. Hung J Ind Chem 1973; 1:307328.
80. Ormos Z, et al. Hung J Ind Chem 1973; 1:463474.
81. Johnson MCR, et al. J Pharm Pharmacol 1975; suppl:80.
82. Schaefer T, Worts O. Control of fluidized bed granulation IV. Effects of binder solution
and atomization on granule size and size distribution. Arch PharmChem Sci Ed 1978;
83. Aulton ME, et al. Mfg Chem Aerosol News 1978; 12:5056.
84. Gorodnichev VI, et al. Pharm Chem J (USSR) 1980; 14(10):7277.
85. Ceschel GC, et al. II Farm Ed Prat 1981; 36(6):281293.
86. Rangnarsson G, et al. Int J Pharm 1982; 12:163171.
87. Meshali M, El-Banna HM, El-Sabbagh H. Pharmazie 1983; 38(5):323325.
88. Hajdu R, Ormos Z. Hung J Ind Chem 1983; 12:425430.
89. Devay A, et al. Acta Pharm Technol 1984; 30(3):239242.
90. Alkan MH, Yuksel A. Drug Dev Ind Pharm 1986; 12(10):15291543.
91. Hontz J. Assessment of Selected Formulation and Processing Variables in Fluid Bed
Granulation, Ph.D. Thesis, University of Maryland, Baltimore, MD, 1987; Dissertation
Abs Int 1987; 6:1655-B.
92. Wan LSC, Lim KS. STP Pharm 1988; 4(7):560571.
93. Wan LSC, Lim KS. STP Pharm 1989; 5(4):244250.
94. Sheskey P, Keary C, Inbasekaran P, Deyarmond V, Balwinski K. Foam technology: the
development of a novel technique for the delivery of aqueous binder systems in high
shear and fluid-bed wet-granulation applications. Poster at AAPS Annual Meeting,
Oct 2630, 2003.
95. Dahl TC, Bormeth AP. Naproxen controlled release matrix tablets: fluid bed granulation
feasibility. Drug Dev Ind Pharm 1990; 16(4):581590.
96. Wan Lucy SC, Heng Paul WS, Muhuri G. Incorporation and distribution of a low dose
drug in granules. Int J Pharm 1992; 88:159163.
97. Higashide F, Miki Y, Nozawa Y, Ishibashi K. Dependence of drug content uniformity
on particle sizes in fluidized bed granulation. Pharm Ind 1985; 47(11):12021205.
98. Davies WL, Gloor WT Jr. Batch production of pharmaceutical granulationin fluidized
bed II: effects of various binders and their concentrations on granulations and compressed
tablets. J Pharm Sci 1972; 61:618.
99. Davies WL, Gloor WT Jr. Batch production of pharmaceutical granulation in fluidized
bed III: binder dilution effects on granulation. J Pharm Sci 1973; 62:170.
100. Seo A, Holm P, Schaefer T. Effects of droplet size and type of binder on the agglomerate
growth mechanisms by melt agglomeration in a fluidized bed. Eur J Pharm Sci 2002;
101. Rawley FA. Effects of the bag shaking cycle on the particle size distribution of granulation.
Pharm Technol 1989; 13(9):7882.
102. Davies WL, Gloor WT Jr. Batch production of pharmaceutical granulation in fluidized
bed I: effects of process variables on physical properties of final granulation. J Pharm
Sci 1971; 60(12):18691874.
306 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
103. Workman J Jr. A review of process near infrared spectroscopy: 19801994. J. Near
Infrared Spectrosc 1993; 1:221245.
104. U.S. Patent USP XXVII 2004.
105. Near Infra Red Spectrometry. European Pharmacopoeia June 2004:59.
106. Choppin G, Buijs K. Near infrared studies of structure of water-II. Ionic solutions.
J Chem Phys 1963; 39:20422050.
107. Osborne BG, Fearn T, Hindle PH. In: Practical NIR Spectroscopy with Applications in
Food and Beverage Industry Analysis, 2nd ed. Harlow, U.K.: Longman, 1993.
108. Fong A, Hieftje GM. Near infrared measurement of relative and absolute humidity
through detection of water adsorbed on a silica gel layer. Anal Chem 1995; 67:
109. Siesler HW, Ozaki Y, Kawata S, Heise HM. Near-Infrared Spectroscopy: Principles,
Instruments and Applications. Weinheim, Germany: Wiley, 2002.
110. Wetzel DL. Near infrared reflectance analysis: sleeper among the spectroscopic techniques.
Anal Chem 1983; 55(12):1165A1175A.
111. McDonald BF, Pebble KA. Some application of near-infrared reflectance analysis in the
pharmaceutical industry. J Pharm Biomed Anal 1993; 11(1112):10771085.
112. Sinsheimer J, Poswalk N. Pharmaceutical applications of the near infrared determination
of water. J Pharm Sci 1968; 57:20072010.
113. Wetzel DL. Near infrared reflectance analysis: sleeper among the spectroscopic techniques.
Anal Chem 1983; 55(12):1165A1175A.
114. Workmann JJ. Review of process and non-invasive near-infrared and infrared spectroscopy
11931999. Appl Spectrosc Rev 1999; 34:189.
115. Callis J, Illman D, Kowalski B. Process analytical chemistry. Anal Chem 1987;
116. Beebe K, Blaser W, Bredeweg R, Chauvel J, Harner R, LaPack M, Leugers A, Martin D,
Wright L, Yalvac E. Process analytical chemistry. Anal Chem 1993; 65:199R216R.
117. Blaser W, Bredeweg R, LaPack M, Leugers A, Martin D, Pell R, Workman J, Wright L.
Process analytical chemistry. Anal Chem 1995; 67:47R70R.
118. Hassel D, Bowman E. Process analytical chemistry for spectroscopists. Appl Spectrosc
1998; 52:18A29A.
119. Workman J Jr, Veltkamp D, Doherty S, Anderson B, Creasy K, Koch M, Tatera J,
Robinson A, Bond L, Burgess L, Bokerman G, Ullman A, Darsey G, Mozayeni F,
Bamberger J, Greenwood M. Process analytical chemistry. Anal Chem 1999;
120. Brittain H. Methods for the characterization of polymorphs and solvates in polymorphism
in pharmaceutical solids. In: Brittain H ed. 1st ed. New York: Marcel Dekker Inc.,
121. Frake PD, Greenhlgh SM, Grierson JM, Hempenstall, Rudd DR. Process control and
end-point determination of fluid bed granulation by application of near-infrared spectroscopy.
Int J Pharm 1997; 151:7580.
122. Rantanen JO, Antikanen JP, Mannermaa, Yliuusi J. Use of near-infrared reflectance
method for measurement of moisture content during granulation. Pharm Dev Technol
2000; 5(2):209217.
123. Rantanen J, Lehtola S, Ra?met P, Mannermaa JP, Ylirussi. On-line monitoring of moisture
content in an instrumented fluidized bed granulator with multi-channel NIR moisture
sensor. Powder Technol 1998; 99:163170.
124. Andersson M, Folestad S, Gottfires J, Johansson MO, Josefson M, Wahlund KG.
Quantitative analysis of film coating in a fluidized-bed process by in-line NIR spectroscopy
and multivariate batch calibration. Anal Chem 2000; 72:20992108.
125. Vazquz ER. Optimization of Drying-End-Points Measurements for the Automation of
a Fluidized-Bed Dryer Using FT-NIR Spectroscopy. M.S. Thesis, University of Puerto
Rico, 2004.
Batch Fluid Bed Granulation 307
 2005 by Taylor & Francis Group, LLC
126. Rantanen J, Ka?nsa?koski M, Tenhunen J, Lehtonen S, Rajalahti T, Mannermaa J,
Yliruusi J. Next generation fluidized bed granulator automation. AAPS PharmSciTech
2000; 1(2):article 10.
127. Rantanen JA, Jorgensen E.Ra?sa?nen,Luukkonen P, Airaksinen S, Raiman J, Ha?nninen K,
Antikainen O, Yliruusi J. Process analysis of fluidized bed granulation. AAPS Pharm
SciTech 2001; 2(4):article 21.
128. Rantanen J, Sampsa JL, Osmo KA, Jukka-Pekka M, Olii ES, Jouko KY. Visualization
of fluid bed granulation with self organizing maps. J Pharm Biomed Anal 2001;
129. Laitinen N, Antikainen O, Airaksinen S, Yliuusi J. At-line particle size analysis with a
novel optical technique during a fluidized-bed granulation process. Poster at AAPS
Annual Meeting, 2002.
130. Jha BK, Tambe SS, Kulkarni BD. Estimating diffusion coefficients of a micellar system
using an ANN. J Coll Interface Sci 1995; 170:392398.
131. Bourquin J, Schmidi H, Van Hoogevest P, Leunberger H. Basic concepts of ANN modeling
in the application to pharmaceutical development. Pharm Dev Technol 1997;
2(2):95109, 111121.
132. Watano S, Takashima H, Miyanami K. Control of moisture content in fluidized bed
granulation by neural network. J Chem Eng Jpn 1997; 30(2):223229.
133. Leane MM, Cumming I, Corrigan OI. The use of artificial neural networks for the selection
of the most appropriate formulation and processing variables in order to predict
the in vitro dissolution of sustained release minitabs. AAPS PharmSciTech 2003;
4(2):article 26.
134. Vandel DA, Davari, Famouri P. Modeling of fluidized bed neural networks. Proceedings
of 32nd IEEE SSST, Tallahassee, FL, March 2000, FAMU-FSU.
135. Quantrille TE, Liu YA. In: Artificial Intelligence in Chemical Engineering. San Diego:
Academic Press, 1991.
136. AAPS/FDA. Scale-up of Oral Solid Dosage Forms, December 1113, 1991.
137. Food and Drug Administration. Immediate Release Solid Dosage Forms. Scale-up and
Post Approval Changes: Chemistry, Manufacturing and Control, In-vitro Dissolution
Testing and In-Vivo Bioequivalence Documentation. Rockville, MD: FDA Center for
Drug Evaluation and Research, November 30, 1995.
138. Gore AY, McFarland DW, Batuyios NH. Fluid bed granulation: factors affecting the
process in laboratory development and production scale-up. Pharm Technol 1985;
139. Bonck JA. Spray granulation. Presented at the AIChE Annual Meeting, November,
140. Jones DM. Factors to consider in fluid bed processing. Pharm Technol 1985; 9(4):50.
141. Matharu AS, Patel MR. A new scale-up equation for fluid bed processing. Poster at
AAPS Annual Meeting, 2003.
142. Rambali B, Baert L, Massart DL. Scaling up of the fluidized bed granulation process.
Int J Pharm 2003; 252(12):197206.
143. Simon EJ. Fluid bed processing of bulk solids. In: Third International Powder Technology
and Bulk Solids Conference, February 20, 1975. Harrogate, U.K.: Heyden & Sons
Publisher, 1975:6373.
144. Kulling W. Method and Apparatus for Removing a Vaporized Liquid from Gas for Use
in a Process Based on the Fluidized Bed Principle. U.S. Patent 4,145,818, March 27,
145. ATEX Directive 94/9/EC of the European Parliament and the Council, March 23,
146. Parikh DM. Fluid bed processing in the 1990s. In: Tabletting and Granulation Yearbook.
Pharm Technol 1996; suppl:4047.
147. International Patent WO 93/08923, 1993.
148. International Patent WO 95/20432, 1995.
308 Parikh and Mogavero
 2005 by Taylor & Francis Group, LLC
149. Glatt Brochure on STRATOS2.
150. Walter KT, et al. Fluidized bed precision granulation with real time process determination.
Poster at AAPS Annual Meeting, 2003.
151. Doelling MK, et al. The development of a microwave fluid bed processor I: construction
and qualification of a prototype laboratory unit. Pharm Res 1992; 9(11):14871492.
152. Doelling MK, Nash RA. The development of a microwave fluid bed processor II: drying
performance and physical characteristics of typical pharmaceutical granulations. Pharm
Res 1992; 9(11):14931501.
153. Luy B, Hirschfeld P, Leuenberger H. Granulation and drying in vacuum fluid-bed system.
Pharm Ind 1989; 51:89.
154. U.S. Patent 3,671,296, June 20, 1977.
155. U.S. Patent 4,034,126, July 5, 1977.
156. U.S. Patent 4,542,043, Sep 17, 1985.
157. Ja?ger KF, Bauer KH. Effect of material motion on agglomeration in the rotary
fluidized-bed granulator. Drugs Made Germany 1982; 25:6165.
158. U.S. Patent 4,740,390.
159. Tu?rkoglu M, He M, Sakr A. Evaluation of rotary fluidized-bed as a wet granulation
equipment. Eur J Pharm Biopharm 1995; 41(6):388394.
160. Li Shun P, Kowarski CR, Feld KW, Grim WM. Recent advances in microencapsulation
technology and equipment. Drug Dev Ind Pharm 1988; 14(2,3):353376.
161. Robinson RL, Hollenbeck G. Manufacture of spherical acetaminophen pellets: comparison
of rotary processing with multiple-step extrusion and spheronization. Pharm Technol
1991; 15(5):4856.
162. Parikh DM. Layering in rotary fluid bed a unique process for the production of spherical
pellets for controlled release. Interphex, New York, May 18, 1991.
163. Hileman GA, Sarabia RE. Manufacture of immediate and controlled release spheres in
a single unit using fluid bed rotor insert. Poster PT 6167 at AAPS Annual Meeting,
164. Iyer RM, Augsburger LL, Parikh DM. Evaluation of drug layering and coating: effect
of process mode and binder level. Drug Dev Ind Pharm 1993; 19(9):98919998.
165. U.S. Patent 5,132,142, 1992.
166. Korakianti ES, Rekkas DM, Dallas PP, Choulis NH. Optimization of the pelletization
process in a fluid bed rotor granulator using experimental design. AAPS PharmSciTech
2000;1(4):article 35.
167. Pis?ek R, Planins?ek O, Tus? M, Src?ic? S. Influence of rotational speed and surface of rotating
disc on pellets produced by direct rotor pelletization. Pharm Ind 2000; 62:312319.
168. Vertommen J. Pelletization in a rotary processor using the wet granulation technique.
In: ACTA Biomedica Lovaniesia Leuven. Belgium Leuven University Press, 1998:166.
Batch Fluid Bed Granulation 309
 2005 by Taylor & Francis Group, LLC
Single-Pot Processing
Harald Stahl
Niro Pharma Systems, Muellheim, Germany
Griet Van Vaerenbergh
Collette N.V., Wommelgren, Belgium
Single-pot processing was developed to provide the means for mixing, granulating,
drying, and blending pharmaceutical granulations in a single apparatus. Although
gory of processors consists of a high- or low-shear mixergranulator (similar to conventional
granulators) and outfitted with a variety of drying options. Initially,
vacuum was combined with a heat-jacketed bowl to provide the means for drying
in the single pot. Today, processors are available that provide vacuum drying with
microwaves or that percolate gas under low pressure into the vacuum chamber
(i.e., processing bowl). Another very interesting improvement is the use of swinging
processing bowls (1).
Single-pot processors for the pharmaceutical industry have been available for
years. They received renewed interest in the mid-1980s when microwaves were
coupled with vacuum to enhance the drying operation. Microwaves applied to drying
pharmaceutical granulations became synonymous with single-pot processing,
and there was anticipation that this technology would eventually become the norm
for granulation processing. Several major pharmaceutical companies purchased production-
scale units following successful trials conducted at vendors pilot facilities.
In the ensuing years, when the technology did not become as popular as expected,
there were rumors that microwave systems could not be validated and suffered from
excessive regulatory hurdles. The reality is neither, and single-pot technology has
continued to evolve during the last decade. It has de-emphasised its association with
microwave drying while continuing to demonstrate its appropriate role in granulation
technology (2).
In a production setting, single-pot processing may offer a number of advantages.
By integrating granulating and drying capabilities into a single unit, capital
investment in equipment and good manufacturing practice floor space may be lower
than the other alternatives. The number of material-handling steps is decreased; consequently,
the total processing time may be shorter while maintaining a high yield
equipment design varies from manufacturer to manufacturer (Figs. 13), this cate-
 2005 by Taylor & Francis Group, LLC
and keeping production support to a minimum. Environmental variables, such as
humidity, are eliminated from the manufacturing process, which may offer advantages
for processing moisture-sensitive formulations. State of the art is to outfit a single-
pot processor with clean-in-place systems, thereby enhancing operator safety by
minimizing exposure to the product both during manufacturing and during cleaning
Figure 2 L.B. Bohle single-pot processor. (Courtesy of the L.B. Bohle Group, Germany.)
Figure 1 UltimaProTM 600 microwave/vacuum single-pot processor with swinging bowl.
(Courtesy of Collette NV, Belgium.)
312 Stahl and Van Vaerenbergh
 2005 by Taylor & Francis Group, LLC
(3). Requirements for solvent recovery systems are lower for single-pot processors
compared with fluid-bed driers. Single-pot processors outfitted with vacuum are
attractive for evaporating events that are explosive or for containing drug substances
with low-exposure limits.
The versatility and compactness of small-scale (325 L) single-pot processors
also make the technology attractive for development and pilot laboratory facilities.
Within the last few years, equipment manufacturers began offering single-pot processors
that can accommodate the batch sizes required during early development
(0.3 g10 kg). The processors can be used as mixerblenders for direct compression
formulations, or as mixergranulators to prepare wet granulations for fluid-bed drying,
or utilized for their full range of capabilities as a single-processing unit for all the
steps required for granulation preparation. Some vendors offer the option of upgrading
their small-scale processors. For example, a user can initially purchase a singlepot
processor with vacuum-drying capabilities and add a microwave drying system
at a later time. Consequently, single-pot processors should be given strong consideration
when equipping a development laboratory or a pilot plant intended to offer
a variety of processing options to the pharmaceutical formulator.
The following text is intended to expose the reader to the drying methods, the
capabilities, and the applications of single-pot processing to pharmaceutical granulations.
Fluid-bed technology, which can also be used to mix, to granulate, and for
drying alternative wet granulation technologies in a single unit, will not be addressed
in this chapter because it is discussed elsewhere in this book.
The steps and sequence of manufacturing pharmaceutical granulations using singlepot
processing are the same as those that use alternative technologies, except that
Figure 3 Zanchetta single-pot processor. (Courtesy of Zanchetta, Italy.)
Single-Pot Processing 313
 2005 by Taylor & Francis Group, LLC
several of the steps are performed in the same product chamber. The majority of production
installations make use of its mixing, granulating, and drying capabilities during
the processing of a single batch. Important to note is the absence of a milling step
in between granulation and drying. The advantage of a reduced number of process
steps is obvious while the drawback is that there is no possibility of breaking lumps
generated during granulation. Single-pot operations therefore require excellent control
of the granulation to prevent the formation of oversized material.
2.1. Dry Mixing
Powders are loaded into the single pot either manually (for development-and pilotscale
units) or by a conveying system (for production-scale units). Vacuum pump(s)
used for the drying operation can also be used to charge the processor. Pneumaticand
vacuum-conveying systems contribute to minimizing operator exposure to the
drug product. The powders are mixed in the dry state until the desired degree of uniformity
is obtained. Depending on the geometry of the processing bowl and the effi-
ciency of its mixing blades, optimal mixing for most processors generally occurs
when the bowl is charged to 5075% of its capacity. Batch size, impeller speed,
and mixing time are the variables that affect the desired degree of blend homogeneity
before the addition of the binder solution.
2.2. Addition of Binder Solution
Once dry mixing is completed, the binder solution is added through a spray lance
connected to a solvent delivery system (such as a pressure pot or peristaltic pump).
For highly viscous binder solutions it is of advantage to use the vacuum system of
the processor for sucking into the processor. Because the single-pot processor is
operating as a granulator at this point, all variables considered during the manufacture
of wet granulations in conventional high-or low-shear mixergranulators are
applicable. Those variables include the rate of binder action, droplet size, and spray
pattern (the last two being determined by the selection of the spray nozzle and the
distance between the nozzle tip and granulation bed). The speed of the main impeller
and the granulating tool (e.g., high-intensity chopper bar), as well as the jacket temperature
should also be controlled during binder addition.
2.3. Wet Massing
Following binder addition, additional energy may be imparted to the granulation
until the desired consistency is obtained. The speeds of the main impeller and granulating
tool, wet-massing time, and jacket temperature are variables that can affect
the physical attributes of the granulation. Like the bowl shape the impeller design
will also affect the amount of shear imparted to the granulation. Granulation end
point may be controlled by process time, temperature of the product bed, and the
energy consumption or torque of the main impeller.
2.4. Drying
After granulation, the material is dried using one of three approaches: (a) vacuum
drying, (b) gas-assisted vacuum drying, or (c) microwave vacuum drying. Details
314 Stahl and Van Vaerenbergh
 2005 by Taylor & Francis Group, LLC
of each drying method are summarized in Drying Methods for Single-Pot
Processors. The product bed is usually stirred at low intensity during the drying process
to facilitate solvent removal and promote uniform drying, as well as to prevent
caking of the granulation on the chambers walls. Agitation may be applied by
slowly tilting the bowl or operating the impeller at low speed, either continuously
or intermittently, throughout the drying stage. Caution must be exercised to avoid
granule breakdown during drying, which may result in unfavorable compression
characteristics (4). Variables for vacuum drying include the level of vacuum maintained
in the bowl, the jacket temperature, and the degree of agitation. In addition
to the parameters listed for vacuum drying, gas-assisted vacuum drying must also
consider the type of drying gas used and its rate of delivery. When microwave
vacuum drying is used, all of the variables used for vacuum drying are applicable,
as well as the level of microwave power used to dry the granulation. If yield is of
greater importance than process time, a very interesting process option is to follow
with the wall temperature very closely the product temperature and use as a source
of drying energy only the introduced microwave energy. This mode of operation
minimizes the amount of material sticking to the walls for the price of a prolonged
drying operation resulting in a reduced throughput. This option is of special interest
for the processing of highly expensive materials.
If required, cooling can be conducted at the conclusion of the drying operation.
The heated water or steam in the bowl jacket, which supplied conductive heat during
the drying process, can be replaced with a glycolwater solution to provide a contact
surface as low as 10C. Another approach to cool the granulation is by purging
a cooling gas into the single pot while agitating the granulation bed.
2.5. Sizing and Lubrication
Once the granulation is dried, it is usually necessary to size it. This may be accomplished
by discharging the material through an in-line mill into a receiving vessel,
where it may be blended with any remaining excipients (e.g., lubricant and flavors).
This process design maintains the containment benefits of the single-pot process.
Alternatively, the remaining excipients may be added to the single-pot processor
and blended with the granulation before discharging and milling. This approach
requires that the lubricant be adequately distributed during milling and material
transfer during the compression operation.
3.1. Conductive Drying
The bowls of single-pot processors are generally jacketed for temperature control,
which minimizes condensation of the granulating solvent and assists in solvent evaporation
during drying. As a result, conductive heating provided by the heatjacketed
lid and the walls of the single pot contributes to the drying process. Its
dependence on the transfer of heat through pharmaceutical powders, which are poor
conductors of heat, prevents its use as the sole mode of drying in single-pot processors.
Eq. 1 addresses the conductive drying component for solvent removal.
Single-Pot Processing 315
 2005 by Taylor & Francis Group, LLC
Heat transfer from the vessel walls to the granulation bed is governed by
Q ? hSDT ?3:1?
where Q ? the energy exchange, h ? the exchange coefficient, S ? the contact surface
of the heated wall, and DT ? the temperature difference between the contact wall
and the granulation.
The rate of drying can be facilitated either by increasing the contact area
between the granulation and the vessel walls (which can be achieved by agitating
the product or utilizing a tilting bowl) or by maximizing the temperature differential
between the vessel walls and product (either through increasing the jacket temperature
or through maintaining the temperature of the product as low as possible during
Eq. 2 is a simple relation that may be of some value for the scaling-up of processes
using conductive heating (5)
tb=ta ? ?A=V?a?A=V?b ?3:2?
where t ? drying time, A ? heat transfer surface (m2), V ? vessel working volume
(m3), a ? refers to pilot scale, and b ? refers to production scale.
This relation accounts for the ratio of the surface area of the jacketed bowl and
the volume of the product requiring drying.
3.2. Vacuum Drying
Single-pot processors using vacuum drying may be considered if the product must be
dried at low temperature (<40C), if solvent recovery is required, or if the potential
for explosion is high. A vacuum is maintained within the vessel, thereby lowering
the temperature at which the granulating solvent evaporates. Because vapors are
removed from the processing bowl, vacuum drying provides a convenient means
for solvent recovery.
De Smet (6) has discussed the theory, advantages, and limitations of vacuum
drying. Aqueous granulations require a large amount of energy during drying, which
is generally supplied by the transfer of heat through conduction from the jacketed
bowl to the product. The amount of energy required for water removal is dependent
on the level of vacuum applied to the vessel and the osmotic pressure of dissolved
substances. As additional material dissolves in the water, the osmotic pressure
increases and additional energy is necessary to drive off the water. Therefore, as
the material becomes drier, the amount of energy necessary to evaporate the water
increases, and the rate of evaporation slows down. Processing times in vacuum driers
are often long, owing to the limited contact of the granulation with the heat from the
jacketed walls, and the slow rate of evaporation of the solvent from the interior of
the granules.
The drying rate of the vacuum component is dependent on the following relation:
V ? ksDP ?3:3?
where V ? evaporation rate, k ? rate coefficient, s ? total surface of granules,
and DP ? the vapor pressure difference between the granules and the surrounding
316 Stahl and Van Vaerenbergh
 2005 by Taylor & Francis Group, LLC
The rate of drying can be facilitated by increasing the level of vacuum (i.e.,
decreasing the pressure within the bowl) to increase the differential between granule
and bowl vapor pressure. Figure 4 is a typical drying curve for a vacuum-drying
process. When moisture content in the granulation is high, the rate of drying is
constant because evaporation of solvent from the product surface readily occurs.
As the level of moisture on the granule surface decreases, water must migrate from
the interior of the granule before evaporation. As a result, the rate of evaporation
progressively decreases.
During vacuum-drying processes, one should be aware of various problems
that could arise. For example, granule damage may occur owing to excessive
attrition as the bed is agitated during drying. Vacuum systems should contain adequate
filtering or blowback to prevent the loss of granulation fines through the
vacuum line, which may compromise the drug uniformity within the processed
A condenser positioned between the processor and vacuum pump, should
always be used, especially for granulations manufactured using organic solvents.
The condensate must be sufficiently cooled to prevent it from being released into
the atmosphere. Also, filters may become blocked owing to condensation forming
on the filter or the entrapment of solid particles. Blockage of the filters reduces
the level of vacuum that can be pulled on the bowl and excessively strains the pump
3.3. Gas-Assisted Vacuum Drying
Accelerating the drying process in vacuum driers is often limited by characteristics of
the product or equipment. Bowl temperature is generally limited by the physicochemical
stability of the product, which can limit the use of higher temperatures
to expedite drying. Increasing the contact area between the product and the vessel
is difficult without significantly altering the design of the equipment. Excessive agitation
of the product can lead to considerable granule attrition, which can lead to poor
granulation flow and compression properties.
Figure 4 Typical curve for vacuum drying. (Courtesy of Manufacturing Chemist).
Single-Pot Processing 317
 2005 by Taylor & Francis Group, LLC
Gas-assisted vacuum drying improves the efficiency of single-pot processors
that use vacuum drying by continuously introducing a small stream of gas through
the granulation to facilitate solvent removal. Drying continues to be performed at
lower temperatures (compared with tray and fluid-bed drying), but at shorter-processing
times than vacuum drying alone.
The gas may be introduced into the unit through openings in the bottom of the
vessel or through the mixing blades (Fig. 5).
Compressed air or nitrogen (mixed with or without air) is the commonly used
gas for these units. The rate of gas flow and the level of vacuum applied to the bowl
can be adjusted for a specific product to produce optimal drying conditions.
The introduction of gas into a vacuum chamber facilitates the drying process
through several actions. The constant flow of gas through the product improves
the transport of moisture from the product to the vacuum-solvent recovery system
(7). Introducing gas into the bowl also increases the vapor pressure driving force
(8). The pressure gradient across the vessel is increased, resulting in a reduction
in the rate at which water molecules recombine, producing a net increase in the
rate of evaporation. This causes the product temperature to be reduced, which
increases the temperature differential between the granules and the bowl wall.
The gas also reduces drying time by increasing the heat transfer coefficient from
the bowl to the bed. In addition to improving the heat transport through the
bed, the gas can reduce or eliminate product sticking to the sides of the vessel walls
because it improves flow and dries the particle surfaces more quickly. As the vessel
wall is the only notable source of drying energy this technology is best used in the
case of:
Figure 5 Gas-assisted vacuum-drying principle. (Courtesy of Collette NV, Belgium.)
318 Stahl and Van Vaerenbergh
 2005 by Taylor & Francis Group, LLC
 Small batch sizes (good surface volume ratio)
 Heat-insensitive materials (allows to operate a higher wall temperature)
 Organic solvents (requires only a fraction of the energy water needs for evaporation).
Additionally, the boiling temperature corresponding to the actual vacuum level
is much lower for organic solvents than for water, generating a larger temperature
difference between product bed and vessel.
3.4. Microwave Vacuum Drying
High-shear granulators with microwave vacuum-drying capabilities provide the fastest
drying rates in the family of single-pot processors. Microwave drying is based on
the absorption of electromagnetic radiation by dielectric materials, the theory of
which has been extensively described (911). Microwaves are a form of electromagnetic
energy similar to radio waves, the frequencies of which fall between 300 and
3000 MHz (between radio and optical waves, (Fig. 6). The two frequencies allocated
for domestic, scientific, medical, and industrial purposes are 915 and 2450 MHz.
Pharmaceutical processors generally use 2450 MHz, because this frequency is more
desirable when used in conjunction with vacuum. Single-pot processors incorporating
microwave drying are constructed of stainless steel because metal is a common
reflector of microwave energy and contains the energy within the processing chamber.
Teflon is essentially inert to microwaves, making it a suitable material for components
required in the processing bowl (e.g., spray lance and temperature probe).
Energy absorption of materials exposed to microwaves is described by Eq. 4
P ? 2pfV2E0Er tan d ?3:4?
where P ? the power density of the material (W/m3), f ? frequency (Hz), V ? voltage
gradient (V/m), E0 ? dielectric permissivity of free space (8.85  1012 F/m), Er ? dielectric constant of the material, and tan d ? loss tangent.
Figure 6 The electromagnetic spectrum.
Single-Pot Processing 319
 2005 by Taylor & Francis Group, LLC
For a constant electric field strength V the term (2pfV2E0) is constant. Therefore,
the power absorbed is proportional to the term (Er tan d), called the loss factor,
which is a relative measure of how easily a material absorbs microwave energy.
Various materials commonly used in pharmaceutical formulations have low
loss factors and only absorb microwave energy at high field strengths. Solvents used
in the granulation process (water, ethanol, isopropanol, and such), however, possess
high loss factors relative to the pharmaceutical powders (10). The dipolar component
of the solvents couples with the high-frequency electromagnetic field producing high
heating rates for the solvent, resulting in its evaporation and subsequent removal
from the processing chamber. Table 1 lists the loss factors for various components
in a typical pharmaceutical granulation (12).
Lucisano and Moss (13) performed a study in which a microwave drying process
was conducted in two different processors, one using fixed-output magnetrons
and the other using a variable-power magnetron. The unit using fixed-output magnetrons
had difficulty obtaining low moisture levels (<0.3%) because the E-field
safety set point was exceeded when the moisture was <1%. This problem did not
occur for the unit using the variable-power magnetron because forward power was
reduced as the E-field increased. During the late stages of drying, the unit was primarily
functioning as a vacuum dryer, as the amount of microwave energy being
introduced into the bowl was minimal. Nowadays, most microwave driers use the
variable output magnetrons.
The use of vacuum during microwave drying lowers the temperature at which
the solvent volatilizes, thereby limiting the temperature to which the material is
exposed. For example, at a vacuum of 45 mbar, water-based granulations will dry
at 31C. Once most of the water is removed from the process, the temperature
of the material will rise as components in the mixture with lower loss factors start
to absorb the microwaves. If too much vacuum is applied to the system, there is a
potential for granule breakdown owing to the excessive pressure between the core
and the surface of the granules. Microwaves are typically applied in a vacuum range
of 30100 mbar. Introducing microwave into a vacuum <30 mbar risks ignition of
the surrounding atmosphere, a condition known as arcing.
Control of the drying process is achieved through the simultaneous measurethe
level of microwave forward power, microwave reflected power, and product temperature
at various times during the drying process. During the initial stage, the product
temperature remains relatively constant as the free solvent is preferentially
evaporated, and the reflected power remains relatively low. The amount of vacuum
applied to the bowl, and to a lesser extent the bowl jacket temperature, will affect the
actual product temperature observed. As drying progresses at a constant rate of forward
power, the amount of absorbed energy decreases as the material dries, thereby
increasing the amount of free energy. As the free energy increases, a corresponding
increase in the reflected power is also observed.
The rise in reflected power is accompanied by an increase in product temperature,
which is simultaneously monitored while the magnetron output power is
reduced. This is necessary because the loss factors for some pharmaceutical components
are so small that very low moisture can be achieved before the temperature
rises. For such materials, the reflected power can rise sharply once most of the solvent
has evaporated, resulting in significant temperature gains.
The rise in temperature and reflected power signifies that the end of the drying
process is approaching. Several factors, such as the loss factors of the formulation
320 Stahl and Van Vaerenbergh
ment of product temperature, forward power, and reflected power. Figure 7 depicts
 2005 by Taylor & Francis Group, LLC
components, the microwave power, and the solvent retention properties of the solids
influence the point at which the previous relation will occur.
For example, lactose has a low loss factor and shows a sharp rise in reflected
power, followed by a slow temperature rise. Conversely, starch has a high loss factor
and demonstrates a fast temperature rise followed by a slow rise in reflected power.
Because of the different technologies incorporated into a single-pot processor, it is
capable of executing many different processes, apart from the standard wet granulation
and drying process, while small modifications or additional options can extend
the flexibility even further.
This chapter discusses some of the possible special processes and applications
in a single-pot processor. Although many of these processes are used in the
pharmaceutical industry, scientific literature on them is rare. The main reason for
this is that many of these processes were developed by the pharmaceutical industry
as product-specific solutions. This does not imply, however, that these processes cannot
be used more widely.
4.1. Melt Granulation
Melt granulation is a process in which the binder solution of the standard wet granulation
process is replaced with a meltable binder such as a wax or polyethylene glycol
(PEG), which is generally added in solid form, and melted during the process by
adding the necessary energy.
The most common production technique for melt granulation uses extruders,
but melt granulation in a high-shear mixer has also been extensively described in
the literature.
Figure 7 Relationship between microwave forward power, microwave reflected power, and
product temperature during microwave vacuum drying.
Single-Pot Processing 321
 2005 by Taylor & Francis Group, LLC
In this process, the necessary energy to melt the binder is provided either
by the mixer arm (mainly in laboratory-scale equipment) or by a heated jacket
If the meltable binder used absorbs microwaves, however (such as PEG for
example), using a single-pot processor equipped with microwave drying can present
major time-savings to the production process.
Providing the melting energy by the impeller or the heated jacket can be a very
time-consuming process, especially in production-scale equipment. Microwaves (as
explained earlier) are an instant source of energy that penetrates into the product
and can provide the energy faster and immediately where needed.
In a comparison between the use of the heated jacket to melt the binder and the
use of the heated jacket supplemented with microwaves, the latter method not only
proved to be more than twice faster in melting the binder used (PEG 3000), but also
the granulation step was reduced threefold (17).
At this moment, there is no literature available about other meltable binders in
combination with microwaves.
However, the results from the previously mentioned study show that it is worthwhile
to consider a single-pot processor for the production of melt granulations.
Another step where a single-pot processor can present a major advantage compared
to the standard production techniques, and especially compared to the process
in a high-shear mixer, is the cooling step. To achieve a stable dry granule from a
melt granulation process, the product needs to be cooled down to room temperature.
In a high-shear mixer, the cooling process is done by circulating cold water or a glycol
water mixture in the bowl jacket. As the contact surface between the product and
the jacket is limited, the cooling process generally takes a long time. If the process is
executed in a single-pot processor equipped with a gas-assisted vacuum-drying system,
this system can be used to pass cold air or even liquid nitrogen through the product
to aid the cooling process and reduce the cooling time considerably. If liquid
nitrogen is used, even a fivefold reduction of the cooling time is achievable (17).
A related chapter in this book should be consulted to gain more understanding of
melt granulation.
4.2. Pellet Production
For the production of spheres or pellets, in most cases an extrusion/spheronization
process is used.
There are, however, many references in scientific literature detailing the production
of pellets using a high-shear mixer, most of which concern melt pelletization
(1521). Taking into account the explanations given earlier on melt granulation, a
single-pot processor can, of course, also be used for this process for the same
Also, for other pelletization processes, not using meltable binders, the use of a
single-pot processor can be advantageous. In scientific literature, there are some
references describing the use of a high-shear mixer for such processes (22,23), but
so far none can be found about single-pot processors. Nevertheless, a standard pellet
formulation often contains microcrystalline cellulose, which needs high water content
to obtain a good granule/pellet quality. Drying the pellets is always a part of
the production process. The advantage of a single-pot processor is that the whole
process of pelletization and drying can be executed in the same equipment, making
322 Stahl and Van Vaerenbergh
 2005 by Taylor & Francis Group, LLC
product transfers redundant and thereby reducing the risk of product loss and contamination
and enhancing containment and operator safety.
To enhance the pelletization/spheronization process, many vendors of highshear
mixers/single-pot processors offer special mixing tools for producing pelletlike
granules. Depending on the geometry of the equipment, this special mixing tool
has either more (up to six) or less (up to two) mixing blades than a standard mixing
tool, which generally has three mixing blades (Fig. 8). All special pellet-mixing tools
have, however, the same purpose: to simulate the product behavior in a spheronizer
and enhance the spheronization process that occurs in the mixer.
4.3. Effervescent Production
The production of effervescent tablets is first of all a conventional solid dosage form
manufacturing process, which has to be taken into consideration, due to the special
characteristics of the product and some unusual features.
For granulation of effervescent products, many different production techniques
can be used, ranging from dry granulation methods over two-step granulation
(granulating acid and alkali phase separately) to one-step granulation using water or
organic solvents.
For the one-step granulation methods, the use of a single-pot processor offers
many benefits. Apart from the overall benefit of eliminating product transfer between
a granulator and a dryer, a single-pot processor allows easy solvent recovery by condensation
in case organic solvents are used as granulation liquid, compared to the
quite complex system for the exhaust gas treatment required for a fluid-bed dryer.
When water is used as granulation liquid for effervescents, the effervescent
reaction will start and cause a chain reaction. The critical point in such a process
is to stop this reaction at the correct time by evaporating the water created by this
reaction. In a single-pot processor, this can be very easily and accurately achieved
by switching on the vacuum-drying system (possibly supplemented with gas-assisted
drying or microwave drying) (24).
Figure 8 Example of a special pelletizing mixer arm. (Courtesy of Collette NV, Belgium.)
Single-Pot Processing 323
 2005 by Taylor & Francis Group, LLC
4.4. Crystallization
Although a high-shear mixer/single-pot processor is mainly intended for powder
processing, there is a possibility of executing crystallization or recrystallization processes
in this type of equipment as well. Starting either from a solution or from a
powder, which is dissolved in a suitable solvent, the speed of the drying process
can be controlled to achieve the desired crystallization process. Using vacuum drying
only, the drying process will be slow and gradual. Temperatures during this process
will remain low, creating an environment for slow crystal growth. When microwaves
are used for drying, the process will be considerably faster and the temperature of the
product will most likely be higher then under pure vacuum conditions. The crystals
that result from such a crystallization process will have different characteristics
from those from the vacuum-drying process. When choosing the appropriate settings
for the drying/evaporation process, combining vacuum and microwave drying, crystals
with specific properties can thus be obtained.
The main advantage of executing a (re) crystallizing process in a single-pot processor
is that the possibility exists of granulating the product at the same time by
varying the mixer speed during the drying process. The resulting product will be suitable
for tableting without the necessity of executing other processing steps (apart
from lubrication).
Because the rate of solvent removal during vacuum drying is dependent on a favorable
surface area/volume ratio, the drying time in a vacuum processor often increases
substantially during scale-up as described by Formula (3.2). Microwave vacuum
drying is relatively insensitive to the surface area/volume ratio and does not suffer
the same inefficiency as vacuum drying during transition from pilot to production
scale. Pearlswig et al. reported successfully scaling a microwave vacuum-drying
process for a moisture-sensitive formulation that required a drying end point of
<0.2% (25). The drying time remained within a 3045 min range throughout the
scale-up from 15 kg (Vactron.75) to 300 kg (Vactron.600), whereas the time for
vacuum drying increased threefold. After additional scale-up to 600 kg in a
Vactron.1200, the drying time rose slightly to a 5055 min range. Poska also
reported attaining equivalent drying times when scaling up in Spectrum processors
for lactosestarch granulation prepared in single-pot processors using vacuum
drying, gas-assisted vacuum drying, or microwave vacuum drying. Although not
as rapid as microwave vacuum drying, gas-assisted vacuum drying can decrease
the drying time by up to 50% compared with that of vacuum drying alone. As the
wall of the vessel is the only source for drying energy, also in this case the scale is
of major importance for the drying time (27).
When performing feasibility trials on a development- or pilot-scale single-pot
processor, it is important to be aware of the maximum energy input capacity of
for each drying method in a pilot- and production-scale single-pot processor are
shown. While for processors equipped with microwaves, the drying time is relatively
independent of the scale, a significant increase in drying time is observed for vacuumand
gas-assisted vacuum-drying processes.
324 Stahl and Van Vaerenbergh
ranging from 65 to 300 L bowl size (26). Figure 9 compares typical drying curves
the corresponding production-scale processor. In Figures 9 and 10, the drying times
 2005 by Taylor & Francis Group, LLC
Figure 10 Comparison of drying curves for different modes of drying in a 600 L Ultima-
ProTM. (Courtesy of Collette NV, Belgium.)
Figure 9 Comparison of drying curves for different modes of drying in a 75 L UltimaProTM.
(Courtesy of Collette NV, Belgium.)
Single-Pot Processing 325
 2005 by Taylor & Francis Group, LLC
As a single-pot processor is often used for highly potent products, or more generally
for contained production, it is important that also cleaning of the machine can be
executed in a contained, automated fashion to eliminate the risk of operator exposure
to the active product.
All single-pot processors on the market nowadays are equipped with a more or
less extensive clean-in-place system. Vendors have made a great effort to optimize
their design enabling easy cleaning in place, which can be validated (3). The focus
has been to eliminate any dead spots in the equipment where the cleaning water cannot
reach and including cleaning spray balls in critical product contact areas such as
the product filter and discharge valve.
Drying of the single-pot processor after the cleaning cycle to prepare the equipment
for the next batch can be done using the systems own vacuum-drying system
and jacketed bowl, making a separate drying unit redundant.
A nice case study of an evaluation of a clean-in-place system on a pilot-scale
single-pot processor is given in Ref (3). In this study it was proved that a complete
changeover from one product to the next can take place in <2 hr.
The stability of pharmaceutical granulations dried by microwaves is comparable with
that provided by alternative methods. Microwaves are nonionizing and do not
possess the amount of energy required for the formation of free radicals or the
Since the introduction of microwave drying at the end of the 1980s, numerous
new and supplemental drug applications that include the use of microwave vacuum
drying of wet granulations have been approved by the Food and Drug Administration
(FDA). We are unaware of any instance in which the FDA required additional
stability or analytical testing beyond that normally required for other methods of
manufacture. Mandal (28), Moss (29), and others (26,30) have also published or presented
data showing the comparability of the physicochemical characteristics of
granulation dried in microwave processors vs. tray driers and fluid-bed driers.
When microwaves were first introduced, some authors (31) concluded that
microwave drying could not be generally recommended because of the inability to
control the microwaves after they enter the drying cavity and the risk of unacceptable
thermal damage to active substances with high loss factors (i.e., high dielectric
constants). The routine production of pharmaceutical products by several installations
of single-pot processors using microwave vacuum drying indicates that their
general concerns can be addressed by the proper selection of formulation components
and process parameters.
Single-pot processors combine established technologies into a single piece of equipment
and, in general, deserve no special regulatory consideration when using them to
develop a new product or to manufacture an approved product. Robin and colleagues
(4) surveyed eight European regulatory agencies in 1992 to determine the
326 Stahl and Van Vaerenbergh
liberation of bound water conditions that foster product instability (Fig. 6).
 2005 by Taylor & Francis Group, LLC
implications of converting from fluid bed drying to microwave vacuum drying within
a single-pot processor. The majority of the agencies required only process validation
data and three suggested limited stability data (up to 6 months of accelerated data).
These requests were no different from those expected for similar types of manufacturing
changes (i.e., change in process or equipment).
Manufacturers considering converting to a single-pot process for an immediaterelease,
solid oral dosage form (tablets, capsules, or the like) with an approved
manufacturing process should consult their appropriate regulatory agencies
governing the practices they use to manufacture their products. For drug products
sold in the United States, manufacturers should refer to FDAs SUPAC IR
marketed products. This document describes the levels of change that may be made
in a manufacturing process and equipment. It outlines the chemistry, manufacturing,
and control tests and documentation for each level of change as well as the appropriate
regulatory filing (Annual Report, Prior Approval Supplement, or other).
For example, a tablet formulation is currently granulated using a high-shear
granulator and dried in a tray drier. A drug manufacturer wishes to replace this process
with a single-pot processor that incorporates high-shear granulation and gasassisted
vacuum drying. This conversion would be viewed as a change in equipment
to a different design and different operating principles (defined as a Level 2 Equipment
Change in the SUPAC IR Guidance document). Such a change requires the
manufacturer to submit a Prior Approval Supplement with up to three batches with
3-months accelerated stability data (depending on the duration of commercial
experience with the product). The submission would also require updated batch
records including the new equipment, and the generation of multipoint dissolution
profiles. The requirements for this conversion are, however, the same as those for
a conversion of a tray drier to a fluid-bed drying process.
Because single-pot processors combine standard engineering approaches into a single
processor, their validation should pose no special problems. Other sources adequately
describe the validation of granulating and drying processes (33), although
validation of the microwave drying system and the approach to process control of
drying end point deserves special mention.
For operational qualification of microwave components, such as forward and
reflected power, and arc detection, we suggest that customers contact the vendors
because of the specialized nature of microwave systems. The cost associated with
the calibration equipment is difficult to justify, and microwave systems should operate
reliably following proper setup and qualification and require no more periodic
maintenance than other granulation approaches.
When microwave processors were first introduced, there was the expectation
that E-field would be a reliable indicator of the drying end point. With experience,
users found that the E-field tends to be too variable, and now view it primarily as
a safety feature monitoring the microwave field within the drying cavity. Most
microwave driers on the market today do not even include E-field monitoring any
more, because of the difficulties with validation of this system.
Product temperature, time, cumulative forward power, and reflected power are
proving to be more reliable indicators of drying end point with verification by some
Single-Pot Processing 327
 2005 by Taylor & Francis Group, LLC
Guidance document (32), which addresses scale-up and postapproval changes for
in-process control that directly measures the moisture content of a product sample.
The industrial processes in operation today all use one of these approaches for endpoint
With the pharmaceutical industry moving toward the use of process analytical
technology (PAT), however, the equipment vendors are also investigating the possibility
of applying this technology to single-pot processors and thereby eliminating
the necessity of taking product samples. The first single-pot processors with PAT will
undoubtedly be introduced into the market shortly.
Users of single-pot processors may use one or all of its processing features (mixing,
granulating, drying, or other). The accompanying data acquisition system must collect,
display, and record the relevant processing conditions and granulation behavior
for each cycle. The degree of sophistication of the system may depend on the venue
in which the processor is used.
In a development setting, the sequence of cycles is often interrupted to collect
samples for analysis, and the user is interested in capturing as much information as
possible to assist in defining a suitable processor for a particular formulation. In
many cases, the control system for a development environment can be limited to a
manual control system. The data acquisition system, however, needs to be more
sophisticated to allow registration of all relevant processing data.
In production, the manufacturing sequence and process parameters are prede-
fined and validated, and information needs are reduced to monitoring critical parameters.
In these settings, an automatic control system with recipes to reduce operator
interventions is indispensable. The data acquisition system may or may not be as
sophisticated as for the development settings, depending on the requirements for
data in the batch records.
For the purposes of trouble-shooting and trend analysis of production
operations, however, the information requirements of development and production
converge, and data acquisition systems for single-pot processors should seek to
address the needs of both types of users.
All data acquisition systems are considered to generate electronic records. If
these electronic records are used as batch documentation, the acquisition system
needs to be compliant with FDA 21 CFR part 11.
Most vendors have addressed these issues by including password control, audit
The primary safety concern during the granulation and drying processes is the prevention
of explosions. Bulk powder, dust clouds, and flammable vapors, all have the
potential to explode. Adequate grounding and ventilation during loading and discharging
of the vessel and controlling the various processing conditions can reduce
the risk of explosion.
In Europe, all equipment used in a potentially explosive atmosphere needs to
be compliant with the ATEX guidelines (34).
Two approaches are generally taken toward explosion protection in single-pot
processors. One consists in removing oxygen from the processing chamber and
328 Stahl and Van Vaerenbergh
trails, and point verification systems (Fig. 11).
 2005 by Taylor & Francis Group, LLC
replacing it with an inert gas (e.g., nitrogen) before any mixing action can take place.
The removal of oxygen reduces the risk of explosion by eliminating one of the elements
necessary to create an explosion. The other approach is to design the equipment in
order to contain the explosion. Executions of 10 or even up to 16 bar of high-shear
granulators and single-pot processors are now becoming available on the market.
Apart from the measures taken to avoid or contain explosions within the processing
chamber, the electrical, electronic, and mechanical parts of single-pot processors
that will be used in potentially explosive atmospheres also need to be explosion
The leakage of microwave energy is a concern for single-pot processors that use
this drying approach. Industrial microwave processors are expected to meet the
guidelines for microwave leakage specified by the Center for Devices and Radiological
Health within FDA and by the American National Standards Institute (35,36).
The guideline is 5 mW/cm2 maximum exposure at a frequency of 2450 MHz at a
distance of 5 cm from any surface of the microwave cavity. Survey meters for the
detection of microwave leakage are relatively inexpensive and should be purchased
by users of single-pot processors that incorporate microwave drying. The survey
meters are calibrated before shipment and returned to the supplier for recalibration
at periodic intervals. Their use should be incorporated in standard operating procedures
for the equipment. Operator readings that exceed the guideline limit are often
indicative of deteriorating seals around the lid cavity.
In addition to energy leakage standards, microwave processors are designed
with safety interlocks to prevent accidental exposure. For example, the magnetrons
can be activated only if the microwave cavity (i.e., bowl of the processor) is operating
under vacuum, usually 30100 mbar. If the vacuum falls outside this range, as in the
unlikely event that an operator inadvertently tries to open the lid during microwave
vacuum drying, the magnetrons are disabled. Vendors also incorporate additional
safeguards to ensure that the microwave power is disabled with access to the bowl.
Because of popular misconceptions about the use of microwave ovens (e.g.,
stainless steel should not be used in a microwave cavity) and electromagnetic radiation
(e.g., all types cause biological effects), a training program should be instituted
Figure 11 Screen of a typical control system for a single-pot processor. (Courtesy of Collette
NV, Belgium.)
Single-Pot Processing 329
 2005 by Taylor & Francis Group, LLC
in any facility that uses microwave drying. This will demystify any unfounded concerns
about the technology and foster a rational approach to a sound safety and
maintenance program.
Over the last 30 years since its introduction, single-pot processing has developed into
a mature and generally accepted production technique.
Even if the technique historically was often used because of its specific advantages
for effervescent production, potent compounds, organic solvents, or multiproduct
facilities, practice has shown that single-pot processing is also attractive for
standard pharmaceutical solid dosage production.
1. Van Vaerenbergh G. The influence of a swinging bowl on granulate properties. Pharm
Technol 2001; 13(3):3643.
2. Lucisano LJ, Poska RP. Microwave technology-fad or the future. Pharm Technol 1990;
3. Van Vaerenbergh G. Cleaning validation practices using a one-pot processor. Pharm
Technol. Europe. Feb 2004.
4. Robin P, Lucisano LJ, Pearlswig DM. Rationale for the selection of a single pot manufacturing
process using microwave/vacuum drying. Pharm Technol 1994; 18:2836.
5. Bellini G, Pellegrini L. Non adiabatic drying. In: Goldberg E, ed. Handbook of Downstream
Processing, Chapman & Hall, 1993.
6. De Smet P. Vacuum drying. Manuf Chem Mar 1989; 3739.
7. LB Bohle, Inc. Technical Report. Granulation and Drying with SYSTEM-VAGAS.
Bristol, PA: LB Bohle, Inc.
8. Niro-Fielder Ltd. Technical Report. AEROVAC System, Accelerated Vacuum Drying.
Eastleigh, Hampshire, U.K.: Niro-Fielder Ltd.
9. Metaxas AC, Meredith RJ. Industrial Microwave Heating. London: Peter Peregrines,
10. Doyle C, Cliff MJ. Microwave drying for highly active pharmaceutical granules. Manuf
Chem Feb 1987; 2332.
11. Waldron MS. Microwave vacuum drying of pharmaceuticals: the development of
a process. Pharm Eng 1988; 8:913.
12. Poska RP. Microwave processing: the development experience revisited. Proceedings of
International Society of Pharmaceutical Engineers Congress, September, 1992.
13. Lucisano U, Moss RA. Vacuum drying vs. microwave-vacuum drying in three pilot-scale
single pot processors using sodium acid pyrophosphate as the model granulation. AAPS
Annual Meeting, San Antonio, TX, Nov 15, 1992.
14. Schaefer T. Melt Agglomeration with Polyethylene Glycols in High Shear Mixers. Ph.D.
thesis, The Royal Danish School of Pharmacy, Copenhagen, 1996.
15. Schaefer T, Holm P, Kristensen HG. Melt pelletisation in a high shear mixer. I. Effects of
process variables and binder. Acta Pharm Nordica 1992; 4:133140.
16. Schaefer T, Mathiesen C. Melt pelletisation in a high shear mixer. IX. Effects of binder
particle size. Int J Pharm 1996; 139:139148.
17. Van Vaerenbergh G. Melt granulation with polyethylene glycol in a one-pot processor.
Pharma. In press.
330 Stahl and Van Vaerenbergh
 2005 by Taylor & Francis Group, LLC
18. Heng PW, Wong TW, Chan LW. Influence of production variables on the sphericity of
melt pellets. Chem Pharm Bull (Tokyo) 2000; 48:420424.
19. Hamdani J, Moe?s AJ, Amighi K. Development and evaluation of prolonged release
pellets obtained by the melt pelletization process. Int J Pharm 2002; 245:167177.
20. Thies R, Kleinebudde P. Melt pelletisation of a hygroscopic drug in a high shear mixer.
Part 1. Influence of process variables. Int J Pharm 1999; 188(2):131143.
21. Voinovich D, Moneghini M, Perisutti B, Franceschinis E. Melt pelletization in high shear
mixer using a hydrophobic melt binder: influence of some apparatus and process variables.
Eur J Pharm Biopharm 2001; 52(3):305313.
22. Vonk P, Guillaume CPF, Ramaker JS, Vromans H, Kossen NWF. Growth mechanisms
of high-shear pelletisation. Int J Pharm 1997; 157(1):93102.
23. Ramaker JS, Albada Jelgersma M, Vonk P, Kossen NWF. Scale-down of a high-shear
pelletisation process: flow profile and growth kinetics. Int J Pharm 1998; 166(1):8997.
24. Stahl H. Manufacturing effervescent tablets. Pharm Technol 2003; 15(4):2528.
25. Pearlswig DM, Robin P, Lucisano LJ. Simulation modeling applied to the development
of a single-pot process using microwave/vacuum drying. Pharm Technol 1994; 18:4460.
26. Poska R. Integrated mixing granulating and microwave drying: a development experience.
Pharm Eng 1991; 11:913.
27. Stahl H. Single pot systems for drying pharmaceutical granules. Pharm Technol Eur
2000; 12(5):2334.
28. Mandal TK. Evaluation of microwave drying for pharmaceutical granulations. Drug
Dev Ind Pharm 1995; 21:16831688.
29. Moss RA. Demonstration-microwave/vacuum drying of pharmaceuticals. AAPS
Annual Meeting, Washington, DC, Nov 17, 1991.
30. Van Scoik K. Microwave vacuum processing in the Vactron 300. AAPS Annual Meeting,
Washington, DC, November 1991.
31. Duschler G, Carius W, Bauer KH. Single-step granulation method with microwaves:
preliminary studies and pilot scale results. Drug Dev Ind Pharm 1995; 21:15991610.
32. Food and Drug Administration. Immediate release solid oral dosage forms: scale-up and
postapproval changes: chemistry, manufacturing, and controls; in vitro dissolution testing;
in vivo bioequivalence documentation; guidance. Fed Reg 1995; 60:6163861643.
33. Berry IR, Nash eds. Pharmaceutical Process Validation. New York: Marcel Dekker,
34. Directive 94/9/EC of the European Parliament and the Council of 23 March 1994 on the
approximation of the laws of the Member States concerning equipment and protective
systems intended for use in potentially explosive atmospheres.
35. Performance Standards for Microwave and Radio Frequency Emitting Products. 21
Code of Federal Regulations, Part 1030. Washington, DC: General Services Administration,
April 1, 2003.
36. IEEE Standard for Safety Levels with Respect to Human Exposure to Radio Frequency
Electromagnetic Fields, 3 kHz to 300 GHSupplement 1999 (IEEE C95.11999). New
York: Institute of Electrical and Electronics Engineers.
Single-Pot Processing 331
 2005 by Taylor & Francis Group, LLC
Extrusion/Spheronization as a Granulation
Ketan A. Mehta
RO?HM Pharma, Degussa Corp., Piscataway, New Jersey, U.S.A.
Gurvinder Singh Rekhi
Elan Drug Delivery Inc., Gainesville, Georgia, U.S.A.
Dilip M. Parikh
Synthon Pharmaceuticals Inc., Research Triangle Park, North Carolina, U.S.A.
Extrusion/spheronization is a multiple step process capable of making uniformly
sized spherical particles. The process is now being widely utilized in the pharmaceutical
industry. This is the revision of the original chapter written by D. Erkoboni
(1). As a pharmaceutical dosage form, pellets are defined as small, free flowing,
spherical, or semispherical units made up of fine powders or granules of bulk drugs
and excipients by variety of processes, extrusion and spheronization being one. It
is primarily used as a method to produce multiparticulates for immediate and controlled
release applications. The major advantage over other methods of producing
drug-loaded spheres or pellets is the ability to incorporate high levels of actives without
producing an excessively large particle. Mehta and Kislalioglu in their study
demonstrated the incorporation of a poorly soluble drug in a pellet matrix up to
40% loading via extrusion/spheronization for controlled drug delivery (2).
Though the process is more efficient than other techniques for producing
spheres, it is more labor and time intensive than the more common granulation techniques.
Therefore, it should be considered as a granulating technique when the
desired particle properties are essential and cannot be produced using more conventional
techniques. However, more recently, pharmaceutical scientists worldwide have
been able to use this method more easily due to advances in extrusion/spheronization
equipment engineering, thus making it simple to use. Pellets offer scientists a great
deal of development flexibility. Chemically incompatible ingredients, for instance,
can be incorporated into single capsule using two pellet types each containing
one of the incompatible ingredients. Pellets of different release characteristics can
be combined to achieve the desired release pattern of the active ingredients. Pellets
are characterized by a low surface area-to-volume ratio compared with powder or
 2005 by Taylor & Francis Group, LLC
granules, which provides excellent coating substrate. Typically, pellets range in diameter
between 0.25 and 1.5 mm. Pellets are normally filled into hard gelatin capsules,
or eventually compressed into tablets which disintegrate into individual
pellets after oral intake.
Spheronization is a process invented by Nakahara in 1964. The patent
describes a Method and Apparatus for Making Spherical Granules from wet powder
mixtures (3). The equipment described in the patent was commercialized by Fuji
Denki Kogyo Co. under the trade name Marumerizer. The process went widely
unnoticed in the pharmaceutical industry until 1970 when two articles were published
by employees of Eli Lilly and Co. Conine and Hadley described the steps
involved in the process including (a) dry blending, (b) wet granulation, (c) extrusion,
(d) spheronization, (e) drying, and (f) screening (optional) (4). Reynolds went on to
further describe the equipment and the mechanics of the process including the movement
of the particles within the spheronizer (5). Both publications cite desirable
product attributes that can be achieved, including good flow, low dusting, uniform
size distribution, low friability, high hardness, ease of coating, and reproducible
packing. Additionally, the resulting pellets offer not only technological advantages
as mentioned before but also therapeutic advantages such as less irritation of the gastrointestinal
tract and a lowered risk of side effects due to dose dumping and reproducibility
of the drug blood levels (6). From the publication of these articles through
present day the interest in extrusion/spheronization has continued to grow. The process
has recently become established in industry but was primarily driven by academia
in the interim. The increased popularity in recent years is, in part, due to a growing
understanding of the effects of process parameters and material characteristics.
In recent times, hot-melt extrusion (HME) has gained subsequent industry and
academic attention and is in the phase of further process maturity for ultimately
gaining wide spread popularity. HME is somewhat widespread in the plastics industry
for the production tubes, pipes, wires, and films. For pharmaceutical systems,
this method has been used to prepare granules, sustained release tablets, and transdermal
drug delivery systems (7). The advantage of HME is that it does not require
the use of solvents and water and few processing steps are needed making the process
somewhat simpler, efficient, and continuous. The disadvantage of HME is that it
may use complicated know-how and typically employs high temperatures around
and over 100C as a processing requirement. However, it has been used as a method
to increase the solubility of a poorly soluble drug, to taste mask a bitter drug and
overall for controlled release dosage form purposes. The bioavailability of the drug
substance has been demonstrated to improve when it is dispersed at the molecular
level in hot-melt extruded dosage forms. Several examples of melt-extruded molecular
dispersions were presented by Breitenbach and Ma?gerlein (8).
A recent patent (9) describes melt- extrusion process as being used to produce
improved controlled release characteristics. The authors described the process which
includes mixing together a therapeutically effective agent, a water-insoluble retardant,
and a binder to form a homogeneous mixture, heating the mixture and thereafter
extruding the mixture into strands. The strands are then cooled, reduced to the
desired sizes.
The number of hot-melt extrusion patents issued for pharmaceutical systems
has increased more than sixfold annually since the early 1980s, with the United
States, Germany, and Japan leading the field. The field of melt extrusion technology
is growing very rapidly and its use to produce films for transmucosal and transdermal
drug delivery applications was presented by McGinity and Repka (10).
334 Mehta et al.
 2005 by Taylor & Francis Group, LLC
Potential applications are many including both immediate and controlled release. In
a paper by Sood et al., extrusion/spheronization method was used to develop controlled
release dosage forms for diltiazem hydrochloride (11). Two or more actives
can easily be combined in any ratio in the same unit. These combination products
can contain actives that are incompatible or have varying release profiles. Spheres
can be used as a method to limit drug migration. Physical characteristics of the active
ingredients and excipients can be modified to improve physical properties and downstream
processing. As an example, a low-density, finely divided active can be pelletized
to increase density, improve flow, and limit dusting (12). Mehta et al. showed
that this technique can be used for the development of zero order controlled release
drug delivery of poorly soluble drugs (13). Functional coatings can be applied easily
and effectively. Dense multiparticulates disperse evenly within the gastrointestinal
tract and can be used to prolong gastrointestinal transit times (14,15) or to improve
tolerance of some compounds. Regardless of the application, care must be taken to
achieve the required sphere or granule properties.
Spheres for controlled release coating applications will likely have significantly
different physical requirements than granules for compression. A product to be
coated for controlled release should have a uniform size distribution, good sphericity
and surface characteristics as well as low friability. Once coated, the sphere should
have the desired release characteristics. Additionally, if the coated spheres are to
be compressed into tablets, they will require sufficient strength to withstand the
forces of compression. Upon disintegration of the tablet, the individual spheres must
retain their original release profile. Physical properties such as flow, density, friability,
porosity, and surface area become important for granules intended for compression
into tablets. The granules should have good deformation and bonding
characteristics to form tablets having desirable physical properties. Drug release
from the final dosage form must meet the target specification.
Product produced using extrusion/spheronization can range from barely
shaped, irregular particles with physical properties similar to a conventional granule
to very spherical particles having properties that are drastically different (16). Tableting
characteristics can be modified by altering the composition of the spherical particles
(17), granulating fluid (18), or the process conditions used to produce them
(19). Compaction studies conducted on spheres similar to those used for controlled
release applications show the bonding and densification that occur during extrusion/
spheronization can alter the deformation characteristics of some materials (18).
Microcrystalline cellulose (MCC), which deforms plastically in the dry powder state,
exhibits elastic deformation followed by brittle fracture once spheronized (17). The
deformation characteristics, coupled with the larger size particles result in reduced
bonding sites and the production of weak compacts. A compaction profile of
The point is not to dwell on the properties required for each application, but
rather to reinforce the fact that each application will have very specific requirements.
One must first understand the properties required and then tailor the process to
yield the desired effects. The effects of process and formulation variables will be
discussed next.
A review of the literature shows that most investigators have tried to understand
small components of this process isolated from other effects. They have
focused on particular formulation or process parameters. It is valuable to have
Extrusion/Spheronization as a Granulation Technique 335
MCC and spheres prepared from MCC is shown in Figure 1.
 2005 by Taylor & Francis Group, LLC
a detailed understanding of the main variables; however, this approach fails to take
into consideration the high degree of interaction that exists between the variables.
The use of statistical experimental design is a valuable tool to understand not
only the main effects but also the interactions that can have a profound effect on
the characteristics of the resulting particles (2022). Additionally, these techniques
are extremely useful during product/process development to understand the effect
of variables and control them to produce product having desired attributes (23).
After pointing out the benefits of design methodology in this application, it should
be understood that, for simplicity, much of the discussion to follow will address the
various topics individually. In reality, however, they truly cannot be isolated from
one another. This chapter will review and discuss the general process, equipment
types, and the effect of process and formulation variables on the properties of
spherical granules.
Extrusion/spheronization is a process requiring at least five units of operation with
an optional sixth screening step. First, the materials are dry mixed (i) to achieve a
homogeneous powder dispersion and then wet granulated (ii) to produce a suffi-
ciently plastic wet mass. The wet mass is extruded (iii) to form rod-shaped particles
of uniform diameter that are charged into a spheronizer and rounded off (iv) into
spherical particles. The spherical particles are then dried (v) to achieve the desired
moisture content and optionally screened (vi) to achieve a targeted size distribution.
cess steps along with critical variables associated with them (24). The end product
Figure 1 Compaction profiles ofmicrocrystalline cellulose powder and spheres. (FromRef. 17.)
336 Mehta et al.
The process flow diagram, shown in Figure 2, has been used to show each of the profrom
each of the steps is shown in Figure 3.
 2005 by Taylor & Francis Group, LLC
4.1. Dry Mixing
During the first step, powders are dry mixed to achieve a uniform dispersion prior to
wet granulation. It is generally carried out in the same mixer used for the granulation;
Figure 2 Process flow chart of the extrusion/spheronization process showing the process
variables for each individual step. (From Ref. 24.)
Figure 3 Product produced by the first four extrusion/spheronization process steps.
(A) Powder from dry mixing; (B) granules from granulation; (C) extrudate from extrusion;
and (D) spheres from spheronization.
Extrusion/Spheronization as a Granulation Technique 337
 2005 by Taylor & Francis Group, LLC
however, if a continuous granulator is used, a separate mixer is required for the
dry mix. This step is typically taken for granted because wet massing follows.
The uniformity of the dry mix, however, can have a significant effect on the quality
of the granulation and, in turn, the spherical particles produced. An uneven distribution
of materials having wide differences in properties such as size and solubility
can result in localized over wetting, at least initially, during the granulation step.
The more soluble and finely divided components can also dissolve and become part
of the granulating fluid. The fluids, rich in soluble compounds, can either remain
as overwet regions or, with continued wet massing, can be redistributed (25).
Sphere uniformity (size and shape) is very much dependent on the uniform distribution
and composition of the granulating fluid which include not only the solvent
but also any dissolved ingredients.
4.2. Granulation
The second step is granulation, during which a wet mass having the requisite plasticity
or deformation characteristics is prepared. With a few exceptions, this step is
similar to conventional granulation techniques used to produce product for compression.
It is typically carried out in a batch type mixer/granulator; however, any
equipment capable of producing a wet mass, including the continuous type, can
be used. Batch type processors include planetary mixers, vertical or horizontal high
shear mixers, and sigma blade mixers. Examples of continuous mixers include the
Nica M6 instant mixer (26) and high shear twin screw mixer/extruders (27). The high
shear twin screw mixer/extruders have mixer/feeders that are capable of shearing
and kneading the feed materials. Dry powders and fluids are fed in through separate
ports and mixed by the action of the extruder blades and screws. The mixer/extruder
is capable of being configured to customize the amount of shear and energy used in
the process by changing the configuration of the mixing blades. This can have an
impact on the properties of the extrudate produced (28). As with the batch processors
it is critical to achieve a uniform level of fluid within the wet mass. The proper
fluid/solids ratio is accomplished by maintaining a steady powder and fluid feed into
the mixer/extruder. Both are critical; however, the powder feed is the most
problematic. Small variations in feed rates can cause significant shifts in the moisture
content of the granulation and, therefore, the quality of the spherical particles produced.
The two major differences in the granulation step, as compared to typical
granulations for compression are the amount of granulating fluid required and the
importance of achieving a uniform dispersion of the fluid. The amount of fluid needed
to achieve spheres of uniform size and sphericity is likely to be greater than that for
a similar granulation intended for tableting. Instruments such as a ram extruder (29)
and a torque rheometer (30) have been used to characterize the flow characteristics
of granulations for use in extrusion/spheronization. They are useful tools in quantifying
the rheological effect of formulation and process variations in the granulation.
The ram extruder has been used to characterize the flow of wet masses through a die,
which has been divided into stages. They are: (a) compression, where the materials
are consolidated under slight pressure, (b) steady-state flow, where the pressure
required to maintain flow is constant, and (c) forced flow, where an increase in force
is required to maintain flow. The three stages are shown in the force vs. displacement
The change from steady state to forced flow is caused by the movement of fluid
under pressure. Extrusion in a ram extruder is continuous, and this phenomenon is
338 Mehta et al.
profile in Figure 4.
 2005 by Taylor & Francis Group, LLC
less likely to be seen in extruders that are discontinuous such as gravity-feed models
Regardless of the mixer used, one must remember that the downstream process
steps of extrusion and spheronization are very dependent on the level of water contained
in the granulation and the quality of its dispersion. High-energy mixers such
as high shear mixers and high shear twin screw mixer/extruders can cause a signifi-
cant rise in temperature. It may be necessary to use a jacket to guard against heat
build-up. High temperatures can result in a greater than tolerable level of evaporation
(33) or an increase in the solubility of some of the solids. A reduction in fluid
will reduce the plasticity of the granulation, while an increase in solubility will
increase the weight ratio of granulating fluid since the solute is then part of that fluid
(34). The water solubility of the drug in the granulation plays a key role in determining
granulation end point for extrusion/spheronization process. A highly water soluble
drug will dissolve in the granulation whereas a highly insoluble drug will have
wetting problems during the granulation step. Upon extrusion and during spheronization
step a granulation containing a highly insoluble drug at a high dose will have
a higher tendency to release moisture due to which the moisture will migrate to the
surface of the extrudates, which might cause interpellet sticking. In order to avoid
this Mehta et al. in their work demonstrated the use of small quantities of talc to
adsorb the surface moisture which helped in the spheronization step without altering
the drug release from the resulting pellets (2).
4.3. Extrusion
The third step is the extrusion step which forms the wet mass into rod-shaped particles.
The wet mass is forced through dies and shaped into small cylindrical particles
having a uniform diameter. The extrudate particles break at similar lengths under
Figure 4 A forcedisplacement profile for a microcrystalline celluloselactosewater mixture
showing the three stages of extrusion on a ram extruder: Compression, steady-state flow, and
forced flow (ram speed, 4 mm/sec; die diameter, 1.5 mm; L/R ratio, 12). (From Ref. 31.)
Extrusion/Spheronization as a Granulation Technique 339
(32). A diagram of a ram extruder is shown in Figure 5.
 2005 by Taylor & Francis Group, LLC
their own weight. The extrudate must have enough plasticity to deform but not so
much to adhere to other particles when collected or rolled in the spheronizer.
Extruders come in many varieties but can generally be divided into three classes
based on their feed mechanism. They include those that rely on a screw, gravity or a
piston to feed the wet mass into the extrusion zone (35). Examples of extruders from
each class are shown in Figure 5. Screw feed extruders include the (a) axial or end
plate, (b) dome, and (c) radial type, while gravity-feed extruders include (d) cylinder,
(e) gear, and (f) radial types. The screw and gravity-feed types are used for development
and manufacturing with the radial varieties being the most popular for pharmaceutical
applications. The piston feed or ram extruder is primarily used in
research as an analytical tool.
Figure 5 Schematic diagrams of extruder types used in extrusion/spheronization.
340 Mehta et al.
 2005 by Taylor & Francis Group, LLC
Screw extruders have either one (single) or two (twin) augers that transport the
wet mass from the feed area to the extrusion zone. During the transport process, the
screws compress the wet mass removing most of the entrapped air. Studies have been
conducted on the ram extruder to understand this compression or consolidation
stage. They have shown the apparent density of the wet mass plug prior to extrusion
is approximately equal to the theoretical apparent particle density, indicating that
nearly all of the voids were eliminated (31). Twin screw extruders generally have a
higher throughput than single screw models, while single screw extruders compress
and increase the density of the extrudate more. Other features that can affect the
density of the extrudate are the spacing of the turnings on the screw and the space
between the end of the screw and the beginning of the die (36). Turnings that are
wide and regularly spaced minimize the amount of compression during material
transport. Screws with closer or progressively closer spacing between the turnings
will result in more compression and produce a denser extrudate. Space between
the screw and the die results in a void into which material is deposited and compressed.
The greater the space, the more compression takes place prior to extrusion.
As material builds up, pressure increases and causes the material to be forced, under
hydraulic pressure, to flow through the die. When space between the screw and the
die is at a minimum, extrusion takes place as material is compressed in the nip,
between the extruder blade and the die.
The primary difference between the various types of screw extruders is in the
extrusion zone. An axial or dome extruder transports and extrudes the wet mass
in the same plane. Axial extruders force the wet mass through a flat, perforated
end plate, typically prepared by drilling holes in a plate. The thickness of the plate
can be more than four times the hole diameter, resulting in high die length to radius
(L/R) ratios. An axial extruder is shown in Figure 6(A).
Dome extruders use a dome or half sphere shaped screen as the die. It is prepared
by stamping holes in metal stock having a similar thickness as the hole diameter.
This results in a die L/R ratio close to 2, however, variations in screen
thickness are possible resulting in a slightly higher or lower ratio. A dome extruder
is shown in Figure 6(B).
Unlike axial and dome extruders, radial extruders extrude the wet mass perpendicular
to the plane of transport. Material is transported to the extrusion zone where
it is wiped against the screen die by an extrusion blade. The mass is forced through
the die by pressure generated at the nip. A screw feed radial extruder is shown in
Figure 6(C).
Figure 6 Types of extruders. (A) Axial; (B) dome and (C) radial. (Courtesy of the LCI
Extrusion/Spheronization as a Granulation Technique 341
 2005 by Taylor & Francis Group, LLC
As with dome type extruders the die is a stamped screen. Due to the shorter die
lengths and the increase number of holes or dies, dome and radial extruders have the
advantage of higher throughput as compared to the axial type.
As with almost every step in extrusion/spheronization, heat build-up during
extrusion is a significant concern. This is especially true of the screw fed extruders.
Axial extruders generate heat due to their long die lengths. Radial extruders can have
a significant heat differential over the width of the screen. Materials fed into the
extrusion zone will have the lowest temperature. However, as material moves to
the front of the zone, the temperature increases due to the longer residence time
of material. Of the screw feed extruders, the dome type has the highest rates and
is least likely to generate significant heat over an extended period.
Gravity-feed extruders include a cylinder, gear, and radial type. The cylinder
and gear both belong to a broader class referred to as roll extruders. Both use two
rollers to exert force on the wet mass and form an extrudate. The cylinder extruder
has rollers in the form of cylinders, one solid and one hollow with drilled holes to
form the dies. The wet mass is fed by gravity into the nip area between the two cylinders
and forced through the dies into the hollow of the cylinder. Gear type extruders
have rollers in the form of hollow gears. The dies are holes drilled at the base of each
tooth. Wet mass is forced through the holes and collected in the hollow of the gears
as the teeth and the base areas mesh. The last type of gravity-feed extruder to be discussed
is the radial type. One or more arms rotate to stir the wet mass as it is fed by
gravity. Rotating blades wipe the wet mass against the screen, creating localized
forces sufficient to extrude at the nip. There is no compression prior to extrusion
which is the major difference between the gravity and screw feed radial extruders.
The primary extrusion process variables are the feed rate, die opening, and die
length. The water content of the granulation is also very critical, since the properties
of the extrudate and resulting spheres are very dependent on the plasticity and cohesiveness
of the wet mass. The process variables and water content have been the
focus of many studies. Harrison, Newton, and Rowe studied the flow of the wet mass
as it is forced through a die (29,31,37,38). They determined that steady-state flow
which results in uniformly sized spherical particles having good sphericity and surface
characteristics. Materials and processes that did not result in steady state, a
condition referred to as forced flow, produced extrudate having surface impairments.
In moderate cases, the surface is rough, while in more severe cases, a phenomenon
commonly referred to as shark-skinning occurs. Examples of smooth extrudate
Forcedisplacement profiles of microcrystalline cellulose (MCC) and water at
various ratios, MCC, lactose, and water at a 5:5:6 ratio, and lactose and water at a
Steady state was possible with the MCC and MCC:lactose samples but not
with lactose alone. As can be seen with the MCC samples, the duration of the compression
stage was water level dependent with no effect seen on the steady state stage.
Additional studies indicated the effect of ram speed (extrusion speed) and die L/R
ratio. An increase in ram speed increased duration of the steady-state stage with
no effect on the compression stage. The L/R ratio had no effect on either compression
or steady state. Wet mass composition, therefore, influenced the ability
to achieve steady state while the water level and ram speed influenced duration.
Higher water levels decreased the force to produce steady-state flow but increased
342 Mehta et al.
A gravity-feed extruder is shown in Figure 7.
(described earlier and shown in Fig. 4) was essential to produce a smooth extrudate,
and shark-skinned extrudate are shown in Figure 8(A) and (B), respectively.
8:2 ratio, developed by Harrison et al., are shown in Figure 9.
 2005 by Taylor & Francis Group, LLC
Figure 7 (A) Gravity-feed rotary extruder; (B) close-up showing the extrusion zone;
(C) operating principle of gravity-feed extruder. (Courtesy of Niro Pharma Systems.)
Figure 8 Scanning electron micrographs showing an example of (A) smooth extrudate and
(B) extrudate having surface impairment, or shark-skinning.
Extrusion/Spheronization as a Granulation Technique 343
 2005 by Taylor & Francis Group, LLC
the duration. Faster ram speeds (extrusion rates) increased the duration of steady
state and increased the force. As discussed in the following text, other investigators
have reported the correlation between extrusion force and sphere quality.
Fechner et al. indicated that for an optimum extrusion process, a mixture of
spongelike and gel-like behavior might be desirable which are possible by the use
of MCC and powder cellulose (39).
Harrison et al. also indicated that a uniform lubricating layer at the die wall
interface must occur to eliminate the slip-stick phenomenon responsible for forced
flow. Development of a lubricating layer was dependent on the length of the die
(a minimum length required), wall shear stress and upstream pressure loss. They
represent the frictional forces at the die wall interface and the estimated pressure loss
at zero die length in the barrel of the ram extruder. The method for deriving these
values is described in Ref. 27. These parameters allow for a quantitative comparison
between formulations and process; however, no specific values can be targeted since
they vary with materials.
Pinto et al. also showed that, at slow ram speeds, water moves toward the die
wall interface and acts as a lubricant resulting in reduced extrusion forces. At higher
speeds water is unable to move rapidly through the mass resulting in higher forces
(40). They indicated that the water content and its distribution are critical in determining
the particle size and sphericity of the product. Lower water content and
higher speed will reduce the size and sphericity of the particles. The extrusion speed
and water content should be adjusted to achieve the desired effect. Other researchers
have investigated the effect of die length using gravity-feed radial extruder. Helle?n
et al. indicated the extrudate became smoother and more bound as the L/R ratio
Figure 9 Forcedisplacement profiles at various moisture contents of mixtures of microcrystalline
cellulose and water: (ad) microcrystalline celluloselactosewater (5:5:6);
(e) lactosewater (8:2); (f) at a ram speed of 4 mm/sec, die diameter of 1.0, and a L/R ratio
of 12. Percentage of moisture content of microcrystalline cellulosewater mixture: a, 59.4;
b, 51.1; d, 45.0. (From Ref. 29.)
344 Mehta et al.
 2005 by Taylor & Francis Group, LLC
of the die was increased (41). Vervaet et al. reported that a higher L/R ratio enables
the use of lower water levels to achieve a more bound extrudate (42). This also
increased the range (drug loading and water level) over which quality spheres could
be produced. They attributed the increased latitude and capability to increased densification
and resulting well-bound extrudate. The average pore diameter and bulk
density reported for extrudate prepared from various MCC:DCP:water ratios at
two L/R ratios are shown in Table 1.
Baert et al. also indicated a similar increase in latitude when a cylinder extruder
having an L/R ratio of 4 was compared to a twin screw extruder having a L/R ratio
of close to 1.8 (43). Other studies have shown that there is an optimal pressure range
over which extrudate capable of yielding acceptable spheres can be produced. Shah
et al. demonstrated the correlation between screen pressure yield and density (44).
A high yield of spheres within a targeted narrow size distribution was produced as
long as the screen pressure was maintained within a given range. The relationship
While many of the researchers have indicated a need for a more cohesive extrudate,
few have expressed a need to remove all surface impairments. Some researchers
have indicated that spheres having acceptable characteristic can be produced from
extrudate having shark-skinning. OConnor and Schwartz have found the presence
of shark-skinning to be advantageous in facilitating the breakage of the extrudate
during the spheronization step (45).
Experimental design studies conducted to concurrently investigate the effect of
extrusion as well as other process and formulation variables have indicated the extrusion
variables to be less significant than granulating fluid level or variables of the
spheronization step. Hasznos et al. determined that extruder speed had little effect
on the size distribution of the final product or moisture change during processing
as compared to the spheronization variables (46). Hilemann et al. indicated that,
when water/MCC ratios are held constant, a change in screen size results in a significant
change in the size distribution (47). However, in a study where water level was
included as a variable, Erkoboni et al. have shown that the effect of screen size on
size distribution is small compared to the effect of a change in water level. A change
in water level can shift the mean size and still result in an acceptable distribution (21).
This is in agreement with earlier work by Malinowski and Smith who also showed
the mean particle size is typically smaller than the size of the screen itself due to
shrinking during the drying step (12). Vervaet et al. have presented an excellent
review of extrusion spheronization (48).
Table 1 Average Pore Diameter and Bulk Density of Extrudate Composed of DCPAvicel
PH-101Water Mixture, Extrudated Using Screens with Different L/R Ratios
Composition DCP
Avicelwater (w/w) L/R ratio of screen
Average pore
diameter (mm) Bulk density (g/mL)
150:380:470 4 0.982 1.132
150:400:450 4 0.992 1.211
150:380:470 2 1.249 0.949
150:400:4502 2 1.292 0.947
Source: From Ref. 42.
Extrusion/Spheronization as a Granulation Technique 345
between yield and screen pressure is shown in Figure 10.
 2005 by Taylor & Francis Group, LLC
4.4. Spheronization
The fourth step in the extrusion/spheronization process is the spheronization step. It
is carried out in a relatively simple piece of equipment. The working parts consist of
a bowl having fixed sidewalls with a rapidly rotating bottom plate or disk. The
rounding of the extrudate into spheres is dependent on frictional forces. The forces
are generated by particle-to-particle and particle-to-equipment interactions. For this
reason the disk is generally machined to have a grooved surface which increases the
forces generated as particles move across its surface. Disks having two geometric
patterns are produced, a cross-hatched pattern with the grooves running at right
angles to one another and a radial pattern with the grooves running radially from
the center. The two varieties are shown graphically in Figure 11.
Some studies have shown the rate of spheronization to be faster with the radial
pattern; however, both plates will result in acceptable product (35).
During the spheronization step, the extrudate is transformed from rod-shaped
pellets into spherical particles. This transition occurs in various stages. Once charged
Figure 10 The effect of extruder screen pressure on the yield of particles within an acceptable
distribution. (From Ref. 44.)
Figure 11 Spheronizer disks having two geometric patterns: (A) a cross-hatched pattern
with the groves running at right angle to one another and (B) a radial pattern with the groves
running radially from the center.
346 Mehta et al.
 2005 by Taylor & Francis Group, LLC
into the spheronizer, the extrudate is drawn to the walls of the extruder due to centrifugal
forces. From here what happens is very much dependent on the properties of
the extrudate. Under ideal conditions, the extrudate breaks into smaller, more uniform
pieces. Within a short period of time, the length of each piece is approximately
equal to the diameter, due to attrition and rapid movement of the bottom plate or
disk. The differential in particle velocity as they move outward to the walls, begin
to climb the walls and fall back onto the rotating bed, along with the angular motion
of the disk results in a rope-like formation (5). Figure 12 shows the rope-like formation
of the extrudates.
This formation can be a critical indicator of the quality of the granulation or
extrudate. As pointed out by Reynolds (5), the disk rotating without movement of
the product indicates an over wet condition. The condition is caused either from
a granulation that was initially over wet or migration of water or a fluid ingredient
to the surface of the extrudate during extrusion or spheronization.
As mentioned, the transformation from cylinder-shaped extrudate to a sphere
occurs in various stages. Two models have been proposed to describe the mechanism
The model proposed by Rowe in 1985 describes a transition whereby the
cylindrical particles (Fig. 13-2A) are first rounded off into cylindrical particles with
rounded edges (Fig. 13-2B), then form dumbbell-shaped particles (Fig. 13-2C), ellipsoids
(Fig. 13-2D), and finally spheres (Fig. 13-2E) (35). The second model proposed
by Baert et al. in 1993 suggests that the initial cylindrical particles (Fig. 13-1A) are
deformed into a bent rope-shaped particle (Fig. 13-1B), then form a dumbbell with a
twisted middle (Fig. 13-1C). The twisting action eventually causes the dumbbell to
break into two spherical particles with a flat side having a hollow cavity (Fig. 13-
1D). Continued action in the spheronizer causes the particles to round off into
spheres (Fig. 13-1E). When the sphere is fractured a hollow particle is revealed
(49). The exact mechanism is likely to be composition dependent. If the extrudate
is overwet, particle growth will occur resulting in a broad size distribution. Under
wet extrudate will not have enough plasticity to further round off in the spheronizer;
and dumbbells that would not deform further.
Figure 12 A characteristic rope-like formation in a spheronizer bowl during operation.
Extrusion/Spheronization as a Granulation Technique 347
and are shown graphically in Figure 13.
the result is the formation of dumbbells. The scanning electron micrographs in Figure
14 show an example of good spheres produced from a sufficiently plastic mass
 2005 by Taylor & Francis Group, LLC
Of the two process steps unique to extrusion/spheronization, the first, extrusion,
is a continuous process while the second, spheronization, is a batch process.
To make the process viable for commercial operations, two systems have been developed
to enable the extruder to continuously feed material to the spheronizer(s). The
Figure 14 An example of (A) good spheres produced from a sufficiently plastic mass and
(B) dumbbells that would not deform further produced from underwet extrudate.
Figure 13 A graphic representation of the two models proposed to describe the mechanism
of spheronization. The model proposed by Baert et al. (38) describes a transition from initial
cylindrical particles (1A) into a bent rope (1B), dumbbell (1C), two spherical particles with a
hollow cavity (1D), and spheres (1E).The model proposed by Rowe (25) describes a transition
from cylindrical particles (2A) into cylindrical particles with rounded edges (2B), dumbbells
(2C), ellipsoids (2D), and spheres (2E). (From Refs. 35,49.)
348 Mehta et al.
 2005 by Taylor & Francis Group, LLC
first system is a semicontinuous shuttle system and the second is a cascade system.
The shuttle system is typically used when uniform particles are required, such as
for controlled release coating applications. The cascade system, however, can be used
for applications where less size and shape uniformity is required, such as granulations
intended for compression.
The shuttle system uses two spheronizers in parallel. It is designed to fill one
spheronizer while the second is in the middle of its cycle, continue to collect extrudate
in a shuttle receptacle while they are both full and operational, and fill the second
after it empties and the first unit is in the middle of its cycle. The shuttle system
operation is shown graphically in Figure 15.
A picture of a spheronizing system having twin spheronizers is also shown in
The cascade operation uses one or more spheronizers that are modified to have
the disks some distance below the discharge chute (36). This results in a spheronization
zone having a fixed volume. The product is continually fed from either the
extruder or a previous spheronizer. As the charge volume grows from incoming
material some product is discharged. The residence time is dictated by the feed rate.
The reduced size and shape distributions are due to the percentage of material that
does not reside in the spheronization zone for the intended period of time. The number
of spheronizers placed in sequence depends on the desired outcome. However, if
only a slight rounding with minimal densification is required, one spheronizer with
a short residence time will be sufficient.
A commercial manufacturing of pellets using the extrusion/spheronization
process can be accomplished by discharging the formed pellets from the spheronizer
in a continuous fluid bed unit. This provides the semiautomatic commercial setup.
Variables in the spheronization step include spheronizer size, charge, disk
speed, and residence time. A number of studies have shown each of the variables
Figure 15 A graphic representation of twin spheronizer shuttle system using two spheronizers
in parallel and shuttle receptacle. (A) When both units are full the shuttle receptacle collects the
extrudate. (B) After one empties, (C) the shuttle box fills it. (DF) The cycle repeats itself for
the second unit. (From Ref. 36.)
Extrusion/Spheronization as a Granulation Technique 349
Figure 16.
Figure 17 illustrates a typical setup in such an equipment configuration.
 2005 by Taylor & Francis Group, LLC
has the potential to play a significant role in influencing the physical characteristics
of the resulting product. Hasznos et al. (46) showed that a higher disk speed and
longer residence time increased the coarse fraction and mean diameter and decreased
the fine fraction. The faster speed and longer time also increased the moisture loss
during the process. Since the moisture loss can reduce the plasticity of the particle,
it can have the same effect as an underwet granulation. The particles may not round
off into spheres and stay as deformed cylinders or dumbbells. Higher spheronizer
Figure 16 Twin spheronizer with extruder. (Courtesy of Niro Pharma Systems.)
Figure 17 Typical semicontinuous pellet production setup. (Courtesy of LCI Corporation.)
350 Mehta et al.
 2005 by Taylor & Francis Group, LLC
charges reduced the moisture loss. They also suggested that an interaction between
spheronizer speed and residence time indicated the total number of revolutions of the
disk was critical. A change in one of the variables could be offset by an opposite
change of the other, as long as the total number of revolutions remained constant
(46). Helle?n et al. showed a similar moisture loss during spheronization. In addition,
they indicated that the major factors influencing the shape of the spheres were the
disk speed and residence time. High speed and long time produced more spherical
particles (26). Wan et al. indicated that a minimum disk speed and residence time
was required to round the cylinder-shaped extrudate. Furthermore, an increase in
speed or time, up to a limit, increased the median diameter of the spheres while
higher speeds and longer times caused a reduction in size. Short residence times at
high disk speeds resulted in small but round particles (50).
A number of investigators have reported the effect of disk speed and residence
time on density. Woodruff and Neussle reported the variables to have no effect on
the density of the spheres as compared to the density of the granulation and extrudate
(16). These results are in conflict with most of the other studies; however, they
are likely due to the use of mineral oil in the formulation. The oil can reduce the frictional
forces at the die wall during extrusion and between particles and equipment
surfaces during spheronization. A number of investigators including Malinowski
and Smith reported that an increase in either disk speed or residence time resulted
in an increase in density (20,21,26,47). Mehta et al. studied the effect of spheronization
time on the pellet hardness and drug release (2). It was concluded that pellet
hardness changed with spheronization time until about 10 min after which the hardness
decreased until the 20 min point. No significant difference in pellet hardness
from 20 to 40 min thereafter was observed. Mehta et al. explained this by the densi-
fication process occurring during the spheronization step. As spheronization time
progresses from time zero to a certain time point t the extrudates are cut into uniform
particles and shaped into spheres by the centrifugal and frictional forces present
in the spheronizer during the operation. These forces act on each and every
particle, making them denser and more spherical with time. However, after a critical
period, no further densification occurs with an increase in spheronization time. In
another study, Mehta et al. showed the effect of spheronization time on the porosity
parameters of the pellets (51). It was summarized that a processing period of
210 min increased the number of pores and the total pore surface area and
decreased the pore diameter. Beyond this time, for up to 20 min of spheronization
time, the porosity was unchanged.
OConnor et al. indicated that the friability of placebo spheres decreased with
increasing residence time while the mean particle diameter decreased (24). Erkoboni
et al. showed an increase in extruder screen size resulted in reduced friability (21).
4.5. Drying
Drying is the final step in the process. This can be accomplished in any dryer that can
be used for conventional type granulations, including tray dryers, column type fluid
beds, and deck type vibratory fluid beds. Each of the drying techniques has advantages;
however, the major differences are based on the rate of water removal. Tray
drying is the slowest of the processes. Fluidized bed dryers result in a much more
rapid drying rate because of the higher air volumes and the potential use of higher
inlet temperatures. Column fluid beds are batch dryers, while the deck type dryers
offer the advantage of a continuous process. Both have been used successfully in
Extrusion/Spheronization as a Granulation Technique 351
 2005 by Taylor & Francis Group, LLC
drying product produced by extrusion spheronization. The drying process must be
chosen based on the desired particle properties. For example, pellets to be dried in
fluid bed equipment will have to withstand fluidization process, resist attrition,
and maintain its integrity.
Tray drying is a slow process in a static bed. Because of this, it can offer the
greatest opportunity for a drug to migrate toward the surface and recrystallize
(52). The more rapid rate in a fluid bed will likely minimize the effects of migration.
This phenomenon can have an effect on a number of particle properties. The increased
active concentration at the surface of the particle can increase the rate of dissolution.
This recrystallization, however, can cause a problem for applications requiring film
coating since the smooth surfaces developed by the spheronization process would be
damaged. Additionally, the crushing strength of tray dried particles will likely be
greater than their fluid bed counterparts. The slow recrystallization in the static
bed allows for crystal bridges to develop as the fluid is removed and the solute
The composition of the wet mass is critical in determining the properties of the
particles produced. This is clearly understood if we look at what material behaviors
are required during each of the process steps. During the granulation step, a plastic
mass is produceda simple enough task if ended there. The materials must form
a plastic mass, deform when extruded and break off to form uniformly sized cylindrical
particles. A minimal amount of granulating fluid should migrate to the surface
during extrusion and the particles should stay discrete during collection. During
spheronization the particles must round off to form uniformly sized spheres. They
must not dry out due to temperature or air volume or grow in size due to agglomeration.
The fact is that a lot is expected from materials used in this process. This
is especially true of formulations containing high percentages of active where low
levels of excipients are used to impart the desired properties to the mass.
The importance of using sphere-forming excipients was noted early on. Conine
and Hadley cited the necessity of using microcrystalline cellulose (4). Reynolds went
on to indicate the need for either adhesive or capillary type binders (5). He cited cellulose
gums, natural gums, and synthetic polymers as adhesives and microcrystalline
cellulose, talc, and kaolin as capillary type binders. Since then much work has been
conducted in an attempt to understand the significance of material properties. Some
of the studies are discussed in the following text.
OConnor et al. studied the behavior of some common excipients in extrusion/
spheronization. The materials were studied as single components using water as the
granulating fluid in an attempt to understand their application in the process. Of the
materials tested, only MCC or MCC with Na-CMC (Na-caboxymethyl cellulose)
was capable of being processed. Others including dicalcium phosphate, lactose,
starch, and modified starch did not process adequately (24).
In an additional study, they investigated the effect of varying drug, excipient,
and excipient:drug ratios. At low drug levels they found the spheronizing excipient
played the most significant role in determining sphere properties. They found that,
for low dose applications, MCC was the best excipient to use since it formed the
most spherical particles. At moderate drug loading (50%), MCC as well as the
two products consisting of MCC coprocessed with Na-CMC (Avicel RC-581 and
352 Mehta et al.
 2005 by Taylor & Francis Group, LLC
Avicel CL-611) resulted in acceptable spheres. At higher loading levels, however, the
MCC did not yield acceptable spheres and the coprocessed materials did. The
spheres produced using Avicel CL-611 were the most spherical. In addition, they
found dissolution to be dependent on the type of excipient used, the solubility,
and concentration of the active. Spheres containing MCC remained intact and
behaved as inert matrix systems, while those containing the coprocessed products
formed a gel plug in the dissolution basket and were described as water-swellable
hydrogel matrix systems. The release profiles for spheres containing each of the excipients
and theophylline in a 50:50 ratio are shown in Figure 18.
Release profiles for spheres containing different drug loads are shown in
An increase in drug load resulted in an increased release rate. Release profiles
for spheres containing actives having different solubilities, including chlorpheniramine
maleate, quinidine sulfate, theophylline, and hydrochlorothiazide are shown
Mehta and Kislalioglu demonstrated the use of polymethacrylate type polymers
such as Eudragit L 10055 and Eudragit S 100 via extrusion/spheronization
in the development of controlled release pellets (13,54). They theorized that for
the development of zero-order controlled release pellets of a poorly soluble drug,
MCC would not be a good choice to form a pellet system via extrusion/spheronization.
This would be due to the fact that MCC being insoluble would form a nondisintegrating
matrix from which it would be difficult for an insoluble drug to be
released. In their work they showed that Eudragit L10055 and Eudragit S 100
can be used as pellet forming and release rate governing polymers for developing
a controlled release drug delivery system with out the use of MCC in the matrix.
Zhou and Vervaet produced matrix pellets by combining microcrystalline
waxes, pregelatinized starches, and hydrolyzed starches with model drugs such as
Ibuprofen, chloroquin phosphate, and others (55). They concluded that the combination
of microcrystalline waxes and pregelatinized starches or maltodextrins is a
flexible system for the production of matrix pellets, even with a high drug concentration.
Additionally, they concluded that the drug release with such a system could be
Figure 18 Dissolution profiles of spheres containing 50% theophylline in different Avicel
MCC types  Avicel PH-101;
, Avicel RC-581; &, Avicel CL-611. (From Ref. 53.)
Extrusion/Spheronization as a Granulation Technique 353
Figure 19.
in Figure 20. An increase in drug solubility resulted in an increased release rate (53).
 2005 by Taylor & Francis Group, LLC
modeled by varying the type and the concentration of the wax and the starch. Tapia
et al. described factors influencing the mechanism of release from sustained release
matrix pellets, produced by the extrusion/spheronization process (56).
Kleinebudde and Jumaa concluded that during the extrusion process, water
content in the extrudate and pellet porosity were increased as the degree of polymerization
of MCC and powder cellulose in the matrix was increased (57).
Millili and Schwartz demonstrated the effect of granulating with water and
ethanol at various ratios. The physical properties of the spheres changed significantly
as the ratio of the two fluids was varied. Spheres could not be formed with absolute
ethanol but were possible with 5:95 water:ethanol. An increase in the water fraction
resulted in a decrease in porosity, friability, dissolution, and compressibility and an
increase in density. The porosity of spheres granulated with 95% ethanol was 54%
Figure 19 Dissolution profiles of spheres containing different concentrations of drug in
Avicel CL-611: , 10%; &, 59%;
`, 80%. (From Ref. 53.)
Figure 20 Dissolution profiles of spheres containing 10% drug in Avicel PH 101: , chlorphenarimine
maleate; , quinidine sulfate; , theophyline; & hydrochlorothiazide. (From Ref. 53.)
354 Mehta et al.
 2005 by Taylor & Francis Group, LLC
while the water granulated product had a porosity of 14%. When greater than 30%
water was used, spheres remained intact throughout the dissolution test. As previously
discussed, water granulated spheres were very difficult to compress while
spheres granulated with 95% ethanol were significantly more compressible than those
prepared using water (18). In contrast, Mehta et al. showed that an increase in granulation
water level increased the total number of pores in the pellet matrix without
changing the pore diameters (51). Additionally they concluded that this direct
increase in porosity increased the dissolution contact angle due to which dissolution
of the poorly soluble drug was increased. Jerwanska et al. concluded that the rate of
drug release increased with increased levels of granulation liquid because of a greater
degree of porosity obtained after drying (58). They also correlated these results with
differences in hardness of the pellets. This was similar to the findings by Mehta et al.
Jerwanska et al. proposed that for a continuous extrusion process, adequate water is
required to bridge the particles together until liquid saturation in the granulation is
achieved. This strategy is necessary to deform the granulation to form extrudates
and consequently shape them into spheres by spheronization. If the granulation water
level is below the liquid saturation point, then the spheres obtained will be hard and
less porous, thereby leading to decreased drug release rates. Above the liquid saturation
point, the hardness and porosity of the pellets are not significantly decreased.
A tablet hardness vs. compression forces profile is shown in Figure 21.
In a later study, Millili et al. proposed a bonding mechanism, referred to as
autohesion, to explain the differences in the properties of spheres granulated with
water and ethanol. Autohesion is a term used to describe the strong bonds formed
by the interdiffusion of free polymer chain ends across particleparticle interfaces
Using a ram extruder, Harrison et al. demonstrated that steady state flow could
not be achieved with lactose. Additionally, they demonstrated the reduced sensitivity
of MCC to small changes in moisture as determined by the force required to induce
plug flow in a cylinder. Comparing MCC to a MCC/lactose blend and 100% lactose,
Figure 21 The effect of varying compression force on the hardness of compacted 16/30-
mesh spheres of 10% theophyllineAvicel PH101:
, spheres prepared by water; &, spheres
prepared by 95% ethanol granulation. (From Ref. 18.)
Extrusion/Spheronization as a Granulation Technique 355
 2005 by Taylor & Francis Group, LLC
they found that, with lactose, small changes in moisture caused large changes in force
while with MCC, larger changes in moisture were required to have similar effects on
the force (29).
Baert et al. used mixtures of microcrystalline cellulose and coexcipients at various
ratios to demonstrate the effect of solubility and the total fluid on extrusion
forces. They showed that if the coexcipient was insoluble, such as dicalcium phosphate,
the force required to extrude increased with increasing levels of coexcipient.
When a soluble excipient such as lactose was used, the force required to extrude
decreased with the addition of the initial amounts of lactose. After a certain level,
however, the reduction in force stopped and began to increase. This was due to
the initial solubilization of lactose and the resulting increase in the total fluid level.
Once the fluid was saturated the remaining lactose was not soluble and the force
began to increase. The increase began at about 10% lactose level for a-lactose and
20% for b-lactose. This was due to the difference in solubility between the two materials
(33). The effects of dicalcium phosphate and various lactose grades on extrusion
force are shown in Figure 22.
Funck et al. showed that low levels of common binders could be used to produce
high drug loaded spheres with microcrystalline cellulose. Materials such as carbomer,
Na-CMC, hydroxypropylcellulose (HPC), hydroxypropylmethylcellulose
(HPMC), povidone (PVP), and pregelatinized starch were used. All materials were
capable of producing spheres of acceptable quality. Dissolution testing showed
spheres containing HPC and HPMC remained intact during testing while spheres
containing starch, PVP, and Na-CMC disintegrated (60).
Lender and Kleinebudde reported that spheres produced with powdered cellulose
had higher porosity and faster dissolution than those made using microcrystalline
cellulose. Spheres could not be produced using only powdered cellulose and
drug; a binder was required. The higher porosity of the spheres prepared from powdered
cellulose may be beneficial for applications requiring compression (61).
Figure 22 Influence of the amount of lactose or dicalcium phosphate dehydrate (% total
weight) on the extrusion forces (N) for mixtures of lactose or dicalcium phosphate dehydrate
Avicel PH 101water after granulation with a planetary mixer. Each end point is the
mean of six values. The SD is lower than 3% for each point. Six different types of lactose were
used: &, a-lactose monohydrate 80 mesh; , a-lactose monohydrate 200 mesh; , a-lactose
monohydrate 325 mesh; &, spray dried lactose DCL 11; , anhydrous b-lactose DCL 21; ,
anhydrous a-lactose DCL 30. One type of dicalcium phosphate dehydrate was used,
. (From
Ref. 33.)
356 Mehta et al.
 2005 by Taylor & Francis Group, LLC
Feilden et al. showed that increasing the particle size of lactose resulted in forced
flow and high extrusion forces, which resulted in poor quality extrudate and spheres having
awide size distribution. Thiswas attributed to the increased pore diameter of themixture
containing the coarse lactose which allowed greater movement of water (62).
Chien and Nuessle (63) showed the use of a surfactant, such as sodium lauryl
sulfate, reduced the migration of drug to the surface of the sphere during drying by
reducing the surface tension of the granulating fluid. The reduction in surface tension
also made it difficult to produce a cohesive extrudate in some cases.
Some miscellaneous observations include the following. Reynolds reported
that excess extrudate friability can be overcome by incorporating more MCC, binder,
or water in the granulation (5). Erkoboni et al. indicated that sphere hardness
was most affected by the level of MCC in the formulation and the level of granulating
fluid used (21). Hileman et al. showed that MCC had a narrower water range
over which quality spheres could be made than MCC coprocessed Na-CMC (47).
Helle?n et al. showed that the surface characteristics were influenced by the water
level with higher water levels giving smoother surfaces (26). Mehta et al. showed that
when concentrations of pellet forming and release rate governing polymers in the
matrix were changed, it altered the dissolution kinetics of a poorly soluble drug (2).
Typically, pellets are produced for administering in a capsule dosage form after manufacturing
with desired modified release properties. These pellets can be compressed
into tablets as well. The tablets are normally intended to disintegrate into discrete
pellets in the gastrointestinal tract and the drug should be subsequently be released
in a controlled manner from the individual pellets. One challenge in the production
of such tablets is maintaining the desired drug release after compaction, as the application
of a compaction pressure can lead to structural changes in the film coating
and consequently, altered drug release. Numerous investigations have been made
into the compaction of pellets, coated or uncoated. It is imperative that the coated
pellets do not lose their release properties during the compression stage. Pellets have
been shown to react differently to compaction and consolidation than powders of the
same material. Wang et al. reported compression of various lactose/microcrystalline
cellulose compositions in powder or pellet form (64). Schwartz et al. (17) demonstrated
the compaction characteristics of MCC processed into spheres are signifi-
cantly different than the original powder. The powder material forms hard
compacts at low compression forces, while the spheres are not compressible and
form soft compacts, even at high forces. They indicated that spheres prepared from
MCC showed a high degree of viscoelasticity over the entire compression range.
Inclusion of coexcipients such as lactose and dicalcium phosphate increase the compactability
by decreasing the viscoelastic resistance or pressure range over which the
spheres behave elastically. A reduction in viscoelastic resistance was seen with
spheres containing both lactose and dicalcium phosphate; however, dicalcium phosphate
had a greater effect. Compaction profiles of spheres containing 10% theophylline
with MCC, MCC/DCP, or MCC/lactose in a 22.5/67.5 ratio are shown in
A similar phenomenon was reported by Maganti and Celik when pellets produced
by rotor granulation were compressed (65). They compared the compaction
behavior of pellet formulations, mainly consisting of MCC, to that of the powders
Extrusion/Spheronization as a Granulation Technique 357
Figure 23.
 2005 by Taylor & Francis Group, LLC
from which they were formed and also found significant differences. The powders
examined were found to compact by plastic deformation and produced strong compacts,
while the pellets exhibited elastic deformation and brittle fragmentation,
resulting in compacts of lower tensile strength. This can be explained by the fact that
the pellets, which are large and spherical in shape as compared to the small, irregular
powder particles they are composed of, have a low surface-to-volume ratio, which
might result in a decreased area of contact between the particles as they consolidate.
Nicklasson (66) investigated the compression behavior of pellets consisting of MCC,
with or without other excipients such as polyethylene glycol and DCP. Deformation
of the aggregates was found to depend on three deformation characteristics, namely,
the capacity for, the mode of and the resistance to deformation. High surface deformation
refers to the great ability of the pellets to conform to the surface of the surrounding
pellets. In pellets containing the soft component, the primary particles can
reposition within the agglomerate and the ability to fill the intragranular pore space
is increased. For pellets containing hard materials, the compaction stress may give
local failure at pellet surfaces. Thus, the material properties of the primary particles
constituting the pellets are important for the compression behavior of pellets. In
number of studies (67,68) various soft materials have been incorporated in pellets
to modify their deformability and compatibility. Nicklasson and Alderborn (69) studied
the modulation of the tableting behavior of pellets through the incorporation of
polyethylene glycol and found that these soft pellets had an increased propensity to
deform and altered mode of deformation to the relatively hardMCC pellets. Iloan?usi
et al. (67) found MCC-based bead formulations incorporating wax to be more compressible
than those made without wax. Salako et al. (68) found that pellets containing
theophylline and MCC were hard and less brittle than the ones containing
glyceryl monostearate which were soft pellets. The soft pellets were found to fracture
under low compression pressures and were able to form a coherent network of
deformable material in the tablets at higher pressures. The hard pellets were unable
Figure 23 The effect of excipients on the compaction profile of spheres. Compaction profiles
of spheres containing 10% theophylline with either MCC, MCCDCP, or MCClactose in
a 22.5:67.5 ratio using the Leunberger model. (From Ref. 17.)
358 Mehta et al.
 2005 by Taylor & Francis Group, LLC
to form such a network at high pressures and found to reduce more in volume
without bond formation than soft pellets.
The size of the pellets can also have a bearing on their compression behavior.
Small pellets have been shown to be less affected than larger ones by the compaction
process (70,71). Smaller beads were significantly stronger, relative to their size, than
larger ones. Researchers also found that larger pellets were much more readily
deformed (71). The application of coating to the pellet core can influence their compression
characteristics. Magnati et al. (72) added a water-based ethyl cellulose coating
(Surelease) to the MCC-based pellets previously used (65) and thereby altered
their deformation characteristics, introducing plastoelastic properties whereas previously
they had been brittle and elastic. The overall ability of the pellets to deform,
both plastically and elastically, increased with an increasing coating level.Miller et al.
(73) investigated the mechanical properties of tablets compressed from pellets coated
with Surelease to that of uncoated pellets and found them to be comparable with the
exception of the diametrical strain, which increased on coating. This was attributed
to the flexibility of the plasticized ethyl cellulose, allowing greater deformation of the
compact to occur before failure.
It has been found that coated pellets can be compressed into tablets while
retaining controlled release of the drug, provided that the effect of excipients and
compression force is considered and determined. The protective effect of an excipient
is dependent on the particle size and the compaction characteristics of the material.
In general, materials that deform plastically, such as MCC and PEG, give the best
protective effect (7477). Yuasa et al. studied the protective effect of 14 different
excipients and were able to correlate the plastic energy percentage to the release rate,
that is, materials that deform plastically were shown to protect the coating best (78).
However, Stubberud et al. found that lactose, a fragmented material, gives better
protection than MCC (79). The compressed induced changes in the structure of a
film coating may depend on formulation factors, such as the type and the thickness
of the coating, the properties and the structure of the substrate pellets, and the
incorporation of excipient particles.
The optimum amount of excipients incorporated in the tablet formulation was
concluded to be 30% (80). Palmieri et al. (81) showed that tablets consisting of maximum
40% coated granules had acceptable release profiles, using MCC as the tablet
excipient. Wagner et al. (82) concluded that a tablet content of 70% (w/w) of pellets,
approximately 1mm diameter, is a critical level resulting in pronounced damage to
the coating. On reducing the pellet proportion to 60% (w/w), tablets fulfilling the
compendial dissolution requirements for enteric-coated products can be prepared.
Tuno?n (83) investigated factors that influence the preparation of modified release
pellets, its compression behavior, influence of inter- and intragranular factors and
release of drug release of pellets. The most frequently encountered explanation for
the loss of modified release upon compaction, in terms of an increased drug release
rate, is the occurrence of cracks in the coating. However, knowledge of why and how
cracks are formed in the coating and techniques to avoid them is highly valuable in
the development of multiple unit tablets containing pellets.
Extrusion/spheronization is a versatile process capable of producing granules or
spheres having unique physical properties. Since it may be more labor and time
Extrusion/Spheronization as a Granulation Technique 359
 2005 by Taylor & Francis Group, LLC
intensive than the more common granulation techniques, it should be considered
as a granulating technique when the desired properties cannot be produced using
more conventional techniques. Potential applications are many including both
immediate and controlled release. Regardless of the application, care must be taken
to understand the desired properties and the formulation and process variables capable
of achieving them. The use of statistical experimental design for formulation
and process development is strongly recommended due to the high degree of interactions
between the variables. Lastly, new technologies such as hot melt extrusion
and spheronization (HME) are gaining considerable interest in the pharmaceutical
drug delivery arena for solving specific problems such as enhancing taste masking,
improving solubility and drug bioavailability and in general for controlled release
drug delivery. Compression of pellets into modified release multiple unit dosage form
is now possible once the proper understanding of the formulation of the core pellets,
type of coating, and protective excipients to maintain the coating integrity of the
pellets is understood.
1. Erkoboni D. Extrusionspheronization for granulation. Parikh DM, ed. Handbook of
Pharmaceutical Granulation Technology. New York: Marcel Dekker, 1997.
2. Mehta KA, Kislalioglu MS, Phuapradit W, Malick AW, Shah NH. Effect of formulation
and process variables on matrix erosion and drug release from a multiunit erosion matrix
of a poorly soluble drug. Pharm Technol 2002; February:2634.
3. Nakahara, US Patent 3,277,520, October 1966.
4. Conine JW, Hadley HR. Preparation of small solid pharmaceutical spheres. Drug Cosmet
Ind 1970; 106:3841.
5. Reynolds AD. A new technique for the production of spherical particles. Manuf Chem
Aerosol News 1970; 41:4043.
6. Mehta KA, Kislalioglu MS, Phuapradit W, Malick AW, Ke J, Shah NH. In vivo release
performance of nifedipine in dogs from a novel Eudragit based multi-unit erosion matrix.
Drug Deliv Technol 2002; 2(2):3842.
7. Young CR, Koleng JJ, McGinity JW. Production of spherical pellets by a hot-melt extrusion
and spheronization process. Int J Pharm 2002; 242:8792.
8. Breitenbach J, Ma?gerlein M. Melt extruded molecular dispersions. Ghebre-Sellassie I, ed.
Pharmaceutical Extrusion Technology. New York: Marcel Dekker, 2003.
9. Oshlack B, et al. US Patent 6,706,281, issued March 16, 2004.
10. McGinity J, Repka M. Hot melt extruded films. Drug Deliv Technol 2004; 4(7):4047.
11. Sood A, Ashokraj Y, Panchagnula R. Use of extrusion spenronization to develop controlled
release dosage forms for diltiazem hydrochloride. Pharm Tech 2004; April:6285.
12. Jalal IM, Malinowski HJ, Smith WE. Tablet granulations composed of spherical-shaped
particles. J Pharm Sci 1972; 61:14661468.
13. Mehta KA, Kislalioglu MS, Phuapradit W, Malick AW, Shah NH. Release performance
of a poorly soluble drug from a novel Eudragit based multi-unit erosion matrix. Int
J Pharm 2001; 213:712.
14. Clarke, GM, Newton JM, Short MB. Comparative gastrointestinal transit of pellet systems
of varying density. Int J Pharm 1995; 114:111.
15. Devereus JE, Newton JM, Short MB. The influence of density on the gastrointestinal
transit of pellets. J Pharm Pharmacol 1990; 42:500501.
16. Woodruff CW, Nuessle NO. Effect of processing variables on particles obtained by
extrusion-spheronization. J Pharm Sci 1972; 61:787790.
360 Mehta et al.
 2005 by Taylor & Francis Group, LLC
17. Schwartz JP, Nguyen NH, Schnaare RL. Compaction studies on beads: compression and
consolidation parameters. Drug Dev Ind Pharm 1994; 20:31053129.
18. Millili GP, Schwartz JB. The strength of microcrystalline cellulose pellets: the effect of
granulating with water/ethanol mixtures. Drug Dev Ind Pharm 1990; 16:14111426.
19. Malinowski HJ, Smith WE. Effect of spheronization process variables on selected tablet
properties. J Pharm Sci 1974; 63:285288.
20. Malinowski HJ, Smith WE. Use of factorial design to evaluate granulations prepared by
spheronization. J Pharm Sci 1975; 64:16881692.
21. Erkoboni DF, Fiore SA, Wheatley TA, Davan T. The effect of various process and formulation
variables on the quality of spheres produced by extrusion/spheronization. Poster
presentation, AAPS National Meeting, 1991.
22. Chariot M, France`s J, Lewis GA, Mathieu D, Phan Tan Luu R, Stevens HNE. A factorial
approach to process variables of extrusion-spheronization of wet powder masses.
Drug Dev Ind Pharm 1987; 13:16391649.
23. Hileman GA, Goskonda SR, Spalitto AJ, Upadrashta SM. Response surface optimization
of high dose pellets by extrusion and spheronization. Int J Pharm 1993; 100:7179.
24. OConnor RE, Holinez J, Schwartz JB. Spheronization I: processing and evaluation of
spheres prepared from commercially available excipients. Am J Pharm 1984; 156:8087.
25. Ojile JE, Macfarlane CB, Selkirk AB. Drug distribution during massing and its effect on
dose uniformity in granules. Int J Pharm 1982; 10:99107.
26. Helle?n L, Yliruusi J, Merkku P, Kristoffersson E. Process variables of instant granulator
and spheronizer: I. Physical properties of granules, extrudate and pellets. Int J Pharm
1993; 96:197204.
27. Kleinebudde P, Lindner H. Experiments with a twin screw extruder using a single-step
granulation/extrusion process. Int J Pharm 1993; 94:4958.
28. Lindberg NO, Tufvesson C, Holm P, Olbjer L. Extrusion of an effervescent granulation
with a twin screw extruder, Baker Perkins MPF 50 D. Influence on intragranular porosity
and liquid saturation. Drug Dev Ind Pharm 1988; 14:17911798.
29. Harrison PJ, Newton JM, Rowe RC. The characterization of wet powder masses suitable
for extrusion/spheronization. J Pharm Pharmacol 1985; 37:686691.
30. Land??n M, Rowe RC, York P. Characterization of wet powder masses with a mixer torque
rheometer. 3. Nonlinear effects of shaft speed and sample weight. J Pharm Sci 1995;
31. Baert L, Remon JP, Knight P, Newton JM. A comparison between the extrusion forces
and sphere quality of a gravity feed extruder and a ram extruder. Int J Pharm 1992;
32. Baert L, Fanara D, De Baets P, Remon JP. Instrumentation of a gravity feed extruder
and the influence of the composition of binary and tertiary mixtures on the extrusion
forces. J Pharm Pharmacol 1991; 43:745749.
33. Ku CC, Joshi YM, Bergum JS, Jain NB. Bead manufacture by extrusion/spheronization
a statistical design for process optimization. Drug Dev Ind Pharm 1993;
34. Rowe RC. Spheronization: a noval pill-making process? Pharm Int 1985; 6:119123.
35. Harrison PJ, Newton JM, Rowe RC. The application of capillary rheometry to the extrusion
of wet masses. J Pharm 1987; 35:235242.
36. Hicks DC, Freese HL. Extrusion and spheronization equipment. Ghebre-Sellassie I, ed.
Pharmaceutical Pelletization Technology, Chap. 4. New York: Marcel Dekker, 1989:71
37. Harrison PJ, Newton JM, Rowe RC. Flow defects in wet powder masses. J Pharm Pharmacol
1984; 37:8183.
38. Harrison PJ, Newton JM, Rowe RC. Congergent flow analysis in the extrusion of wet
powder masses. J Pharm Pharmacol 1984; 37:8183.
39. Fechner P, Wartewig S, Fueting M, Heilmann A, Neubert R, Kleinebudde P. Properties
of microcrystalline cellulose and powder cellulose after extrusion/spheronization as
Extrusion/Spheronization as a Granulation Technique 361
 2005 by Taylor & Francis Group, LLC
studied by Fourier transform Raman spectroscopy and environmental scanning electron
microscopy. AAPS PharmSci 2003; 5(4):Article 31.
40. Pinto JF, Buckton G, Newton JM. The influence of four selected processing and formulation
factors on the production of spheres by extrusion and spheronization. Int J Pharm
1992; 83:187196.
41. Helle?n L, Ritala J, Yliruusi P, Palmroos A, Kristoffersson E. Process variables of the
radial screen extruder: I. Production capacity of the extruder and the properties of the
extrudate. J Pharm Tech Int 1992; 4:5060.
42. Vervaet C, Baert L, Risha PA, Remon JP. The influence of the extrusion screen on pellet
quality using an instrumented basket extruder. Int J Pharm 1994; 107:2939.
43. Baert L, Remon JP, Elbers JAC, Van Bommel EMG. Comparison between a gravity feed
extruder and a twin screw extruder. Int J Pharm 1993; 99:712.
44. Shah R, Kabadi M, Pope DG, Augsberger LL. Physico-mechanical characterization of
the extrusion/spheronization process. Part I. Instrumentation of the extruder. Pharm
Res 1994; 11:355360.
45. OConnor RE, Schwartz JB. Extrusion and spheronization technology. Ghebre-Sellassie
I, ed. Pharmaceutical Pelletization Technology, Chap. 9. New York: Marcel Dekker,
46. Hasznos L, Langer I, Gyarmathy M. Some factors influencing pellet characteristics made
by an extrusion/spheronization process. Part I. Effects on size characteristics and moisture
content decrease of pellets. Drug Dev Ind Pharm 1992; 18:409437.
47. Hileman GA, Goskonda SR, Spalitto AJ, Upadrashta SM. A factorial approach to high
dose product development by an extrusion/spheronization process. Drug Dev Ind Pharm
1993; 19:483491.
48. Baert L, Remon JP. Influence of amount of granulating liquid on the drug release rate
from pellets made by extrusion spheronization. Int J Pharm 1993; 95:135141.
49. Wan LSC, Heng PWS, Liew CV. Spheronization conditions on spheroid shape and size.
Int J Pharm 1993; 96:5965.
50. Mehta KA, Kislalioglu MS, Phuapradit W, Malick AW, Shah NH. Effect of formulation
and process variables on porosity parameters and release rates from a multi unit erosion
matrix of a poorly soluble drug. J Control Release 2000; 63:201211.
51. Dyer AM, Khan KA, Aulton ME. Effect of the drying method on the mechanical and
drug release properties of pellets prepared by extrusion. Drug Dev Ind Pharm 1994;
52. OConnor RE, Schwartz JB. Spheronization II: Drug release from drugdiluent mixtures.
Drug Dev Ind Pharm 1985; 11:18371857.
53. Mehta KA, Kislalioglu MS, Phuapradit W, Malick AW, Shah NH. Multi-unit controlled
release systems of nifedipine and nifedipine: Pluronic F-68 solid dispersions: characterization
of release mechanisms. Drug Dev Ind Pharm 2002; 28(3):275285.
54. Zhou F, Vervaet C, Remon JP. Matrix pellets based on the combination of waxes,
starches and maltodextrins. Int J Pharm 1996; 133:155160.
55. Kleinebudde P, Jumaa M, El Saleh F. Influence of degree of polymerization on behavior
of cellulose during homogenization and extrusion/spheronization. AAPS PharmSci 2000;
2(2):Article 21.
56. Jerwanska E, Alderborn G, Newton JM, Nystrom C. The effect of water content on the
porosity and liquid saturation of extruded cylinders. Int J Pharm 1995; 121:6571.
57. Millili GP, Wigent RJ, Schwartz JB. Autohesion in pharmaceutical solids. Drug Dev Ind
Pharm 1990; 16:23832407.
58. Maganti L, Celik M. Compaction studies on pellets I. Uncoated pellets. Int J Pharm
1993; 95:2942.
59. Funck JAB, Schwartz JB, Reilly WJ, Ghali ES. Binder effectiveness for beads with high
drug levels. Drug Dev Ind Pharm 1991; 17:11431156.
60. Linder H, Kleinebudde P. Use of powdered cellulose for the production of pellets by
extrusion/spheronization. J Pharm Pharmacol 1994; 46:27.
362 Mehta et al.
 2005 by Taylor & Francis Group, LLC
61. Fielden KE, Newton JM, Rowe RC. The influence of lactose particle size on the spheronization
of extrudate processed by a ram extruder. Int J Pharm 1992; 81:205224.
62. Vervaet C, Baert L, Remon JP. Extrusionspheronization: a literature review. Int
J Pharm 1995; 116:131146.
63. Tapia C, Buckton G, Newton JM. Factors influencing the mechanism of release from
sustained release matrix pellets, produced by extrusion/spheronization. Int J Pharm
1993; 92:211218.
64. Chien TY, Nuessle NO. Factors influencing migration during spheronization. Pharm
Technol 1985; 4:4448.
65. Wang C, Zhang G, Shah NH, Infeld MH, Malick AW, McGinity JW. Compaction properties
of spheronized binary granular mixtures. Drug Dev Ind Pharm 1995; 21:753779.
66. Nicklasson F. Compression Mechanics of Pharmaceutical AggregatesStudies on the
Tabletting of Spheronized Aggregates with Varying Composition and Porosity Ph.D.
Thesis, Uppsala University, Sweden, 2000.
67. Iloan? usi NO, Schwartz JB. The effect of wax on compaction of microcrystalline cellulose
beads made by extrusion and spheronization. Drug Dev Ind Pharm 1998; 24:3744.
68. Salako M, Podczeck F, Newton JM. Investigations into the deformability and tensile
strength of pellets. Int J Pharm 1998; 168:4957.
69. Nicklasson F, Alderborn G. Modulation of the tableting behavior of microcrystalline cellulose
pellets by the incorporation of polyethylene glycol. Eur J Pharm Sci 1999; 9:5765.
70. Haslam JL, Forbes AE, Rork GS, Ripkin TL, Slade DA, Khossravi D. Tableting of controlled
release multiparticulates, the effect of millisphere size and protective overcoating.
Int J Pharm 1998; 173:233242.
71. Johansson B, Nicklasson F, Alderborn G. Effect of pellet size on degree of deformation
and densification during compression and on compactability of microcrystalline cellulose
pellets. Int J Pharm 1998; 163:3548.
72. Maganti L, Celik M. Compaction studies on pellets II. Coated pellets. Int J Pharm 1994;
73. Miller RA, Leung EMK, Oates RJ. The compression of spheres coated with an aqueous
ethylcellulose dispersion. Drug Dev Ind Pharm 1999; 16:14111426.
74. Be?chard SR, Leroux JC. Coated palletized dosage form: effect of compaction on drug
release. Drug Dev Ind Pharm 1992; 18:19271944.
75. Torrado JJ, Augsburger LL. Effect of different excipients on the tableting of coated particles.
Int J Pharm 1994; 106:149155.
76. Beckert TE, Lehmann K, Schmidt PC. Compression of enteric-coated pellets to disintegrating
tablets. Int J Pharm 1996; 143:1323.
77. Haubitz H, Mehnert W, Fro?mming KH. Preparation of theophylline multiple units
tablets. Pharm Ind 1996; 58:8386.
78. Yuassa H, Takashima Y, Omata T, Kanaya Y. Studies on stress dispersion in tablets III.
Suppression of fracture of coated film by an excipients during the preparation of tablets
containing particles. STP Pharm Sci 2001; 11:221227.
79. Stubberrud L, Eriksson M, Kordnejad K, Graffner C. Watersolid interactions. IV.
Influence of moisture sorption on the compaction of film-coated particles. Pharm Dev
Technol 1998; 3:141151.
80. Lehmann K, Petereit HU, Dreher D. Scnellzerfallende Tabletten mit gesteuerter Wirkstoffabgabe.
Pharm Ind 1993; 55:940947.
81. Palmieri GF, Wehrle? P, Martelli S. Drug release from compressed Eudragit RS 30D
coated beads. STP Pharm Sci 1996; 6:118121.
82. Wagner KG, Krumme M, Beckert TE, Schmidt PC. Development of disintegrating multiple-
unit tablets on a high-speed rotary tablet press. Eur J Pharm Biopharm 2000;
83. Tuno?n A? . Preparation of Tablets from Reservoir Pellets with an Emphasis on the Compression
Behavior and Drug Release. Ph.D. Thesis, Uppsala University, Sweden, 2003.
Extrusion/Spheronization as a Granulation Technique 363
 2005 by Taylor & Francis Group, LLC
Effervescent Granulation
Guia Bertuzzi
IMA S.p.A, Solid Dose Division, Bologna, Italy
Effervescent granulation is an important step in the production of fizzy dosage forms
that, most of the time, cannot be avoided to achieve the desired characteristics of the
effervescent tablets. It is also very critical because it can affect the stability of the
final dosage forms.
The first effervescent preparations were described over two centuries ago in the
official compendia; they were in powder form for use as cathartic salts. Later, in
1815, a patent described a combination of neutral salt or powder which possesses
all the properties of the medicinal spring of Seidlitz in Germany, under the name
of Seidlitz Powders, which contains sodium potassium tartarate, sodium bicarbonate,
and tartaric acid, in the proportions 3:1:1, respectively (1). Effervescent
granules and tablets have become more and more popular as the dosage form
because they are readily soluble and easy to consume just by drinking the glass of
water where they are dissolved.
To state the growing interest in such forms, in the 1980swhen electronic
bibliography searching was not available yetthe results of a literature search about
effervescent forms were published so as to help scientists working on new
developments (2).
According to the 4th edition of the European Pharmacopoeia, the effervescent
forms are defined as those granules or tablets that are to be dissolved in water before
administration to patients. Effervescent tablets or granules are uncoated and generally
contain acidic substances and carbonate or bicarbonate which reacts rapidly to
release carbon dioxide when dissolved in water. Disintegration of the tablets usually
occurs within 2 min or even less, due to the evolution of carbon dioxide.
Effervescent forms have many advantages over conventional pharmaceutical
forms. They substitute liquid forms when the active ingredient is less stable in liquid
form because they can be administered only by first dissolving the tablet in water.
Active ingredients that are not stable in liquid form are often more stable in the
effervescent form. Their administration is easy and is particularly helpful to patients,
for instance, children, who are not able to swallow capsules or tablets. They have
a pleasant taste due to carbonation, which helps to mask the bad taste of certain drugs.
 2005 by Taylor & Francis Group, LLC
This could help to avoid the gastric side effect of certain drugs. In certain cases they
can shorten drug absorption rate in the body when compared to traditional tablets,
with a quicker therapeutic effect (3). They are easy to use and appeal to consumers
more than the traditional dosage forms because of their color and fizzy appearance.
Disadvantages of such solid dosage forms are more related to production technology
even if processing methods and equipment are the same as the conventional
ones. The general product requirements are similar to conventional granules: particle
size distribution and shape, uniformity of active distribution, to produce satisfactory
free-flowing granules, capable of being tableted using high-speed rotary tablet press.
However, it is also necessary to focus attention on some aspects of the procedure,
including compression and packaging because effervescent dosage formsirrespective
of whether it is a sachet or a tabletare larger and consequently difficult to produce.
The pharmaceutical industry has faced many problems, especially in the preparation
of effervescent tablets as it is certainly the application for which the choice
of the processing equipment is at least as important as the formulation design.
Effervescence is the evolution of gas bubbles from a liquid, as the result of a chemical
reaction. The most common reaction for pharmaceutical purpose is the acidbase
reaction between sodium bicarbonate and citric acid:
3NaHCO3 ?aq? 252g?3mol?
? H3C6H5O7 ?aq? 192g?1mol?
)3H2O?aq? 54g?3mol?
? 3CO2
132g?3mol? ? 3Na3C6H5O7 ?aq? 258g?1mol?
This reaction starts in presence of water, even with a very small amount as catalyzing
agent, and because water is one of the reaction products, it will accelerate the
rate of reaction, leading to difficulty in stopping the reaction. For this reason the
whole manufacturing and storage of effervescent products has to be planned by
minimizing the contact with water. Looking at the stechiometric ratios in the reaction
it is quite easy to understand the reason why effervescent tablets are so large.
Recently, some effervescent systems have been prepared that act as penetration
enhancers for drug absorption, not only in oral forms but also in some topical forms,
such as skin or vaginal applications. In these cases the reaction takes place directly
after administration, in the mouth due to saliva (4), upon the wounds due to blood
serum (5), or when formulated in a suppository. The effervescence can be provoked
by the moisture of the vaginal mucosa to treat vaginal infections or by simply adjusting
the pH.
There are other forms in which effervescence is based on a reaction different
from carbon dioxide formation. Effervescence is due to reactants that evolve oxygen
or other gases, which are safe for human consumption even if they are not suitable
for oral administration but can be employed in preparations for external use such as
antibacterial for dental plate cleaning.
The criteria to choose the raw materials for effervescent products are not very different
from those for conventional tablets, since in both cases good compressibility and
compactability are the targets to be achieved.
366 Bertuzzi
 2005 by Taylor & Francis Group, LLC
The intrinsic characteristics of effervescent forms bring some considerations
that limit the choice of the raw material, including the selection of the active ingredient.
Moisture content of the raw material is a very significant aspect, because it
affects compressibility and stability of the tablets. To avoid premature effervescent
reaction during the process or once the tablets are formed, raw materials with very
low moisture content have to be used.
Since an effervescent tablet is required to dissolve within 2 min or less in a glass
of water (about 100 mL), the solubility of raw materials and their rate of solubility
are other significant parameters. The active ingredient must be soluble, water dispersible,
or at least solubilized by salt formation during the dissolution in the glass of
water. The rest of the excipients, such as additives like sweeteners, coloring agents,
flavors, also have to be water soluble.
For all the above considerations the list of the excipients has not changed since
many years. However, the physical properties of these raw materials have recently
improved. Many different grades for each material are now commercially available,
also including also some preformulated grades, which are highly recommended for
direct compression. Formulators must choose the excipient grade on the basis of
the active ingredient, tablet size, and the process technology available. Therefore,
the ultimate use of the effervescent granules or tablets mainly affects the choice
of the raw materials.
To design an effervescent formula it is necessary to consider the stechiometric
ratios in the reaction, and the carbon dioxide solubility in water, which is 90 mg/
100mL of water (in STP conditions). The suggested ratio between acid and alkaline
components is about 0.6 but sometimes it might be required to increase the acid
source to get a pleasant taste. In fact, the alkaliacid ratio controls both the effervescence
capacity and the taste of the solution to be administered.
When the solubility of the active ingredients is not pH dependent, the alkaliacid
ratio can be optionally selected. This ratio can also be determined according to the pH
that is required for dissolving the active ingredient. In fact, when the active solubility
increases at the acid side, the pH of the solution is lowered by adding an excess of
the acidic agent. Conversely, an excess of alkaline source must be added when the active
ingredient is more soluble at higher pH (6). However, another approach that can be
used to increase the active ingredient solubility is to increase the volume of carbon
dioxide to be generated by increasing the alkaline component in the formulation.
As far as other excipients, such as diluents or binders, are concerned there is a
very little freedom for the formulator to experiment, because of the large dimension
of the tablet required for effervescent systems. In addition, compressibility cannot be
enhanced by additional binders for effervescent dosage form, because of the larger
size of the tablet.
In the latest development in effervescent forms, some formulations have been
designed to control the rate of effervescence, so to obtain a rapid, intermediate, or
slow rate. The rate control is strictly related to the acidalkaline components ratio,
but the chemical properties of the effervescent excipients or their combinations can
have influence on it, especially when a slow rate of effervescence is required.
Because of the nature of the effervescence reaction, additional excipients are sparingly
used as the alkaline and acid ingredients also act as fillers to get a tablet bulk.
Effervescent Granulation 367
 2005 by Taylor & Francis Group, LLC
They are in a so large an amount that effervescent tablets are much larger than conventional
ones. In case it is necessary to add a filler, sodium bicarbonate is the material
of choice due to its lower cost and because it does not influence final pH of the
solution and it also increases the effervescence effect. Sodium chloride and sodium
sulfate are other possible fillers. They are high-density crystalline powders that are
very compatible with the other ingredients.
Additives are added in a small amount to obtain tablets that are more attractive
to users. Flavors, colors, and sweeteners are used as usual in all the formulations.
4.1. Acid Materials
Necessary acidity for effervescence can be provided by three main sources: food
acids, acid anhydrides, and acid salts. Citric acid, tartaric acid, and ascorbic acid
are the most commonly used food acids because they are odorless, have a nice taste,
are not expensive, and are easy to handle.
4.1.1. Citric Acid
Citric acid is the more often used acidic ingredient because of its good solubility and
pleasant taste. It is commercially available in powder form and is either colorless or
exist as white crystals. The particle-size grades are: coarse, medium, fine, powder
(only anhydrous). It is very soluble in water and soluble in ethanol (7). It can be used
as monohydrate or anhydrate, depending on the selected equipment, technology,
and process conditions. It is very hygroscopic; however, the anhydrous form is less
hygroscopic than the monohydrate (8). However, caking of the anhydrous may
occur upon prolonged storage at humidity >70%. The monohydrate melts at
100C, and releases the water of hydration at 75C. For this reason, it can be used
as a binder source in hot melt granulation.
4.1.2. Tartaric Acid
It is very soluble in water and very hygroscopic compared to citric acid. In the effervescence
reaction with sodium bicarbonate it behaves like citric acid in producing an
evident effervescence. It must be used in a higher amount to get the proper stoichiometric
proportions, because it is a diprotic acid, while citric acid is a triprotic one.
In terms of compressibility it is comparable to citric acid (8).
4.1.3. Ascorbic Acid
It is white in crystalline form and light yellow in fine powder. It is not hygroscopic
and this may be helpful in production because it is easier to handle. It is freely soluble
in water (1 g in about 3 mL) and absolute ethanol (9). If exposed to light, it gradually
darkens. Its behavior in the effervescent reaction with sodium bicarbonate is
comparable to the other acids (citric and tartaric) in terms of release rate of carbon
4.1.4. Acid Anhydrides
Anhydrides of food acids are potential acid source as they are precursors of the corresponding
acid on hydrolyzation. The effervescent effect is strong and sustained by
the continuous production of acid in the solution. Water has to be avoided during
368 Bertuzzi
 2005 by Taylor & Francis Group, LLC
the whole process when anhydrides are part of a formulation. Otherwise they would
be hydrolyzed to the corresponding acid before their use in the product (10).
4.1.5. Acids Salts
Sodium dihydrogen phosphate and disodium dihydrogen pyrophosphate are acid
salts that have been used in effervescent formulation. They are soluble in water, producing
acid solution and react quickly with alkaline sources. They are commercially
available either as granules or powder.
4.1.6. Other Less Frequent Sources of Acid
 Fumaric and nicotinic acids, which are not hygroscopic, but have low water
 Malic acid has recently been introduced in effervescent formulations because
of its smooth and light taste. It is highly hygroscopic and soluble but has
less acid strength than tartaric or citric acids.
 Acetylsalicylic acid, though an active ingredient that is very commonly administered
in effervescent preparations, cannot be used as an acid source because
of its low water solubility.
 Adipic acid deserves to be mentioned even if it does not function as an acid
source because of its low water solubility. It has given good results as
a lubricant for effervescent calcium carbonate tablets (11).
4.2. Sources of Carbon Dioxide
Solid carbonates salts are the most popular source for effervescents; bicarbonate
forms are more reactive than carbonates.
4.2.1. Sodium Bicarbonate
Sodium bicarbonate is the major source of carbon dioxide in effervescent forms and
is able to provide a yield of 52% of carbon dioxide. It is commercially available in
five grades according to particle size, from free-flowing uniform granule to fine powder,
which are odorless and slightly alkaline in taste. When heated, the bicarbonate is
converted into anhydrous sodium carbonate. This reaction is time and temperature
dependent. Ninety percent of the conversion is achieved within 75 min at 93C, but
dehydration starts at 50C, which must be considered as a critical temperature in
processing (12).
Being a nonelastic material sodium bicarbonate has a very low compressibility
but this issue has been overcome since its production by spray-drying technique.
Direct compressible grades are now available but contain some additives such as
polyvinylpyrrolidone (PVP) or silicone oil.
4.2.2. Sodium Carbonate
Sodium carbonate is commercially available in three different forms, all very soluble
in water: anhydrous, monohydrate, or decahydrate (13). It is more resistant to the
effervescent reaction and in some formulations can be used as a stabilizing agent
in amounts not exceeding the 10% of the batch size. It is used as a stabilizing agent
in certain effervescent formulas because it absorbs moisture preferentially, preventing
the effervescent reaction from starting. Of course the anhydrous form is the
Effervescent Granulation 369
 2005 by Taylor & Francis Group, LLC
preferred choice for this purpose. Recently, a particular grade of sodium bicarbonate
has been produced as round-shaped particles, coated with a carbonate layer to
increase bicarbonate stability (14).
4.2.3. Potassium Bicarbonate and Potassium Carbonate
They substitute the sodium salts when sodium ion is not required (15). They are less
soluble than the corresponding sodium salts and are more expensive.
4.2.4. Calcium Carbonate
Precipitated calcium carbonate occurs as fine, white, odorless, and tasteless powder
or crystals. Its water solubility is very poor, and it is not soluble in ethanol or isopropanol.
It is a high density powder, and thus not suitable for compression. It is
normally used as a drug in effervescent tablets for patients who suffer from calcium
deficiency. It can also be used as alkaline source because it provides stability to the
effervescent system (16).
4.2.5. Sodium Glycine Carbonate
Sodium glycine carbonate provides a light effervescence reaction, but causes rapid
disintegration of the tablets, so it has been applied in the preparation of fastdissolving
sublingual tablets. It is much more compressible than the other alkaline
compounds and has been found suitable for direct compression (17).
4.3. Binders
The use of a binder in effervescent formulations is limited by the fact that any binder,
even if water soluble, will retard the tablet disintegration. Therefore, the amount of
binder in a given formula will be a compromise between desired granule strength and
desired disintegration time.
As will be described in Section 5.1.2, water itself is an effective binder for effervescent
granules when granulated with all the components together. A small amount
of water, very finely distributed on the powder bed, acts as a binder by partially dissolving
the raw materials so that they can be agglomerated. Other solvents, ethanol
and isopropanol are not binders themselves, but can be used as granulating liquids to
dissolve dry binders.
Binders for dry granulation, such as lactose, mannitol, dextrose, are almost
inappropriate, because they would be effective only in larger amount than that
allowed by an effervescent formulations. The binder choice in wet granulation is also
limited by the method of production and consequently by the amount of granulating
In case of granulation of both the alkaline and acid components together with
water, it would not make sense to put a binder in the formulation because the small
water amount will never be able to dissolve the binder.
The most popular binder for effervescent tablets is polyvinylpyrrolidone
(PVP). Types K25 and K30 are preferred for their water solubility and dissolution
rate which is an important parameter for the final purpose of effervescent tablets.
PVP is effective at low percentage in the formula, starting from 2% and over. It is
feasible either for dry or wet granulation. It is soluble in water, alcohols, and
hydroalcoholic liquids (18).
370 Bertuzzi
 2005 by Taylor & Francis Group, LLC
4.4. Lubricants
Because tableting is a critical step in effervescent production, selecting the lubricant
is one of the most important issues. Lubrication of the effervescent mixture is quite
problematic because of the chemicalphysical nature of the lubricants. Most of the
lubricants, because of their low water solubility, inhibit the tablet disintegration,
which, as already said, must be very rapid in case of effervescent tablets. The effervescent
tabletsmainly for marketing reasonsare often required to provide a clear
transparent solution, that is, without any insoluble scum forming on the water
surface, or any residue left. In selecting a lubricant, proper attention must be given
to its solubility in water along with its compatibility with the active ingredient.
Many different lubricants have been tested for a long time to establish the most
appropriate one for effervescent tablets (19), including the opportunity to carry out
external lubrication of the granules directly in the dies of the tablet-press.
Lubricant substances which are reported in literature as suitable for effervescent
manufacturing because they are water soluble are: sodium benzoate, sodium
acetate, L-leucine, and Carbowax 4000. A very recent application is a combination
of calcium and potassium sorbates, micronized polyethylene glycol with calcium
ascorbate or trisodium citrate (20). Combination of spray-dried L-leucine and polyethylene
glycol 6000 has been reported as a successful lubricant in the literature
(21). Other less soluble lubricants have been used in formulating effervescent tablets,
however, a balance should be found between compression efficiency and water
solubility. Magnesium stearate is also employed but the most suitable, commercially
available type is its combination with sodium lauryl sulfate, a surface-active agent
that helps in its dispersion (22).
4.5. Additives: Sweeteners, Coloring Agents, and Flavors
Coloring agents can include all the dyes soluble and suitable for food, such as those
the F.D.&C. ones and all the natural coloring substances in amounts varying in the
range of 0.13.5% of the total weight of the formulation.
Flavors can be selected from synthetic flavors or natural extracts. Lemon,
orange, and other fruit essences are particularly suitable to obtain organoleptic
requirement in amounts varying from 0.5% to 3.0% of the total weight of the
Manufacturing conditions are crucially important even with regards to stability of
the products after they have been packed. Almost all the raw materials used for effervescent
tablet manufacturing are hygroscopic, and hence moisture absorption from
the air must be prevented to avoid the effervescent reaction prior to use of the
mixing or granulating, lubricating, tableting, and packaging) can be carried out in
completely closed and integrated handling system, consisting of intermediate bulk
containers (IBCs), tumblers for IBCs, docking and dosing stations.
In this case only the packaging area will be ventilated with low moisture content
of air. Otherwise traditional open handling systems can be employed but the
whole facility has to be conditioned with air at minimum level of moisture content
Effervescent Granulation 371
The whole production process as shown in Figure 1 (dosing of the ingredients,
 2005 by Taylor & Francis Group, LLC
(23). In fact, the suggested conditions throughout the plant are: relative humidity
(RH) below 20% and uniform temperature at 21C. It is known, however, that
25% RH at controlled room temperature (25C) is enough to avoid instability caused
by atmospheric moisture (24).
Manufacturing effervescent drugs on a large scale is usually done using a semicontinuous
procedure, by paying attention to synchronize all the process steps, in
order to achieve the largest production throughput. A continuous process flow, with
continuous feeding of raw materials and collection of granules, can be performed by
extrusion of the wet mass and drying in a continuous fluid bed dryer. Granulation of
effervescent mixtures must, most of the times, be executed in batches and is definitely
the most critical step of this particular kind of pharmaceutical manufacturing, as it
strongly influences the characteristics of the final forms, granules or tablets, and consequently
the following steps of production. Lubrication of the granulation,
Figure 1 Integrated production plant.
372 Bertuzzi
 2005 by Taylor & Francis Group, LLC
compression of tablets, and packaging of effervescent tablets should be carefully
planned to produce a suitable effervescent product.
The critical issues discussed earlier, for the compression of effervescent granules,
are related to the low compressibility of the majority of the raw materials,
large dimensions of the tablets, poor content of binder in the formulation, and diffi-
culties of lubricating the mixture. For all these reasons the tableting phase has to be
carried out carefully, and even the choice of the tablet press is of great importance.
To overcome lubrication difficulties, some suppliers of tablet press have developed
equipment that is capable of carrying out external lubrication of the granules. Antiadherent
materials are sprayed directly into the dies of the tablet press, during the
pause phase of compression, so as to prevent sticking of granules upon dies and
punches. External lubrication is not as good a method as using lubricant substances
to disperse in the granules. It is not considered the best solution for lubrication
because it is considered less compliant to the standard GMP (good manufacturing
practices) rules; In addition, the assembling and disassembling of the tablet press
is complex. An alternative method to obtain better tablets, facilitating the compression,
is to tablet the granules while they are still slightly wet, or not dried yet as we
are still considering very low moisture content of <1%. Tablets are then dried and
stabilized by means of a process step in a static ventilated oven.
Packaging operations for both granules and tablets must be done, as already
mentioned, in a low humidity environment. The critical aspects about packaging
of effervescent drugs are obviously related to stability of the tablets and granules
and the main objective is to protect them, as much as possible, not only during
packaging operations but also after they are packed, so as to impart a reasonable
shelf-life. The oldest packaging method for effervescent preparations consisted of
wrapping the acid and alkaline components separately, to avoid the premature effervescent
reaction until use. All the effervescent drugs can be packed in individual dose
units, in airtight containers made of protective aluminum foil or plastic laminates.
Tablets can also be packed by stacking them one by one in plastic or metal tubes,
which have almost the same diameter of the tablets so as to minimize the air which
remains in contact with the tablets. The tubes must be resealed every time, after taking
out each tablet.
Certain tablets are also wrapped in an aluminum foil before packaging in the
tube and this seems to be the best solution for long-term stability. Patented types of
tubes containing silica-gel in the internal side of the cap are the most recent invention
For tablets that are packed in strips, it is essential that the packaging machine
has a fine control over the temperature of the welding unit, so that an accurate sealing
of the strips can be obtained and overheating phenomenon that could provoke
release of residual water from the tablets can be avoided (24).
5.1. Granulation Methods
Two main granulation methods have been known for a long time. The 1911 edition
of the British Pharmacopoeia reported a detailed description of the manufacturing
procedure (26).
Effervescent granules are made by mixing citric and tartaric acids with the
medicament, and the sodium bicarbonate with the sugar when present; and then
thoroughly mixing the one with the other, and granulating the resulting mixture
by stirring in a pan heated to between 93 and 104, passing through sieves of
Effervescent Granulation 373
 2005 by Taylor & Francis Group, LLC
a suitable size, and drying at temperature not exceeding 54. This method for
preparing the granules yields satisfactory results but the following alternative
method has also been suggested: Mix the sodium bicarbonate, the sugar, and
the medicament, pass the mixture through a No 20 to No 30 stainless steel sieve,
subject the mixed acids to the same process, and thoroughly mix the two sifted
powders. Place the mixed powders in layers on suitable dish, pan or glass tray,
heated to 75 to 85, if required but not exceeding the higher temperature. When
the mass, after being suitably kneaded and compressed, has assumed a uniform
plastic condition, suitable for granulation, rub it to a No 5 to No 10 stainless
steel sieve, according to the size of granules desired, and dry the granules at
a temperature not exceeding 50.
Only a few aspects have substantially changed in the modern methods. Two
main methods can be executed with various types of granulation equipment:
a. Single-step method. All the components of an effervescent formula are
granulated together, handled with care, running the process in an contained
manner to maintain the stability of the mixture until the step is
completed. This single-step method is normally carried out using dry
granulation, hot-melt granulation, and certain wet granulation processes.
b. Multistep method. The alkaline and the acid components are granulated
separately, then mixed together, just before the tableting or packaging
step. Usually these are also typical applications of wet granulation technologies.
5.1.1. Dry Granulation
Dry granulation by roller compactor is definitely the most appropriate method for its
simplicity, low costs, and higher product throughput. The number of operations and
space required are less and consequently, air ventilation is reduced. On the other
hand, not all the excipient grades are suitable for such a technology. More sophisticated
grades, and thus expensive, pre-prepared by raw material bulk suppliers are
An alternate process to dry granulation is direct compression of the blend of all
the raw materials, in the attempt to avoid operating and stability problems. It would
be the ideal process for effervescent tablet manufacturing, but its application is limited
to a few cases, for example, when the active ingredient cannot be granulated
(e.g., when it is already included in a complex like with a cyclodextrine), or contains
some water of crystallization.
5.1.2. Wet Granulation
Despite some disadvantages, wet granulation is still the most preferred method for
effervescent granulation. As required for conventional tablets, this method assures
homogeneous granules, suitable for compression, and is able to provide uniform
tablets either in terms of weight or active ingredient content.
For this technique, it is necessary to use a granulating liquid that might interact
with the powders initiating the effervescent reaction. Hence it is essential to handle
the process with great care.
A wet granulation process in two separate steps is mostly recommended and
suitable for conventional equipment like high-shear granulator-dryer and fluid bed
processor. Acidic and alkaline ingredients are granulated separately. The two granules
are then mixed together, just before adding the lubricant for tableting. Water,
374 Bertuzzi
 2005 by Taylor & Francis Group, LLC
alcohols, or hydroalcoholic solutions can be used as a binding liquid because this
process is a standard wet granulation process. The usual procedure is to granulate
only one of the effervescent sources and add the other one in powder in the final
blending. All these possibilities are illustrated in Figure 2.
In certain cases only the acid components of the formulation are granulated
and then mixed with the sodium bicarbonate, preferably when it is of fine granular
grade. Other additives such as flavors and lubricants can then be added and mixed.
This approach increases productivity and reduces costs since one granulation step
has been eliminated.
The peculiar process that distinguishes effervescent granulation consists in a
single-step granulation of all the components of the formulation, which can be performed
either with nonreactive or reactive liquids with reference to the effervescence
5.1.3. Single-Step Granulation
Single-step granulation process provides dry effervescent granules directly by granulating
the acid and the alkaline materials together. It is possible to use only water
as the granulating liquid, thus controlling the effervescent reaction to granulate.
A nonreactive liquid like absolute ethanol or isopropanol can also be used, but in
this case it is necessary to use a binder to agglomerate the raw material particles. Process with Water. A very small amount of water, less than 1% of
the batch size, can be used to initiate the effervescent reaction. Some carbon dioxide
is released, but some water too, which acts as the binding liquid as well. It takes a few
minutes (from 5 to10 min) to obtain wet granules. The effervescent reaction rate
rapidly increases and becomes difficult to stop; it must however be terminated very
quickly otherwise no product will remain. An immediate start of the drying of the
granules will control and stop the effervescent reaction. For single-step technology,
both fluid bed granulator and high-shear granulator-dryer are suitable.
Figure 2 Alternative granulation processes.
Effervescent Granulation 375
 2005 by Taylor & Francis Group, LLC
In high-shear granulator-dryer technology it is possible to suddenly switch to
drying phase by creating vacuum inside the bowl, just after the granulation phase
when wet granules are well massed. Vacuum is created in a few seconds, which
immediately provokes the decrease of the water boiling point down to about
20C. At the same time the bowl is heated up to provide more energy for water
evaporation. The water released on the granules surface is removed within a few
seconds and the effervescent reaction stops.
The application of microwave radiation, combined with vacuum inside the
bowl of the high-shear granulator (27), can also be used to stop the effervescent reaction
and to dry the effervescent granules (28).
Drying of effervescent granules is a shorter process than drying conventional
granules granulated with water, because of the very small amount of water involved
in the process. However, drying still remains a critical step since it is very hard to
remove, even the small quantity of water, from hydrophilic or hygroscopic materials.
Typical drying time is about 50 min for 20 kg batch of effervescent granulation under
vacuum. See Figure 3 for comparison of drying of conventional and effervescent
granules by vacuum technology.
Drying rate for effervescent granules is lower because the moisture content to
be removed is in a range below 2% of the batch size. Consequently, the drying time,
passing from pilot scale to industrial scale, does not increase as much as for conven-
The great advantage of this method for effervescent tablet preparation lies in
the installation of equipment that does not require explosion-proof systems. An
example of effervescent aspirin produced in 600 L high-shear granulator-dryer
The formulation consists of:
Anhydrous citric acid 116.6 kg
Sodium bicarbonate 154.2 kg
Sodium carbonate 39.2 kg
Acetylsalicylic acid 50 kg
Figure 3 Drying rate of effervescent granules, compared to conventional granules.
376 Bertuzzi
tional granules (Fig. 4).
equipped with vacuum and tilting bowl (29) is shown in Figure 5.
 2005 by Taylor & Francis Group, LLC
First, the effervescent system is granulated with very small amounts of water
(24 mL/kg) and sprayed in very fine droplets. The acetylsalicylic acid is added
later, in the final blending, after granulation is completed. The results of three
Figure 4 Drying time scale up for high-shear granulator-dryer.
Figure 5 Rotocube high-shear granulator-dryer equipped with vacuum and tilting bowl.
Effervescent Granulation 377
batches are reported in Table 1.
 2005 by Taylor & Francis Group, LLC
Table 1 Granules Produced with Single-Step Technology in a High-Shear Granulator-Dryer
Results of samples
size (kg) kg %
time (min)
neutralizing power
0.1N acid/tablet)
Batch #1
granulated with
720mL of water
(2 mL/kg)
360 352.2 97.80 30 0.048%
60 <0.01%
90 <0.01% 244.6 6.1
Batch #2
granulated with
1440mL of water
(4 mL/kg)
360 336.4 93.44 30 <0.01%
60 <0.015%
<0.01% 244.6 6.4
Batch #3
granulated with
1440mL of water
(4 mL/kg).
Note: no tilting
of the bowl while
360 323.5 89.86 60 0.075% 229.8
<0.01% 245.5 6.25
378 Bertuzzi
 2005 by Taylor & Francis Group, LLC
The lower yield of the third batch is due to product sticking to the walls of the
bowl, which could not be discharged. The reason why it happened is related to the
set parameters during drying phase: in batch #3, drying was performed by keeping
the bowl static, instead of tilting it as in the other batches. The batch size in actual
production was then increased to 644 kg in 1400 L equipment.
All the components of an effervescent mixture can be granulated together in
a single-step process in a conventional fluid bed granulator-dryer. Granulation occurs
when water is sprayed on the fluidized bed, initiating the effervescent reaction. The
reaction is stopped when water is not sprayed anymore and drying phase is carried
out with warm dry air. It is quite understandable that this method is difficult to
control and reproduce (30). A subsequent patent application (31) describes an
improvement of the method reproducibility, which can be achieved by controlling
the air humidity (it has to be less than 1 g/m3), using an hydroalcoholic solution
instead of water. Even better results in controlling the effervescent reaction are
achieved when spraying and drying phase are combined together.
It is, however, very difficult to reproduce such a process. Therefore, an alternative
procedure to manufacture effervescent granules has been invented using a rotor
fluid bed spray-granulator (32).
Warm air, which is the only method available for drying in fluid bed technology,
is not capable of stopping the reaction all at once as it happens by applying
vacuum inside the processing bowl. Therefore, the only way to proceed is to
minimize in some way the contact between the two components of the effervescent
system. An intelligent and brilliant hypothesis put forward by Professors P. Gautier
and J.M. Aiache (32) is to alternate the granulation of the acid materials with that of
alkaline materials while still using a single equipment such as a rotary fluid bed system.
In the literature a vitamin C formulation is reported to explain this technique
but other active ingredients can also be used (33).
The process consists of two or three continuous steps to produce effervescent
spheres by layering the acid components over alkaline spheres or vice versa. The
binding liquid is, however, a hydroalcoholic solution in which PVP the binder must
be previously dissolved.
The first step is the granulation of alkaline components in the rotary fluid bed.
In the second step, the granulating solution is sprayed in combination with the acidic
powders, which deposit on the alkaline spheres creating an external acid layer separated
by a neutral layer of the binder. When agglomeration is completed, drying
phase with hot air starts with no interruptions (32). Process with Alcohol or Hydroalcoholic Solution. As reported in the
previous example, it is sometimes preferable to granulate with a hydroalcoholic solution
to initiate a lighter effervescence so as to keep the reaction under better control
during the process. Use of alcohols is indispensable in case a binder like PVP is
included in the formulation. In fact, the amount of water required to dissolve
PVP to obtain the binding action will be too high, and it will not be possible to keep
the effervescent reaction under control.
As already suggested for a conventional granulation process that requires the
use of an inflammable liquid, it is convenient to install fully explosion-proof equipment
with an accessory solvent-recovery utility that will limit the emission of vapors
in the atmosphere. Solvent recovery will be more advantageous while drying under
vacuum than with a fluid bed.
Effervescent Granulation 379
 2005 by Taylor & Francis Group, LLC
Certain high-shear granulator-dryers are equipped to achieve 99% of solvent
recovery by installing a tank to collect the condensate, and by bringing down the
possible residual exhaust with a shower of water.
5.1.4. Hot Melt Granulation
Hot melt process is an alternative technology to wet granulation (which is discussed
in detail in another chapter of this book). Agglomeration of the particles of a powder
mixer can be achieved by melting hydrated citric acid so as to release the hydration
water which acts as the granulating liquid. Once granules are formed it is necessary
to cool them to achieve the proper hardness and mechanical stability. There are two
different techniques:
 Surface hot melt granulation (SHMG), which consists of mixing all the raw
materials together in a blender and then drying the mix in a tray oven at
90C. Water is then released from citric acid and other ingredients to form
granules (34). Unfortunately, batch reproducibility of this method is very
 Hot melt granulation is normally carried out in high-shear granulator-dryer
with the capability to heat up the bowl. In certain cases the released water
of hydration of citric acid initiates the effervescent reaction to produce
additional water which acts as the binding liquid. However, this process
for obvious reasons is difficult to control. The same process has been
applied to fluid bed spray-granulator where low melting point polymers,
like polyethylene glycols (PEGs) or polyoxyethylene glycols can also be
used as binders (35). Hot Melt Extrusion. Hot melt extrusion is a recent patented method
to produce effervescent granules, especially dedicated to produce granules having a
controllable rate of effervescence (36). The formulations for this technology must
contain a hot-melt extrudable binder. Preferred binders are the polyethylene glycols
with molecular weight in the range 10008000 Da, but some other polymers have
been investigated. Binder percentage varies according to the formulation in a range
of 2040% of the total weight. There are two main extrusion techniques to carry out
in extruders that must be equipped with a solid conveying zone, multiple separate
temperature controllable heating zones, and an extrusion die:
a. A blend of all the ingredients, including the active ingredient of the formulation,
is hot-melt extruded at high temperature, in order to melt or
soften the binder. The extrudate is then ground or chopped to obtain
effervescent granules.
b. The acidic agent and the hot-melt binder are formulated in the right proportions
to obtain an eutectic mixture that has decreased the melting
point temperature. This binary mixture is separately melted, the alkaline
agent is added as powder in the next step. The melted mixture is then
extruded, chopped, or ground as in the previous method.
To control the effervescence rate of the final dosage form, it is possible to
adjust some parameters like the temperature and the rate of extrusion. The temperature
range selected can be critical because degradation of the active agent may occur
alongside decomposition of the effervescent components. This range is usually from
about 50C to about 120C.
380 Bertuzzi
 2005 by Taylor & Francis Group, LLC
Table 2 Manufacturing Process for Effervescent Tablets
Dry granulation Granulation by heating Wet granulation
Slugging or
Surface hot
melt granulation
Hot melt
(two steps)
(one step)
Equipment Blender?tablet
press compactor
Blender Blender?tray
Single-pot mixergranulator-
bed dryer
of steps
6 3 6 5 6 4
time (h)
23 10 22 14 15 12
Advantages Fast process No product
No granulating
No product
No product
No granulation Totally closed No granulating
No granulation Short time
Disadvantages Raw materials
at low residual
Raw materials
for direct compression
at low residual
moisture (more
Loss of
Difficult to
carry out
to clean
to clean
Dusty process Long time Difficult
to clean
High manpower
Dusty process
High manpower
Note Obsolete Rarely applied
Effervescent Granulation 381
 2005 by Taylor & Francis Group, LLC
The rate of extrusion is related to the time of materials exposure to high temperature
which is usually less than 5 min.
The hot-melt extrusion technology in some cases can be run as a continuous
process, having a higher throughput than batch hot-melt granulation process per
All the previous sections provided an overview of all the possible technologies
to manufacture effervescent granules but how to choose the most appropriate technique
for a certain formulation. An interesting study to figure out the best produc-
Three technologies were evaluated: dry granulation, hot melt granulation, and
wet granulation. The choice of the process technology was strictly related to the physical
properties of the raw materials, namely particle size, density, flowability, and
moisture content. Moisture content is definitely the most significant parameter since
the powder mixture has less than 0.20.3% of moisture content. It would be stable
but difficult to tablet and pointing such a situation, a wet granulation process is
The use of high-shear granulator-dryer has been found as the more economic
and flexible production method by other researchers (38) as well. This technology
allows us to use a wider range of excipient grades, avoiding problems related to
particle size or moisture content of the raw materials. Even if the granules produced
by this technology appear finer, their flow properties are good despite the large
fraction of fine-sized particles. Tableting properties are always in line within the
2. NCF. December 1984 pp. 1922; May 1985 pp. 2123.
3. Moller PL, Norholt et al. J Clin Pharmacol 2000; 40(4):370378.
4. US Patent 6,200,604, 1999.
5. Rapp M. PCT Int Appl 2000.
6. US Patent 6,077,536, 1998.
7. Handbook of Pharmaceutical Excipients. Washington/London: The American Pharmaceutical
Association/The Pharmaceutical Society of Great Britain, 1986:7880.
8. Schmidt PC, Bro?gmann B. Dtsch Apoth-Ztg 1987; 127:991997.
9. Handbook of Pharmaceutical Excipients. Washington/London: The American
Pharmaceutical Association/The Pharmaceutical Society of Great Britain, 1986:68.
10. Repta AJ, Higuchi T. J Pharm Sci 1969; 58:11101113.
11. Romaco Zanchetta R&D Laboratory archive.
12. Handbook of Pharmaceutical Excipients. Washington/London: The American Pharmaceutical
Association/The Pharmaceutical Society of Great Britain, 1986:263265.
13. Handbook of Pharmaceutical Excipients. Washington/London: The American
Pharmaceutical Association/The Pharmaceutical Society of Great Britain, 1986:436438.
14. SPI Pharma Group. Fall 2001.
15. Duvall RN, Gold, G. Miles, Inc., USA, 1990, 6 pp.
16. US Patent 6,242,002, 1999.
17. US Patent 6,284,272, 1999.
18. Handbook of Pharmaceutical Excipients. Washington/London: The American Pharmaceutical
Association/The Pharmaceutical Society of Great Britain, 1986:234239.
382 Bertuzzi
tion method for effervescent tablets was presented by Laugier and Rona (Table 2)
 2005 by Taylor & Francis Group, LLC
Homan P. Pharm J 2001; 267:911936 (http://www.pharmj.com).
19. Strickland WA Jr, Higuchi T, Busse L. J Am. Pharm. Assoc. (Sci. Ed.), 1956; 45:5155.
20. Daher LJ. Bayer Corporation, USA. US Patent 1999, 6 pp.
21. Rotthaeuser B, Kraus G, Schmidt PC. Pharmazeutisches Institut, Eberhard-Karls-Univ.,
Tuebingen, Germany. Pharmazeutische Industrie 1998; 60(6).
22. Rudnic E, Schwartz JB. Oral Solid Dosage Forms, Chapter 45, Tablets, 861, Remington.
23. Armandou J-P, Mattha AG. Pharm Acta Helv 1982; 57:287289.
24. Mohrle R. Effervescent tablets. Lieberman HA, Lachman L, eds. Pharmaceutical Solid
Dosage Forms. Vol. 1. New York: Marcel Dekker, 1980:225258.
26. The British Pharmaceutical Codex. Published by direction of the Council of the Pharmaceutical
Society of Great Britain, 1911
27. Collette News Vol. 2, Issue 1, May 2001, GEA Powder Technology Division.
28. Aiache JM, Cardot JM. Utilisation des micro-ondes dans la fabrication des formes
pharmacutiques. Conf. INSA Roeuen, 1999.
29. Romaco Zanchetta R&D Laboratory Archive.
30. Coletta V, Kennon L. J Pharm Sci 1964; 53:15241525.
31. Patent EP 673,644.
32. Gauthier P, Aiache J-M. Pharm Technol Eur 2001; 13(10):32 (see also 34
33. US Patent 6,210,711, 1999.
34. Yanze FM, Duru C, Jacob M. Laboratoire de pharmacie galenique, pharmacotechnie
et de Biopharmacie, Universite Montpellier I, UFR de Sciences Pharmaceutiques,
Montpellier, France, Die pharmazie 12/2000.
35. Yanze FM, Duru C, Jacob M. Drug Dev Ind Pharm 2000; 26(11):11671176.
36. US Patent 6,488,961, 1999.
37. Laugier M, Rona R. PMPS Spring 2002:2628.
38. Pearlswig DM. Simulation Modeling Applied to the Single Pot Processing of
Effervescent Tablets. Masters Thesis, North Carolina State University, Raleigh, NC.
Effervescent Granulation 383
 2005 by Taylor & Francis Group, LLC
pp. 3637).
p. and
Melt Granulation and Pelletization
T. W. Wong
Faculty of Pharmacy, University of Technology MARA, Selangor, Malaysia
W. S. Cheong and P. W. S. Heng
National University of Singapore, Singapore
Melt granulation and melt pelletization are agglomeration processes that have
gathered increasing interest in the pharmaceutical industry for the concept of utilizing
a molten liquid as a binder. Unlike the conventional use of aqueous or organic
solvents as binders, the binding liquid in melt processes remains as a constituent
of the formulation. However, the basic principles in melt agglomeration processes
are relatively similar to those of wet agglomeration processes with solvents except
that the interpretation of the melt agglomeration processes is not complicated by
an evaporation of the molten binding liquid.
1.1. Overview of Agglomeration
Agglomeration is one of the key pharmaceutical manufacturing processes concerning
the conversion of fine solid particles into larger entities by agitating the fine solid particles
in the presence of a binding liquid using equipment such as tumbling drum,
fluid bed granulator, and high-shear mixer. The agglomeration is considered as
a wet process because a binding liquid is required for the wetting of solid particles
prior to agglomeration. The agglomerates formed have a specific mean size and
shape, improved flowability and mechanical strength, narrow bulk density and porosity
values, as well as, a modified drug release rate (1,2). Agglomerates can be classified
into two types: granules and pellets, based on their physical characteristics.
Granules are irregularly shaped agglomerates which have a rather wide size distribution,
typically within the range of 0.12 mm. Pellets are spherical agglomerates with
a narrow size distribution, within the range of 0.52 mm. Traditionally, the binding
liquid required for agglomeration of fine solid particles is aqueous or organic solvent.
The solid particles are held together by the binding liquid which provides
temporary binding force by means of liquid bridges, and subsequently by solid
bridges of the residual solute after the removal of solvent by evaporation. The hardness
of solid bridges largely determines the strength of the dried agglomerates (3).
 2005 by Taylor & Francis Group, LLC
The solid bridges can also be formed by the resolidification of molten material. This
is the basis by which meltable materials can be used as a binder in agglomeration.
1.2. Development of Melt Processes
The concept of using molten binding liquid for agglomeration instead of the conventional
binding liquid prepared from aqueous or organic solvent that has to be evaporated
off after agglomeration is completed was reported in some research
findings in the late 1970s and early 1980s (46). However, the concept was not widely
studied until more than a decade later and the period thereafter, using high-shear
mixer and fluidizing granulator as the processors for melt agglomeration (727).
High-shear mixer has gained greater popularity for use in melt agglomeration than
other equipment such as coating pan (4), drum granulator (28,29) or extruder
(3034) for it provides intense shear forces from the high-speed rotation of the
impeller. The impeller mixing action facilitates a more homogenous distribution of
molten binding liquid, leading to the formation of agglomerates with narrower size
distribution and uniform drug content. In addition, the meltable binder can be
brought to melting by the frictional heat generated from the high shearing forces
of the impeller, without external heat supply (8). Highly spherical pellets can be produced
by a continuous agglomeration and rounding effect of the high-speed impeller
rotation in a high-shear mixer.
The characteristics of melt agglomerates such as size, size distribution, and
shape change progressively with variations in processing and formulation parameters.
The melt agglomeration process can be broadly classified as melt granulation
or melt pelletization, based on the physical properties of the formed agglomerates
(35). For simplicity of presentation, melt agglomeration will be used as the term
representing both melt pelletization and melt granulation, and melt agglomerates
as the term for their resultant products in this discussion.
1.3. Requirement of Melt Agglomeration
Melt agglomeration is carried out by adding the binder to the fine solid particles
either in the form of a molten liquid or in the form of a solid that melts during
the process. Generally, an amount of 1030% w/w of binder, with respect to that
of fine solid particles, is used. The melting temperature of solid binder can be
achieved by external heat supply or heat generated from interparticulate friction during
high-speed mixing in the processor. A meltable binder suitable for melt agglomeration
has a melting point typically within the range of 50100C. Meltable binders
with melting point lower than 50C are generally unsuitable as the end products are
liable to melting, softening, or sticking during handling and storage. On the other
hand, binders of high melting temperatures are not desirable due to greater risks
of thermal instability to the drugs used when very high heat is required for the melting
of the binders. The meltable binder can be classified as hydrophilic or hydrophobic.
Examples of hydrophilic binders include polyethylene glycols and poloxamers.
Hydrophobic binders include fatty acids, fatty alcohols, waxes, and glycerides.
Hydrophilic meltable binders are used to prepare immediate-release dosage forms
while the hydrophobic meltable binders are preferred for prolonged-release formulations.
One or more meltable binders may be employed in the formulation of a batch
of melt agglomerates. The binding action of meltable binders can be manifested
through liquid bridges formed by the molten liquid. Alternatively, a softened
386 Wong et al.
 2005 by Taylor & Francis Group, LLC
semisolid can act as a binder (6,36). The local melting zone on the surface of a
semisolid binder could promote particle aggregation and subsequently agglomerate
growth. Nonetheless, a higher binder content up to 50% w/w was possible as there
was less binding liquid available for agglomerative action for a given weight of meltable
binder (6). Furthermore, the process of melt agglomeration is less sensitive
to the level of binder when softened semisolid is used as a binder. Frequently, the
process of agglomeration which employs softened semisolid binder is termed as
thermoplastic agglomeration.
The fine solid particles can be organic or inorganic in origin and in melt
agglomeration studies, inert substances commonly employed are lactose and dicalcium
phosphate. The melting point of fine solid particles should be at least 20C
higher than that of the maximum processing temperature. This prevents excessive
softening of the solid particles as they form the support for the molten binding liquid
during the building-up process of agglomerate structure. The fine solid particles may
be constituted by one or more components. The physicochemical characteristics of
both meltable binders and nonmeltable fine solid particles have crucial effects on
the outcome of a melt agglomeration and will be discussed in Section 3.
1.4. Advantages of Melt Agglomeration
The growing interest in melt agglomeration has largely been attributed to various
advantages of melt agglomeration which are not attainable using the conventional
wet agglomeration techniques. With an appropriate selection of meltable binders,
both immediate-and prolonged-release agglomerates can be prepared using a onestep,
one-processor approach (14,3742). Melt agglomeration is applicable for processing
of water-sensitive materials such as effervescent excipients and hygroscopic
drugs, thus omitting the use of either aqueous or organic solvent (43,44). With melt
methods, organic solvent, flame-proof facilities, and solvent recovery equipment are
not required, giving rise to lower cost of operation, and reduced ecological and toxicological
hazards. In addition, the formative process of melt agglomerates does not
require a drying phase and this shortens the total processing time. Melt agglomeration
is an ideal model for the elucidation of an agglomeration process as the process
of particle binding is not complicated by an evaporation of molten binding liquid
unlike the aqueous and nonaqueous wet agglomeration. Moreover, the physicochemical
properties of the molten binding liquid are potentially modifiable by incorporating
an additive (45).
Despite the various advantages offered by melt agglomeration, the technique is
not suitable for processing of heat-labile materials due to the involvement of elevated
temperatures. Nonetheless, successful agglomeration of volatile substances with
almost complete prevention of material loss by vaporization has been reported (46).
The probable reason is that the agglomerate surfaces was covered by layers of solidified
molten binder and loss of volatile substances was further avoided since the
melt process was in a closed chamber. In addition, Lactobacillus acidophilus bacteria
were found to have a higher survival rate in formulations prepared by continuous
melt technology than in formulations prepared using conventional wet techniques,
in spite of the molten mass temperature rising to more than 100C (47). Clearly,
the detrimental effects of heat were considerably reduced by the absence of moisture
in melt agglomeration.
The growth processes of melt agglomerates are highly sensitive to formulation,
processing, and equipment variables. Uncontrollable melt agglomeration is
Melt Granulation and Pelletization 387
 2005 by Taylor & Francis Group, LLC
frequently encountered when such variables are not optimally adjusted. In view of
challenges in controlling the melt agglomeration process, melt agglomeration has
remained an interesting field of research with the potential for practical and wider
application in the pharmaceutical industry.
1.5. Characterization of Melt Agglomerates
Over the past decade, many researchers have examined the effects of formulation,
processing, and equipment variables on melt agglomeration process through physicochemical
characterization of the formed agglomerates. This has contributed to
a better understanding of the formation and growth processes of melt agglomerates.
Similar to agglomerates produced by wet techniques, the most common method
employed for the characterization of melt agglomerates is the determination of mean
agglomerate size and size distribution by sieving. The size and size distribution of
melt agglomerates generally fit a log-normal mathematical model. They can be represented
by geometric weight mean diameter (dgw) and geometric standard deviation
(sg), respectively, as described by Sch?fer and Worts (48):
log dgw ? Pwi  log di
Pwi ?1:1?
log Sg ? ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffififfiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi Pwi?log di?2  ?Pwi  log di?2
Pwi s ?1:2?
where di is the mean diameter and wi is the weight of agglomerates of sieve
fraction i.
Alternatively, the size and size distribution of melt agglomerates can be
expressed as the model independent mass median diameter and span, respectively.
The mass median diameter is defined as agglomerate size corresponding to the
50th weight percentile of the cumulative agglomerate size distribution and span
being the quotient of difference between the agglomerate sizes corresponding to
the 90th and 10th weight percentiles of cumulative size distribution to mass median
diameter. The percentages of fines and lumps, generally described with respect to
agglomerate fractions smaller than 90250 mm and larger than 20004000 mm,
respectively, are also reported as an indirect quantitative input on size distribution
of melt agglomerates.
Apart from size and size distribution of melt agglomerates, shape, surface morphology,
specific surface area, moisture content, density, porosity, tensile strength,
friability, flow, and packing properties have often been determined in elucidation
of the mechanism and kinetics of melt agglomeration. The distribution of binder
content in melt agglomerates has a strong bearing on the homogeneity of melt
agglomerate growth. The actual binder content in melt agglomerates has also been
measured directly through experiments or estimated indirectly via the application
of mathematical modeling. A direct measurement of polyethylene glycol content
was determined using near-infrared spectrophotometry (20). The content of binder
in melt agglomerates was also determined with the aid of the high-performance
liquid chromatography analytical method (49). The polyethylene glycol content
was indirectly measured from the true density values of milled agglomerate fraction
determined by means of a gas displacement pycnometer (50) and the polyethylene
388 Wong et al.
 2005 by Taylor & Francis Group, LLC
glycol content was calculated using Eq. 1.3 as follows:
?1  x?
rl ?
rb ?
ra ?1:3?
where x is the fraction of polyethylene glycol, rl, rb, and ra are true density values of
lactose, binder, and melt agglomerates, respectively. The content of polyethylene glycol
can also be determined through gravimetric analysis whereby the weight of binder
is obtained by weight subtraction of nonbinder fraction from that of the melt
agglomerates (51,52). In the case of complexation and crystallite behavior of drug,
binder, and fine solid particles, the melt agglomerates have often been subjected to
evaluation by differential scanning calorimetry and x-ray diffractometry. In vitro
dissolution studies have been conducted to examine the drug release characteristics
of melt agglomerates. El-Shanawany (53) and Voinovich et al. (39) found that the
in vivo drug release attributes of melt agglomerates were in accordance with the
in vitro findings. Nonetheless, a quantitative relationship between the in vivo and in
vitro performances of melt agglomerates is still generally lacking.
Another important feature of melt agglomerates is their performance stability
during storage. It was found that the melt agglomerates could maintain their dissolution
profile after a year of storage at 25C and 60% relative humidity (40). The
release of sulfamethazine from tablets made of melt granules remained unchanged
after 2 years of storage at 25C in closed containers (37). Nevertheless, the rate
and extent of drug release were found to increase with agglomerates stored at an
elevated temperature of 40C and at a relative humidity of 75%, probably owing
to the softening of the melt matrices (41,54). On the basis of the meltable nature
of agglomerates, stability testing is preferred to be conducted at an appropriate
choice of storage temperature which will not bring about marked alterations to
the physicochemical state of the agglomerates.
The mechanism of melt agglomeration is similar to that of wet agglomeration,
except that the formation and growth processes of melt agglomerates are not complicated
by binding liquid losses via evaporation during the agglomeration process.
The growth process of melt agglomerates is dependent on the interplay between the
size enlargement and size reduction processes. The likelihood of an agglomerate to
grow in size or experience breakages is a result of the balance between externally
applied mechanical forces and agglomerate strength. The agglomerates will grow
in size if they have sufficient strength to withstand the impact of externally applied
forces or vice versa. The strength of agglomerates is affected by the relative magnitude
of capillary, frictional, and viscous forces. The capillary forces aid agglomerate
consolidation by pulling the solid particles together while the viscous and
frictional forces resist both consolidation and dilation of the solid particle assembly.
The mechanisms involved in the formation and growth of melt agglomerates
have been categorized into different stages and will be discussed in the following
Melt Granulation and Pelletization 389
 2005 by Taylor & Francis Group, LLC
2.1. Stages of Agglomerate Growth
The mechanism of wet agglomeration has traditionally been subdivided into nucleation,
coalescence, layering, abrasion transfer, crushing, and other concomitant
events such as snow balling and onion skinning, based mainly on the elementary
growth mechanism suggested by Sastry and Fuerstenau (Fig. 1) (55). An alternative
description of the agglomeration mechanism has also been proposed (2,56). The process
of agglomeration consists of a combination of three phases: wetting and nucleation,
consolidation and growth, together with the steps of attrition and breakage
2.1.1. Wetting and Nucleation
Nucleation is the initial phase of agglomeration in which nuclei or small agglomerates
of loose and porous structure are formed after the primary particles are wetted
by a binding liquid droplet. The primary particles are bound by liquid bridges in the
Figure 1 Elementary agglomerate growth mechanisms. (Adapted from Ref. 55.)
390 Wong et al.
(Fig. 2).
 2005 by Taylor & Francis Group, LLC
pendular state. The degree of liquid saturation of an agglomerate can be increased
either by continuous addition of liquid or through the consolidation of the agglomerate.
Two nucleation mechanisms, namely immersion and distribution, are proposed
by Sch?fer and Mathiesen (50), based on the process of melt agglomeration
agglomeration is a function of the ratio between the sizes of primary particles and
molten binder droplets (25,50,52).
Nucleation by immersion occurs when the size of the molten binder droplets is
greater than that of the fine solid particles. Immersion proceeds by the deposition of
fine solid particles onto the surfaces of molten binder droplets. The propensity of
nucleation by immersion method is promoted by large binder droplet size, high binder
viscosity, and low shearing forces. High binder viscosity and low shearing forces
reduce the opportunity of molten binder droplets to break down, thus keeping their
size comparatively large and several times the size of the fine solid particles.
In nucleation by distribution method, a molten binding liquid is distributed
onto the surfaces of fine solid particles. The nuclei are formed by the collision
between the wetted particles. The formed nuclei have a loose structure with
entrapped air unlike those produced by the immersion method. Generally, small binder
droplet size, low binder viscosity, and high shearing forces are favorable conditions
for nucleation by the distribution method.
The nucleation phase is characterized by the disappearance of fines (35), as
a consequence of coalescence between the wetted primary particles or the primary
particles with the formed nuclei. The resultant nuclei would undergo consolidation
under the impact of the externally applied mechanical forces and acquire sufficient
Figure 2 A recent approach in describing the agglomeration mechanism. (Adapted from
Ref. 2.)
Melt Granulation and Pelletization 391
(Fig. 3). The dominance of either mechanism in the nucleation process of melt
 2005 by Taylor & Francis Group, LLC
strength to resist further breakdown by impact forces and will be able to grow into
bigger agglomerates.
2.1.2. Consolidation and Growth
Agglomerate growth refers to progression in the size of the formed nuclei, typically
by binary coalescence and layering. The number of nuclei is progressively reduced
with an increase in size of the resultant agglomerates by coalescence. In contrast,
the number of agglomerates remains largely unchanged when size enlargement proceeds
by layering. The layering growth stage generally takes place after the agglomerates
have attained a certain size and rigidity and is associated with the reduced rate
of coalescence (57).
The probability of successful fusion between the collided nuclei is dependent on
the liquid saturation state of nuclei. During the process of agglomeration, the agglomerates
are consolidated by the agitation forces. The reduction in agglomerate pore size
and number promotes the migration of binding liquid from agglomerate core to surfaces,
thus enhancing the surface plasticity and propensity of agglomerate growth by
coalescence. The rate and extent of consolidation of an agglomerate is governed by
interparticulate frictional, capillary, and viscous forces. The interparticulate frictional
and viscous forces resist the consolidation process of agglomerates. Conversely, the
capillary forces promote the consolidation of agglomerates by pulling the particles
together (58). The net extent of consolidation for an agglomerate is dependent on
the relative magnitude of these forces, which is a function of agglomerate size, viscosity,
and surface tension of binding liquid, as well as the physicochemical properties of fine
solid particles (17,20,45,52,59,60).
The degree of liquid saturation of an agglomerate may be increased through
consolidating the agglomerates by agitation. Alternatively, it is attainable by adding
an additional volume of binding liquid to the existing agglomerates, making the liquid
bridges of nuclei progress from pendular to funicular and followed by the capillary
state where the level of liquid saturation can be as high as 80100%. The degree of
Figure 3 Modes of nucleation mechanism: (A) distribution and (B) immersion. (Adapted
from Ref. 50.)
392 Wong et al.
 2005 by Taylor & Francis Group, LLC
liquid saturation reaches and probably exceeds 100% prior to agglomeration when a
cohesive powder, highly viscous binding liquid, or an excessive amount of binding
liquid is employed. At the droplet state when excessively high liquid saturation is
attained, there is considerable risk of agglomerate overwetting and uncontrollable
agglomerate growth may result. In contrast to the conventional wet agglomeration
processes, the level of liquid saturation of melt agglomerates cannot be reduced via
the evaporation of solvent. The effects of excess surface wetness are counteracted
by the application of a very high impeller speed which aids in the dispersion of lumps
and large agglomerates or through the use of porous fine particles which could
accommodate binding liquid in pores and crevices, away from the agglomerate
2.1.3. Attrition and Breakage
Attrition and breakage refer to the phenomenon of agglomerate fragmentation in the
are solidified by tray cooling to ambient temperature without the need for drying by
a tumbling process. Consequently, breakage is known to have a more essential role
in affecting the resultant properties of the melt agglomerates during the agglomerative
The propensity of agglomerate breakage is largely governed by the strength of
the agglomerates. Rumpf (61) described the strength of moist agglomerates in their
static funicular and capillary states as follows:
st ? SC
1  e
cos y ?2:1?
where st is the tensile strength, S is the liquid saturation, C is a constant, e is the
agglomerate porosity, g is the surface tension of binding liquid, d is the surface mean
particle diameter, and y is the contact angle between the solid particles and the binding
liquid. The strength of moist agglomerates is promoted by an elevation in surface
tension of the binding liquid, liquid saturation, and wetting state of solid particles, as
well as a reduction in agglomerate porosity and mean size of solid particles.
The application of a binding liquid with low viscosity and surface tension
values increases the propensity of agglomerates to breakage (17). This deters the
agglomerates from consolidation. The regular breakdown of agglomerates aids in
the distribution of binding liquid across the solid particle bed ensuring homogenous
agglomerate growth. Nevertheless, it prevents a complete buildup of structure which
is needed for size enlargement process. In the case of spheronization, the regular
breakdown of agglomerates is translated to the absence of a plastically deformable
coalesced structure which is needed for rounding through shaping of agglomerate
surfaces by particle rearrangement.
The formation and growth processes of melt agglomerates are governed by formulation,
processing, and equipment variables to a greater extent than those of conventional
wet agglomerates as a result of the involvement of high impeller speeds and
high shearing forces for agglomeration and viscous molten binding liquid. Many
Melt Granulation and Pelletization 393
 2005 by Taylor & Francis Group, LLC
dry and wet states respectively (2). In melt agglomeration, the formed agglomerates
researchers had studied the influences of these variables, through the characterization
of melt agglomerates as previously illustrated, for a better understanding and
control of the melt agglomeration processes.
The effects of processing variables such as mixing time, mixing speed, mixer
load, jacket temperature, and method of binder addition on melt agglomeration have
been extensively investigated (8,11,13,43,52,6267). Typically, an increase in mixing
time or mixing speed promotes agglomerate growth through squeezing the molten
binding liquid from agglomerate core to surfaces by means of a densification process,
thus increasing the degree of liquid saturation of agglomerates and their propensity
to grow by binary coalescence following the collision between two or more plastically
deformable surfaces. Under the continuous stirring action of impeller rotation,
the deformable surfaces of these agglomerates can be rounded via intra-agglomerate
rearrangement of particles leading to the formation of spherical pellets.
The influences of equipment variables on melt agglomerate growth are more
marked with high-shear mixers than with low-shear mixers and fluid bed granulators.
One main reason is that the intensity of shearing forces is greater in high-shear
mixers. The high shearing forces promote a more even distribution of molten binding
liquid making the size distribution of melt agglomerates becoming narrower. The
level of shearing forces generated in a high-shear mixer is dependent on the geometry
of impeller blade, design of processing chamber, and relative dimension between the
blade and chamber. The construction of both impeller blade and processing chamber
can have a significant impact on the flow pattern of processing material. The use of a
truncated cone-shaped lid (11) and an inner wall lining made of polytetrafluoroethylene
(68) reduced the adhesion of mass undergoing agglomeration onto the wall of
the processing chamber. Variation in the curvature of impeller blade and the distance
from its base to the floor of the chamber likewise brought about changes in the flow
pattern of the wet mass (12,18). The pattern of material flow is related to the
mechanical force distribution and homogeneity of melt agglomeration. It has
a strong bearing on the disparity in size and shape of the formed agglomerates.
In the case of formulation variables, the effects of size, size distribution, shape,
density, and packing properties of fine solid particles have been reported (10,69,70).
The use of solid particles of mean size smaller than 10 mm is usually problematic in
melt agglomeration because a very high level of liquid saturation is needed to overcome
the high agglomerate strength resulting from the cohesiveness of small particles
in order to provide sufficient agglomerate deformability for growth. This, in turn,
can lead to a potentially uncontrollable melt agglomerative process. Generally, solid
particles with size ranges between 20 and 25 mm are preferable for the production of
melt agglomerates (10). Excessively large solid particles, with a mean size of over
100 mm, produce mechanically weak agglomerates or fines. The excessively large or
small solid particles will tend to produce irregularly shaped melt agglomerates as
such aggregates are susceptible to breakdown under the influence of shear forces
or have rigid particle linkages within the agglomerate which inhibit particle deformation
and rearrangement. The effects of size, size distribution, shape, density, and
packing property of fine solid particles on melt agglomeration are interdependent.
Size, size distribution, shape, and density affect the packing geometry of solid particles,
bound by molten binding liquid, within the agglomerate. The strength of particulate
interlocking, state of liquid distribution, and saturation of agglomerates can
be altered via the modification in these properties of fine solid particles.
The effects of binder volume, binder rheology, binder surface property, and
binder particle size on melt agglomeration have largely been reported in relation
394 Wong et al.
 2005 by Taylor & Francis Group, LLC
to the influences of other formulation or processing variables (15,17,20,36,45,70). The
growth of melt agglomerates is promoted predominantly by an increase in viscosity,
tack, and specific volume as well as a decrease in surface tension of the molten binding
liquid. The viscosity, tack, specific volume, and surface tension govern the intraagglomerate
mobility of molten binding liquid and state of liquid saturation of melt
agglomerates. The influences of viscosity, tack, and surface tension of molten binding
liquid on melt agglomeration are affected by product temperature, mixing speed, and
physicochemical properties of the fine solid particles. The effect of binder particle size
on melt agglomeration is less marked except when high-viscosity binder is used and at
early agglomerative phases (50). At the early agglomerative phase, the dimension of
droplets of a very viscous molten binding liquid can remain large and unreduced by
the high shearing forces, resulting in larger agglomerates for further growth.
The growth processes of melt agglomerates are sensitive to formulation, processing,
and equipment variables. A better understanding of the mechanism of melt agglomerate
formation aids in the development of methods to control and monitor the inprocess
changes of product attributes. In conventional wet agglomeration process,
the instrumentation of equipment to monitor and control the operation has been studied
by several investigators. The incorporation of sensors and suitable instrumentation
of the equipment enables changes in agglomerate properties during the
agglomeration process to be monitored such that the end point of the process can
be accurately predicted.
4.1. Image Processing/Infrared Spectroscopy/Bed Height
In fluid bed agglomeration process, the control of agglomerate growth through monitoring
agglomerate moisture content using infrared spectroscopy (7173) or agglomerate
size and shape using an image processing system (74,75) has been investigated.
The application of bed height control to monitor agglomerate growth in fluid bed
agglomeration process has also been studied (76,77). Of all instrumentation techniques,
the image processing system can potentially be adopted as a tool to control the
melt agglomeration process. Such a system requires a heat-resistant optic device capable
of withstanding the processing temperatures associated with the melt agglomeration
process of up to as high as 120C. The infrared spectroscopy or equivalent
techniques are less suitable for use in monitoring the melt agglomerate growth process.
Unlike the wet agglomeration process, the meltable binder is incorporated with
the processing material prior to mixing and shearing in the melt processor to bring
the binder to its melting temperature. Practically, there is no need for monitoring the
changes of binding liquid content added. Melt agglomeration in high-shear mixer
proceeds by the active centrifugal massing of the processing materials. In contrast
to the fluid bed agglomerator, the bed height of powder particles in the processing
chamber is less affected by gravitational forces and the upward air flow velocity
required for particle suspension. For a given material load, the bed height is expected
to have minimal effects on the changes in product attributes. It is envisaged that the
measurement of bed height would not be useful in reflecting the in-process product
Melt Granulation and Pelletization 395
 2005 by Taylor & Francis Group, LLC
4.2. Torque/Current/Power Consumption
For wet agglomeration in high- or low-shear mixers, the popular method for instrumentation
involves an indirect measurement based on the changes in rheological
properties of moistened mass which are related to the growth propensity of agglomerates.
The process of extrusion in extrusionspheronization had been monitored
using the measurement of power consumption of the extruder motor or extrusion
force on the screen (7884). Low-shear mixers had employed torque (8590), power
consumption (85,9193), or current (94) measurements as the tool for end-point
determination. The instrumentation technique for high-shear mixers involved the
measurement of torque on impeller shaft (95,96), power consumption (80,95110),
current (111,112) or motor slip (113116). Technically, power and current consumption
are the simpler techniques for monitoring the agglomerate growth in a highshear
mixer. It was easier to equip a high-shear mixer with a power consumption
meter than a sensor for torque or slip measurement (117). Torque generated had
been reported to provide a more descriptive profile of the agglomeration process
than power consumption. Nevertheless, there were only small differences between
the two methods for process monitoring (95,96,117). The measurement of torque
is a more expensive method and technically more difficult due to problems associated
with the transmission of signals from the sensor (118). Holm et al. (98) showed that
the power consumption profile of an agglomeration process was related to plasticity
and liquid saturation of the processing material in addition to that of the interparticulate
friction within the agglomerates. Since plastic deformation is important for
agglomerate growth, power consumption may then be related to the agglomeration
A slow and continuous binding liquid addition is a prerequisite for obtaining
a power consumption that is suitable for use in process control of wet agglomeration
the first phase, a wetting of powder mixture takes place without any observable rise
in the power consumption.
The subsequent addition of binding liquid brings about particle agglomeration
and a sharp rise in the magnitude of power consumption. The power consumption of
the agglomeration process levels off as the agglomerates gradually become dense and
saturated with the binding liquid. During the fourth phase, the power consumption
rises due to overwetting and excessive growth of agglomerates by coalescence. The
power consumption falls in the fifth phase and it is associated with the formation
of a suspension upon the addition of an excessive amount of the binding liquid.
The optimum end point of the process lies within the third phase of the power consumption
tracing. In the case of melt agglomeration process, the addition of meltable
binder is almost instantaneous. This lack of the typical power consumption profile
makes power consumption measurement less suitable for monitoring melt agglomeration
end point as in the case of the wet agglomeration process. In laboratory
high-shear mixers, an increase in melt agglomerate size is not reflected in the power
consumption signals except at an excessively high impeller speed (11,69). The
deformability of the agglomerates produced by a molten binding liquid is relatively
low in comparison to that of agglomerates formed when using an aqueous binding
liquid. The sensitivity of power consumption measurement is not sufficiently discriminating
to reflect small changes in the plasticity of melt agglomerates.
During the premelt agglomeration phase of the powder mixture, the level of
396 Wong et al.
(118). Five characteristic phases of power consumption are identifiable (Fig. 4). In
power consumption is low and with minimal fluctuations in its magnitude (Fig. 5).
 2005 by Taylor & Francis Group, LLC
The magnitude of the power consumption falls to a value slightly below that of the
values obtained during the premelt agglomeration phase when the binder particles
begin to melt upon the influence of frictional heat generated by the impeller rotation
(9). Such observation could be related to the lubricating effect of molten binding
liquid on the solid particles undergoing agglomeration. The power consumption rises
sharply following the complete melting of the binder particles. The molten binding
liquid is distributed rapidly throughout the powder mixture. The level of power consumption
becomes lower upon the complete distribution of the molten binding liquid
and increases as the agglomerate growth takes place.
Generally, a higher level of power consumption is recorded when a higher
impeller speed or a more viscous molten binding liquid is employed for the preparation
of melt agglomerates. The rise in power consumption is delayed when a viscous
meltable binder is used (119). The growth process of melt agglomerates is not easily
interpretable from the record of power consumption when the growth propensity of
melt agglomerates is strongly governed by the viscosity of the molten binding liquid.
This is ascribed to greater difficulties in distributing viscous molten binding liquid
and a slower rate of agglomerate densification. A viscous binder hinders the consolidation
of agglomerates as a viscous liquid resists plastic deformation and its migration
from core to the surface of the agglomerate during consolidation. The power
consumption measurements are not directly comparable between high-shear mixers
of different processing capacities. However, the values of power consumption can be
correlated to the changes in size of the agglomerates produced using a large-scale
high-shear mixer, albeit such relationship is not attainable when using a laboratory
scale high-shear mixer (12).
Figure 4 Power consumption profile of a wet agglomeration process. (Adapted from
Ref. 118.)
Melt Granulation and Pelletization 397
 2005 by Taylor & Francis Group, LLC
4.3. Energy Consumption
The applicability of both processing power and energy consumption as tools to
monitor agglomerate growth in a melt agglomeration process has been studied
(19,65). Postmelt specific power consumption, postmelt specific energy consumption,
and average postmelt specific power consumption were computed. The postmelt specific
power consumption was calculated by dividing the postmelt power consumption
read at specific postmelt processing time, with the total weight of processing material.
The postmelt specific energy consumption was obtained by integrating the area
of postmelt specific power consumption against postmelt processing time curve. The
average postmelt specific power consumption was represented by the quotient of
postmelt specific energy consumption to that of the corresponding length of postmelt
processing time.
4.3.1. Melt Agglomerate Size
The increase in the melt agglomerate size was more closely correlated to the postmelt
specific energy consumption than that of the average postmelt specific power consumption
or postmelt specific power consumption. This is because specific energy
consumption can be considered as the amount of work done for the formative
process of consolidating the primary particles into melt agglomerates. It is more indicative
of the stage of melt agglomerate growth. The postmelt specific energy
consumption and postmelt processing time were linearly correlated and the formation
of larger melt agglomerates at longer processing time involved a greater level
Figure 5 Profile of power consumption at premelt and postmelt phases. (Adapted from
Ref. 19.)
398 Wong et al.
 2005 by Taylor & Francis Group, LLC
of energy consumption. The postmelt specific energy consumption and melt agglomerate
size were linearly related and such a relationship was not dependent on the effect
of impeller speed except that of the binder concentration and size of solid particles. In
relating the influences of binder concentration and size of solid particles on the outcome
of agglomeration with respect to postmelt specific energy consumption, an
equation was established. The processes of melt agglomerate formation and growth
were described using a modified macroscopic population model. The model of melt
agglomeration, under the specified processing conditions, was summarized as:
d ? f 1?s??1 ? f 2?c?Emelt ?4:1?
where d refers to mean melt agglomerate size, Emelt is the postmelt specific energy
consumption, f1(s) and f2(c) represent polynomial relationships with size of solid
particles and binder concentration as independent variables, respectively. Using
Eq. 4.1, the theoretical value of melt agglomerate size was found in good agreement
with the experimental data obtained from melt agglomeration runs employing various
binder concentrations and different sizes of primary particles. Nonetheless, it
should be emphasized that the application of postmelt specific energy consumption
for the control of melt agglomeration process would be inappropriate in processes
encountering an excessive level of material adhesion or when large amounts of
lumps are formed. The effective weight of processing material is markedly reduced
in the presence of an excessive level of material adhesion. This will give rise to inaccuracies
in computing the postmelt specific energy consumption level. With the
excessive formation of lumps, the rope-like flow pattern of the agglomerates will
be affected. The movement of agglomerates, bound within the processing chamber,
is expected to be variable. As a result, the predictability of the melt agglomeration
process using postmelt specific energy consumption will be substantially reduced by
the changes in product attributes from one batch to another. Practically, Eq. 4.1 is
useful for predicting the end point of an optimal or near-optimal melt agglomeration
Frequently, the end point of an agglomeration process is governed by the size
of the formed agglomerates. The geometric-weight mean diameter of melt agglomerates
is calculated using Eq. 1.1 described by Sch?fer and Worts (48). In the case of
melt agglomerates with size distribution which cannot be fitted into a log-normal
relationship, the mass median diameter of agglomerates shall be adopted for representing
of the size of melt agglomerates produced.
4.3.2. Melt Agglomerate Shape
The shape and tensile strength of melt agglomerates may be used to characterize the
agglomeration process besides the more generally used size of melt agglomerates.
Similar to the measurement of size, both shape and tensile strength can be determined
using quick, simple, and accurate methods. The sphericity of melt agglomerates
is best evaluated using the image analysis system. The image analyzer consists of
a computer system connected to a video camera mounted on a stereomicroscope.
The average projected area and perimeter of melt agglomerates are determined from
the digitized images of agglomerates. By transformation of these basic parameters,
melt agglomerate sphericity is calculated. The sphericity value gives a measure of
melt agglomerate roundness. A sphericity value, determined by Eq. 4.2, of unity
Melt Granulation and Pelletization 399
 2005 by Taylor & Francis Group, LLC
describes a perfect circle. The agglomerate sphericity is defined as:
Sphericity ?
4p  Area
?Perimeter?2 ?4:2?
The melt agglomeration process can be divided into two spheronization phases,
a fast initial rate followed by a slow rate (65). The change in spheronization rate was
associated with that of agglomerate size, porosity, and flow pattern of the processing
material. The study established a biexponential mathematical model to relate the
agglomerate sphericity with postmelt specific energy consumption. The relationship
of agglomerate sphericity with postmelt specific energy consumption was found to be
independent of the effect of various production variables, such as postmelt impeller
speed, binder concentration, and size of solid particles. The biexponential mathematical
model has an equation as follows:
1  Sphericity ? ?0:4107 exp?0:0118Emelt? ? ?0:0533 exp?0:0004Emelt?
Overall, the characterization of the melt agglomeration end point using the
biexponential mathematical model is simple and without the need to consider the physicochemical
properties of the processing materials used. As the process of agglomeration
is aimed at making highly spherical pellets, predictive methods based on
sphericity will help to capture the ideal end point together with predictive methods
based on size.
4.3.3. Melt Agglomerate Tensile Strength
The formative process of melt agglomerates is accompanied by changes in the size,
shape, and porosity of melt agglomerates (19,65