Nikolaos A. Peppas (Editor) - Hydrogels in Medicine and Pharmacy-Fundamentals-CRC Press (1986)

Hydrogels
in
Medicine
and
Pharmacy
Volume I
Fundamentals
Editor
Nikolaos A. Peppas
Professor
School of Chemical Engineering
Purdue University
West Lafayette, Indiana
Boca Raton London New York
CRC Press is an imprint of the

Taylor & Francis Group, an informa business
CRCPress
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
Reissued 2019 by CRC Press
© 1986 by Taylor & Francis Group, LLC

CRC Press is an imprint of Taylor & Francis Group, an Informa business
No claim to original U.S. Government works
This book contains information obtained from authentic and highly regarded sources. Reasonable
efforts have been made to publish reliable data and information, but the author and publisher cannot
assume responsibility for the validity of all materials or the consequences of their use. The authors and
publishers have attempted to trace the copyright holders of all material reproduced in this publication
and apologize to copyright holders if permission to publish in this form has not been obtained. If any
copyright material has not been acknowledged please write and let us know so we may rectify in any
future reprint.
Except as permitted under U.S. Copyright Law, 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 (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC),
222 Rosewood Drive, Danvers, MA 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.
A Library of Congress record exists under LC control number:
Publisher's Note
The publisher has gone to great lengths to ensure the quality of this reprint but points out that some
imperfections in the original copies may be apparent.
Disclaimer
The publisher has made every effort to trace copyright holders and welcomes correspondence from
those they have been unable to contact.
ISBN 13: 978-0-367-24923-6 (hbk)

ISBN 13: 978-0-429-28509-7 (ebk)
Visit the Taylor & Francis Web site at http://www.taylorandfrancis.com and the
CRC Press Web site at http://www.crcpress.com
PREFACE
Hydrogels are cross-linked macromolecular networks swollen in water or biological fluids.

Since the discovery by Wichterle and Lim of the Czechoslovakian Academy of Sciences of
the remarkable biomedical properties of poly(2-hydroxyethyl methacrylate), a wide range
of hydrophilic polymers have been examined as potential candidates for replacement of soft
tissue or for other biomedical applications. More recently, hydrogels have become excellent
carriers for release of drugs and bioactive macromolecules either in their equilibrium swollen
state or as dynamically swelling systems.
The biocompatibility of hydrogels is attributed to their ability to simulate the natural tissue
due to their high water content and their special surface properties. In addition, hydrogels
are versatile materials since they can be rendered more or less hydrophilic by copolymerization
of two or more monomers. Their major disadvantage, i.e., their relatively low mechanical
strength, can be overcome either by cross-linking, by formation of interpenetrating networks,
or by crystallization which induces crystallite formation and drastic reinforcement of their
structure.
Although the polymer and biomedical literature contains numerous publications on the
preparation, structure, and applications of hydrogels, there have been very few efforts to
methodologically present their characteristics in a book. The 1976 book by J. D. Andrade
entitled Hydrogels fo r Medical and Related Applications, a collection of well-written papers
presented at an American Chemical Society meeting, has taken its place as a classic in
hydrogel literature. However, the absence of other books in this area is notable. It was
therefore, with great pleasure and enthusiasm that I accepted the invitation extended by CRC
Press, upon suggestion by Professor Jindrich Kopecek of the Czechoslovakian Academy of
Sciences, to edit a book on the biomedical and pharmaceutical applications of hydrogels.
The 20 chapters of this book are divided into three volumes. The first concentrates on
fundamental aspects of hydrogel chemistry and physics, the second on polymers used as
hydrogels, and the third is dedicated to a presentation of the biomedical and pharmaceutical
applications of hydrogels. Despite the large number of subjects that have been included in
this book, it is impossible to cover all the applications of hydrogels in the biomedical area.
Yet, these 20 chapters offer a very good introduction to the uninitiated scientist, and an in-
depth analysis of the main applications of hydrogels to the expert researcher. The first four
chapters of Volume I are devoted to an introduction to the preparation and properties of
hydrogels in a general sense, and they contain a reasonable number of references which
substantiate the equations and molecular theories presented. Our effort in these chapters is
not to give an exhaustive analysis of the plethora of research contributions written on the
structure of hydrogels, but rather, to direct the reader to those structural parameters that are
most important in biomedical applications. Thus, Chapter 1 presents the main methods of
preparation of hydrogels and discusses their basic network structure. Chapter 2 includes a
detailed analysis of the theoretical and experimental determination of the structural parameters
that are needed for an adequate understanding of the properties of hydrogels. In Chapter 3
we offer a complete analysis of the diffusive properties of hydrogels both from the classical
point of view (statistical mechanics and free-volume theories) and in terms of scaling con-
cepts. Chapter 4 includes an expert analysis of the surface properties of hydrogels with
emphasis on the experimental methods of analysis of their surface.
The next two chapters of Volume I and the five chapters of Volume II discuss fundamental
studies related to the hydrogel structure. These are exhaustive literature reviews which point
out the most important findings and they discuss the main structural characteristics of specific
categories of hydrogels. Since hydrogels are excellent carriers for immobilization of biom-
olecules and cells, Chapter 5 offers a throrough analysis of the main reactions and experi-
mental results of these immobilization techniques. Chapter 6 discusses the protein adsorption
on hydrogel surfaces. The five chapters of Volume II address the preparation, structure,
thermodynamics, physical, and mechanical properties of four of the most important biomed-
ical hydrogels, i.e., poly(vinyl alcohol), poly(2-hydroxyethyl methacrylate), polyethylene
glycol) and their copolymers, and cellulose derivatives.
In Volume III we have classified polymeric hydrogels according to their biomedical
applications and we offer ample evidence of their diverse biomedical uses. Thus, Chapters
1 and 2 address questions related to their biocompatibility and blood compatibility. Appli-
cation of hydrogels as contact lens materials continues to be a major area of research, as
discussed in Chapter 3, whereas their use as artificial tendons (Chapter 4) is one of the
newer areas of application for hydrogels. Additionally, more chapters discuss the use of
hydrogels for pharmaceutical applications both as equilibrium swollen gels (Chapter 5) and
as swelling-controlled release systems (Chapter 6).
Bioerodible hydrogels, discussed in Chapter 7, have attracted considerable research interest
in pharmaceutical and biomedical uses in recent years. Hydrogels with certain surface
functional groups have a tendency to interact with and adhere to mucus, thus becoming
important carriers for a variety of medical applications, as discussed in Chapter 8. Finally
the third volume concludes with a brief description of hydrogels as materials for other
medical uses such as artificial vocal cords, skin, cartilage, etc.
The natural uneveness of a multi-authored book has been alleviated by a delicate editing
effort, where extensive cross-listing of chapters, equations, and theroies lead to an integral
presentation of the material of the book. Thus, the various chapters of the book should not
be read as isolated reviews, but as portions of a general effort to present the theory and
practice of hydrogels in a methodological way. Repetition of certain ideas in two or more
chapters has been allowed only where absolutely necessary for the continuity of the book.
The distinguished colleagues and co-authors of this book must be thanked not only for
replying to my request to contribute their critical evaluation in their corresponding area(s) of
expertise, but also for agreeing to the proposed editing changes that made this book more
“ readable” . Their contributions to the book are invaluable and most appreciated.
Special thanks are due to the editorial staff of CRC Press, and especially to Mr. Chris
Richardson, for their assistance during all stages of production of this book. The School of
Chemical Engineering of Purdue University offered generous support during the preparation
and typing of several of the chapters of this book. Special mention must be made of the
contributions of my graduate student Antonios G. Mikos who, in addition to his co-authorship
of one Chapter, did the preparation of many of the figures of this book. Finally, I am indebted
to a number of my former and present graduate students for their contributions not only to
this book but also to our research work on hydrogels in general. They are B. D. Barr-
Howell, Dr. R. E. Benner, Jr., Dr. G. W. R. Davidson, III, N. M. Franson, B. Gander,
Dr. T. W. B. Gehr, P. J. Hansen, R. S. Harland, M. H. Honey, Dr. T. H. Kim, Dr. R.
W. Korsmeyer, Dr. L. M. Lucht, S. R. Lustig, D. L. Meadows, A. G. Mikos, Dr. D. R.
Miller, J. G. Mounts, H. J. Moynihan, C. T. Reinhart, C. C. R. Robert, Dr. S. Segot-
Chicq, J. L. Sinclair, M. J. Smith, R. A. Sorensen, and Dr. W. H. Yang. To them and
the person who introduced me to hydrogels, Professor E. W. Merrill of M. I. T., this volume
is dedicated.
Nikolaos A. Peppas
THE EDITOR
Nikolaos A. Pep pas, Sc.D., is a Professor of Chemical Engineering at Purdue University.

A native of Greece, he received his Diploma in Chemical Engineering from the National
Technical University of Athens, Greece in 1971 and his Sc.D. in Chemical Engineering
from the Massachusetts Institute of Technology in 1973. After 2 years of military service
at the Research Center of National Defense of Greece and 1 year of Research Associate at
the Arteriosclerosis Center of M .I.T., he joined Purdue University as an Assistant Professor
in 1976, was promoted to Associate Professor in 1978, and to Professor in 1981. In 1982—
83, he was Visiting Professor at the University of Geneva, Switzerland and at the California
Institute of Technology. In 1986, he was Visiting Professor at the University of Paris (Paris-
Sud). Dr. Peppas’ research interests include investigation of the structure of polymeric
networks, diffusion in membranes, transport mechanisms in glassy polymers, surface prop-
erties of polymers, biointerfacial phenomena, and iomedical applications of polymers (es-
pecially hydrogels) including blood compatible materials, materials for reconstruction of soft
tissues, and carriers for controlled release of drugs and biomacromolecules.
Dr. Peppas is the Editor of Biomaterials, a Contributing Editor of Polymer News, and a
member of editorial boards of the Journal of Applied Polymer Science, the Journal of
Controlled Release, and Biomedical Polymers. He is a member of the New York Academy
of Sciences, the American Physical Society, the American Chemical Society (where he
served as the 1982— 85 National Program Committee Chairman of the Division of Industrial
and Engineering Chemistry, and in various other positions of other divisions), the American
Institute of Chemical Engineers (where he serves as the First Vice-Chairman of the Materials
Engineering and Sciences Division of Chairman of the New Materials Subcommittee), the
Society of Plastics Engineers (where he serves as a Director of the Medical Plastics Division),
the Controlled Release Society (where he served as a governor and Vice-President and was
elected the 1986— 87 President-elect), the Society of Rheology, the Society for Biomaterials,
the Adhesion Society, the American Society for Artificial Internal Organs, the International
Society for Artificial Organs, the American Association of Pharmaceutical Scientists, A A AS,
ASEE, the Biomedical Engineering Society, the North American Membrane Society, the
European Society on Membrane Science and Technology, the Greek Chemical Engineers
Association, and Sigma Xi.
Dr. Peppas is the author or editor of seven books or special issues, has authored or co-
authored 175 scientific articles, 65 papers in proceedings volumes, and 40 abstracts, and has
given 240 lectures. He was the recipient of the 1984 Materials Engineering and Sciences
Award of AIChE, the 1982 Zyma Foundation Award for the Advancement of Medical and
Biological Sciences, the 1980 Western Electric Fund Award of ASEE, the 1978 and 1985
A. A. Potter Engineering Awards of Purdue, the 1978, 1980, 1982, and 1985 R. N. Shreve
Awards of the School of Chemical Engineering of Purdue University, and the 1982 AIChE
Best Counselor Award.
CONTRIBUTORS
Barbara D. Barr-Howell, M.S. Steven R. Lustig, M.S.

Research Engineer School Research Assistant of Chemicals
Amoco Research Center Engineering
Naperville, Illinois Purdue University
Wayne R. Gombotz, B. A.
Research Assistant Antonios G. Mikos, M.S.
Department of Bioengineering School Research Assistant of Chemical
University of Wshington Engineering
Seattle, Washington Purdue University
Allan S. Hoffman, Sc.D.
Professor of Bioengineering Nikolaos A. Peppas, Sc.D.
Department of Bioengineering Professor
University of Washington School of Chemical Engineering
Seattle, Washington Purdue University
Thomas A. Horbett, Ph.D.
Resident Associate Professor Buddy D. Ratner, Ph.D.
Department of Chemical Engineering, Associate Professor
Center for Bioengineering Department of Chemical Engineering,
University of Washington Center for Bioengineering
Seattle, Washington University of Washington
Seattle, Washington
TABLE OF CONTENTS
Chapter 1
Preparation Methods and Structure of Hydrogels.................................................................... 1
Nikolaos A. Peppas and Antonios G. Mikos
Chapter 2
Characterization of the Cross-LinkedStructure of Hydrogels.................................................27
Nikolaos A. Peppas and Barbara D. Barr-Howell
Chapter 3
Solute Diffusion in Hydrophilic Network Structures.............................................................. 57
Nikolaos A. Peppas and Steven R„ Lustig
Chapter 4
Hydrogel Surfaces........................................................................................................................ 85
Buddy D. Ratner
Chapter 5
Immobilization of Biomolecules and Cells on and Within Synthetic Polymer
Hydrogels......................................................................................................................................95
Wayne R. Gombotz and Allan S. Hoffman
Chapter 6
Protein Adsorption to H ydrogels.............................................................................................127
Thomas A. Horbett
Index 173
Volume I: Fundamentals 1
Chapter 1
PREPARATION METHODS AND STRUCTURE OF HYDROGELS
Nikolaos A. Peppas and Antonios G. Mikos
TABLE OF CONTENTS
I. Introduction........................................................................................................................ 2
II. Hydrogel Preparation by Chemical Cross-Linking...................................................... 2

A. Cross-Linking of Polymers................................................................................... 2
B. Copolymerization/ Cross-Linking Reactions...................................................... 6
1. Reaction Characteristics.......................................................................... 6
2. Kinetic M echanism ................................................................................. 8
3. Molecular Weight Distribution...............................................................9
4. Gelation.............................................................. 11
5. Vitrification............................................................................................. 11
III. Hydrogel Preparation by Radiation ..............................................................................12

A. Effect of Ionizing Radiation on Polym ers.................................................... 12
1. Types of Ionizing R adiation................................................................ 12
2. Effects of Ionizing Radiation on Macromolecules............................. 13
3. Gas Evolution..........................................................................................14
4. Effect of O xygen................................................................................... 15
5. Cross-Linking Reaction.........................................................................15
B. Irradiation of Polymers in Solution................................................................ 17
IV. Network Structure and Defects.......................................................................................19
V. Semicrystalline H ydrogels.............................................................................................20
References 24
2 Hydrogels in Medicine and Pharmacy
I. INTRODUCTION
Hydrogels are water-swollen networks (cross-linked structures) of hydrophilic homopol-

ymers or copolymers. They are three-dimensional and the cross-links can be formed by
covalent or ionic bonds. Often, weaker forces such as van der Waals forces and hydrogen
bonds can serve as cross-links, thus forming swollen networks which behave as hydrogels.
Finally, semicrystalline, uncross-linked hydrophilic polymers may form hydrogels upon
swelling since the crystallites act as physical cross-links and do not dissolve in water.
Hydrogels may be classified in various ways depending on their chemical or physical
structure. A common classification, especially useful in biomedical applications, includes
neutral hydrogels, ionic hydrogels, and swollen interpenetrating polymeric networks (IPN).
Even before the landmark work of Wichterle and Lim1 which established the importance
of poly(2-hydroxyethyl methacrylate) hydrogels as excellent candidates for contact lens
applications, investigators had recognized the potential of hydrogels to be used in medical
and pharmaceutical applications. The classic book edited by Andrade2 in 1976 remains a
reliable source of much information in the area of biomedical applications of hydrogels.
Major research contributions on the preparation, structure, properties, and biomedical ap-
plications of hydrogels have appeared in the last 25 years. Although numerous research
groups have contributed to this area, one must recognize the special contributions of a dozen
leading groups. The Czechoslovakian group at the Academy of Sciences (O. Wichterle, J.
Janacek, B. Sedlacek, J. Kopecek) contributed much of the early work on the structure,
physical, and mechanical properties of hydrogels, especially hydrogels of PHEMA. The
research groups of the University of Washington (A. S. Hoffman, T. Horbett, B. D. Ratner),
the University of Utah (J. D. Andrade, S. W. Kim, D. E. Gregonis), and MIT/Harvard (E.
W. Merrill, E. Salzman, I. V. Yannas) have been instrumental in studying the surface
properties of hydrogels and investigating their blood compatibility and other biomedical
properties. The research groups of other investigators at the Retina Foundation (M. F.
Refojo), University of Toronto (M. V. Sefton), Research Triangle (H. Yasuda), and Purdue
have advanced our knowledge of the structure of hydrogels and have investigated various
other medical applications. During the same period, research in European institutions such
as the University of Strathclyde (N. Graham), University of Aston (B. J. Tighe), and
University of Naples (L. Nicolais, C. Migliaresi), in Israel (A. Silberberg) and in Japan (I.
Sakurada, T. Ikeda) has addressed many questions related to the preparation, structure, and
medical applications of hydrogels.
Preparation of swollen, hydrophilic, polymeric networks may be achieved by the following
techniques:
1. Cross-linking of a homopolymer or copolymer in solution or in the solid state with

subsequent swelling in water or in a biological fluid
2. Simultaneous copolymerization and cross-linking of one or more monofunctional and
one multifunctional monomer followed by swelling in an appropriate penetrant
Typical hydrophilic monomers used in the preparation of biocompatible hydrogels are sum-
marized in Table 1. In this chapter, we will give a brief overview of some of the methods
of preparation of hydrogels and we will discuss their molecular structure.
II. HYDROGEL PREPARATION BY CHEMICAL CROSS-LINKING
A. Cross-Linking of Polymers
One of the two main techniques of preparation of hydrogels is that of direct cross-linking
of linear or branched homopolymers or copolymers using a small amount of a cross-
Table 1
MONOMERS FOR PREPARATION OF HYDROGELS
Monomer Name Abbreviation Chemical structure
Hydroxyethyl methacrylate HEMA CH,=C(CH 3)COOCH,CH,OH

II Hydroxyethoxyethyl methacrylate HEEMA CH,=C(CH,)COOCH 2CH,OCH,CH,OH
Ill Hydroxydiethoxyethyl methacrylate HDEEMA CH,=C(CH 3 )COOCH 2CH,OCH,CH:i20CH,CH,OH
IV Methoxyethyl methacrylate MEMA CH,=C(CH,)COOCH 2CH,OCH 3
V Methoxyethoxyethyl methacrylate MEEMA CH,=C(CH 3 )COOCH,CH 20CH 2CH 20CH 3
VI Methoxydiethoxyethyl methacrylate MDEEMA CH,=C(CH 3 )COOCH,CH,OCH,CH 20CH 2CH 20CH 3
VII Ethylene glycol dimethacrylate EGDMA CH,=C(CH 3 )COOCH 2CH 200CC(CH,)=CH,
VIII N-vinyl-2-pyrrolidone NVP CH,=CHNCOCH 2CH,CH,
IX Methacrylic acid MA CH,=C(CH,)COOH
X Vinyl acetate VAc CH,=CHCOOCH 3
~
12'
~
~
:-:-
~
;:::
~
~
~
iS
1::;"'
CM
linking agent. This reaction is carried out in solution (usually aqueous solution for biomedical
applications), although suspension polymerization may be used for the production of hy-
drogels as microparticles.
Since most hydrophilic polymers have pendent hydroxyl groups, difunctional or poly-
functional agents that condense with organic hydroxyl groups may be used as cross-linking
agents. These include various aldehydes such as formaldehyde, acetaldehyde, and glutar-
aldehyde,3'7, maleic and oxalic acid, dimethylurea, polyacrolein, diisocyanates, divinyl sul-
fate, and ceric redox systems.
The solvent for these reactions is usually water, but alcohols such as methanol, ethanol,
and benzyl alcohol may be used, provided that upon production of the network structure,
the reaction solvent can be exchanged with water to form the final hydrogels.
An important parameter often used for identification of the final cross-linked structure is
the (nominal) cross-linking ratio, X, which is defined as the ratio of moles of cross-linking
agent to the moles of polymer repeating units. For difunctional cross-linking agents forming
networks (see also Figure 1), the number average molecular weight between cross-links,
Mc, may be determined from the simple relation
Mr
( 1)
2X
where Mr is the molecular weight of the repeating unit. For application of this equation, it
is assumed that all the cross-linking agent has reacted with the polymer.
The cross-link is usually a chemical bridge of molecular weight much smaller than the
molecular weight of the chains between two consecutive cross-links. However, in the Flory
representation of an ideal network, it is customary to depict a cross-link as a point, a
volumeless species with respect to the rest of the chain. Thus, a difunctional cross-linking
agent reacting with two linear chains to cross-link them forms in effect one tetrafunctional
cross-link in the network (see also Figure 1).
It is usually assumed that cross-links are produced randomly and that the number of cross-
links is proportional to the cross-linking ratio. Statistical theories of cross-linking have been
presented by Flory,8 10 Charlesby,11 and Saito.12
A typical reaction scheme for this type of cross-linking is shown below
fC H 2-C H }n + PY -> fC H 2-C H }k + XY

l l
X P
fC H -C H i,
End-linking and cross-linking reactions may also occur in the absence of cross-linking
agents if a free radical initiator can be used which forms free radicals in the backbone chain.
Chain fracture
~c h 2- c h -c h 2- c h ~ -» ~ C H 2-C H -C H -C H ~ + H’
X X X X
~c h 2- c h -c h -c h -c h 2~ -> ~ c h 2- c h -c h - + c h =c h 2~
X X X X
FIGURE 1. Schematic representation of the cross-linked structure of a hydrogel, including tetrafunctional (A)
and multifunctional (B) cross-links, chain ends (C,D), entanglements (E), loops (F), and unreacted polymer chains
(G). The distance between two consecutive cross-links, when expressed as molecular weight, is the molecular
weight between cross-links, Mc. Upon swelling by a penetrant (•) the space between cross-links (H) is available
for diffusion, solute flow, etc.
Cross-linking by side-chain fracture
~c h 2- c h -c h ~c h 2- c h -c h -c h -c h 2- c h ~
X X X X
~c h 2- c h -c - c h -c h 2- c h ~
X X X
End-linking by main-chain fracture
~c h 2- c h -c h c h 2- c h -c h -c h -c h 2- c h -
X X X X
c h 2
CHX
CH
c h 2
c h 2
Figure 1 presents a schematic representation of a hydrogel produced by this technique.

Ideal tetrafunctional cross-links (A) are the usual cross-links produced, but multifunctional
cross-links may be present especially when using a multifunctional cross-linking agent. Since
the reaction is almost never carried out to 100% conversion, unreacted polymer chains (G)
may be found in the hydrogel which may be removed by extraction (leaching). Network
defects such as entanglements (E) and loops (F) are often present.
B. Copolymerization/Cross-Linking Reactions
7. Reaction Characteristics
Copolymerization/cross-linking reactions are also used to produce polymer gels, especially
poly(hydroxyalkyl methacrylates). The chemistry of typical copolymerization/cross-linking
reactions has been discussed by Gregonis et al.13 and from a more industrial point of view
by Pedley et al.14
Initiators that can be used in these reactions include both radical and anionic initiators.
Azobisisobutyronitrile (AIBN) is widely used although Gregonis et al.13 discuss various
advantages of the use of azobismethylisobutyrate and other similar chemicals. Various
peroxides including benzoyl or cumyl peroxide can be used as well.
Solvents may be added during the reaction to decrease the viscosity of the solution. For
preparation of hydrogel microparticles, novel suspension polymerization techniques are also
available.15
Chain transfer is a typical problem in these reactions. For example, it has been reported16
that during the polymerization of HEM A in the absence of a cross-linking agent, chain
tranfer to the polymer may be observed which leads to the formation of a cross-linked
PHEMA.
Table 2 summarizes the main polymerization/copolymerization methods for the production
of polymer for hydrogels and discusses various disadvantages of the techniques.
The classical gelation theory1718 can be applied to monomethacryl-dimethacryl monomer
copolymerization/cross-linking reactions if all the methacryl groups have equal reactivities.
By analogy to monovinyl-divinyl monomer copolymerization/cross-linking reactions two
effects may give rise to deviations:19 the decreased reactivity of pendent methacryl groups
as they are shielded by the rest of the molecule and the internal cyclization reactions. A
kinetic model20 is presented for the free radical monomethacryl-dimethacryl monomer co-
polymerization/cross-linking reaction. The model can be applied for low cross-linking ratios,
X, where the cross-linking reactions are intermolecular.21
Table 2
POLYMERIZATION METHODS FOR PREPARATION OF BASIC POLYMERS FOR
HYDROGELS
Polymerization Problems related to polymer

method Important features preparation and purity
Bulk (mass) Only initiator and monomer needed; cross-linking agent High viscosity, difficult agitation lead
can be added to nonuniformity of product; residual
monomers
Solution Initiator, solvent and monomer needed; easy agitation; Chain transfer frequently gives broad
controlled heat transfer; polymer soluble or insoluble in MWD products; removal of solvent
solvent
Suspension Initiator, solvent, monomer, and suspending agent needed;
cross-linking agent can be added; polymer produced in
spherical or irregular particles depending on monomer/
suspending agent interfacial tension
Emulsion Initiator, solvent, monomer, suspending agent, and emul- Residual emulsifier, etc.
sifier needed
Gaseous Reaction in gaseous phase; high pressure; unknown Pure polymers; technique not applied to
kinetics many systems
Plasma Glow discharge; unknown kinetics New technique; ultrapure polymers;
high cost of manufacture
Volume I: Fundamentals
2. Kinetic Mechanism
The copolymerization/cross-linking mechanism consists of four steps: initiation, propa-
gation,and cross-linking, termination by combination and disproportionation, and chain trans-
fer to monomer. Two cases can be considered for a symmetric dimethacryl monomer, e.g.,
ethyleneglycol dimethacrylate (EGDMA), in regards to the reactivities of the methacryl
groups. First, the reactivities of the methacryl groups of the dimethacryl monomer are
independent of each other and second, the reactivities of the methacryl groups of the di-
methacryl monomer are dependent on each other. Here, the case of the reactivities of the
methacryl groups of the dimethacryl monomer being independent of each other is considered.
It is assumed that no intracross-linking reactions take place, no depolymerization reactions
occur, and the penultimate effect22 is negligible.
A representative reaction scheme is presented below. The actual number of reactions
involved at each step is shown in parenthesis.
Initiation (six such reactions)
I —» 2 A
A + M, —> P],0,o
Propagation and cross-linking (ten such reactions)
^ pi i
P p .q .r + M, Pp+Lq,r
P
x p.q .r
-4- p
' A x.y,z
—> O
x p ,q ,r + 1
-4- P
' 1 x , y — 1 ,z + 1
Termination by combination (three such reactions)
p p ,q ,r + Ap x.y.z —» mA Xp + x,q + y , r4
Termination by disproportionation (three such reactions)
k .d„
P p,q,r + P x ,y ,z M p ,q ,r + M x ,y ,z
Chain transfer to monomer (ten such reactions)
k f„
P p, q,r M ] > ^ p .q .r P 1.0.0
k f,z
P p .q .r + P x, y,z M p, q,r + P x ,y -I.z + 1 + Qo.0,1
Here I is the initiator, A is a molecule with initiated radical, M, is the monomethacryl

monomer, e.g., HEMA, and M2 is the dimethacryl monomer, e.g., EGDMA. We have
considered primary polymer chains, i.e., the polymer chains which would result if all cross-
links were severed.17 The symbols Pp q r and Qp q r represent living polymer chains with
monomethacryl and dimethacryl monomer terminal groups, respectively, and Mp q r is a dead

polymer chain. Three subscripts, p,q,r, are used to describe a primary chain; they refer to
the number of monomethacryl units, pendent methacryl groups, and cross-links per chain,
respectively.
3. Molecular Weight Distribution

The rate equations resulting from the aforementioned mechanism are as follows:
Initiator, I
r, = - k dI ( 2)
Initiated radical, A
rA = 2fkdI - [k„M, + ki2(2M2 + M3)]A (3)
Monomethacryl monomer, M,
rM, = [kj,A + (kpn + kf|l)P + (kp2l + kf2l)Q]M, (4)
Dimethacryl monomer, M2
rM2 = _ 2[ki2A + (kpi2 + kf|2)P + (kP22 + kf22)Q]M2 (5)
Total living polymer, P
rP = [kSlA + (kp21 + kf2l)Q]M, - [(kpi2 + kfl,)(2M2 + M3) ( 6)

+ (ktC|1 + kld|l)P + (klci2 + kId|2)Q]P
Total living polymer, Q
rQ = [ki2A + (kpi2 + kf|2)P](2M2 + M3) - [(kP2l + kf21)M, (7)

+ (ktcl2 + ktd|2)P + (ktC22 + kId22)Q]Q
Here, M3 is the concentration of pendent methacryl groups, defined as:
q(Pp.q.r + Qp.q.r + Mp,q,r) (8)

p = 0 q —0 r = 0
The terms P and Q are the concentrations of total living polymer with monomethacryl and
dimethacryl monomer terminals, respectively, defined as:
p ^ i i 2 pP.q, (9)
p= 0 q= 0 r= 0
q ^ i i i QP.q,r no)
p= 0 q= 0 r= 0
The moments of the molecular weight distribution of cross-linked copolymer for the total
polymer are defined as follows:
»k.j.k = 2 S X P'q'rk(Pp,q,r + Qp.q.r + Mpqr) (11)

p = 0 q = 0 r= 0
where i,j, and k are integers. From Equations 8 and 11 it is obvious that
<l»o.,.o = M, (12)
Based on the definition of the distribution moments, we obtain the rate equations for the
zeroth and first distribution moments. The rate equation for the concentration of total polymer
is
r,ooo = [kuM, + ki2(2M2 + M3)] A - l- ktcllP2 - ktcl2 PQ - ktC22Q2
+ [kfllM, + kfl2(2M2 + M3)]P + [kf21M, + kf22(2M2 + M3)]Q (13)
The rate equation for the concentration of monomethacryl units in the polymer is
V o .o - IKA + (kpu + k fi.)p + ( k P2. + kf2|)Q]M, (14)
The rate equation for the concentration of pendent methacryl groups in the polymer is
r*oio = [ki2A + (kpi2 + kfl2)P + (kp22 + kf22)Q](2M2 - M3) (15)
Finally, the rate equation for the concentration of cross-links is
M'O.O.I 2[k12A + (kpi2 + kfl2)P + (kp22 + kf22)Q]M3 (16)
The moments of the molecular weight distribution of the cross-linked copolymer are used
to evaluate important averages, e.g., the number average degree of polymerization (linear
polymer), Vn, the number average molecular weight (linear polymer), Mn, and the number
average molecular weight between cross-links, Mc. The equations are
4 V o ,o i,o ~F l|i0Q |
yn = (17)
^0,0,0
w iî,o,o W 2(i|iQj Q + l/2l[/001)

Mn = (18)
*K),0,0
^0,0,1
(19)
*1*0,0,0
Mn
Mc = ( 20 )
1 + v.
Here, w, and w2 are the molecular weights of the monomethacryl and dimethacryl monomer,
respectively and ve is the average number of effective cross-links per chain, i.e., the cross-
links resulting from intermolecular cross-linking reactions.
4. Gelation
The gel point corresponds to the incipient formation of an infinite network. Before the
gel point, all the polymerization steps may be assumed reaction controlled (no gel effect),
which implies that the reaction rate constants are constant. After the gel point, this assumption
is not valid particularly for the cross-linking, the termination, and the chain transfer to
pendent methacryl group reactions, because they become diffusion controlled.23 The diffusion
of the polymer chains is segmental and not translational because they are joined with each
other. The propagation rate constants may be unchanged, because monomer diffusion is not
affected much by the gelation.
The gel point can be calculated using the proposed kinetic model. It corresponds to an
average number of effective cross-links per (linear) chain equal to two.
*1*0,0,1 _ 2
= ( 21 )
^0,0,0
Beyond the gel point, the proposed kinetic mechanism for the monomethacryl-dimethacryl
monomer copolymerization/cross-linking reaction does not change. The polymerization rate
constants of the diffusion-controlled steps do change. This change can be described using
a mobility factor, |x, defined as:
polymerization rate of cross-linked chains

ix = ——— —------------------------------------ ;------ (22)
polymerization rate of primary chains
The mobility factor depends on the mobility of the cross-linked chains. In other words, it
is a function of the average number of cross-links per chain, ve.
|x = fx(ve) (23)
Clearly, as the average number of cross-links per chain increases, the mobility factor decreases.
5. Vitrification
Vitrification is defined as the transition from a liquid (or rubbery) to a glassy state. In
general, the vitrification can occur before or after the gel point. At glass transition, the
monomer diffusion coefficient falls approximately four orders of magnitude, which results
essentially in the cessation of the polymerization reaction. Therefore, the cessation of the
polymerization reaction is not necessarily an indication of total conversion.
The condition for vitrification is that the reaction temperature, T, be equal to the glass
transition temperature of the polymer system.
T = T (24)
Here, Tg is the glass transition of the polymer solution or the swollen gel, depending on
whether vitrification occurs before or after gelation. The glass transition temperature of the
polymer system, assuming free volume additivity, is:24
T = + otp(l ~ v, - (25)
g ot|V, + a 2v 2 + a p(l - t>, - V2)
Here, a is the difference between the volume coefficient of expansion at the liquid (or
rubbery) and the glassy state, v is the volume fraction and the subscripts, 1,2, and p refer
to the monomethacryl monomer, the dimethacryl monomer, and the polymer, respectively.
Before gelation, the polymer consists of a mixture of linear and branched macromolecules.
The glass transition temperature of the polymer may be approximated to that of the linear
copolymer, that is a weighted average of the glass transition temperature of the
homopolymers.25,26
After gelation, the sol fraction is assumed to be zero. The glass transition temperature of
the cross-linked polymer depends on the degree of cross-linking. Thus, as the degree of
cross-linking increases, the glass transition temperature also increases.27 The shift in the
glass transition temperature is caused by (1) the degree of cross-linking and (2) the copolymer
effect, and it is related to the degree of cross-linking28 through Equation 26
~ TBd _ Xc
(26)
TgP 1 - Xc
Here, T 'gp is the glass transition temperature of the linear copolymer (when the cross-links
themselves are absent) and Xc is the mole fraction of monomer units which are cross-linked
in the polymer, which is
____*K),o,i_______
X, = (27)
^1,0,0 "f ^0,1,0 d" 4*0,0,!
Therefore,
^gp___^gp —. ____ 1_______ /^)o\

rr i , , Uoj
T gp 4 h ,o ,o 4*o,i,o
If the cross-linking ratio is small, the glass transition temperature of the linear copolymer,
T 'gp, can be approximated to the glass transition temperature of the homopolymer corre-
sponding to the monomethacryl monomer, e.g., PHEMA.
Figure 2 shows the effect of cross-linking ratio and reaction time on the Mc during a
typical copoly merization/cross-linking reaction of HEM A and EGDMA after the gel point.
High cross-linking ratios result in small average molecular weights between cross-links.
III. HYDROGEL PREPARATION BY RADIATION
Discussion of the effects of ionizing radiation on polymeric materials and their solutions,
study of the mechanisms and analysis of the results, can be found in books by Chapiro29
Bovey30 Charlesby,31 and Nikitina et al.32 and in review articles by Chapiro33 and Feng and
Freeman.34 In more recent reviews, Hoffman35 and Ikada et al.36 discuss various radiative
methods of preparation of hydrogels and point out the advantages of these techniques, which
include lack of impurities such as unreacted monomers, cross-linking agents, and emulsifiers.
A. Effect of Ionizing Radiation on Polymers

7. Types of Ionizing Radiation
Ionizing radiation covers a whole range of different types of radiations, some of which
are of primary source and others of secondary source. Ionizing radiation is electromagnetic
radiation of moving particles, which carry enough energy to ionize simple molecules either
in air or in water. Electron beam irradiation and y-radiation are used to cross-link polymers.
Artificially accelerated electron beams are available from several systems which deliver them
with energy from 0.5 to 20 MeV.
Volume 1: Fundamentals 13
50000
GP
1 : X -- 0.005
37500
2: X -- 0.010
3: X -- 0.020
u
1~ 25000
12500
0
0 1 2 3 4
t, hr
FIGURE 2. Effect of reaction time and cross-linking ratio, X, on the number average molecular weight between
cross-links, M, for the system 2-hydroxyethyl methacrylate (HEMA) and ethyleneglycol dimethacrylate (EGDMA),
cross-linked at 70°C, using I% mol AIBN as initiator. The cross-linking ratio, X, is defined as the ratio of mol
EGDMA per mol HEMA. GP designates the time of incipient gelation (gel point).
Electron accelerators, such as the commonly used Van de Graaff accelerator, substitute
for isotopes emitting 13-rays, which are generally not used in radiation polymer chemistry
because of technical problems. The penetration of fast electrons is much lower than that of
-y-rays. Commercially available Van de Graaff accelerators can deliver electrons with energy
2 to 10 Me V. According to the setting of the accelerator, the dose rate can vary from I 0 3
to 2 X 105 rad/sec. The term "dose" is used to describe the radiation received by a substance
placed in a radiation field. The concept of dose implies that energy is transferred from the
radiation source to the polymer and dose is therefore expressed in erg/g of irradiated polymer
or other similar units. One rad corresponds to absorbed energy of 100 erg/g of material.
The-y-rays are emitted by radioactive isotopes and they cover a broad range of energies.
6 °Co is the most commonly used source of two monochromatic -y-beams with energies of
1.17 and 1.33 MeV. Gamma-rays have very high penetrating power and the dose rate can
vary between 5 to I 00 rad/sec. The half-life of 60 Co is 5. 3 years, so that the dose rate should
be calculated as a function of time from the time of installment of a new unit.
2. Effects of Ionizing Radiation on Macromolecules

Kinetic ally the absorption of radiation by macromolecules causes ionization of the polymer
species, if the energy absorbed is greater than the ionization potential. Ionization is followed
by free radical formation. Hydrogel radicals formed in the previous step can react with
polymer radicals and no net effect is observed. This step is called recombination. However,
in almost every case polymer free radicals can react to form polymer chains or networks
(cross-linking), while other polymer chains may split, forming shorter chains (degradation).
Polymers either cross-link or degrade according to their chemical structure. According to
Chapiro,29 polymers that are cross-linked when irradiated include polyethylene, polypro-
pylene, polystyrene, poly acrylates, polyacrylamide, poly(vinyl chloride), PVA, poly(N-
vinyl-2-pyrrolidone), polysiloxane, polyesters, and polyamides. Polymers that are degraded
include poly isobutylene, poly methacrylate, polymethacylamide, poly(vinylidene chloride),
cellulose and its derivatives, and some fluorinated polyethylenes. A number of theories have
been advanced in order to account for the fact that some polymers cross-link, while others
degrade. It seems that large side groups in a long linear chain lead to degradation. As a
rule, among vinyl polymers those with the structure
-(C H 2-C )-n
cross-link, while polymers with the structure
-(C H 2-C )-n
degrade. The steric hindrance presented by a tetrasubstituted carbon in the chain is shown
by a somewhat higher heat of polymerization, so that a definite correlation exists between
the heat of polymerization and the tendency of cross-linking. Thus, the higher the heat of
polymerization, the more likely that the polymer will cross-link upon irradiation. Chapiro
reports that polymers with heat of polymerization higher than 16 kcal/mol have a tendency
to cross-link. Both cross-linking and degradation may occur simultaneously and many authors
have considered the radiation effects as a kinetic phenomenon, where the rate of cross-
linking or degradation predominates.
3. Gas Evolution
Gas evolution is a common phenomenon during the irradiation of polymers. For example,
in many cases, hydrogen formed by two hydrogen radicals is evolved, which, in the case
of hydrogels, is entrapped in the polymeric materials. The total yield of gases evolved after
irradiation of different polymers has been reported by Petrov and Karpov.37
When bulky samples of polymer are subjected to irradiation, the effect of gas evolution
is less pronounced, since oxidation is diffusion controlled. Oxidative degradation becomes
noticeable at very low dose rates, when the rate of diffusion of oxygen into the polymer is
higher than the rate of consumption of this gas by chemical reactions. Minimal attention
has been paid to this phenomenon in a number of earlier studies and this presumably accounts
for some of the discrepancies reported in the literature concerning the behavior of certain
polymers under irradiation.
4. Effect o f Oxygen
A very important factor in the production of cross-linked networks via irradiation is the
presence of oxygen during the cross-linking process. Many polymers which can be expected
to form cross-linked structures, degrade if irradiated in air under slow dose rates. A possible
explanation for this degradation is the formation of weak peroxidic bonds in the main
polymeric chain. These bonds decompose and cause oxidative degradation of the main chain.
Thus, a polymer radical can react in the following way:
Reaction with oxygen
-C H -C H 2- + 0 2 -C H 2-C H 2-
0 2
Decomposition and rearrangement
O
II
-C H -C H 2- ^ ~ C + -OCH2-
02 H
In many cases, breakdown of the polymer is more rapid in the presence of oxygen. Infrared
measurements indicate that hydroxyl and carbonyl groups are formed. At low dose rates,
oxygen can diffuse into the bulk of the polymer fast enough to provide sufficient oxygen
for peroxide formation. At high dose rates, the oxygen is rapidly used and it cannot be
replaced.38 The degradation mechanism and its kinetics have been analyzed by Sakurada et
al.39,40 and Grassie.41
In the case of polymers that can be affected by oxygen, irradiation in vacuo is necessary
for the production of a cross-linked network. Free radicals produced during irradiation of
polymers may become trapped in the irradiated material, especially when the temperature
is kept below the glass transition temperature. The lifetime of radicals is rather long (102
to 104 sec) in solid state.42 If the irradiated polymer is partially crystalline, then the trapped
radicals are even more firmly trapped. Nitta et al.43 used area measurements of ESR ab-
sorption spectral curves of irradiated polymers to determine the concentrations of produced
radicals.
5. Cross-Linking Reaction
During the process of cross-linking by irradiation, the molecular weight increases (pregel
reaction), in the presence of a good solvent a “ sol” extractable phase is produced (gel
formation), and a cross-linked “ gel” phase is eventually formed.
The quantity G, characteristic of the efficiency of radiation energy to form a cross-link
on a polymer chain, is defined as the number of initiations per 100 eV of energy absorbed
by the polymer. Radiation cross-linking of polymers is thought to occur by “ hot” H atoms,
that react preferably with CH2 groups, which then combine with adjacent radicals.29
The term end-linking is used to designate reactions of gel formation by the main-chain
fracture of long chain polymers. After this scission, the chain can form a network of infinite
extent (gel). Two end groups produced at a fractured site can react with neighboring mol-
ecules. The distinction between this process and cross-linking is shown in Figure 3. By
removal of a side chain or hydrogen, a radical is formed in the chain, which can be reacted
by a lateral link (cross-linking), or by the fractures of a degraded main chain (end-linking).
initial conditions
crosslinking
endlinking
FIGURE 3. Schematic representation of cross-linking and end-linking during irradiation of

a polymer.
Single main-chain fractures can only produce branching, for network formation multiple
sites of degradation are needed.
The variation of the molecular weight distribution in a polymer after irradiation is described
by a series of differential equations.12 The solution of these equations gives the conversion
at the gel point and the number and weight average molecular weights. Problems involving
cyclization and branching have been solved, showing that gel formation is retarded by
cyclization. It has been found that for any initial distribution, the gel formation occurs only
when the ratio of the two probabilities of degradation and cross-linking is smaller than four.
Complicated mechanisms where cyclization and end-linking occur simultaneously are dis-
cussed by Saito.4445
The variation of viscosity before and after gelation has been studied by Saito.46 G-values
of cross-linking and degradation can be calculated from the intrinsic viscosity in the period
of irradiation. It was shown that the intrinsic viscosity can decrease in certain cases,46 even
if it reaches infinity at the gel point. Such a behavior of the intrinsic viscosity is possible,
only if the initial molecular size distribution is sufficiently broad. The contribution of a
branched polymer to the intrinsic viscosity was studied by Katsuura.47 When impurities
hinder cross-linking, the relations between physical quantities of an irradiated polymer
containing impurities and its structure can be reduced to those of a polymer with no impurities,
if use is made of a reduced radiation dose instead of the real dose.
B. Irradiation of Polymers in Solution

Analogous phenomena have been observed during the irradiation of solutions of polymers.
Some polymers cross-link, others degrade. Irradiation of a polymer in solution can lead to
cross-linking at lower dose than for a solid polymer, because of the effect of the radiolysis
products on the polymer chains and the lower viscosity of the solution which allows free
radical diffusion.
A typical reaction mechanism is presented. In this mechanism, PH is the polymer and
SH is the solvent:
Activation
PH PH*
SH -> SH*
Free radical formation
PH* p + H*
SH* -> S’ + H*
Gas evolution
H* + H* -> H2
Recombination
P* + H* -> PH
S’ + H* -> SH
Energy transfer
PH + SH* PH* + SH
PH* + SH -> PH + SH*
Radical transfer
PH + S* -> P + SH
P* + SH -> S‘ 4 - PH
Cross-linking
P + P ^ P - P
Degradation
P
Am +1 AP*n
Irradiation of polymers in solution presents a new kinetic aspect since chemical changes
may result either from direct or indirect effects on the polymer via the solvent. The direct
1ooo~--~--~------~------~
-100
u
~
-2:w CRITICAL
CONCENTRATIO N
(/)
0
0
10
..J
w
(9
1
01 1 10 100
MOLE Ofo POLYMER
FIGURE 4. Relation between the minimum dose for gelation of a polymer

solution due to irradiation and the polymer concentration.
effect on the polymer molecule is independent of polymer concentration in the solution,

whereas the indirect effect is dependent upon the concentration. The concentration depend-
ence can account for the large changes produced in dilute solutions by relatively small
radiation doses. Protection against the indirect effect can be altered by additives which can
react with the radicals formed in the solvent.
Many polymers in solution undergo simultaneous intermolecular cross-linking and deg-
radation when irradiated. The relation between minimum dose for gelation and polymer
concentration is shown in Figure 4. No relation exists between the radiation sensitivity of
the solvent and the rate of cross-linking of dissolved polymers. However, cross-linking
seems to be slightly favored in poor solvents. A mechanism was proposed48 in which the
formation of macroradicals and low molecular weight radicals from the solvent by direct
action were the primary steps of the cross-linking reaction.
Kiran and Rodriguez 49 studied the effects of concentration on the gelation dose for different
degrees of polymerization. Especially interesting is their study on the mechanisms of cross-
linking and structures obtained at doses lower than the gelation dose. In unirradiated solutions
below their critical concentration, the polymer chains are not overlapping and segmental
mobility is high. At low irradiation doses, intramolecular linking and chain scission are
favored, the former decreasing the segmental mobility. While the viscosity is increasing,
intermolecular cross-links are formed and three-dimensional networks (microgels) appear.
However, the distance between such units becomes larger with continued radiation and the
system becomes diffusion controlled. Finally, intramolecular cross-linking (within the in-
polymer
molecule macroradical
')'-rays
RADICAL FORMATION
polymer radical CIWN SCISSION

chain site
INTRAMOLECULAR
CROSSLINKING
INTERMOLECULAR CROSSLINKING FOLLOWED

BY INTRAMOLECULAR CROSSLINKING
FIGURE 5. Schematic representation of radiation-induced charges in polymer chains in solution.
termolecularly linked networks formed in the previous way) and chain scission are favored.
Microgel units approach tight units and the solutions display turbidity (see also Figure 5).
IV. NETWORK STRUCTURE AND DEFECTS
The physical and other properties of hydrogels depend on the structure of the polymeric
networks that have been produced by the previous cross-linked reactions. Most hydrogels
for biomedical applications are noncrystalline and the structure of their cross-linked networks
can be investigated by a number of classical physicochemical techniques.
We review here the current status of our knowledge of noncrystalline polymer networks.
The term noncrystalline is preferred to the term amorphous because networks often contain
localized ordered structures or nonhomogeneous structures not suggested by the common
Flory picture of a randomly cross-linked mass of macromolecular chains. The most acceptable
definition of an ideal network is a collection of Gaussian chains between multifunctional
junction points (cross-links).
In nonswollen networks the chains may be (1) in an unstrained state (normal degree of
coiling), (2) in a supercoiled state, and (3) in an expanded state, resulting from a cross-
linked network swollen in a mixture of monomers, subsequently polymerized.
Real polymer networks always deviate from the accepted definition of an ideal Gaussian
network. Imperfections can result from deviations of the original conditions of cross-linking,
from cross-linking of already existing cross-linked networks and from end-linking.50
Imperfections are generally
• Pre-existing order
• Network defects
• Inhomogeneities
• Phase separation structures
Under the term pre-existing order, there are included many imperfections like crystallites,
exhibiting three dimensional order, nonrandomly oriented segment sequences, artificially
oriented chains and associations of similar groups yielding micellar, and globular orders
(supermolecular order). Cross-linking leads to fixing of structural features of the state existing
at the moment of reaction.50 Extensive experimental support for the existence of equilibrium
chain ordering in amorphous polymers is now available.50,51
Clustering (aggregation or association) as a result of dissimilar parts within one chain
molecule is well known especially in the field of biopolymers. Association of various
dissimilar parts of chains in solution as well as in bulk can be another reason for pre-existing
order in hydrogels. During aging of solutions, polymolecular micelles and parallel cylinders
with ordered chains and sheets are gradually developed which affect the final network.
Network defects include closed loops, unreacted functionalities, and permanent chain en-
tanglements as shown in Figure 6. Thus, the effect of defects is by no means simple.
Macro- and microsyneresis is the result of another imperfection, i.e., phase separation.
Phase separation occurs when the critical value of cross-linking density is exceeded because
the amount of the solvent in the gel exceeds the maximum swelling capacity of the gel in
the same solvent and under the same conditions (see also Chapter 2).
The final network is a composite one with imperfections because cross-links were intro-
duced in different unit volumes of the network phase (macrosyneresis). Slow relaxation of
the network in comparison with the rate of establishing local polymer-solvent phase equilibria
is one of the reasons for the phenomenon of microsyneresis. In both cases, the nonequilibrium
states of the network chains are locally stressed. Microsyneresis is expected to occur more
in lightly cross-linked networks, whereas macrosyneresis is expected in densely cross-linked
networks. In many cases, microsyneresis may be accompanied by strong hysteresis effects
(supercooling and superheating) and/or the formation of a new phase.
The aforementioned defects and imperfections are of utmost importance in the dynamic
and equilibrium swelling characteristics of a polymer network. In general, the equilibrium
swelling properties are a function of the hydrophilicity of the macromolecular structure and
the cross-linked structure as defined by the degree of cross-linking, number average molecular
weight between cross-links and other parameters that will be discussed in Chapter 2. Table
3 offers the main swelling characteristics of certain commonly used biomedical hydrogels.
In Chapters 2 through 6 of Volume I of this book, various authors will address the
characterization of the cross-linked structure of hydrogels and the structure of the polymers
from which they are produced. Table 4 presents a summary of well-known physicochemical
techniques with which one can determine the structure of hydrogels.
V. SEMICRYSTALLINE HYDROGELS
In many biomedical applications, hydrogels with improved mechanical strength and di-
LOOPS
ENTANGLEMENTS
UNREACTED FUNCTIONALITY
FIGURE 6. Network defects.
mensional stability are required. These materials are usually semicrystalline hydrogels pro-
duced by heat treatment (annealing) of noncrystalline hydrogels above their Tg.
In principle, crystallization of polymers can occur by several processes:
1. Crystallization from the melt usually produces spherulites. Spherulites are crystallites
formed from growth with a preferred chain orientation relative to a center (nucleus).
Polarized light reveals that the polymer chains are oriented tangentially around each
nucleus, the crystalline phase consists of a multitude of crystallites, and the spherulites
do not exceed 100 IJ.m in diameter.
2. Crystallization from dilute solutions often yields lamellar single crystals.
3. Crystallization of polymers in polymer-diluent systems is the typical method of prep-
aration of semicrystalline cross-linked networks and it will be discussed below.
Crystals are about lOO A thick while the length of a polymer chain is much longer. The
Table 3
EFFECT OF CHEMICAL STRUCTURE ON EQUILIBRIUM SWELLING OF SELECTED
HYDROGELS2
Main Structures: 1 4CH,- -CH CH -CHEG DM A f

-[C H -C H h 4CH2—CHEGDMAf
| m r r €
1 m Y Z
Y Y
Copolymer (III)
Homopolymer (I) Copolymer (II)
Equilibrium degree of
Homo- or copolymer Main features swelling (%)
PVA I, X=H, Y=OH 60—95 [f(Mc)J

PHDEEMA II, X C H 3, YCOOJCH2CH2OJ2 c h 2c h 2o h >90
PMDEEMA II, X C H 3, Y=C00[CH2CH20 ]2 c h 2c h 2o c h 3 >90
PDHPMA II, X C H 3 Y C O O CH2CH(OH)CH2OH 70—95
PHEEMA II, c c h 3, y c o o c h 2c h 2o c h 2c h 2o h 80
PMEEMA II, x c h 3 y c o o c h 2c h 2 o c h 2c h 2 o c h 3 62
Hydrogels in Medicine and Pharmacy
P(HEMA-co-NVP) III, X C H 3, Y CO O CH2CH2OH, W=H, Z= pyrrolidone ring 54— 56

P(HEMA-co-MA) III, x =w c h 3, y c o o c h 2c h 2o h , z c o o h 46
PHEMA II, x c h 3, y c o o c h 2c h 2o h 32—42 [f(Mc)J
P(HEMA-co-MMA) h i , x =w c h 3, y c o o c h 2c h 2o h , z c o o c h 3 10—40 [f(m,n)J
PMMA I, x c h 3, y c o o c h 3 0— 1
a Selected data from References 13, 14, 52, 53.

Table 4
EXPERIMENTAL TECHNIQUES FOR
DETERMINATION OF THE STRUCTURE OF
POLYMERS AND HYDROGELS
Structural parameter Major experimental techniques
Average molecular weights Dilute solution viscometry

(Mn,Mw,Mv,Mz) Membrane osmometry
Light scattering
Sedimentation
End-group analysis
Gel permeation chromatography (GPC)
Molecular weight distribution Gel permeation chromatography (GPC)
Degree of crystallinity Differential scanning calorimetry (DSC)
Thermomechanical analysis (TMA)
X-ray diffraction
Infrared spectroscopy
Glass transition temperature Differential scanning calorimetry (DSC)
Dilatometry
Dynamic mechanical experiments
Thermal stability Thermogravimetric analysis (TGA)
Degree of cross-linking Swelling studies
Rubber elasticity analysis
Degree of branching Viscometry
Light scattering
nature of the surface of single crystals is dependent on the crystallization conditions. The
kinetics of isothermal crystallization of hydrogels can be represented by the Avrami equation
X(t) = 1 - exp( —ktn) (29)
where X(t) is the crystallinity as a function of time and k and n are constants. The exponent
n is an integer number whose value depends on the time dependence of the nucleation rate
and on the growth morphology.
In polymer crystallization, short chains which cannot fold are rejected from the crystalline
phase and participate in the amorphous material. The critical chain length which is required
for folded-chain crystallization depends on the temperature, i.e., the degree of supercooling.
Molecular entanglements do not interfere seriously with folded-chain crystallization. The
rate of a primary polymer crystallization process is strongly temperature-dependent. The
thickness of the crystal, determined from low angle X-ray scattering, electron microscopy,
or melting point data, decreases with increasing crystallization undercooling.
The tendency of a polymer to crystallize is enhanced by regularity and polarity. Thus
generally, linear, isotactic polymers are semicrystalline while atactic, branched, nonlinear
polymers are not. However the isotactic form is not always the most crystalline. A classical
example is PVA for which more will be discussed later. Polarity favors crystallizability.
The structural effects of the network on the degree of crystallinity of hydrogels will be
discussed in greater detail in Chapter 7, using poly (vinyl alcohol) as a typical example of
crystallizable hydrogel.
REFERENCES
1. Wichterle, O. and Lim, D., Hydrophilic gels for biological use, Nature (London), 185, 117, 1960.
2. Andrade, J. D., Ed., Hydrogels for Medical and Related Applications, ACS Symposium Series, Vol. 31,
ACS, Washington, D.C., 1976.
3. Ogasawara, K., Nakamura, N., and Matusawa, S., The heterogeneous formalization of PVA accom-
panying crosslinking reaction, Angew. Makromol. Chem., 24, 89, 1972.
4. Sakurada, I., Shiki, Z., and Sakaguchi, Y., Intramolecular cyclization of PVA prepared by various
procedures, Kobunshi Kagaku, 21, 298, 1964.
5. Kawase, H., Intermolecular crosslinking of PVA with various monoaldehydes, Kogyo Kagaku Zasshi, 74,
1228, 1971.
6. Kawase, H., Intermolecular crosslinking in the acetalization of PVA with acetaldehyde, Kogyo Kagaku
Zasshi, 74, 1458, 1971.
7. Peppas, N. A. and Merrill, E. W., PVA hydrogels, J. Polym. Sci. Polym. Chem. Ed., 14, 441, 1976.
8. Flory, P. J., Constitution of three-dimensional polymers and the theory of gelation, J. Phxs. Chem., 46,
132, 1942.
9. Flory, P. J., Molecular size distribution in three-dimensional polymers. I. Gelation, J. Am. Chem. Soc.,
63, 3083, 1941.
10. Flory, P. J., Molecular size distribution in three-dimensional polymers. II. Trifunctional branching unit,
J. Am. Chem. Soc., 63, 3091, 1941.
11. Charlesby, A., Gel formation and molecular weight distribution in long chain polymers, Proc. R. Soc.
London, Ser. A, 222, 542, 1954.
12. Saito, O., Statistical aspects of infinite network formations, Polym. Eng. Sci., 19, 234, 1979.
13. Gregonis, D. E., Chen, C. M., and Andrade, J. D., The chemistry of some selected methecrylate
hydrogels, in Hydrogels for Medical and Related Applications, Vol. 31, ACS Symposium Series, Andrade,
J. D., Ed., American Chemical Society, Washington, D.C., 1976, 88.
14. Pedley, D. G., Skelly, P. J., and Tighe, B. J., Hydrogels in biomedical applications, Br. Polxm. J., 8,
71, 1976.
15. Mueller, K. F., Heiber, S., and Flankl, W., U S. Patent 4,224,427, 1978.
16. Nyquist, E. B., Hydroxy monomers, in Functional Monomers, Yocum, R. H. and Nyquist, E. B., Eds.,
Marcel Dekker, New York, 1973, 299.
17. Flory, P. J., Principles of Polymer Chemistry, Cornell University Press, Ithaca, N.Y., 1953.
18. Stockmayer, W. H., Theory of molecular size distribution and gel formation in branched polymers. II.
General crosslinking, J. Chem. Phys., 12, 125, 1944.
19. Minnema, L. and Staverman, A. J., The validity of the theory of gelation in vinyl-divinyl copolymeri-
zation, J. Polym. Sci., 29, 281, 1958.
20. Mikos, A. G., Takoudis, C. G., and Peppas, N. A., Kinetic modeling of copolymer/cross-linking
reactions, Macromolecules, in press, 1986.
21. Storey, B. T., Copolymerization of styrene and /?-divinylbenzene. Initial rates and gel points, J. Polym.
Sci. Part A, 3, 265, 1965.
22. Odian, G., Principles of Polymerization, McGraw Hill, New York, 1971.
23. Atherton, J. N. and North, A. M., Diffusion-controlled termination in free radical copolymerization,
Trans. Faraday Soc., 58, 2049, 1962.
24. Kelley, F. N. and Bueche, F., Viscosity and glass temperature relations for polymer-diluent systems, J.
Polym. Sci., 50, 549, 1961.
25. Wood, L. A., Glass transition temperatures of copolymers, J. Polym. Sci., 28, 319, 1958.
26. Pochan, J. M., Beatty, C. L., and Pochan, D. F., Different approach for the correlation of the Tg of
mixed amorphous systems, Polymer, 20, 879, 1979.
27. Fox, T. G. and Loshaek, S., Influence of molecular weight and degree of crosslinking on the specific
volume and glass transition temperature of polymers, J. Polym. Sci., 15, 371, 1955.
28. Nielsen, L. E., Crosslinking-effect on physical properties of polymers, J. Macromol. Sci. Rev. Macromol.
Chem., 3, 69, 1969.
29. Chapiro, A., Radiation Chemistry of Polymeric Systems, Interscience, New York, 1962.
30. Bovey, F. A., The Effects of Ionizing Radiation on Natural and Synthetic High Polymers, Interscience,
New York, 1958.
31. Charlesby, A., Atomic Radiation and Polymers, Pergamon Press, New York, 1960.
32. Nikitina, T. S., Zhuravskaya, E. V., and Kuzminsky, A. S., Effects of Ionizing Radiation on High
Polymers, Gordon & Breach, New York, 1963.
33. Chapiro, A., Radiation chemistry of polymers, Radial. Res. Suppl., 4, 179, 1964.
34. Feng, P. Y. and Freeman, E. S., Assessing radiation effects in polymers, in Physical Methods in Mac-
romolecular Chemistry, Vol. 1, Carroll, B., Ed., Marcel Dekker, New York, 1969.
35. Hoffman, A. S., Applications of radiation processing in biomedical engineering. A review of the preparation
and properties of novel biomaterials, Radiat. Phys. Chem., 9, 207, 1977.
36. Ikada, Y., Mita, T., Horii, F., Sakurada, I., and Hateda, M., Preparation of hydrogels by radiation
techniques, Radiat. Phys. Chem., 9, 633, 1977.
37. Petrov, I. Y. and Karpov, V. L., Investigation of gas evolution from the action of nuclear radiation on
polymers, Proc. All Union Conf. Radiat. Chem. Acad. Sci. U.S.S.R., 1, 263, 1958.
38. Collinson, E. and Swallow, A. J., The radiation chemistry of organic substances, Chem. Rev., 56, 471,
1956.
39. Sakurada, I., Okamura, S., and Kawasaki, S., The kinetics of degradation of macromolecular linear
molecules. I. Theoretical equations, J. Soc. Chem. Ind. Jpn. Suppl., 45, 415, 1972.
40. Sakurada, I., Okamura, S., and Kawasaki, S., The kinetics of degradation of macromolecular linear
molecules. III. Oxidative degradation, J. Soc. Chem. Ind. Jpn. Suppl., 45, 416, 1972.
41. Grassie, N., Polymer degradation, in Macromolecular Science, Vol. 8 , Bawn, C. E., Ed., Butterworths,
London, 1969, 278.
42. Watanabe, T., A remark on the chemical kinetics in polymers during irradiation, J. Phys. Soc. Jpn., 13,
1063, 1958.
43. Nitta, I., Onishi, S., Ikeda, Y., and Sugimoto, S., Radical production curve of irradiated polymers as
determined by ESR, Ann. Rep. Jpn. Assoc. Radiat. Res. Polym., 2, 314, 1960.
44. Saito, O., On the effects of high energy radiation to polymers. I. Crosslinking and degradation, J. Phys.
Soc. Jpn., 13, 198, 1958.
45. Saito, O., On the effect of high energy radiation to polymers. II. Endlinking and gel fraction, J. Phys.
Soc. Jpn., 13, 1451, 1958.
46. Saito, O., On the effect of high energy radiation to polymers. III. Viscosity, J. Phys. Soc. Jpn., 13, 1465,
1958.
47. Katsuura, K., On the viscosity of irradiated polymers. III. Effect of branching, J. Phys. Soc. Jpn., 15,
2310, 1960.
48. Henglein, A., Crosslinking of polymers in solution under the influence of 7 -radiation, J. Phys. Chem.,
63, 1852, 1959.
49. Kiram, E. and Rodriguez, F., Effects of 7 -radiation on aqueous polymer solutions. A comparative study,
J. Macromol. Sci. Phys., 7, 209, 1973.
50. Gordon, M. and Ross-Murphy, S. B., The structure and properties of molecular trees and networks, Pure
Appl. Chem., 43, 1, 1975.
51. Lipatova, T. E., Structural interpretation of the mechanism of polymer network formation, Pure Appl.
Chem., 43, 27, 1975.
52. Refojo, M. F., Contact lenses, in Encyclopedia of Chemical Technology, Vol. 6 , Interscience, New York,
1979, 720.
53. Peppas, N. A. and Merrill, E. W., Crosslinked hydrogels as swollen elastic networks, J. Appl. Polym.
Sci., 21, 1763, 1977.
Chapter 2
CHARACTERIZATION OF THE CROSS-LINKED STRUCTURE OF

HYDROGELS
Nikolaos A. Peppas and Barbara D. Barr-Howell
TABLE OF CONTENTS
I. Study of Hydrogel Networks......................................................................................... 28
II. Equilibrium Swelling T heory........................................................................................28

A. Theoretical M odels.............................................................................................. 28
1. Gaussian Models....................................................................................28
2. Non-Gaussian Models.......................................................................... 34
III. Rubber Elasticity Theory ...............................................................................................38
IV. Experimental Methods of Determination of Mc ......................................................... 44

A. Swelling Studies..................................................................................................44
B. Rubber Elasticity Studies...................................................................................51
V. Determination of the Mesh S iz e ...................................................................................54
VI. Conclusions...................................................................................................................... 54
Acknowledgments.......................................................................................................................55
References 55
I. STUDY OF HYDROGEL NETWORKS
The theoretical study of a hydrogel network has the purpose of revealing the structure
and configuration of its chains by use of appropriate theoretical models.
The knowledge of the cross-linking density (in other words the effective number of cross-
linked subunits, ve, and/or the number average molecular weight between cross-links, Mc)
is of great importance because of its effect on the mechanical and physical properties of the
produced materials and their behavior in practical applications. Characterization of the
hydrogel network structure is a complex procedure because of the many types of possible
networks, including, regular, irregular, loosely cross-linked, highly cross-linked, and im-
perfect networks (see also Chapter 1). Because of these variations in the network, only
average values for the cross-linking density and the molecular weight between cross-links
can be obtained experimentally or theoretically.
Methods of studying network structures include
• Chemical methods
• Swelling methods
• Elastic modulus tests
• Creep experiments
• Dynamic mechanical tests
• Shift of the glass transition temperature due to cross-linking
In order to determine the cross-linking density of a hydrogel, use is made of mainly two
theories, the equilibrium swelling theory and the rubber elasticity theory. The published
literature in these areas is voluminous. Therefore, in this chapter we have taken more of a
tutorial approach instead of reviewing all the existing literature. As for the methods of
analysis, we concentrate only on references discussing the structural analysis of biomedical,
especially hydrophilic, swollen networks.
II. EQUILIBRIUM SWELLING THEORY
A. Theoretical Models
7. Gaussian Models
Perhaps the most well-known theoretical model for determining the number average
molecular weight between cross-links, Mc, is the model of Flory and Rehner.1 Development
of the model is based on two important assumptions. First, the cross-linked polymer chains
can be represented by a Gaussian distribution, i.e., the end-to-end distance between the
chain ends is much smaller than the contour length of the chain. Second, the cross-links
are, on the average, tetrafunctional. A simplified derivation of this model12 is given here.
If an uncross-linked polymer is soluble in or thermodynamically compatible with a liquid
penetrant, then the same polymer when cross-linked will be swollen by this liquid.
Three types of data can be obtained from a typical equilibrium swelling experiment:
1. The amount of uncross-linked polymer, which is extracted and separated from the gel
is known as the sol fraction
2. The molecular weight distribution of the chains in the sol fraction
3. The amount of cross-linked (and therefore swollen) polymer, known also as the gel
fraction.
The degree of swelling of the gel fraction is typically expressed either as the equilibrium
volume or as the equilibrium weight swelling ratio, Q and q, respectively. The equilibrium
FIGURE 1. Swelling of a polymeric network by a thermodynamically compatible liquid (A). The distance
between two cross-links (B) is the number average molecular weight between cross-links, Mc.
volume swelling ratio, Q, is defined as the volume of the equilibrium swollen gel divided
by the volume of the same gel before swelling (or as often called, in the relaxed state).
Similarly, the equilibrium weight swelling ratio, q, is the corresponding ratio of the weights
of the two gel states. Typical values of Q for highly swollen gels vary from 5 to 100 and
sometimes even 1000. Typical values of moderately swollen hydrogels vary from 1.5 to 5.
As the network is swollen by the absorption of solvent (see also Figure 1), the chains
between the cross-links assume a rather elongated configuration, so that a force opposite to
the elastic retractive force of the chains develops. As swelling proceeds, this force increases
whereas the thermodynamic force of dilution decreases. Finally, a state of equilibrium
swelling is reached, at which both forces are equal.
When proceeding to an evaluation of the cross-linking density of a swollen network, one
considers expressions of the ordinary free energy of mixing, AGmix, and the elastic free
energy, AGel, consequential to the expansion of the network structure. These terms are
expressed via thermodynamic and elastic parameters of the system polymer-swelling agent.
The total change in free energy, AG, results from mixing a pure penetrant (swelling agent)
and an amorphous, isotropic polymer network. The change in free energy may be expressed
by the combination of the free energy of mixing, AGmix, and the elastic free energy, AGel.
AG = AGmix + AGel (1)
The relationship for AGmix of a polymer and penetrant2 is
AGmix = kT[n,lnv, + n2lnv2 + Xin,v2] ( 2)

where n ,, and n2 are the moles of swelling agent and polymer, v, and v 2 are the corresponding
volume fractions, k is the Boltzmann constant, T is the temperature, and Xi is the Flory
polymer-solvent interaction parameter.
For a cross-linked system without separate (uncross-linked) polymer chains (n2 = 0)
AGm,x = kT[n,lm>, + Xin,u2] (3)
The deformation process must occur without appreciable change in the internal energy
and therefore the elastic free energy is defined by Equation 4
AGel = -TA Se, (4)
where ASel represents the change of entropy resulting from the deformation process. For the
condition of isotropic swelling (a s = a x = a y = a z), the elastic free energy is given by
Equation 5.
AGel - (kTve/2)(3a~ - 3 - In a 3) (5)
Here, ve is the effective number of chains in the network and a s is the expansion factor
which expresses the linear deformation of a network structure due to isotropic swelling.
The chemical potential of a solvent, jjl , - |x,°, in a swollen gel is given by Equation 6.
m ~ |x? = N(dAGmix/dn,)TP -f N(dAGe/d a s)Tp(das/dn,)TP (6)
where N is the Avogadro’s number.

In order to evaluate the term (das/dn,)T P, the expansion factor, a s, must be known in
terms of the number of moles of swelling agent, n,. This relationship is given by Equation
7.
a 3 = i/u2 = (Vo + n,V ,/N)/V0 (7)
Here, V0 is the molecular volume of the relaxed network (network before swelling) and V,
is the molar volume of the swelling agent. The evaluation of the derivative of a s with respect
to n, results in the following relationship.
(das/dn,)T P = V ,/3afV0N (8)
Use of this relationship in Equation 6 and evaluation of the two remaining derivatives results
in Equation 9.
p,, - |i° = RT[ln(l - v 2) + u2 + Xi^2 + VjOW oXv^3 - v 2/ 2 ) (9)
We define the equilibrium polymer volume fraction, v 2tS, as the concentration at which the
activity of the swelling agent is one, or where p,, = |x,°. Then, the system is at the equilibrium
state.
—[ln( 1 - v 2nS) + v2,s + Xiv 2,s] = V,(ve/V q)( v '2/3s - v 2 J2) (10)
By replacing the term ve with more familiar quantities such as the number average molecular
weight between cross-links, Mc, and the number average molecular weight, Mn, of the
polymer before cross-linking, one may write:
ve = v(l - 2MC/Mn) (11)
where
v = V/UMC (12)
Here, V is the total volume and v is the specific volume of the polymer.
Substitution of these relationships into Equation 10 results in Equation 13.
- f ln ( l - v 2J + u2,s + X .< J = (V,/EMC)(1 - 2 M jM n)(v'£ - v 2J2) (13)
Upon rearrangement, this equation becomes the familiar Flory and Rehner1 model.
J _ = _2_ _ (U/V,)[ln(l - u 2s) + v 2s + Xi^2J

Mc Mn [vl2« - v j l \
It should be noted that the factor (1 -2M C/Mn) is the correction for network imperfections
resulting from chain ends; this factor reduces to one for perfect networks.
The previously presented equation of Flory and Rehner1 applies to swollen networks of
cross-linked polymers where the cross-links are introduced in the solid state (cross-linking
of a solid polymer). Peppas and Merrill3 developed a comparable model which takes into
account situations where the cross-links are introduced in the swollen state (cross-linking of
a solution) (see also Figure 2). The derivation of this model is given here, where v 2 r is
defined as the relaxed polymer volume fraction (polymer volume fraction immediately after
crosslinking but before swelling) and v 2i. is the swollen polymer volume fraction (polymer
volume fraction after equilibrium swelling).
As in the derivation of the model of Flory and Rehner,1 the change in free energy is
composed of a mixing term and an elastic retractive term.
AG = AGmix + AGel (15)
For a cross-linked network and solvent system, the free energy of mixing is given by Equation
16.
AGmix = kTCn.lm;, + x.n,u2) (16)
Here, n, is the number of moles of swelling agent, and v 2 are the volume fractions of
swelling agent and polymer respectively, and Xi is the Flory polymer-solvent interaction
parameter.
The elastic retractive term is
AGel = -T A S e, (17)
The expansion factor, a s, is defined as
1/3
as (18)
where V0 is the volume of the gel is the relaxed state. The contribution to the deformation
from the relaxed state to the swollen state (where a x, a y, a a are the three-dimensional
components of a s) can be used to determine the elastic contribution to entropy.
(a}
(b)
,..
(c)
FIGURE 2. Three situations of commonly performed cross-linking. Cross-linking of a solid polymer (a); Cross-
linking of a polymer solution (b); Copolymerization and simultaneous cross-linking during the reaction of two
monomers, one of which serves as a cross-linking agent (c).
ASel = C - y e [a* + a* + a 2z - 3 - ln(axa a z)] (19)
By definition ASel is
^ sw e lle d ^ re la x e d (2 0 )
Assuming an isotropic material (a s = a x = a y = a z) and recalling that in the relaxed state

a s = 1 and in the swollen state a 3 = Vs/Vr, ASel is defined by Equation 21
3kv
ASel = ----- Y [«s - 1 - lnas] (21)
which when substituted into the definition of AGe} results in Equation 22.
. 3kTve
AGel = [c£ - 1 - ln a j (22)
At equilibrium, the chemical potential of the swelling agent in the gel equals the chemical
potential of the solvent in the bulk.
dAGmix + dAGe,
M-i - Mm (23)
dn, dn,
Evaluation of the derivatives needed for the above equation yields
dAGm
= RT[ln(l - v 2J + v 2s + x,u2,s (24)
dn,
and
dAGei = RTV, / J _____ 2_\ r / M i a _ l /V2,s\ ]

(25)
an, V w c Mn/ V2rL U 2,y 2 \ u 2r/ J
Substitution of these relationships into Equation 23 yields Equation 26.
RTV / 1
RT[ln(l - v 2J + u2.s + Xlv U + (= - = o (26)
u \M C
This equation, upon rearrangement, yields the Peppas and Merrill3 model.
1 _ 2 (u/V,)[ln(l - v 2 s) + u2 s +
(27)
This expression reduces to the Flory and Rehner1 model when u2 r = 1, i.e ., when the cross-
links are introduced in the solid state.
The previous Equations 14 and 27 apply to two typical situations of cross-linked polymers
prepared by cross-linking of polymer chains either in the solid state or in solution. There
are, however, certain situations of production of polymer networks from monomers by
.... .... ~
I
I
I
I
Cn-2
Co
FIGURE 3. A macromolecular chain of the -c-c-c- type in

a cross-linked polymer.
simultaneous copolymerization and cross-linking of a monomer with one double-bond with

one that has two double-bonds and serves as the cross-linking agent. In these situations, the
value of M" required in Equations 14 and 27 is not available. Methods for solving this
problem are discussed in Section IV of this chapter.
2. Non-Gaussian Models
Equations 14 and 27 were derived with the assumption that a Gaussian chain distribution
was applicable in the network. There are, however, many situations of highly cross-linked
hydrogels, where this assumption is not valid. Highly cross-linked macromolecular networks
must be analyzed using non-Gaussian distribution models. Many non-Gaussian models exist
including models by Fixman et al. 4 •5 Kovac, 6 Peppas and Lucht/ and Galli and Brumage. 8
The developments of the non-Gaussian distribution models by Kovac 6 and Peppas and Luchf
are similar and are presented here in a simplified manner.
Consider a macromolecular chain consisting of N + 1 repeating units which are connected
by N bond vectors ~, with i = I, 2, ... N (see Figure 3). Then the end-to-end distance
of this chain, r, can be written as
r (28)
As this chain, which is part of a highly cross-linked macromolecular network, is extended

either due to application of external forces or due to swelling, a retractive equilibrium force,
f, is applied on the junctions defining the end-to-end vector r. Then, the distribution and
partition functions, 'll(t) and Z(t), respectively, can be written as
'll(f) exp(r · f - E)/Z(f) (29)
Z(f) f exp(r · f - E)db (30)

where E is the potential of the macromolecular chain.910 The statistical average of r can be
shown to be
r = d[lnZ(f)]/df (31)
The exact distribution function is expanded according to Fixman and Kovac5 in a series of
Hermitian polynomials orthogonal to the function exp( —g) with
g = - r • f + g(f) (32)
Fixman and Kovac5 chose g (f) in such a way as to make
W ) = exp(-g)/Z (f) (33)
Then g(f) can be obtained as
g(f) = ^2b iX (34)

= il b i - b i+il2 = ^21 i=
E iN 2
For a freely jointed10 chain a = 0 and p is given by
P = — [1 + (1 + 4f2b2/9)1/2] (35)
2b
Using the above relationships we obtain
r = - Nb2f[l + (1 + 4f2b2/9)12] - 1 (36)
which leads to
TAS = - I f(r)dr' + C = - Nln(l - r 2) + C (37)

l
The above analysis of Fixman and Kovac5 is applied to the swelling of macromolecular
networks6,7 by considering the total change of the partial molar Gibbs free energy with
respect to the swelling agent due to the sum of the changes in the partial molar Gibbs free
energy resulting from the expansion of the network, AGel and from the mixing of the swelling
agent with the chains of the network, AGmix. At equilibrium, the change in the partial molar
free energy must be zero.
AG = AGel + AGmix - O (38)
The total change in entropy of the network is the sum of the changes in entropy of the
chains, assuming the chains act independently.
One may now consider an isotropic cross-linked network composed of a total of M chains,
M/3 directed along each of the x,y,z axes. Then, Equation 37 may be rewritten as
TAS., = TASX + TASy + TASZ = y [jN (ln(l - X2) + ln(l - Y2) + ln(l - Z2)] (39)
where X,Y,Z are the end-to-end distances in each direction. It is further assumed that the
deformational extensions are affine,10 where the unstrained chains have the equilibrium mean
square end-to-end distance of the freely rotating chain, Fq, such that
X - a x(F2)172 (40a)
Y - a y(F2)172 (40b)
Z = a z(F2)172 (40c)
and
(F2) = Nb2 (40d)
Extension is assumed to be isotropic (a s = a x = a y = a z). The chain extension, a s, is

converted to equilibrium polymer volume fraction, v 2, by Equation 41.
a s3 = — = Q (41)
^2
We may also define the number of chains, M, in terms of the average molecular weight
between cross-links, Mc, the volume of the polymer, V, and the specific volume of the
macromolecule, v.
M = 3F=r (42)
vM c
The total change in free energy due to entropic changes of the network upon swelling is
described by Equation 43.
AG„ - 4H„ - TAS„ . + ,43,
Note that in Equation 44, the change in enthalpy of the chains due to the expansion of the
network is assumed to be zero.
The change in partial molar free energy due to mixing only can be described by the Flory-
Huggins equation2
AGmix = RT[ln(l - v 2) + u2 + Xi^|] (44)
Here, AGmix is the change in partial molar free energy on mixing, v 2 is the polymer volume
fraction in the swollen system, and Xi is the polymer-solvent interaction parameter.
By substituting Equation 44 and the partial derivative of AGel with respect to the number
of moles of swelling agent as obtained from Equation 43 into Equation 38, the modified
Gaussian network equations of Peppas and Lucht7 is obtained.
(WV,)[ln(l - u2 s) + u2 s + x^2
J _ _ _2_ ■J[> - s ’* ! ] ’
(45)
Mc ~ Mn
[< : - iv ,,] [ i +
The term N is defined as the number of bond vectors per chain and is expressed by Equation
46, where Mr is the molecular weight of the repeating unit and A. is the number of links per
repeating unit.
AMC
(46)
Mr
For all vinyl polymers, A. has the value of two.

As in the Flory and Rehner and model,1 the Peppas and Lucht model7 was derived for
swollen networks of cross-linked polymers where the cross-links were introduced in the
solid state. For polymers cross-linked in solution, 11 Equation 45 is modified to give Equation
47, according to the recent work of Peppas et a l.1214
(WV,)[ln(l - v2s) + v 2s + < ] 3

_j_ _2_
(47)
Mc ~ Mn
Here v 2s is the polymer volume fraction after equilibrium swelling and u2 r is the polymer
volume fraction after cross-linking but before swelling (relaxed volume fraction).
The non-Gaussian model of Kovac6 is obtained by a derivation similar to that shown for
the Peppas and Lucht7 model. It was developed to provide a better statistical treatment of
short chains and large deformations. The final form of the Kovac model is given by Equation
48
v, r ur2/3i-i
Mc = ^ [ln(l - v 2) + U2 + (4g)
A third non-Gaussian model was developed by Galli and Brumage.8 This model was
developed for an unmodified freely jointed chain making no approximations other than
carrying terms only up to the order of 1/N3. A simplified development of this model is given
here.8
The initial step in the development was to obtain the entropy of deformation of a network
that has been formed by random cross-linking of v chains whose end-to-end distances are
distributed according to the exact distribution model derived by Galli and Brumage.8 This
is accomplished using Flory’s2 derivation except that the exact distribution function for the
freely jointed chain is used rather than the Gaussian distribution. Equation 49 is obtained
from this analysis.
AS = kv/2 {lnu2“ 1 + 3 - 3/(2N) - 1/(2N2) - 23/(100N3)
+ u 2_2/3[ —3 + 3/N - 6/(5N2) - 24/(25N3)] 4- u - 4/3[-3 /(2 N ) + 39/(10N2) - 63/(25N3)]
+ u2
-6/3[ —11/(5N2) + 221/(25N3)] + 8/3[ —513/(100N3)]} (49)
The next steps are also similar to the Flory2 development that relates the free energy of
dilution, AGmix, as given by the Flory-Huggins equation, and the free energy of the elastic
network, AGel, to obtain the expression for the average molecular weight between cross-
links. Following this development results in the final model of Galli and Brumage.8
[■ 1 JL _ JL
N + 5N2 + 25N3 ]
1 i l _ _!!_
N + 5N2 _ 25N3
11 221
5N2 + 25 N3 ]
+ v2- 5/3 (50)
The term T is defined by
(51)
and N is number of links in a freely jointed chain all of which have the same length 1.
All of the modified models approach the Gaussian model of Flory and Rehner1 when the
number of bond vectors between cross-links becomes large (usually more than 100). The
effect of junction functionality greater than 4 on the value of Mc has been discussed by
Barr-Howell and Peppas.12
Computer simulations of the cross-linked structure of polymers as a function of the
physicochemical parameters appearing in the various models are instructive of the dependence
of Mc on these parameters. For example, Figure 4 shows the effect of the Xi factor and the
equilibrium polymer volume fraction, v 2 s, on the Mc. These values have been calculated
using Equation 45 and are taken from the unpublished work of Lucht and Peppas.15 It is
evident that for moderately or poorly swollen hydrogels, i.e., hydrogels with d 2,s > 0.70,
the value of Mc is not very much dependent on the thermodynamic nature of the swelling
agent.
Figure 5 shows the dependence of Mc on the equilibrium polymer volume fraction and
the size of the repeating unit. Finally, Figure 6 shows the effect of the degree of swelling,
Q, which is the reciprocal of v2<s, on the molecular weight between cross-links, Mc calculated
by a Gaussian (Equation 14) and a non-Gaussian (Equation 45) model.
III. RUBBER ELASTICITY THEORY
When a cross-linked network acting as an elastomer is stretched, it reaches an equilibrium

strain while the stress remains constant. The rubber elasticity theory can be used to explain
such equilibrium states thermodynamically. For basic references in the development of the
rubber elasticity theory see Flory,2 Meares,11 and Treloar.16
For the system of a sample of length L, where a linear force f is applied, the work done
by the system is
dW = —fdL + PdV (52)
where dL is a small extension and dV is the increase of volume.

Thus the first law of thermodynamics can be written as
dE = TdS - fdL + PdV (53)
The Helmholtz free energy is defined as
A = E - TS (54)
10000~~----~----~----~----r----;r----;r----,
)(. 0.8
......z
......
;;r:
Ill
a:
c
:5
:::
..J
0
2
.4 .5 .6 .7 .8 .9 1.0
TRUE VOLUME FRACTION IN SWOLLEN NETWORK, ~.S
FIGURE 4. The effect of the equilibrium polymer volume fraction, u 2 " and the
X1 factor on M, calculated with the modified Gaussian distribution incorPorated in
Equation 46 using a hypothetical polymer with u = 0.769 cm 3/g, molecular weight
of repeating unit M, = !50 and a molar volume of swelling agent V, = 80.56
cm 3/mol.
Differentiation and substitution of results from Equation 53 yields
dA = -PdV + fdL- SdT (55)
Finally, it can be seen that
f- ( -a A) (56)
aL T.v
The assumption of constant volume in isothermal deformations is reasonable for amorphous

polymers.
For the correlation of the applied force and the free energy, the free energy must be
calculated from the characteristics of the network to which the stress was applied. 17 For this
purpose, consider a (x,y ,z) rectangular coordinate system and a subunit having its one end
5000
(f)
:::..:::
z
_J
(f)
(f) 4000
0
a::
u
z
w ...
~ 3000
1-
w
ID
1-
I
(.!)
2000
w
~
a::
<l:
a
_J
1ooo
w
_J
0
~
0.0
0 0.4 0.5 0.6 0.7 0.8 0.9
EQUILIBRIU M VOLUME FRACTION
FIGURE 5. The effect of the equilibrium polymer volume fraction and the size of
the repeating unit of the polymer, M, on tvC calculated with the modified Gaussian
distribution incorporated in Equation 46 using a hypothetical polymer with u = 0.769
cm'lg and a swelling agent with x, = 0.4 and V, = 80.56 cm'lmol. Curve 1 is for
M, = 100, curve 2 for M, = 250, curve 3 for M, = 400, and curve 4 for M, =
550.
at the origin of the system and the other end at a distance of ri, with coordinates (xi, yi, zJ.
After a deformation, the subunit will attain a new position at which the end of it will be at
a distance rand its new coordinates will be (x,y ,z). The deformations on the three axis are
a,, aY, a, and they are defined below.
X y z
-a -a (57)
X·
I
y Yi z z,
Thus the network consists of ve active subunits of r; mean square length in the unstrained
state and f 2 mean square length in the deformed state and:
(58)
z
t!:J 2500
3:
I-
W 2000
CD
I-
I
<.!)
1500
w
3:
a:: 1000
<{
_.J
:::J
u
w 500
_.J
0
~
1.50 2.00 2.50 3.00 3.50

DEGREE OF SWELLING
FIGURE 6. The effect of the volume degree of swelling, Q, on M" as
calculated by a Gaussian (Equation 14, curve I) and a non-Gaussian (Equation
46, curve 2) model. The calculations are for a hypothetical polymer with u
= 0.769 cm'lg, M, = 150 and a swelling agent with x, = 0.4 and V, =
80.56 cm'lmol.
For an isothermal process ~lA = - W so that
(59)
From Equations 58 and 59, llAe 1 becomes
(60)
Now consider elongation in one direction
L
a- (61)
Lo
since
(62)
It can be seen that
ay = az — a (63)
Using this relationship in Equation 60 results in
AAe] (64)
From Equation 56
(65)
From Equations 64 and 65
f ( 66 )
Using the known relationship L0 = V/A0 in Equation 66 results in Flory’s final equation.2
_f_
T (67)
A0
Equation 67 was developed for isothermal deformations of a sample. It is valid for cross-
linked networks where the cross-links have been introduced randomly into an amorphous
polymer in the absence of any solvents. This Gaussian equation requires that the degree of
cross-linking not be excessive, deformation occur at constant volume, the contributions of
the effective subunits to the total retractive force be additive, and deformation be a reversible
process in a thermodynamic sense.
Earlier theories, like Flory’s1819 first development had considered that the parameter [rf/
If] was equal to one.
Equation 67 can also be written in the following way according to Treloar.16
2 M ,\ / i
T ( 68 )
m J v or
In this way it can be applied to the case of cross-linked swollen networks, where the density
p is known.
The parameter [rf/fg] is called the “ front factor” and enters the expression of the Helmholtz
free energy. The term rf is the squared end-to-end distance for the network subunits in the
undeformed isotopic state and If is the corresponding squared end-to-end distance for the
undeformed subunits in the absence of cross-links.
The above-mentioned equation does not apply to the case of hydrogels, where the cross-
links were introduced in the presence of solvent. For this case, another expression was
developed by Silliman20 and modified by Peppas and Merrill.21 The derivation of this equation
follows.
From the derivation of the rubber elasticity theory, the change of the Helmholtz free
energy of a network is given as:
v„kT
AAel = (a 2 + a 2 + a 2 - 3) (69)
2 (I)
where a x, a y, a z represent the three dimensional changes due to combination of swelling

and deformation. Thus
V,
(70)
vr
The network is formed at volume Vr and the polymer, of volume fraction of u 2 r, is sub-
sequently swollen by the diluent to a volume Vs. For the deformation due to elongation, the
constant volume assumption is used. Thus
vr
(71)
V.
The swelling ratio is defined as
V.
Q - (72)
and
a xa ya K = aQ (73)
For a simple elongation in the direction of the z-axis, the term a can be written as
_ Lfs (74)
For isotropic swelling

0
L
L™o,s _= I/V \\ '1/3
v sA _ Q 1/3
(75)
L„ KvJ
Then
«x = «Q' (76)
and
(77)
tty az a 1/2
Making the above substitutions into Equation 69 we obtain Equation 78.
kT ve Q2 3 / r 2\ / , 2 3
AAe, = (78)
( ! ) ( “' + ; - q ^>)
The elastic retractive force for the swollen network is
1 /3 A \
(79)
f “ W n
Taking the derivative of Equation 78 with respect to x and substituting the result into the
above equation yields in Equation 80.
RT veQ23 /r?
a (80)
L0, cry
Since
Vs
L, (81)
A~
and
f
T (82)
A0,s
the final equation becomes
T (83)
The equation is analogous to an equation derived by Verkhotina et al.22 for cross-linked

swollen polymers and which states that:
pRT
mc = hr* (84)
r: v 3/2
The difference between the two equations is not significant and either one can be used
to determine the molecular weight between cross-links.
In the application of the rubber elasticity theory to the study of swollen cross-linked
networks, it should be remembered that factors like finite extensibility of the chains, stress
crystallization, and the presence of microcrystalline regions violate the basic assumptions
of theoretical development and cause non-Gaussian behavior.
IV. E X PER IM EN TA L M ETHODS OF D ETERM IN ATION OF M c
The experimental methods of determination of the number average molecular weight

between adjacent cross-links of hydrogels, Mc, and other related parameters are rather
straightforward, although not often used in the analysis of the hydrogel structure. Here we
review simple procedures of the determination of important physicochemical parameters
which are needed to calculate Mc. Although, the experimental techniques for measuring
Mc are scattered in the literature, recently Mark23 has written a good review on this subject.
All the experimental studies are done on uniform, thin films of hydrogels. Often, these
studies require determination of physicochemical parameters of the gel immediately after
crosslinking and before swelling.
A. Swelling Studies
The following measurements are made for each hydrogel sample:
• Wa r, the sample weight in air after cross-linking,

• Wn r, the sample weight in a nonsolvent after cross-linking,

• Wa s, the sample weight in air after swelling,
• Wn s, the sample weight in a nonsolvent after swelling,
• Wa d, the sample weight in air after drying.
These measurements are used to determine the volume of the hydrogel sample after cross-
linking (but before swelling), Vg r, and after equilibrium swelling, Vg s. The equations used
to calculate the hydrogel volume before and after swelling are based on Archimedes’ buoy-
ancy principle.
V, (85)
V ( 86)
Here pn is the density of the nonsolvent at the temperature of experimentation. The volume
of the dry polymer, Vp, is calculated using Equation 87.
W
Vp = —— (87)
Pp
Here pp is the density of the polymer.

The polymer volume fractions of the hydrogel sample in both the relaxed and the swollen
state, u 2 r and u2 s, respectively, are calculated using the following relationships:
v 2 ,r ( 88 )
v 2 ,s (89)
In the derivation of these relationships for the volume fractions, addivity of volumes is
assumed.21,24 For hydrogels which are cross-linked in the solid state, v 2 r is one.
Before the above data can be used in either the Gaussian on non-Gaussian models to
determine the molecular weight between cross-links, information about the polymer/solvent
thermodynamic interaction parameter, Xi> must be obtained. This parameter is a function
of both polymer volume fraction and the temperature of swelling.
The thermodynamic interaction parameter is expressed in Flory’s theory2 by the polymer-
solvent interaction parameter, Xi, which is defined as:
BV,
(90)
RT
where V, is the molar volume of the solvent, T is the temperature, and B represents the
interaction energy density which is characteristic of the solvent-soluble pair. The definition
of B is given by Equation 91.
zAw,
B = (91)
Here Aw12 is the change in energy for the formation of an unlike contact pair, V 's is the
molecular volume of a segment, and z is the total number of segments. The factor AwI2 is
independent of T and therefore no entropy contribution is contained in Aw12. Thus Xi is
inversely proportional to T.
The Xi factor may also be related to two more thermodynamic parameter k , and v|/, under
conditions of dilute polymer concentration (1 to 5%). In the thermodynamic theory of dilute
polymer solutions, the changes in enthalpy and entropy are expressed as
AH, = RT k ,u 2 (92)
AS, = Rv|i,u2 (93)
where v 2 is the volume fraction of the solute (polymer), k , is the enthalpic parameter, and
vpi is the entropic parameter. Within the limits of Flory’s2 theories, Xi can be related to k ,
and iJ/, via Equation 94.
X. = k, - ih + 1/2 (94)
The 0-temperature is defined as:
k ,T
(95)
At the temperature T = 0 , k , = i|/j, and Xi = 0.5. At the 0-temperature, the segment-

solvent interactions are zero. If Xi <0.5, the solvent is considered to be a thermodynamically
“ good” solvent and if Xi > 0.5, the solvent is considered to be a “ poor” solvent.
The thermodynamic interaction parameters for specific polymer/solvent pairs have been
determined by Rehner,25 Sheehan and Bisio,26 and Orwoll.27
If the thermodynamic interaction parameter is not known, it can be determined experi-
mentally using membrane and vapor osmometry, freezing point depression, light scattering,
dilute solution viscometry, and inverse gas chromatography.22
Osmotic pressure, it , can be used to determine Xi using Equation 96.
X, = —[t t V,/RT + ln(l - v 2) + u2(l - l/x ,)]/^2 (96)
Solving the above equation for tt and expanding in a power series yields
TT = (RT/V,)^(1/ x ,)v 2 + (1/2 - x M + ^ v \ + ••••] (97)
The volume change of mixing is negligible and v 2 can be written as
u2 = c2/p2 (98)
The dependence of Xi on concentration can be represented by a power series. These as-

sumptions lead to Equation 99.
tt /RTc 2 = 1/M2 + [(p2)2V ,]-1 - x ,) c 2 + [(p2)3V,]-' ^ - x ,)c? + ....... (99)
The interaction parameter can be evaluated from the second virial coefficient or directly
from Equation 96.
Table 1
APPLICATION OF A GAUSSIAN MODEL (EQUATION 27)
TO THE DETERMINATION OF THE STRUCTURAL
CHARACTERISTICS OF LOOSELY CROSS-LINKED
POLY(VINYL ALCOHOL) HYDROGELS8
Polymer Cross-linking
Cross- vol. density
linking fraction Swelling Mc Px x 104
ratiob v_ s ratio Q (g/mol) (mol/cm3)
0.030 0.112 ± 0.004 8.94 484 ± 426 26.21

0.020 0.093 ± 0.005 10.72 1086 ± 34 12.68
0.016 0.080 ± 0.006 12.44 1984 ± 96 6.40
0.012 0.055 18.12 4041 3.14
0.011 0.055 ± 0.003 18.07 4157 ± 49 3.05
0.010 0.052 ± 0.003 19.16 4891 ± 56 2.59
0.009 0.051 ± 0.004 19.41 4846 ± 77 2.62
0.008 0.045 22.04 6162 2.06
0.007 0.044 ± 0.002 22.87 6024 ± 44 2.11
0.006 0.040 ± 0.002 24.83 7426 ± 52 1.71
0.005 0.033 ± 0.001 30.34 9552 ± 48 1.33
a PVA hydrogels produced by reaction of 10% aqueous PVA solution with glutar-
aldehyde (see Peppas and Benner31) at 80°C for 10 min, and swollen in water at
37°C. Standard deviation reported where at least four samples were prepared.
b X = mol cross-linking agent/mol PVA repeat.
The pressure, P ,, of solvent vapor in equilibrium with the solution may be used to determine
Xi with the application of Equation 100, as discussed by Bonner.28
X, = {ln[P,/(l - v2)P°] - v 2(\ - l/Xl)}A>i (100)
Freezing point depression of the polymer29 can also be used to determine the interaction
parameter using Equation 101
( 1/Tm2 - l / T ^ / u , = [(R/V,)(AHf 2/V2)](l - xdh) (101)
Finally, inverse gas chromatography28 30 can be used to determine Xi- Here the polymer
is loaded in the column and becomes the stationary phase. The solvent is injected into the
column along with an inert carrier gas. The solvent has a tendency to be absorbed by the
polymer; this tendency is a function of Xi • In terms of the retention volume, Vg, the interaction
parameter can be determined using Equation 102.
Xl = ln(RTV2/p°V,Vg) - (1 - 1/Xl) - P?(Bn - V,)/RT (102)
The factor B ,, is the gaseous second virial coefficient for the diluent vapor.
Once the polymer volume fractions of the hydrogel sample in both the relaxed and the
swollen state are known along with the thermodynamic interaction parameter, it is possible
to determine the molecular weight between cross-links using the appropriate Gaussian or
non-Gaussian equation.
Table 1 shows typical results of the calculation of the number average molecular weight
between cross-links, Mc, for poly (vinyl alcohol) hydrogels cross-linked in solution by glu-
teraldehyde, as reported by Peppas and Benner.31 In this table, we present the type of
information that is typically necessary for application of hydrogels in medicine or phar-
0 0.01 0.02 0.03 0.04

Molar ratio, X
FIGURE 7. Dependence of number average molecular weight, M" of poly(vinyl alcohol) hydrogels on
the molar cross-linking ratio, X. (From Peppas, N. A. and Benner, R. E., Jr., Biomaterials, I, 158, 1980.
With permission).
maceutics. Thus, X is the nominal cross-linking ratio, i.e., the molar ratio of the cross-
linking agent to the moles of repeating units of the polymer, and u 2 ,, is the polymer volume
fraction at equilibrium swelling with deionized water at 37°C. The equilibrium volume
swelling ratio, Q, is an important parameter defined according to Equations 41 or 72. Finally,
in addition to Mc, one usually reports the degree of cross-linking or cross-linking density,
p,, which is defined according to Equation 103.
p, =-M (103)
u c
The calculations of M, have been done using Equation 45, since the PV A hydrogels described
here are rather loosely cross-linked and the cross-links have been introduced in solution.
Figure 7 shows a typical dependence of Mc on the nominal cross-linking ratio, X, for
these gels and Figure 8 depicts the corresponding plot of the cross-linking density as a
function of X. In general, as X increases the cross-linking density increases linearly and the
Me decreases, i.e., the gel becomes more cross-linked.
Figure 9 shows the relation between the equilibrium polymer volume fraction, u 2 .,, and
the cross-linking ratio, X, and Mc. Finally, Figure 10 depicts the effect of X on the swelling
ratio, Q. In general, as the gel becomes more cross-linked, the equilibrium polymer volume
fraction increases and the swelling ratio decreases.
Most amorphous hydrogels can be analyzed in a similar way because they are usually
highly swollen. The previous theories and methods of structural evaluation cannot be applied
to semicrystalline hydrogels since the crystallites usually act as additional cross-links but of
finite volume. Some approximations can be made if the degree of crystallinity is not very
high, e.g., x(t) < 0.10 and the crystallites are of small size. This point was discussed by
Patterson et al. 32 who determined the average molecular weight between all types of junctions
(cross-links, entanglements and crystallites) for poly(vinyl chloride) networks plasticized by
(swollen in) diethyl hexyl phthalate.
Volume/: Fundamentals 49
..,
E 2
u
'.!!en
0
E
•Q
)(
Q..
0 001 002 003 004

Molar ratio, X
FIGURE 8. Dependence of the cross-linking density, p, (mol!cm 3) on the molar cross-linking ratio, X.
(From Peppas, N. A. and Benner, R. E., Jr., Biomaterials, I, 158, 1980. With permission).
0 2000 4000 6000 8000
010
008
>
·-
N
006
I
0.04
0
{
001 002 003
J
0.04
Molar ratio, X
FIGURE 9. Dependence of the equilibrium polymer volume fraction, u 2 .., on the molar ratio X and the
~1_, (From Peppas, N. A. and Benner, R. E., Jr., Biomaterials, I, 158, 1980. With permission).
30
() 20
0
0...
0
c
Cl)
3: 10
(/)
0 0.01 0.02 0.03 0.04

Molar ratio, X
FIGURE 10. Dependence of the equilibrium volume swelling ratio, Q, on the molar cross-linking ratio,
X. (From Peppas, N. A. and Benner, R. E., Jr., Biomaterials, 1, 158, 1980. With permission).
Table 2
APPLICATION OF A NON-GAUSSIAN MODEL (EQUATION 47) TO THE
DETERMINATION OF THE STRUCTURAL CHARACTERISTICS OF HIGHLY
CROSS-LINKED PHEMA HYDROGELSa
Equil.
Cross-linking Relaxed volume Swollen volume degree Cross-linking
ratio fraction fraction swelling Me density
X X 103 (v,) (v,.J (Q) (g/mol) ( x 103 mol/cm3 )
5.0 0.6517 ± 0.0087 0.5958 ± 0.0021 1.678 1700 ± 80 0.759 ± 0.03

5.0 0.6287 ± 0.0080 0.5631 ± 0.0104 1.776 3700 ± 1200 0.367 ± 0.09
10.0 0.6583 ± 0.0161 0.5869 ± 0.0023 1.704 2050 ± 90 0.625 ± 0.02
25.0 0.6639 ± 0.0203 0.6207 ± 0.0153 1.611 1100 ± 240 1.1118 ± 0.19
50.0 0.6972 ± 0.0376 0.6431 ± 0.0059 1.555 800 ± 70 1.156 ± 0.09
12.8 0.6652 ± 0.0120 0.5810 ± 0.0017 1.721 2350 ± 130 0.548 ± 0.03
• PHEMA hydrogels produced by reaction of2-hydroxyethyl methacrylate (HEMA) with ethylene-

glycol dimethacrylate (EGDMA) at 60°C for 12 hr, and swollen in water at 37°C. All values
were calculated for M. = 75,000 (from Peppas et al.").
For highly cross-linked, amorphous hydrogels it is necessary to apply the non-Gaussian

models discussed earlier. This is the case with poly(2-hydroxyethyl methacrylate) (PHEMA)
hydrogels, as discussed recently by Peppas et al. 13 Typical values of the structural charac-
teristics of some PHEMA hydrogels are presented in Table 2. Figure 11 shows the effect
of the nominal molar cross-linking ratio, X, on Mc for these PHEMA hydrogels. Figure 12
depicts the difference between the Gaussian and the non-Gaussian prediction of Mc for the
same hydrogels.
5000
(J)
::X::
z........ 4500
_.J
(J)
(J)
4000
E>
0::
u 3500
z
w
w 3000
~
I-
w
CO 2500
I-
I
(.!) 2000
........
w
~
1500
0::
a:
_.J
:::> 1000
u
w
_.J
E> 500
:E
0
0.000 .010 .020 .030 .040 .050 .060
MOLAR CROSSLINKING RATIO
FIGURE 11. Influence of the molar cross-linking ratio, X, on M, for highly cross-
linked PHEMA hydrogels. (From Peppas, N. A., Moynihan, H. J., and Lucht, L.
M., J. Biomed. Mater. Res., 19, 397, 1985. With permission).
An additional problem encountered when analyzing the structure of some hydrogels is

related to the manner by which they have been prepared, for example, PHEMA hydrogels
are prepared by copolymerization/cross-linking of HEMA with EGDMA, a chemical agent
with two double bonds which serve as the cross-linking bridges between PHEMA chains.
In these situations, the value of Mn required in all the models for the calculation of Me is
not known a priori. One method to bypass this problem is to conduct polymerization
experiments of the main monomer (here HEMA) in the absence of cross-linking agent, and
to determine M" of the uncross-linked polymer under the same experimental conditions. It
is, however, interesting to note that our recent studies have shown 13 that, at least for PHEMA
hydrogels, the calculated value of Me is independent of M" if a value of M" of 75,000 or
higher is used.
B. Rubber Elasticity Studies

Mechanical tests are typically carried out using a tensile tester. The hydrogel samples are
cut to a dumbbell shape using an appropriate die. The dimensions of the samples and the
set-up of a typical tensile experiment are discussed elsewhere. 21 One of the advantages of
the dumbbell-shaped sample is that the ultimate fracture occurs in the center of the sample
and is not affected by the stress concentration at the jaws. In order to prevent slippage of
the sample in the jaws as well as breakage at the jaws, serrated plastic grip faces may be
5000
(f)
~
z 4500
.........
_j
(f)
(f)
4000 \
0
a::: \
u 3500 \
z
w ~
w 3000
3:
1-
w
en 2500
1-
\
:r::
(!) 2000 \
.........
w
\
3: \
"' "
1500
a:::
a:
-- - -- --
_j
1000 ""-.
::l .............. ......._
u
w
_j
0 500
:L
0
0.000 .010 .020 .030 .040 .050 .060
M~LRR CR~SSLINKING RAT!~
FIGURE 12. Comparison of the value of M, of PHEMA hydrogels as calculated

by a non-Gaussian (--, Equation 47) and the corresponding Gaussian model
( ---, Equation 27) (From Peppas, N. A., Moynihan, H. J., and Lucht, L. M., J.
Biomed. Mater. Res., 19, 397, 1985. With permission).
glued to the jaws. For hydrogels, the entire set-up must be immersed in a waterbath equipped
with a thermoregulator.
The tensile tester is initially calibrated. The swollen samples are cut, the average thickness
determined, and they are transferred to the tester. The initial gauge length is recorded, and
then the run performed until the sample ruptures. If the ultimate failure does not occur in
the middle of the sample, the run should be discarded.
The results from the tensile test data can be used to calculate several important parameters.
First, the stress T can be calculated using Equation 104.
F
(104)
where the load, F, is the weight used in the tester, and A0 is the initial area of the sample.
Second, the strain at break, E, is calculated using Equation 105.
(105)
30
2.0
1/)
a.
...
1/)
1/)
Gl
'-
en
1.0
0 0.5 1.0
Extension a - 11a 2
FIGURE 13.:_ Tensile stress, T vs. the extension factor (a - lla') for PVA hydrogels with Mc = 2100
(line l) and Mc = 7400 (line 2) (From Peppas, N. A. and Benner, R. E., Jr., Biomaterials, l, 158, 1980.
With permission).
where Lllb is the change in length at the breakpoint and 10 is the initial length. Finally, the
elongation a is calculated using Equation 106.
1 + E (106)
10
Once data have been obtained for the same sample at several different loads, the results
I
are plotted as T vs. (a - --;). The points are fitted on a straight line and the slope can then
a-
be used in any of the rubber elasticity theory models to calculate Mv More extensive analysis
of these tests can be found elsewhere 21 and especially in the excellent work of Mark and
his collaborators. 33 - 37
A typical set of mechanical testing data which can be used for the determination of Me
is shown in Figure 13. These data are for chemically cross-linked PVA hydrogels. 31 It must
be noted that often the values of Me obtained from mechanical tests are not the same
with those obtained from swelling studies, probably because of the disentanglement process
occurring during mechanical testing, as discussed by Peppas and Merrill21 and Peppas and
Benner.31
V. D ETER M IN A TIO N OF THE MESH SIZE
A structural parameter that is often needed when applying hydrogels in medicine and
pharmacy and especially when determining their diffusive characteristics is the mesh size.
The mesh size broadly defines the space between macromolecular chains in a cross-linked
network (see Figure 1) and it is usually characterized by the correlation length, or distance,
£, between two adjacent cross-links. An approximate method for its calculation is straight-
forward and has been discussed by Korsmeyer and Peppas38 and Peppas et al.13
The end-to-end distance of the freely rotating polymer chains between the network cross-
links (r2)12, is calculated using Equation 107
/1 — cos0\ '2
(f? )
1/2
(107)
\1 + cosO/
The number of links, N, between two cross-links is determined from Equation 97. For a
backbone chain consisting of carbon atoms cos 0 = — 1/3, since 0 = 109.5°, and 1 =
1.54 A; thus the expression for (r2)1/2reduces to
/ M \ 1/2
(r?)l/2 = 21( = ^ ) (108)
The values of the end-to-end distance of polymer chains in the unperturbed state, is
calculated through the Flory characteristic ratio, or rigidity factor, Cn.
Cn (109)
Nl2
Finally, the size of the macromolecule in the solvated (swollen) state, £, is determined
from Equation 110
€ = <*(f2)1/2 = © i/2 (HO)
Typical values of the correlation length of the mesh size, £;, for PHEMA hydrogels are
reported in Figure 14.
VI. CONCLUSIONS
In conclusion, the analysis of the cross-linked structure of hydrogels can be done by a

variety of experimental techniques, prominent among which are swelling and mechanical
tests. Care must be taken to analyze the structure by applying the appropriate Gaussian or
non-Gaussian chain model depending on whether the network is loosely or highly cross-
linked. Corrections for cross-linked structures prepared in solution or for junction function-
ality higher than four are available.
50
45
40
35
o(I
30
w
N
1--4
(f)
25
/
:::r: /~
(f) 20 ~
w /,
'E
tf
15
10
0
500 1500 2500 3500 4500
M~LECULAR WEIGHT BETWEEN CR~SSLINKS
FIGURE 14. Correlation length of the mesh size, ~. as a function of the number
average molecular weight between cross-links, M" for the PHEMA hydro gels reported
in Table 2. Dashed curves denote 95% confidence limits (From Peppas, N. A.,
Moynihan, H. J., and Lucht, L. M., J. Biomed. Mater. Res., 19, 397, 1985. With
permission).
ACKNOWLEDGMENTS
Most of the models discussed here have been developed during our work sponsored by
the Department of Energy, through grants Nos. DE-FG22-80PC30222 and DE-FG22-
83PC60792.
REFERENCES
I. Flory, P. J. and Rehner, R., Jr., Statistical mechanics of crosslinked polymer networks. I. Rubberlike
elasticity, J. Chem. Phys., 11,521, 1943.
2. Flory, P. J., Principles of Polymer Chemistry, Comell University Press, lthaca, N.Y., 1953, 576.
3. Peppas, N. A. and Merrill, E. W., PVA hydrogels: reinforcement of radiation-crosslinked networks by
crystallization, J. Polym. Sci. Polym. Chem., 14, 441, 1976.
4. Fixman, M. and Alken, R., Polymer conformational statistics. I. Probability, J. Chem. Phys., 58, 1553,
1973.
5. Fixman, M. and Kovac, J., Polymer conformational statistics. III. Modified Gaussian models of stiff
chains, J. Chem. Phys., 58, 1564, 1973.
6 . Kovac, J., Modified Gaussian model for rubber elasticity, Macromolecules, 11, 362, 1978.
7. Peppas, N. A. and Lucht, L. M., Macromolecular structure of coals. I. The organic phase of bituminous
coals as a macromolecular network, Chem. Eng. Commun., 30, 291, 1984.
8 . Galli, A. and Brumage, W. H., The freely jointed chain in expanded form, J. Chem. Phys., 79, 2411,
1983.
9. Reed, T. M. and Gubbins, K. E., Applied Statistical Mechanics, McGraw-Hill, New York, 1973.
10. Flory, P. J., Statistical Mechanics of Chain Molecules, Interscience, New York, 1969.
11. Meares, P., Polymers: Structures and Bulk Properties, Van Nostrand, London, 1965.
12. Barr-Howell, B. D. and Peppas, N. A., Importance of junction functionality in highly crosslinked
polymers, Polym. Bull., 13, 91, 1985.
13. Peppas, N. A., Moynihan, H. J., and Lucht, L. M., The structure of highly crosslinked PHEMA
hydrogels, J. Biomed. Mater. Res., 19, 397, 1985.
14. Peppas, N. A., Lucht, L. M., Moynihan, H. J., Gobran, D., and Sinclair, G., Equilibrium and Dynamic
Swelling of Highly Crosslinked Polymeric Networks, Proc. IUPAC, 1983, 29.
15. Lucht, L. M. and Peppas, N. A., Macromolecular structure of coals. VIII. The effect of physicochemical
parameters on Mc, in press, 1986.
16. Treloar, L. R. G., The Physics of Rubber Elasticity, Clarendon Press, Oxford, 1958.
17. Ciferri, A., Present status of rubber elasticity theory, J. Polym. Sci., 54, 149, 1961.
18. Flory, P. J., Rabjohn, N. and Shaffer, M. C., Dependence of elastic properties of vulcanized rubber on
the degree of crosslinking, J. Polym. Sci., 4, 225, 1949.
19. Flory, P. J., Rabjohn, N. and Shaffer, M. C., Dependence of tensile strength of vulcanized rubber on
degree of crosslinking, J. Polym. Sci., 4, 435, 1949.
20. Silliman, J. E., Network Hydrogel Polymers-Application in Hemodialysis, Sc.D. thesis, Massachusetts
Institute of Technology, Cambridge, 1972.
21. Peppas, N. A. and Merrill, E. W., Crosslinked hydrogels considered as swollen elastic networks J. Appl.
Polym. Sci., 21, 1763, 1977.
22. Verkhotina, L. N., Gembitskii, L. S., and Gubenkova, E. N., Thermoelastic properties and gel formation
of PVA in different media, Vysokomol. Soedin. Ser. A, 15, 1350, 1973.
23. Mark, J. E., Experimental determination of crosslink densities, Rubber Chem. Technol., 55, 762, 1982.
24. Janacek, J., and Hasa, J., Structure and properties of hydrophilic polymers and the gels. VI. Equilibrium
deformation behavior of PHEMA and PHEEMA networks prepared in the presence of a diluent and swollen
with water, Collect. Czech. Chem. Commun., 31, 2986, 1966.
25. Rehner, R., Jr., Literature references to values of the polymer-solvent interaction parameter, J. Polym.
Sci., 46, 550, 1960.
26. Sheehan, C. J., and Bisio, A. L., Polymer/solvent interaction parameter, Rubber Chem. Technol., 39,
149, 1966.
27. Orwoll, R. A., The polymer-solvent interaction parameter x, Rubber Chem. Technol., 50, 451, 1977.
28. Bonner, D. C., Vapor-liquid equilibria in concentrated polymer solutions, J. Macromol. Sci. Rev. Ma-
cromol. Chem., 13, 263, 1975.
29. Flory, P. J., Thermodynamics of crystallization in high polymers. IV. A theory of crystalline states and
fusion in polymers, copolymers, and their mixtures with diluents, J. Chem. Phys., 17, 223, 1949.
30. Smidsrod, O. and Guillet, J. E., Study of polymer-solute interactions by gas chromatography, Macro-
molecules, 2, 272, 1969.
31. Peppas, N. A. and Benner, R. E., Jr., Proposed method of intracordal injection and gelation of polymer
solution in vocal cords: polymer considerations, Biomaterials, 1, 158, 1980.
32. Patterson, K. G., Padgett, S. J., and Peppas, N. A., Microcrystallinity and three-dimensional network
structure in plasticized PVC, Colloid Polym. Sci., 260, 851, 1982.
33. Mark, J. E., The use of model polymer networks to elucidate molecular aspects of rubberlike elasticity,
Adv. Polym. Sci., 44, 1, 1982.
34. Mark, J. E., Recent studies of rubberlike elasticity, Macromol. Chem. Suppl., 2, 87, 1979.
35. Sharaf, M. A. and Mark, J. E., The effects of crosslinking and strain of the glass transition temperature
of a polymer network, Rubber Chem. Technol., 53, 982, 1980.
36. Mark, J. E., Rubber elasticity, J. Chem. Educ., 58, 898, 1981.
37. Queslel, J. P. and Mark, J. E., Characterization of elastomeric network structures using the effects of
swelling on stress-strain isotherms and the extents of swelling at thermodynamic equilibrium, Polym. Bull.,
10, 119, 1983.
38. Korsmeyer, R. W. and Peppas, N. A., Effect of the morphology of hydrophilic polymeric matrices on
the diffusion and release of water soluble drugs, J. Membr. Sci., 9, 221, 1981.
Chapter 3
SOLUTE DIFFUSION IN HYDROPHILIC NETWORK STRUCTURES
Nikolaos A. Peppas and Steven R. Lustig
TABLE OF CONTENTS
I. Generalized MassTransfer Theory for Hydrogels....................................................... 58

A. Continuum Approach......................................................................................... 58
1. Analysis of Mass Transport byMulticomponent Diffusion
Equations................................................................................................ 58
a. Generalized Stefan-Max well Equations................................ 58
b. Generalized Fickian Equations..............................................60
2. Irreversible Thermodynamics.............................................................. 60
B. Simplifications for Ordinary Diffusion in Hydrogels................................... 61
1. Fickian D iffusion...................................................................................61
2. Nernst-Planck A nalysis.........................................................................61
3. Hydrogel Permeability: Kedem and KatchalskyDevelopment.........62
II. Diffusion in Nonporous Hydrogels..............................................................................63

A. Determination of Solute Diffusion Coefficients..............................................63
1. Statistical Mechanics...............................................................................64
2. Free-Volume T heories.......................................................................... 65
3. Scaling Theories.................................................................................... 72
B. Effect of Polymer Structure on Solute D iffusion.......................................... 76
III. Diffusion inPorous Hydrogels.......................................................................................78
IV. Conclusions.......................................................................................................................80
Acknowledgments........................................................................................................................ 80
References 80
I. GENERALIZED MASS TRANSFER THEORY FOR HYDROGELS
A. Continuum Approach
7. Analysis o f Mass Transport by Multicomponent Diffusion Equations
Transport properties of gases and liquids in hydrogels are described by several theories
which owe their development to classical and statistical mechanics. The accurate evaluation
and practical application of transport coefficients have been possible only for very few
systems. These systems are usually simple and contain monatomic molecules with binary
collisions. The kinetic theory of dilute gases has been used to calculate the transport coef-
ficients. An excellent summary of this analysis can be found in Chapters 7 and 8 of Hirsch-
felder et al.1The entire development is based upon the Boltzman integrodifferential equation
for the distribution function f(r,v,t). The equation has been solved for simple systems using
various approximations, such as the Chapman and Enskog series analysis.1
Analysis of complex systems including solutions, multicomponent systems, and mem-
branes (including hydrogels) is not possible with simple kinetic theories. Therefore, further
analysis was developed for the study of transport phenomena in dense systems. A good
summary can be found in Chapters 9 and 10 of Hirschfelder et al.1
a . Generalized Stefan-Maxwell Equations

As a result of the analysis described above, a set of mathematical expressions may be
used to describe the relation between fluxes and forces in multicomponent diffusion in
hydrogels. The most general of these forms is referred to as the Stefan-Maxwell equations.
Numerous authors have derived different forms of the equations, although all the derivations
are fundamentally identical. The derivations are based on the solution of the Boltzman
equation with appropriate perturbations and the conservation equations. A preferred analysis
of the Stefan-Maxwell equations can be found in the book by Chapman and Cowling.2
The form of the Stefan-Maxwell equations used by Lightfoot,3 Cussler,4 and Krishna and
Standart5 is presented here. Consider a system of n-independent components under isothermal
conditions. Diffusion of species i through the multicomponent system where the hydrogel
is considered as one of the components may be described by the Stefan-Maxwell equations,
presented as Equations 1 to 4, and henceforth designated as the S-M equations,
( 1)
with
( 2)
and
(3 )
and
S di = 0 (4)
In these equations, subscript i designates the important diffusing species and subscript j
designates anyone of all other diffusing components. Subscript k may be a reference com-
ponent, although one may choose to use any arbitrary species without loss of generality. A
common practice is to designate the component with the highest concentration as k, (e.g.,
solvent for diffusion in solutions, or the hydrogel for diffusion through hydrogels). Also
Vj and vk are the velocities of components j and k respectively, while terms are the
multicomponent diffusivities of the pairs ij. Finally, dj designates the generalized driving
force for diffusion. For a system that has chemical potential, pressure, electrostatic potential,
and gravity gradients acting on it, may be written according to Equation 5.
n n
cRT di = C; VT P |xs + (c,V, - «i)VP + (CjV; - w, ^ ckvk)F V(J> + co, X gk (5)

k= 1 k= 1
Here (jl , P, 4>, and g refer to the chemical potential, pressure, electrostatic potential, and
gravity forces, respectively; ch vi? o^, and refer to the concentration, specific volume,
mass fraction, and ionic charge of each solute (diffusing species), respectively; F is the
Faraday constant, c is the total concentration of the system, R is the gas constant, and T is
the absolute temperature.
Equation 1 summarizes (n-1) rate expressions known as the S-M equations and is a direct
result of the solution of the Boltzman equation with fluxes expressed in terms of velocities.
However, contrary to other conventional diffusion expressions (such as the simplified Fickian
diffusion expression), in Equation 1 the forces are presented explicitly in terms of the fluxes.
Equation 2 is a direct result of the symmetry of the diffusion coefficient matrix. Equation
4 is a result of the use of the Gibbs-Duhem theorem. Equation 5 is an explicit relation of
the driving forces to all possible gradients that can be observed in the diffusion process.
Equation 1 must be recognized as the S-M relations for isothermal processes, because
temperature effects have not been included in the expressions. Nonisothermal multicom-
ponent diffusion may be described by Equation 6 where the additional term describes tem-
perature gradients and includes the thermal diffusion coefficients, DT.
j^k jî
Several texts present simplified forms of the S-M equations for the case of ideal ther-
modynamic systems where the only gradient is a concentration gradient. For example, Bird,
Stewart, and Lightfoot6 simplified the S-M equations in the form of Equation 7.
Vk) (7)
J=l u \i
jî
Apart from the use of the mole fraction gradients instead of chemical potential gradients,
the major difference between Equation 7 and Equation 1 is that the former is expressed in
terms of the so-called binary multicomponent diffusion coefficients, Dy. A very important
characteristic of is that they are much more concentration dependent than the multicom-
ponent diffusivities, of the more rigorous S-M relations.3
Because of the complex nature of the S-M equations, many investigators have used
simplifications of the S-M relations to describe their systems. The three simplifications that
have appeared in the literature most often are generalized Fickian diffusion, binary Fickian
diffusion, and Nemst-Plank diffusion.
b. Generalized Fickian Equations

Generalized Fickian diffusion may be described by Equation 8.
c2 n
I = - 2 MjMj D m V x ; (8)
P j =I
Here, p is the density, M-xand Mj are the molecular weights of component i and j respectively,
and the other terms are as defined before. Equation 8, the generalized Fickian expressions
(henceforth designated as GF relations), closely resemble the S-M equations except for the
fact that in the GF equations, the fluxes are explicit functions of the forces. Actually,
Equation 8 includes only chemical potential gradients. Similar expressions can be written
for the flux due to pressure gradients (pressure diffusion), the flux due to gravity gradients
(forced diffusion), and the flux due to thermal gradients (thermal diffusion). The total fluxes
due to these forces are the summation of all the individual fluxes.6
2. Irreversible Thermodynamics
In the last 40 years, the theory of irreversible thermodynamics has been vigorously
developed and tested. The theory has been used to treat irreversible processes in a detailed
macroscopic fashion, which was not possible with its classical forebear. The theory of
irreversible thermodynamics might well have appeared in its macroscopic form long ago if
the necessary concepts of entropy production, linear laws, and symmetry of coefficients had
been generalized. However, it was not until the statistical mechanical investigations of
Onsager7 that the present macroscopic form of the equations appeared.9
The structure of thermodynamics that has arisen through contributions of Clausius, Helm-
holtz, Gibbs, and others contain severe limitations when applied to actual physical phenom-
ena. These limitations are related to questions of equilibrium and continuity. Classical
thermodynamics recognized the existence of nonequilibrium states but it can neither provide
definitive information when the system or medium is not in equilibrium, nor be instrumented
in subdividing the system into subsystems in equilibrium. For these reasons, classical ther-
modynamics make quantitative statements only for equilibrium states.
The theory of irreversible processes (henceforth designated as TIP) has been a partial
answer to the questions raised by classical thermodynamics. The development of TIP was
first discussed by Onsager. The work of Onsager was expanded by Meixner,10,11 Eckart,12
Prigogine,13 and deGroot14 who developed modified expressions of TIP to thermal diffusion,
thermoelectricity, viscous flow, and heat transfer. To facilitate the thermodynamic descrip-
tion of various irreversible phenomena the assumptions and general formalism of TIP will
be discussed now.
The dissipation function, a ', due to the increase of entropy in an irreversible process can
be written in the form of Equation 9.
Ter' = S J.X, (9)

i
Here, the fluxes, Ji? are flows of matter, heat or electricity, and the generalized forces, Xi9
are gradients of temperature, chemical potential, electrical potential, etc.
Another relationship between Jj and X x that has been known for many years is that, in
simple cases, the forces and fluxes are linearly related to each other. Linear relationships
have been verified experimentally; they are still consistent with the theoretical development
of Onsager. For cases where there is more than one irreversible process occurring, each flux
Ji has not only been found to be linearly related to its conjugate force Xi9 but is also linearly
related to all forces in the summation of the expression for Ter'. The general form of the
flux J, is
J; = 2 Lu X, ( 10)
i
where are the general linear coefficients or the phenomenological coefficients.

Meixner11pointed out that the fluxes or forces can be chosen arbitrarily to a certain extent.
He established a set of requirements for determining the set of conjugate variables after one
set of variables is chosen. The first rule is that the product of any flux and its conjugate
force have the dimensions of entropy production. The second rule states that any transfor-
mation of fluxes and forces for a given system does not alter the sum of the products. The
choice of fluxes and forces in irreversible thermodynamics is similar to the choice of co-
ordinate systems for the solution of problems in classical mechanics. The thermodynamic
solution to an irreversible process produces useful correlations between observable phenom-
ena only when the fluxes and forces correspond to those that lend themselves to experimental
determination.15
Onsager was the first to propose that for a system of n independent components, if J4and
Xj are chosen from the expression of Ter' (Equation 9) and are independent, the phenom-
enological coefficients L {j of the linear laws satisfy the symmetric relation
L, = Ljj (11)
for all components i and j. Equation 11 is usually referred to as “ the Onsager reciprocal
relations.” 78
Since the contributions of Onsager, TIP has been used by many investigators to describe
processes in many different areas. For example, the application of TIP to ion-selective
membranes1618 has helped to understand and quantify the phenomena of ion transport through
a membrane or hydrogel. Another closely related application is electrolytic systems; TIP
has been used to describe binary19 and ternary systems.20 Systems that have reactions taking
place in a membrane have also been described using TIP.21
B. Simplifications for Ordinary Diffusion in Hydrogels

1. Fickian Diffusion
A simplification of the GF relations for binary systems leads to the well known Fickian
analysis described by equation 12.
ji = cDy Vx, ( 12)
This equation is applicable to binary systems exhibiting thermodynamic ideality. The mul-
ticomponent diffusion coefficient, Dxj, used in the GF equations now becomes the binary
concentration dependent diffusion coefficient, Dy.
2. Nernst-Planck Analysis
The second important simplification of the S-M equations was developed by Nemst and
Planck3 to analyze systems where the diffusion of one component is the important transport
process, and all the other components can be collectively called the mixture. Therefore, the
Nernst-Planck equations (henceforth designated N-P) apply to pseudobinary systems; they
are conveniently described by Equation 13.
E Nj
Ni = - c D imdi + (13)
J-1 D„
Here, N designates molar fluxes with respect to a stationary coordinate system and the other
terms are as defined before.
The diffusion coefficient appearing in the N-P equations, Dim, is the pseudobinary diffusion
coefficient which is defined according to Equation 14 in terms of the multicomponent
diffusion coefficients, D^, as
(14)
The N-P equation have been used for the description of multicomponent diffusion in
hydrogels.
3. Hydrogel Permeability: Kedem and Katchalsky Development

The general equations of TIP are valid for all irreversible processes. However, the physical
form of the forces and fluxes in Equation 9 is very important for efficient use of these
equations in engineering problems. The product of the forces and fluxes must be in entropy
units and the forces and fluxes should be easily measurable as was earlier mentioned. These
criteria must also be fulfilled for all membrane systems. The first to define the fluxes and
forces commonly used for membrane and hydrogel systems were Kedem and Katchalsky.23
The Kedem-Katchalsky theory will be henceforth referred to as the K-K theory. By analogy
to Equation 9 we write:
To-' = Js Ajxs + Jw A|x, (15)
Here the subscripts s and w denote solute and solvent, respectively and A|x is a chemical
potential difference across a hydrogel membrane.
The fluxes and forces defined in Equation 15 may be transformed into simpler quantities.
Chemical potentials can be related to the concentration and the applied hydrostatic pressure
for an ideal solution. With this in mind, K-K were able to transform the gradients of chemical
potentials to gradients of concentration and hydrostatic pressure. Concentration may in turn
be transformed into osmotic pressure by van’t H offs law. A second assumption in the K-
K development is that of small concentration differences. The assumption allows the log-
arithmic average concentration to be approximated by the arithmetic average concentration.
Finally, the solute and solvent fluxes may be related to the total volume flow and the
diffusion flow. Consequently, phenomenological equations relating flows and forces take
the form
Jv = LpAP + LpD At t (16)
and
JD — LpDAP + L d At t (17)
The flux Jv is known as the total flow and JD as the exchange flow of the system. For this
system, Onsager’s reciprocal relation is
Kedem and Katchalsky further simplified their equations by using the reflection coefficient
defined by Staverman24
(19)
where vw and vs are the velocities of the water and solute, respectively. The reflection
coefficient is an adequate measure of membrane selectivity and can be easily measured
experimentally. As it is presently defined, LD, the solute permeability is still rather difficult
to measure. To facilitate an accurate measurement of LD, another convenient coefficient
may be defined. The new coefficient, go , is the coefficient of solute permeability at zero
volume flow, or
( 20 )
The parameters defined by Kedem and Katchalsky have important practical significance.
The parameters Lp, a , and go are much easier to measure than the parameters that had been
measured before them Lp, LD, and LpD, i.e., the hydraulic permeability, the solute perme-
ability, and the permselectivity, respectively. Therefore, these parameters are preferred in
current investigations to characterize the hydrogels used.
II. D IF F U S IO N IN N O N P O R O U S H Y D R O G E L S
In the next two sections research related to polymer structure and solute diffusion through
hydrogels will be reviewed. Basic information about the physical theories which establish
the dependence of the solute diffusion coefficient on the polymer structure will also be
discussed. Finally, several contributions will be analyzed which describe solute diffusion
through hydrogels in terms of phenomenological coefficients.
In terms of solute diffusion there are three general types of polymeric hydrogels.
Macroporous hydrogels — These are hydrogels with large pores, usually greater than
0.1 |jLm. They are used in some separations and other industrial applications. Membranes
with pores as small as 200 A may be included in this category. Although diffusion occurs
through these hydrogels, convection may be the predominant mass transfer mechanism.
Microporous hydrogels — The pore size of these systems ranges from 50 to 200 A. The
pores are slightly larger than the solute. However, under these conditions, the diffusion path
of the solute through the pores is restricted (hindered diffusion).
Nonporous hydrogels — As the term implies, these are hydrogels that do not have a
porous structure. The spaces between macromolecular chains (henceforth referred to as the
mesh) are the only available areas for diffusion of solutes. Therefore, molecular diffusion
is the only mode of mass transport and convection is usually negligible.
A. Determination of Solute Diffusion Coefficients

Several theories have been developed to predict the diffusion coefficient of a solute in a
continuum, including hydrodynamic theories,25 the Eyring theory,26 and the Enskog theory.1
None of the aforementioned theories is directly applicable to hydrogels since the continuum
postulate fails to describe swollen polymer systems at the microscopic level. Diffusion of
even relatively small molecules is strongly affected by the polymer structure.22’27,28 There
are three current developments for the prediction the solute diffusion coefficient dependence
on polymer structure: (1) the statistical mechanical model of diffusion, (2) free-volume
theory, and (3) scaling concepts.
FIGURE 1. Polymer microstructure for solute diffusion in a noncrystalline

hydrogel.
1. Statistical Mechanics
Pace and Datyner have applied statistical mechanics to describe diffusional transla-
tion.29 31 They propose that noncrystalline polymer regions possess an approximate order
with chain bundles that are locally parallel along distances of several monomers. The faint
halos observed in X-ray diffraction patterns of noncrystalline polymers form the basis of
this proposition. The motion of the solute may be more facile in the direction parallel to
the chains than across the chains, as shown in Figure 1. Movement along the axis of locally
parallel chains is relatively fast and determines the effective jump length. Movement across
the axis of chains requires an activation energy to produce sufficient chain separation. The
description applies statistical mechanics to the average diffusional jump length, X, and the
diffusional jump frequency, v. The theory of stochastic processes suggests26,32 that the
diffusion coefficient follows from equation 21
X2v
( 21)
6
The model permits successful calculation of the diffusional activation energy from first
principles in terms of independently determinable molecular parameters. The approach has
been extended to complex (nonspherical) molecules where the diffusion coefficient is de-
pendent on the molecular shape and concentration. Equation 22 shows the functional form
for the concentration dependence of the penetrant for moderate concentrations (c < 10 3
mol/cm3)
D(c,T)
= (1 + f,c)2(l + f2c2) ( 22)
D(0,T)
where f, and f2 are constants for a particular temperature and polymer-penetrant pair. Pace
and Datyner have compared the predicted coefficients for simple penetrants in vinyl, non-
vinyl, and similar polymers.
2. Free-Volume Theories
The free-volume theories are based on the observation that polymers demonstrate the
presence of a substantial portion of void space or unoccupied volume. The size of these
voids may be several times the dimensions of the monomer. Free volume may result from
packing irregularities and long-range monomer interactions which give rise to the excluded
volume effect. Chain mobility further induces temporary voids resulting from Brownian
motion. The free-volume theories postulate that a penetrant must jump from one void to the
next.33 An increase in temperature raises the thermal motions of both the polymer chains
and the penetrant resulting in an increase in free-volume and the penetrant jump frequency.
Free-volume theories generally neglect cooperative motions in which the random chain
motions might amplify the penetrant translation.34 Likewise, the addition of a plasticizer
would also induce chain mobility and, if the plasticizer is not bound to the chains, provide
additional free-volume in which a penetrant is free to diffuse.
The Cohen and Turnbull35 free-volume theory assumes that molecular segments are able
to exchange free-volume without altering the total energy of the system. The theory describes
only the isotropic effect of structural parameters on the transport of small molecules in
polymers. Fujita34 and Yasuda and Lamaze36 later generalized this approach. Reinhart and
Peppas28 extended the theory to include the effects of cross-linking. The free volume theory
is successful in systems which are sufficiently concentrated to permit significant contact
between effective chains. Other works have made further refinements for dilute solutions
and films.37'40 Li and Gainer39 refined the theory in an attempt to explain an increasing solute
diffusion coefficient upon increasing polymer concentration in solutions.
The free-volume development originates with a modified form of Eyring’s rate theory. If
cooperative motions are neglected,26 the diffusing molecule requires a free energy of acti-
vation, AG + , since there exist translational barriers. The solute diffusion coefficient in the
gel, D3 21, is given by Equation 23.
Here k is the jump length, k is the Boltzman constant, h is the Planck constant, and AH +
and AS+ are the enthalpy and entropy of activation, respectively. Similarly, the same form
also describes the solute diffusion coefficient in the pure solvent, D3<1. The assumption is
made that the hydrogel is sufficiently swollen so that the enthalpic contribution to the diffusion
coefficient is the same for the diffusion of the penetrant in the hydrogel as for its diffusion
in pure solvent. It is also assumed that the average jump distance for the molecule is the
same in both media. Under these conditions, the ratio of the solute diffusion coefficient in
the gel to that in the pure solvent is, then,
T\ e A S ^ 21/k
^3,21 _ ~____
e AS3 ,/k
(24)
D3.i
The activation entropy can be expressed as a probability of activation, P +, by the Planck

equation (Equation 25).
S + = k TlnP+ (25)
Upon substituting the respective probabilities from Equation 25 into Equation 24, the nor-
malized diffusion coefficient can be expressed as a ratio of translational probabilities as in
Equation 26.
The probability P^21, contains two important contributions: (1) the probability of finding,
in the medium an opening or unhampered space as an area, equal to or larger than the size
of the solute and (2) the conformational probability of the network forming a volume
sufficiently large enough for the passage of the jumping solute. Therefore, the translational
probability in the gel, P3+21, can be broken into two components: (1) the probability of moving
through the free volume, P3 21 and (2) the probability of moving through the gel mesh,
P^21. Since these actions are simultaneous, the overall probability is simply the product of
the aforementioned probabilities.
p+ p - pi
D3.21 r 3,21 _ r 3 ,2 1 r 3,21
p+ ~ p+ (26)
^3,1 r 3,1 r 3,1
The basis for the free-volume theory is the homogeneously swollen polymer consisting
of the volume fraction of water, v l9 and of polymer, u2. The polymer is impermeable to the
solute so that diffusion may only proceed through the water phase. The free-volume in the
hydrogel, Vf 21, has contributions from the water and the polymer, Vf>1, and Vf 2, respectively.
Vf,2i — UiVf.i + 'u2Vf>2 (27)
The solute is soluble in the solvent and there is bound water immediately neighboring the
chains.41 The free volume contribution from the polymer is typically very small.42
The probability of finding the free volume between V' and V' + dV' is the average
distribution of free-volume, P(V'),
yV'
P(V') = £ exp (28)
* f Vf
where y is a factor (V2 ^ y ^ 1) which corrects for the overlap of free volume associated
with two or more molecules.
The solute can diffuse through the gel if it finds a succession of holes larger than the
solute size (see also Figure 2). The conformal probability of forming a hole sufficiently
large for the solute in one jump, v*, is given by
yv*
P(v*) = P(V') dV' = exp (29)
Jv *
’ vT
The required free volume is the cross-sectional area times the jump length.
v* = it r2\ (30)
where r is the hydrodynamic radius of the solute. Yet, the solute can only pass through this
volume if it is not blocked by the cross-linked chains of the network. Peppas and Moynihan
Volume l: Fundamentals 67
Crosslink
FIGURE 2. Schematic representation of the screening effect of the cross-linked

structure of a cross-linked hydrogel on solute diffusion. The “ blobs” around the
macromolecular chains define a region available for diffusion (mesh). The charac-
teristic length of this region, is the mesh size.
have also considered the probability of finding an adequate opening between the chains in
moderately swollen systems.43
Following the analysis of Yasuda et al.,44 and Peppas and Reinhart,27 Equations 31 and
32 provide the probabilities of fin ; the required free-volume in the gel and solution,
respectively.
“ 7V*
p-
r 3,21 = e Vf-21 (31)
yv*
+ = e
rP 3 , 1 (32)
The ratio of these probabilities is then
r 3,21 <7vr)f c * (33)

P + = e
r 3,1
The free-volume has contributions from the solvent and polymer as shown in Equation
27. Since the diffusion coefficient of the solute in a dry polymer network is several orders
of magnitude smaller than that in a swollen gel, it is reasonable to assume that the free
volume available for solute diffusion is the volume of water in the gel,27 35 44 then Vf3] ~
u, Vf The result after combining Equations 28 and 33 is, then
p- —-
LM 1 = e Q -i (34)
P+
r 3,1
where
_ 7 Tt t 2\
(35)
" v fJ
and
(36)
Here, Q is the volume degree of swelling for the gel and Y is a structural parameter, near
unity, which is proportional27 to r2. In effect, Y is a scale factor for the ratio of the molar
volume displacement per diffusional jump of solute to the free-volume contribution per mole
of solvent.
With these assumptions, the ratio of the solute diffusion coefficients, in the hydrogel,
D312 and the pure solvent, D3 ,, is the product of a sieving probability factor, P^21 and the
free volume factor, i.e.,
= P U e (Q- ') (37)
Several investigators have made contributions by their interpretation of the probability of

moving through the gel mesh, P^21. This probability describes the sieving mechanism whereby
the distance between effective chains in the gel may obstruct the passage of the solute.
Yasuda et al.44 described the sieving mechanism using a differential hole distribution
function f(q). The volume fraction of holes of cross-section q yields the probability 4>(q)
for the existence of a hole with a cross section larger than or equal to q.
P3.21 = <Kq) = Jqf f(x)dx (38)
They showed the effect of several distribution functions including the following: (1) all
holes have equal cross-sections, i.e., a Dirac delta function; (2) the probability for any hole
with a cross-section q ^ 2q0 is equal but there are no larger holes; and (3) the probability
for the hole is Gaussian.
Using Colton et al.45 data, they showed that the effective cross section of the solute, irr2,
is the proper parameter which controls the diffusion of a solute in the gel, and that the
sieving mechanism is insignificant up to the size of albumin (Mw = 66,000) for their high
swollen cellulosic membranes.
Peppas and Reinhart27,28 and Meadows and Peppas46 described the sieving mechanism in
terms of the network mesh size. They postulated that the sieving probability is proportional
to the volume fraction of mesh sizes larger than a critical mesh size, M* (see also Figure
3). They argued that if the effective chains are far apart so that no entanglements exist,
FIGURE 3. Cross-linked structure of a polymer gel, showing effective chains of

the structure defined by cross-links. The effective area for diffusion of a solute is
characterized by an average mesh size, The smaller solutes, illustrated as dark
circles, must pass between the macromolecules.
diffusion is not obstructed by the chains, and the upper limit of Mc is the number average
molecular weight of the uncross-linked polymer, Mn. The lower limit of Mc is the value
M* below which the solute is always screened. They concluded that the sieving probability
is described by
P U = f(n> (39)
where
= mc - m;
(40)
n m „ - m ;
Later, Reinhart and Peppas28 demonstrated that a screening functionality of the form
f(n) = n2 accurately fits diffusion data of bovine serum albumin and other solutes through
some highly swollen in poly (vinyl alcohol) gels (see also Chapter 7).
Recently Vrentas, Duda, and their collaborators47'59 presented a more exact treatment of
the free-volume theory. Their contributions include the calculation of the specific free volume
for polymer-penetrant systems and the definition of conditions under which the mutual
diffusion coefficient can be deduced solely from the free-volume treatment. They defined
the simplifications required to reconcile their treatment with the previous free-volume theory
versions by Cohen and Turnbull35 and Fujita34 which fail at high polymer concentrations
and low solute molecular weights. The highlights of their analysis are presented here.
The specific free-volume, VF1, is the quantity used for the prediction and correlation of
viscosities and self-diffusion coefficients. It is defined as the difference between the specific
volume of the equilibrium liquid structure of a material, V, and the specific volume of the
equilibrium liquid structure at 0°K, V0. The specific free-volume is also considered to be
the sum of the specific interstitial free-volume, VF1, and the specific hole free-volume, VFH.
The former represents the specific volume of temporary voids caused by the nonharmonic
vibrations of Brownian motions. The latter represents the specific volume of interstitial
vacancies caused by instantaneous packing irregularities.
Vp = v - v 0 = VFI + VFH (41)
The reader might consider the distinction between VFI and VpH in terms of the deGennes
blob model60 of polymer structure, (see also Figure 2). The reader should recall that in the
blob model, chain portions between cross-links are confined to “ spheres of influence” . The
interstitial free volume refers to those instantaneous voids created inside a blob and the hole
free-volume refers to the isolated space between blobs.
Following the ideas of Turnbull and Cohen35 and Vrentas and Duda58-59 one can assume
that substantially more energy is required to redistribute the interstitial than the hole free-
volume. Thus, it is the specific hole free-volume which is available for molecular transport.
It is also assumed that the sum of the specific occupied volume and the specific interstitial
free-volume increases linearly with temperature as follows:
i r d (v Fi + Vo)i
Vp, + VD L dT JP “■
(42)
where the expansion coefficient, a c, is assumed to be independent of molecular weight, and

varies between zero and the liquid thermal expansion coefficient. The analysis is extended
to binary systems by assuming an additivity of volumes for the contributions of (VFI + V0)
for each component. Thus,
V f h - V - (VH + y o)M (43)
where
(Vp, + VC)M = V°(0) 0), e x p ( £ a c, d l ) + V«(0) co2 exp^jT a C2 d l ) (44)
Here (ol is the mass fraction of the ith species and V° (0) is the specific volume of pure
component i at 0°K.
The free-volume calculation is then used in the Cohen and Turnbull analysis as described
later for the self-diffusion coefficient of a one component, simple liquid system.
o , = C„, « p ( - ^ - ) = D„, (45)

where
Do1 - X2 — (46)
and D01 is assumed constant since it has a much weaker temperature dependence than the
exponential term.
The treatment is now generalized for a two-component system consisting of a penetrant
(subscript 1) and polymer (subscript 2). The concentration dependence of the self-diffusivities
is as follows:
D, = Dol expj
7(co,v; + (o2gv;)i
(47)
v FH I
a r 7(a>1v ; + (o2^v;)]
D2 = exPi (48)
(N*/N)M VFHC
where M2 is the polymer molecular weight, N* is the effective number of segments in each
polymer chain, N is the number of freely orienting segments in a polymer molecule, A is
a constant, V° is the specific critical hole free-volume of the ith species, and the quantity
£ is defined by Equation 49.
€ = v;/v; = ViMj/V^Mj (49)

Here V* is the critical molar volume of the ith species and Mj is the molecular weight of a
jumping unit. It is also assumed that the critical amount of local hole specific free-volume
necessary for a jump to take place is approximately equal to the specific occupied volume
of the liquid. Thus,
v ; - v?(0) (50)
v ; = V°(0) (51)
Vrentas and Duda57 then extended the analysis for the prediction of mutual diffusion
coefficient, D, in three instances. In a first approximation the authors suggest
(P tx2 + P 2xt) / dfr \

(52)
RT Vdln^/x^
where xt is the mole fraction of the ith species. In the case of trace amounts of solvent,
P = Pi = Doiexp (53)
In the case where > P 2, then P T = P t and
q _ D t P2^2P i / d m \
(54)
RT \dPi / t ,p
This free-volume theory, when combined with the Flory expression for chemical potential
and the Bueche expression for N*/N could then be used to predict the temperature, con-
centration, and molecular weight dependence of D in the amorphous, two component sys-
tems.47 49 The treatment is not valid at very low polymer concentrations at temperatures
significantly greater than the glass transition temperature for at least one of the components,
and nonequilibrium, viz swelling, temperature, and concentration conditions.
In a later publication, Vrentas and Duda55 refined the theory by approximating all the
thermal expansion coefficients by average values in the temperature range under consider-
ation. In addition, the quantity D01 was given an Arrhenius temperature dependence as
follows:
Do1 = D0e E/RT (55)
where the parameters D0, E, and £ must be determined experimentally by a nonlinear

regression analysis. The mutual diffusivity is given by
k v ; + a>2£ v ; r
D = Do1(1 - Uj)2(l (56)
r FH/7
where
— <o ,(K21 + T - Tgl) + — w 2(K22 + T - Tg2) (57)
and
^,V?
v (58)
cojV? + (o2V°
and the parameters K n/y, K21, K 12/y, and K22 are structural parameters defined earlier.48
They summarize how all parameters can be obtained if the following data are available:
density data for the pure polymer and pure solvent as a function of temperature, viscosity
data for the pure polymer and pure solvent as a function of temperature, at least three values
of the diffusivity for the polymer-solvent system at two or more temperatures, and sorption
equilibrium data for the polymer-solvent system at a single temperature or other thermo-
dynamic data from which Xi can be determined. It was shown that the free-volume theory
can accurately predict diffusional behavior over wide ranges of temperature and concentra-
tion. Later contributions include the description of diffusion coefficients for polymer-solvent-
solvent systems57 and for molten polymers with trace amounts of solvents.58
3. Scaling Theories
More recently, Lustig and Peppas61 64 utilized scaling concepts to determine the mor-
phological effects of the solute and gel on the solute diffusivity. They presented the novel
scaling law by Equation 59
P3.21 _
(59)
D3., " ( ‘ “ i)
where the modified equality sign (= ) indicates that any constants of proportionality which
may be system dependent only have been omitted. Here r is the longest dimension of the
solute. The analysis applies for cases where the gel and solute are chemically inert and the
solute concentration is low. At these concentrations, the volume of solute in the gel is small
with respect to the system-free volume available for diffusion in the equilibrium swollen
state. Thus, only the polymer and solvent contribute to the free-volume. The solute can be
spheroidal or ellipsoidal in shape and small with respect to the average length between cross-
links in the polymer gel. Thus, it is not likely that the solute would become entangled in
the macromolecular mesh.
In terms of the free-volume analysis presented earlier, the sieving mechanism is described
in terms of the relative sizes of the solute and the network screen. The sieving mechanism
is given by the probability of a solute of size such that it might pass through an opening of
size £. It is desired to know whether the ratio of the sizes, r/£, or the ratio of the areas, (r/
£)2, best describes this probability. The parameter r characterizes the solute size by means
of the diameter of an equivalent sphere for the solute. In reality, the solute may not be
spherical but, in general, ellipsoidal. Figure 4 ellucidates this problem by showing two
solutes of equal cross-sectional area but different ellipsoidal shape. The hydrodynamic radius
considers an average of the major and minor axes of the ellipsoid. This figure illustrates
that solute 1 stands a better probability of passing through the opening than solute 2.
Therefore, one should conclude that the probability scales to the ratio of the sizes. In
summary,
PL.2 = 1 - ^ (60)
The data presented in Figures 5 and 6 provide support for the scaling law with data from
various references.28,45,65 68 Figure 5 is a plot of the ratio of solute diffusion coefficients vs.
the right-hand side of Equation 59. Each line presents data for solute diffusion coefficients
determined experimentally for hydrogels swollen to equilibrium in water. Figure 6 is a
similar plot of the left hand side of Equation 59 vs. the right, where the data are for different
sizes of solutes through two gels.
In cases where the mesh size, £, is unknown, it is possible to extend the previous scaling
analysis for semidilute, isotropic gels. In the semidilute state, the mesh size obeys the
known60 scaling law;
€ = € 3c 2- 3/4 (61)
and for the polymer, assuming constant molar volume,
1
v 7 = — = c7\ = c7 (62)
Therefore,
€ = Q34 (63)
Combination of Equations 59 and 63 leads to the final scaling law for the semidilute state.
_ D 3.i
[1 - rQ -34] e( (64)
^3,1
The reader should appreciate the importance of determining the volume degree of swelling
for characterizing membranes and gels.
FIGURE 4. Sieve mechanism of hydrogel structure showing that solutes of

different shapes but identical cross-sectional area must pass through the same
average mesh.
Both structural parameters, £ and Q, of the gel are concentration dependent. By definition,
the volume degree of swelling is the reciprocal of the polymer volume fraction. Schaefer69
developed a unified model for the mesh size concentration dependence in gels. In gels, the
correlation length is the network mesh size. As gels become more concentrated, different
molecular pair correlations dominate. Each range of pair correlations has a characteristic
concentration dependence. Figure 7 gives a schematic representation of the correlation length
concentration dependence. The sharp breaks are a consequence of the model and are not
expected in real systems. The transitions shown occur at typical concentrations; the actual
FIGURE 5. Dependence of normalized diffusion coefficient, D312 /D3 ,, on the term (l-r/£)
exp[-l(Q -l)]. Data presented are for salicylic acid (curve 1) through PHEMA, progesterone
(curve 2) through P(HEMA-co-MEMA) hydrogels cross-linked with EGMA, progesterone
(curve 3) through P(HEMA-co-MEMA) hydrogels cross-linked with TEGDMA, theophylline
(curve 4) through PVA, and bovine serum albumin (curve 5) through PVA.
transitions for a real system vary as a function of molecular weight between cross-links and
chain rigidity.
Therefore, it is possible to show the concentration dependence of the normalized solute
diffusivity in a typical gel. The concepts summarized by Equation 59 and the unified theory
for the mesh size correlation length concentration dependence can be combined.63 The theory
used to develop the relation in Equation 59 has been supported with the data up to 1984 for
solutes smaller than the mesh size and polymer volume fractions between 0.05 and 0.74.
In the dilute regime, the polymer chains do not touch each other and therefore, the
characteristic gel network does not exist. As the polymer chains become more concentrated,
the chains gel (“ blob” model). Then, the network structure begins to pose barriers to solute
and Equation 59 should apply. It is apparent that a dramatic decrease in the diffusivity occurs
as the result of: (1) size screening as illustrated by the effect of increasing the solute size
and (2) free-volume availability as illustrated the effect of increasing the polymer volume
fraction. The review article of Nystrom and Roots70 addresses some of these points.
The practical utility of scaling concepts is somewhat limited since the analysis omits
constants of proportionality. The ability of scaling laws to illustrate trends from relatively
simple logical arguments imparts their predictive value. The experimentalist would use them
as semitheoretical correlations for data. The omitted constants for Equation 59 are a function
of the polymer-solvent pair. Once one determines them for a gel and one solute, one can
predict the effect of: (1) increasing the solute size by a known factor and (2) increasing the
polymer cross-linking by a known amount, on the solute diffusivity and rate of solute release.
FIGURE 6 . Dependence of the normalized diffusion coefficient, D3J 2/D 3 ,, on the term
rQ , 4exp[-l(Q -l)]. Data presented are for various solutes through heparinized PVA (curve
1) and Avisco wet gel (curve 2).
B. Effect of Polymer Structure on Solute Diffusion

For nonporous, homogeneous, polymeric hydrogels, Fick’s law, Equation 12, is more
conveniently written as follows:
K,Aclw = dM,
(65)
8 Adt
Here the solute partition coefficient, Ki? describes the equilibrium ratio of the saturation
concentration of the solute in the hydrogel to that in the surrounding release medium, e.g.,
water, and Aclw represents the solute concentration difference of the solutions on either side
of the membrane.
The steady state flux, j i? per unit area of exposure, A, can be also expressed as dM/Adt,
where Mj is the quantity of solute diffusion at time t. It is apparent that solute diffusion
rates are time independent and that they can be controlled by geometric factors (thickness
and surface area of the membrane) and physicochemical parameters (solute diffusion coef-
ficient in the polymer, Dim, and solute partition coefficient, Kj). The last two terms are often
combined as DimKi or DimKj/8 to describe the hydrogel permeability coefficient, Pim. The
choice of polymer determines hydrogel permeability and therefore diffusion rate of each
solute.
From a diffusion point of view, thermodynamic interactions between the polymer and the
diffusing species may be important. The nature of intensive variable gradients responsible
-( /)
:!::
c: Dilute
I
I
I
I
::::s I
-
>.
10 2 I
"-
0
"- !Semi-Dilute
..c I
-..
"- I
0 I
~
I
I
Cl,) I
N
10 I Theta
Cf) I
I
..c:
(/)
Cl,) v* V
~ I
I
I
I
I
10- 4 10- 3 10- 2 10- 1
Polymer Volume Fraction, v2

FIGURE 7. Effect of the polymer volume fraction, u 2 , on the mesh size, I;, for different regimes
of polymer dilution.
for transport, such as chemical potential, pressure, electrostatic potential, and temperature
gradients must also be investigated and controlled.
The solute partition coefficient, K;, is largely a measure of solute preference for the
polymer relative to the surrounding diffusion media. Therefore, this is a purely thermody-
namic parameter. If the solute has chemical groups similar to that of the polymer, the
partition coefficient will be high, whereas if the structures are different, K; will be low.
Paul et al. 71 discussed various aspects of the definition and measurement of the solute partition
coefficient in hydrogels. Weiss et al. 72 · 73 described various experiments of solute transport
in heterogenous hydrogels.
From a materials point of view, optimum diffusive conditions can be achieved by con-
trolling the crystalline phase, porous structure, degree of swelling, additive concentration,
mesh size of the cross-linked macromolecular chains, and thermodynamic transitions related
to macromolecular relaxation phenomena, namely glassy/rubbery transitions in the presence
of a solute and a swelling agent, usually water. 64 · 74
Regardless of the accepted theory, the degree of cross-linking directly affects the solute
diffusion through the polymer network due to the change of the number of junctions. The
cross-linking density, p,, the number average molecular weight between cross-links, Me,
and the mesh size ~. are parameters describing the network structure. They are all affected
by the number of cross-links within the hydrogel. This change in junction point hinders or
facilitates chain mobility and increases or decreases the structural screening effect of the
membrane for solute diffusion. The barriers due to the junctions in a membrane cause a
decrease in the normalized solute diffusion coefficient. Similar effects are observed with
increased branching or filler content. 74
The existence of the crystalline phase in semicrystalline networks requires that significant
modifications be made to properly describe the solute transport that takes place through a
network. It is in the crystalline region of the membrane that one may find extreme sensitivity
to changes in chain arrangement. The reduction of the available space for diffusion due to
the presence of crystallites results in an increased characteristic length of diffusion. Under
ideal conditions, the partition coefficient, the solute diffusion coefficient, and the solute
permeability are independent of position in a specimen and constant within each phase.
Peterlin75 showed that the transport properties of the crystalline region are negligible when
compared to those of the amorphous phase. Also, the distribution of both phases and the
restraint imposed upon the amorphous phase by the crystals are considered the main con-
tributions of anisotropy on transport properties of the semicrystalline hydrogel. The aniso-
tropy of the permeability of the solute through the amorphous phase is considered small and
negligible while the crystalline phase is considered impermeable to solutes.
Peterlin analyzed the solute diffusion coefficient through semicrystalline membranes by
analogy to diffusion through catalytic pores. He presented the expression of Equation 66
'3,2$
( 66 )
B
where D3 21 is the solute diffusion coefficient through an amorphous membrane of the same
polymer without crystallites, i|/ is the detour ratio, and B is the blocking factor. With the
exception of the somewhat sketchy work of Peterlin, there has been no serious attempt until
now to quantify the effects of the degree of crystallinity of swollen networks on solute
diffusion through them.
In a recent unpublished work, Harland and Peppas offered a new model for solute diffusion
in semicrystalline hydrogels where the diffusion coefficient, Dc, is expressed as
D 3.21 (1 - v c)
Dr = (67)
Here, v c is the volume fraction of crystalline polymer in the equilibrium swollen hydrogel
and t is the tortuosity factor of the semicrystalline hydrogel.
III. D IF F U S IO N IN P O R O U S H Y D R O G E L S
Briefly, solute diffusion through porous hydrogels (macroporous and microporous) in the
absence of convection may be described by Fickian or multicomponent diffusion equations.
For diffusion through pores, the diffusion coefficients refer to solute diffusion through the
solvent-filled pores. The structure of the pores in the membrane is incorporated into the
diffusion coefficient by means of the void fraction (porosity), e, and the tortuosity, t . When
some solubility of the solute in the polymer exists, a pore wall partition coefficient, K ,
must be incorporated into the diffusion coefficient. Thus, the final form of the diffusion
coefficient is
eKn
Dlm = D: ( 68 )
where Diw is the diffusion coefficient of the solute through solvent and Dim is the effective
diffusion coefficient of the solute through the hydrogel.
In the special case of microporous systems where the size of the diffusing species is of
the same order of magnitude as the diameter of the pore, special simplified expressions have
been developed to describe the transport process. The Faxen76 theory, the equation of
Renkin,77 and the Quinn-Anderson theory78 are most often used to describe the pore solute
diffusion process. These theories are based on a hydrodynamic analysis of solute diffusion
through porous systems in the absence of pressure gradients.
The general form of the Renkin equation77 is
~ = (1 - X)2(l - 2. KMX + 2.09X3 - 0.95X5) (69)
where
X (70)
Here, Dim and Diw are the diffusion coefficients of the solute in the hydrogel and in water,
respectively. The radius of the solute is represented by rs, while the radius of the pore is
represented by rp. Further corrections of this theory have been discussed by Colton et al.79
and Satterfield et al.80
Several of these theories have been modified and applied to a variety of porous hydrogels,
especially poly(2-hydroxyethyl methacrylate) by Refojo and Leong,81 Ratner and Miller,82
Migliaresi et al.,83 and Wisniewski and Kim.84 However, we must insist that the application
of solute diffusion theories for porous gels to nonporous hydrogels (with molecular mesh
size less than 100 A), an unfortunate and very frequent practice, creates major misconceptions
about the mechanism of diffusion, the values of the solute diffusion coefficients, and the
exchange capabilities of these hydrogels.
Finally, there is a variety of empirical models which have evolved out of efforts to fit
experimental results of the solute diffusion to all types of hydrogels, porous and nonporous.
Prominent among them is the model of Davis85 which expresses the solute diffusion coef-
ficient as
5 ^ = exp [ - ( 5 + 10-4 M{) c j (71)
where M4is the molecular weight of the diffusing solute and cmis the polymer concentration
(in g/g) of the gel.
Other models have been proposed by Paine and Scherr86 and by Ogsdon et al.87 The last
model suggests that
^ = exp [ - ( 1 + - ) v l/2c ,,2m] (72)

Diw rp
where v is the specific volume of the polymer and cm is the polymer concentration in the
gel.
Various modeling aspects of solute diffusion in continuously swelling hydrogels will be
discussed in Volume II, Chapter 7.
Finally, it must be mentioned that the literature is full of a variety of correlations between
the diffusion coefficient and many types of structural or solute parameters, which are often
theoretically unacceptable and of minimal general value. The reader is cautioned to avoid
using simplistic correlations which are not based on some theoretical development, whether
this is a statistical analysis, a free volume theory, or a scaling analysis.
IV . C O N C L U S IO N S
In conclusion, there are several models which can be used to describe solute diffusion in
hydrogels. Classical mechanics and irreversible thermodynamics theories can be used to
describe the transport process. The solute diffusion coefficient can be written in terms of
various structural parameters of the hydrogel, including the swelling ratio, degree of crys-
tallinity, degree of swelling, branching, etc. Theoretical frameworks that can be used to
develop relevant correlations include statistical mechanics, the free-volume theory, and
scaling concepts.
ACKNOW LEDGM ENTS
We wish to acknowledge important contributions by Ronald Harland, David Meadows,

and Richard Korsmeyer. This work was supported by NSF Grant No. CPE-8207381.
REFERENCES
1. Hirschfelder, J. O., Curtiss, C. F., and Bird, R. B., Molecular Theory of Gases and Liquids, John
Wiley & Sons, New York, 1954.
2. Chapman, S. and Cowling, T. G., The Mathematical Theory of Non-Uniform Gases, Cambridge University
Press, London, 1939.
3. Lightfoot, E. N., Transport Phenomena and Living Systems, John Wiley & Sons, New York, 1974.
4. Cussler, E. L., Multicomponent Diffusion, Elsevier, Amsterdam, 1976.
5. Krishna, R. and Standart, G. L., Mass and energy transfer in multicomponent systems, Chem. Eng.
Commun., 3, 201, 1979.
6 . Bird, R. B., Stewart, W. E., and Lightfoot, E. N., Transport Phenomena, John Wiley & Sons, New
York, 1960.
7. Onsager, L., Reciprocal relations in irreversible processes, Phys. Rev., 37, 405, 1931.
8 . Onsager, L., Reciprocal relations in irreversible processes. II, Phys. Rev., 38, 2265, 1931.
9. Miller, D., Transport properties of binary electrolyte solutions and their relation to the thermodynamics of
irreversible processes, Chem. Rev., 60, 15, 1960.
10. Meixner, J., Thermodynamics of irreversible processes, Z. Phys. Chem. Abt. B, 53, 235, 1943.
11. Meixner, J., Thermodynamics of irreversible processes, Ann. Phys., 43, 244, 1943.
12. Eckart, C., Thermodynamics of irreversible processes: simple fluids and fluid mixtures, Phys. Rev., 8 ,
267, 1940.
13. Prigogine, I., Etude thermodynamique des phenomenes irreversibles, University of Liege, Belgium, 1947.
14. de Groot, S. R., Thermodynamics of Irreversible Processes, North-Holland, Amsterdam, 1951.
15. Katchalsky, A. and Curran, P. F., Nonequilibrium Thermodynamics in Biophysics, Harvard University
Press, Cambridge, Mass., 1967.
16. Lorimer, J. W., Boterenbrood, E. I., and Hermans, J. J., Transport processes in ion-selective mem-
branes, Trans. Faraday Soc., 21, 141, 1956.
17. Kramer, H. and Meares, P., Correlation of electrical and permeability properties of ion-selective mem-
branes, Biophys. J., 9, 1006, 1969.
18. McCallum, C. and Paterson, R., Transport properties of a quartenary ammonium ion exchange membrane,
J. Chem. Soc. Earoday Trans. 2, 70, 2113, 1974.
19. Miller, D., Application of irreversible thermodynamics to electrolyte solutions. I, J. Phys. Chem., 70,
2639, 1966.
20. Miller, D., Application of irreversible thermodynamics to electrolyte solutions. II, J. Phys. Chem., 71,
616, 1967.
21. Cussler, E. L., The origin and symmetry relations for membrane diffusion, J. Membr. Sci., 1,319, 1976.
22. Peppas, N. A. and Meadows, D. L., Macromolecular structure and solute diffusion in membranes: an
overview of recent theories, J. Membr. Sci., 16, 361, 1983.
23. Kedem, O. and Katchalsky, A., Thermodynamics analysis of the permeability of biological membranes
to nonelectrolytes, Biochem. Biophys. Acta, 27, 229, 1958.
24. Staverman, A. J., Theory of measurement of osmotic pressure, Rec. Trav. Chem., 70, 344, 1951.
25. Daniels, F. and Alberty, R. A., Physical Chemistry, John Wiley & Sons, New York, 1955.
26. Glasstone, S., Laidler, K. J., and Eyring, H., Theory of Rate Processes, McGraw-Hill, New York,
1941.
27. Peppas, N. A. and Reinhart, C. T., Solute diffusion in swollen membranes. I. A new theory, J. Membr.
Sci., 15, 275, 1983.
28. Reinhart, C. T. and Peppas, N. A., Solute diffusion in swollen membranes. II. Influence of crosslinking
on diffusive properties, J. Membr. Sci., 18, 227, 1984.
29. Pace, R. J. and Datyner, A., Statistical mechanical model for diffusion of simple penetrants in polymers.
I. Theory, J. Polym. Sci. Polym. Phys., 17, 437, 1979.
II. Applications in non-vinyl polymers, J. Polym. Sci. Polym. Phys., 17, 453, 1979.
III. Applications in vinyl and related polymers, J. Polym. Sci. Polym. Phys., 17, 465, 1979.
32. McGregor, R., Diffusion and Sorption in Fibers and Films, Vol. 1, Academic Press, New York, 1974.
33. Hopfenberg, H. B. and Stannett, V., The diffusion and sorption of gases and vapors in glassy polymers,
in The Physics of Glassy Polymers, Harvard, R. N., Ed., John Wiley & Sons, New York, 1973, 504.
34. Fujita, H., Diffusion in Polymer-Diluent Systems, Fortschr. Hochpolym. Forsch., 3, 1, 1961.
35. Cohen, M. H. and Turnbull, D., Molecular transport in liquids and gases, J. Chem. Phys., 31, 1164,
1959.
36. Yasuda, H. and Lamaze, L. E., Permselectivity of solutes in homogeneous water-swollen polymer
membranes, J. Macrom. Sci. Phys., 5, 111, 1971.
37. Vrentas, J. and Duda, J. L., Molecular diffusion in polymer solutions, A.l.Ch.E.J., 25, 1, 1979.
38. Sundelof, L. O., Ber. Bunsenges. Phys. Chem., 83, 329, 1979.
39. Li, S. U. and Gainer, J. L., Ind. Eng. Chem., Fund., 7, 433, 1968.
40. Spriggs, H. D. and Gainer, J. L., Ind. Eng. Chem., Fund., 12, 291, 1973.
41. Ling, G. N. and Horne, R. A., Eds., Water and Aqueous Solutions: Structure, Thermodynamics and
Transport Properties, John Wiley & Sons, New York, 1972.
42. Ferry, J. D., Viscoelastic Properties of Polymers, John Wiley & Sons, New York, 1980.
43. Peppas, N. A. and Moynihan, H. J., Solute diffusion in swollen membranes. IV. Theories for moderately
swollen networks, J. Appl. Polym. Sci., 30, 2589, 1985.
44. Yasuda, H., Peterlin, A., Colton, C. K., Smith, K. A., and Merrill, E. W., Permeability of solutes
through hydrated polymer membranes. III. Theoretical background for the selectivity of dialysis membranes,
Makromol. Chem., 126, 177, 1969.
45. Colton, C. K., Smith, K. A., Merrill, E. W., and Farrell, P. C., Permeability studies with cellulosic
membranes, J. Biomed. Mater. Res., 5, 459, 1971.
46. Meadows, D. L. and Peppas, N. A., Solute diffusion in swollen membranes. III. Non-equilibrium ther-
modynamics aspects of solute diffusion in polymer network membranes, Chem. Eng. Commun., 31, 101,
1984.
47. Vrentas, J. S. and Duda, J. L., Diffusion in polymer-solvent systems. I. Reexamination of the free-
volume theory, J. Polym. Sci. Polym. Phys., 15, 403, 1977.
48. Vrentas, J. S. and Duda, J. L., Diffusion in polymer-solvent systems. II. A predictive theory for the
dependence of diffusion coefficients on temperature, concentration, and molecular weight, J. Polym. Sci.
Polym. Phys., 15, 415, 1977.
49. Vrentas, J. S. and Duda, J. L., Diffusion in polymer-solvent systems. III. Construction of Deborah
number diagrams, J. Polym. Sci. Polym. Phys., 15, 441, 1977.
50. Vrentas, J. S. and Duda, J. L., Diffusion of large penetrant molecules in amorphous polymers, J. Polym.
Sci. Polym. Phys., 17, 1085, 1979.
51. Duda, J. L., Ni, V. C., and Vrentas, J. S., An equation relating self-diffusion and mutual diffusion
coefficients in polymer-solvent systems, Macromolecules, 12, 459, 1979.
52. Vrentas, J. S., Liu, H. T., and Duda, J. L., Effect of solvent size on diffusion in polymer-solvent
systems, J. Appl. Polym. Sci., 25, 1793, 1980.
53. Ju, S. T., Duda, J. L., and Vrentas, J. S., Influence of temperature on the diffusion of solvents in
polymers above the glass transition temperature, Ind. Eng. Chem., Prod. Res. Dev., 20, 330, 1981.
54. Duda, J. L., Vrentas, J. S., Ju, S. T., and Liu, H. T., Prediction of diffusion coefficients for polymer-
solvent systems, A.I.Ch.E.J., 28, 279, 1982.
55. Vrentas, J. S., Duda, J. L., and Hsieh, S. T., Thermodynamic properties of some amorphous polymer-
solvent systems, Ind. Eng. Chem., Prod. Res. Dev., 22, 326, 1983.
56. Vrentas, J. S., Duda, J. L., and Li, L. W., Concentration dependence of polymer self-diffusion coef-
ficients, Macromolecules, 16, 261, 1983.
57. Vrentas, J. S., Duda, J. L., and Ling, H. C., Self-diffusion in polymer-solvent systems, J. Polym. Sci.
Polym. Phys., 22, 459, 1984.
58. Vrentas, J. S., Duda, J. L., and Ling, H. C., Free-volume theories for self-diffusion in polymer-solvent
systems. I, J. Polym. Sci. Polym. Phys., 23, 275, 1985.
59. Vrentas, J. S., Duda, J. L., Ling, H. C., and Hou, A. C., Free-volume theories for self-diffusion in
polymer-solvent systems. II, J. Polym. Sci. Polym. Phys., 23, 289, 1985.
60. deGennes, P. G., Scaling Concepts in Polymer Physics, Cornell University Press, Ithaca, N.Y., 1979.
61. Lustig, S. R. and Peppas, N. A., Scaling concepts in controlled release, Proc. Symp. Controlled Rel.
Bioact. Mater., 11, 104, 1984.
62. Peppas, N. A. and Lustig, S. R., The mathematics and physics of solute transport in continuously swelling
hydrophilic polymers, Proc. Symp. Controlled Rel. Bioact. Mater., 12, 30, 1985.
63. Lustig, S. R. and Peppas, N. A., Scaling laws for solute diffusion in equilibrium swollen gels, Polym.
Prepr., 26(2), 72, 1985.
64. Peppas, N. A. and Lustig, S. R., The role of crosslinks, entanglements, and relaxations of the macro-
molecular carrier in the diffusional release of biologically active materials: conceptual and scaling rela-
tionships, Ann. N.Y. Acad. Sci., 446, 26, 1985.
65. Zentner, G. M., Cardinal, J. R., and Kim, S. W., Progestin permeation through polymer membranes.
II. Diffusion studies on hydrogel membranes, J. Pharm. Sci., 67, 1352, 1978.
6 6 . Wood, J. M., Attwood, D., and Collett, J. H., Characterization of PHEMA gels, Drug Dev. Ind. Pharm.,
93, 93, 1983.
67. Korsmeyer, R. W. and Peppas, N. A., Effect of the morphology of hydrophilic polymeric matrices on
the diffusion and release of water-soluble drugs, J. Membr. Sci., 9, 211, 1981.
6 8 . Babb, A. L., Muarer, C. J., Fry, D. J., Popovich, R. P., and Ramos, C. P., Methods for the in vivo
determination of membrane permeabilities and solute diffusivities, Trans. ASAIO, 14, 25, 1968.
69. Shaefer, D. W., Unified model for the structure of polymers in semidilute solutions, Polymer, 25, 387,
1984.
70. Nystrom, B. and Roots, J., Molecular transport in semidilute macromolecular solutions, J. Macromol.
Sci. Rev. Macromol. Chem., 19, 35, 1980.
71. Paul, D. R., Garcin, M., and Garmon, W. E., Solute diffusion through swollen polymer membranes,
J. Appl. Polym. Sci., 20, 609, 1976.
72. Weiss, N. and Silberberg, A., Permeation in heteroeneous gels, Polym. Prepr. Am. Chem. Soc., Div.
Polym. Chem., 16(1), 289, 1975.
73. Weiss, N., van Vliet, T., and Silberberg, A., Permeability of heterogeneous gels, J. Polym. Sci. Polym.
Phys., 17, 2229, 1979.
74. Peppas, N. A. and Gurny, R., Relation entre la structure des polymeres et la liberation controlee de
principes actifs, Pharm. Acta Helv., 58, 2, 1983.
75. Peterlin, A., Dependence of diffusive transport on morphology of crystalline polymers, J. Macromol. Sci.
Phys., 11, 57, 1975.
76. Faxen, H., Die Bewegung einer starren Kugel langs der Achse eines mit zahrer Flussigkeit gefullten Rohres,
Arkiv. Mat. Astron. Fys., 17, 27, 1923.
77. Renkin, E. M., Filtration, diffusion and molecular sieving through porous cellulose membranes, J. Gen.
Physiol., 38, 225, 1954.
78. Anderson, J. L. and Quinn, J. A., Restricted transport in small pores: a model for steric exclusion and
hindered particle motion, Biophys. J., 14, 130, 1974.
79. Colton, C. K., Satterfield, C. N., and Lai, C. J., Diffusion and partitioning of macromolecules within
finely porous glass, A.I.Ch.E.J., 21, 289, 1975.
80. Satterfield, C. N., Colton, C. K., and Pitcher, W. H., Restricted diffusion in liquids with fine pores,
A.I.Ch.E.J., 19, 628, 1973.
81. Refojo, M. F. and Leong, F. L., Microscopic determination of the penetration of proteins and polysac-
charides into PHEMA and similar hydrogels, J. Polym. Sci. Polym. Symp., 6 6 , 227, 1979.
82. Ratner, B. D. and Miller, I. F., Transport through crosslinked PHEMA hydrogel membranes, J. Biomed.
Mater. Res., 7, 353, 1973.
83. Migliaresi, C., Nicodemo, L., Nicolais, L., and Passerini, P., Physical characterization of PHEMA gels,
J. Biomed. Mater. Res., 15, 307, 1981.
84. Wisniewski, S. and Kim, S. W., Permeation of water-soluble solutes through PHEMA and PHEMA
crosslinked with EGDMA, J. Membr. Sci., 6 , 299, 1980.
85. Davis, B. K., Diffusion in polymer gel implants, Proc. Natl. Acad. Sci. U.S.A., 71, 3120, 1974.
86. Paine, P. L. and Scherr, P., Drag coefficients for the movement of rigid spheres through liquid filled
cylindrical pores, Biophys. J., 15, 1087, 1975.
87. Ogston, A. G., Preston, B. N., Wells, J. D., On the transport of compact particles through solutions of
chain-polymers, Proc. R. Soc. London Ser. A, 333, 297, 1973.
Chapter 4
H Y D RO GEL SURFACES
Buddy D. R atner
TABLE OF CONTENTS
I. Introduction...................................................................................................................... 86
II. Chain Mobility at HydrogelSurfaces............................................................................ 86
III. Water Structuring at HydrogelSurfaces.........................................................................89
Acknowledgments........................................................................................................................ 92
References 92
I. IN TROD UCTION
The surface region of hydrated hydrogel polymers represents a unique environment dif-
ferent from the surfaces of almost all other materials. This surface region is characterized
by high chain mobility, gradients of composition, heterogeneous chain lengths, amphiphilic
character, and unique water structuring. These are all properties that also contribute to the
difficulty in precisely analyzing and characterizing hydrogel surfaces. This chapter will
attempt to sum up some of the generally accepted ideas about hydrogel surfaces and will
consider aspects of hydrogel surfaces that are presently the subject of some controversy.
The elements that together contribute to the complex nature of these hydrogel surfaces,
particularly when they are immersed in biological media, are represented schematically in
Figure 1. These elements are
1. The cross-linked hydrogel polymer chains that exhibit the ability to move in response
to the environment, to interact with chemical components present in the system, and
to structure water.
2. The ions, low molecular weight organics, and proteins that are fractionated by the
hydrogel in various ways resulting in gradients extending from the bulk solution
concentration to the bulk polymer concentration (within the gel).
3. The water that can be present as free molecules, molecules bound to the polymer
chains or to the diffusants mentioned in (2) above, and in structured ice-like units.
The water structure can be induced by the polymer chains or the diffusants in the
system.
These three considerations form the primary foci for research on hydrogel surfaces. This
chapter will concentrate on studies in these areas.
Three other areas of investigation should also be mentioned. First, there is a large body
of literature relating to the chemical modification of hydrogel surfaces and the chemical
modification of nonhydrogel surfaces to give them hydrogel character (see Chapter 5). There
is, in fact, so much work in this area that a review of this topic would distort the intended
emphasis of this chapter, i.e., the nature of the unaltered surfaces. The reader should see
References 1 to 7 for a sampling of work in this area. The second relevant area of study
relates to theoretical treatments of hydrogel surfaces. Work in this area attempts to model
various aspects of these complex surfaces, particularly in terms of Gaussian chain statistics
(see, for example, Chapter 2 and Reference 8 or 9). A general conclusion from such
theoretical treatments is that the thickness of the diffuse gel interface increases with increasing
solvent strength.9 However, this chapter will restrict its scope and concentrate on experi-
mental studies. Finally, there have been a few studies dealing with the micromechanical
properties of hydrogel interfaces. The deformability of these interfaces can absorb energy
from a fluid stream and alter the expected frictional resistance to flow.1011
II. C HAIN M O BILITY AT H Y D RO GEL SURFACES
Water acts as a plasticizer for many hydrophilic polymers. If polymers can exhibit hydrogel
character (see Reference 12 for a review of factors that are advantageous for hydrogel
behavior), the backbone and side chains will exhibit relatively unrestricted rotational mobility
at room temperature. Translational mobility will be somewhat (but certainly not completely)
hindered depending upon the extent of covalent or noncovalent cross-linking in the gel (see
Chapter 1).
This chain mobility becomes important in understanding the nature of the hydrogel surface
when two factors are considered: the strongly dipolar (or amphiphilic) character of most
BULK GEL DIFFUSE SURFACE BULK WATER
POLYMER CHAINS
~
0 FREE WATER MOLECULE
G> BOUND WATER MOLECULE
~ IONS
0 LOW MOLECULAR WEIGHT

ORGANICS
PROTEINS
FIGURE I. A schematic diagram illustrating the principle interactive

components at the hydrogel-biological fluid interface.
hydrogel monomers and the thermodynamic driving force that attempts to direct the system
toward the minimization of the interfacial energy. Thus, consider Figure 2. This schematic
diagram presents· the accepted model for surface transitions that occur with high! y polar
hydrogel polymers. For a hydrogel polymer like poly(2-hydroxyethyl methacrylate) (PHEMA),
if the polymer surface is in contact with water (Figure 2A), hydroxyl groups will be oriented
out toward the water phase. If the polymer contacts a hydrocarbon solution (Figure 2B),
the backbone methyl groups will be preferentially oriented to the interface. The hydrogen
bonding that can occur across the interface when hydroxyl groups interact with water sig-
nificantly reduces the interfacial energy. Similarly, the van der Waals interactions between
hydrocarbon liquids and methylene or methyl groups result in an energy minimization. For
the case of a hydrogel polymer surface contacting air or vacuum, the presence of unsolvated
OH OH c c~
OH OH CH CH 3
OH OH CH CH~
OH OH CH CH~
OH OH CH CH~·
OH OH CH CH 3 .
OH OH CH H~
OH OH >eH
OH OH <CH
(A) (B)
FIGURE 2. Schematic representation of the model for the surface chemistry change that takes
place depending upon whether a slab ofpoly(2-hydroxyethyl methacrylate) is immersed in (A) water,
or (B) a hydrocarbon liquid.
hydroxyls would be energetically unfavorable. Since the interaction energy of a van der
Waals-type bond is significantly lower than that of a hydrogen-bond-type interaction, it will
be energetically preferable to expose "bare" hydrocarbon moieties as opposed to "bare"
hydroxyls. The high chain mobilities typically observed in this class of polymers make these
energetically driven chain conformational alterations possible. Experimental observations
that support this model will now be discussed.
In a 1975 publication, Holly and Refojo' 3 observed that fully hydrated PHEMA gels with
water contents between 31 and 42% showed water contact angles measured in air that were
between 60 and 80°. This suggests substantial hydrophobic surface character. 13 Receding
contact angles measured in the same study were much lower, but were still not zero. This
contact angle hysteresis could not be explained by surface roughness, dissolution of material
from the gel, or chemical changes induced by the liquid in the gel. When underwater contact
angles of air bubbles on PHEMA surfaces were measured, high contact angles were noted.
Thus, the water contact angle in this situation was low, an indication that the surface had
substantial hydrophilic character. These observations led to the proposal that the PHEMA
chains could undergo significant rotation directing to the surface either backbone methyl
groups or hydroxyl groups. This explanation could account for the behavior of a surface
that could show either hydrophilic or hydrophobic character depending upon the environment
to which it was exposed.
In a follow-up study on this subject, 14 the same authors reviewed previously published
studies that suggested that as early as 1938, this chameleon-like character for hydrogel
surfaces had been observed on gelatin. Additional studies were presented demonstrating by
contact angle methods various aspects of this surface reversibility for a few hydrogel systems
including PHEMA, poly (glyceryl methacrylate), poly(hydroxyethyl acrylate), and certain
graft hydrogels and copolymers. In particular, it was observed that the Zisman critical surface
measurement on hydrated hydrogels did not suggest hydrophilic surface character, probably
because the measurements were made with liquids that were immiscible with water. Such
liquids would tend to orient the hydrophobic groups of the hydrogel toward the interface
emphasizing the hydrophobic aspects of these polymers. Similar observations that also
indicated a rather hydrophobic behavior for hydrogel surfaces when studied by conventional
contact angle methods in air, were made in a later study.15 The parameter that did reflect
the expected hydrophilic nature for these highly hydrated hydrogel surfaces was the contact
angle of a captive drop of water in the presence of a hydrophobic liquid (^-octane).14 This
contact angle decreased with increasing water content in the gel.
Electron spectroscopy for chemical analysis (ESCA)16 has been used to directly observe
these changes in polymer functional group orientation in the surface 30 to 100 Aof hydrogels.
Radiation grafted PHEMA and polyacrylamide hydrogels on silicone rubber were studied
by ESCA in both a dry state and in a hydrated condition.17 Because the ESCA analysis is
performed under high vacuum, the hydrated specimens were actually examined in a frozen
form at liquid nitrogen temperature. For the acrylamide grafts in the hydrated frozen con-
dition, a carbon Is spectrum indicative of acrylamide was observed and a nitrogen signal
was measured. Upon drying the surfaces in situ, i.e., raising the temperature, the surface
spectra changed and were found to be characteristic of only silicone rubber; nitrogen was
no longer detected. Rehydration of the specimen resulted in a reappearance of the hydrogel
graft at the surface, as indicated by the Cls peak shape and the nitrogen signal. Similar
observations were made for PHEMA grafted to silicone rubber. Representative ESCA spectra
from these experiments are presented in Figure 3. Two explanations for these results are
possible:
1. The hydrogel layer has sufficient molecular mobility that, in order to minimize inter-
facial energy, it can either come to the surface to hydrogen bond with the water or it
can bury itself in the highly flexible poly(dimethyl siloxane) matrix thereby exposing
only low surface energy silicone rubber at the surface.
2. Low molecular weight uncross-linked silicone oligomers are present in this polymer
system. For the hydrated surface, they reside in the bulk of the polymer allowing the
hydrogel to interact strongly with the aqueous phase. Upon dehydration, these oli-
gomers rise to the surface and cover over the hydroxyl groups.
Either scenario is consistent with the concept of hydrogel surfaces behaving in a chameleon-
like manner to minimize interfacial energy.
Many other studies have appeared in the literature that illustrate this surface rearrangement
in response to the environment in contact with the surface. For other examples of this
phenomenon, the reader should see References 15, 18 to 20, and 21.
The surface mobility of hydrogels forces one to critically review surface studies performed
on dry hydrogels. Still, it is probable that interesting compositional information on these
types of systems can be gleaned from carefully performed ESCA studies on dry hydrophilic
polymers.22'25
III. W A TER STRU CTU RIN G AT HY D RO G EL SURFACES
Within liquid water, the water molecules can interact (or not interact) with each other to
form a distribution of organizational units ranging from monomeric water molecules to
highly structured or “ ice-like” units.2627 As early as 1973, it was hypothesized that the
1,0
A =
5K IOK~C
No. of No of
Counts Counto
Acrylamide on Silicone Rubber IlK HEMA on SoliC<III-
41<
Hydrated H,.W ..d
$
31( 61<
: ··. ...·...-:· ~
:>Q
2K 4K
.;.· ... .·· ...·.. :.•. ..:·-·....: ..... 2K..,.··········•:............................. ...............•.. :·'
"';:;;-
IK ·.. ·........ ·· ;::;·
D D
~
I:>.
;::;·
s·
295 290 285 2llD 295 290 285 2llD
BINDING ENERGY (ev) BINDING ENERGY (ev) "'l:l;::,
D I:>.
"tl
;::,...
B No. of
No. of ~
Counts
Counts
Acrylamide on Silicone Rubber HEMA on Silicone Rubber
~
l:l
Dehydrated Dehydrated
20K 20K
~
15 K 15K
IOK 10<
5K 5K
0 0
300 295 285 280 300 295 290 285 280

BINDINGZfNERGY (ev) BINDING ENERGY (ev)
FIGURE 3. Carbon Is ESCA spectra for radiation grafted hydrogel surfaces: (A) Acrylamide grafted to silicone rubber, hydrated (160K), (B)
Acrylamide grafted to silicone rubber, dehydrated (303K), (C) 2-hydroxyethyl methacrylate grafted to silicone rubber. hydrated (160K), and
(D) 2-hydroxyethyl methacrylate grafted to silicone rubber, dehydrated (303K).
special biological behavior observed with hydrogels may be related to their abilities to
structure water.28 29 Water structuring arguments have also been invoked to explain other
aspects of the functioning biological systems.30'36 Many measurements have been made of
water structuring within hydrogels.37 42 On the other hand, some convincing data have been
obtained by dielectric and heat capacity methods suggesting that little water structuring
within polymers occurs.43 The majority of the papers in the literature on this subject do
present strong data suggesting the presence of structured water in hydrogels. Thermodynamic
interpretations and other arguments based on experimental data present a sufficiently strong
case for some degree of reorganization of water within and at the surface of polymeric
systems that it is worthwhile for this paper to review some of the evidence relating to this
structuring. However, when appropriate, the controversy that now exists concerning struc-
turing arguments will be discussed.
If full agreement does not exist with regard to structuring of water within polymers, then
the situation is further complicated for the understanding of the structuring of water at
polymer and hydrogel interfaces, since this measurement at such surface-localized zones is
a significantly more difficult experimental problem. Still, it is not unreasonable to speculate
that this water structuring or reorganization due to perturbation by the surface does exist
and can influence interactions with the hydrogel surface. A number of suggestions concerning
water structuring at hydrogel surfaces have been offered,24 36 but little hard data have been
presented.
Reasonable data on water interactions at surfaces can be obtained using rigid, organized
inorganic surfaces because of the availability of good analytical techniques to study such
surfaces. To set the stage for the discussion of hydrogel surfaces, let us consider some of
the conclusions from research in this area. Many studies have been published that deal with
the binding of water molecules at such rigid surfaces, particularly to highly specific surface
sites.44 Where the surface sites are arranged in precise geometric arrays, e.g., at a specific
single crystal face, one can readily envisage that adsorbed water would also be structured
in a highly organized fashion. The geometric organization of a single layer of water molecules
at ordered surfaces can be measured by such techniques as low energy electron diffraction
(LEED). Information on the composition of a surface and the interactions occurring at a
surface can be obtained from Auger and photoelectron spectroscopy, electron energy loss
spectroscopy and infrared spectroscopy. Results from LEED, Auger, and IR studies suggest
that, for water molecules at clean ordered surfaces, e.g., the Fe 100 surface, phenomena
such as site-to-site mobility, site vacancies, hydrogen bonding, and chemical reaction, e.g.,
hydroxyl formation, are observed.45 However, there is significant variation in behavior from
surface to surface. Thus, for platinum 100 surfaces, the spectrum for adsorbed multilayers
has been described as ice I-like, based on high resolution electron energy loss studies.46 In
any event, it is reasonable to say that the behavior observed for water molecules at the
surface is substantially different from that in the liquid far from the surface. However, as
multiple layers of adsorbed water build up on surfaces, significant controversy exists as to
how far into the bulk water phase surface-perturbed water molecules will exert their influence.
Thus, some authors propose that the perturbations occurring at a surface induce a structure
that is propagated out to 0.1 |Jim ,47 while other workers report evidence that no thick stagnant
layers are found at the water-solid interface48 or that surface induced changes in molecular
mobility do not reach beyond the first layer of water molecules.49 To settle this controversy,
better experimental techniques for the study of multilayer water films at surfaces will be
required.
An interesting perspective on the water-polymer interface has been presented by Adam-
son.44 Based largely upon thermodynamic arguments developed from contact angle and
adhesion studies, he challenges a number of common assumptions and presents a model for
the structure of water at polymer surfaces. He concludes that structural perturbations of
adsorbed water at the polymer surface simply cannot be ignored since finite contact angles
could not exist if the vapor in equilibrium with the surface was not of a significantly different
structure from the bulk liquid water in the drop. In addition, he concludes, based upon
observations of contact angle hysteresis on agar gels containing 99% water, that contact
angle hysteresis may be due to short range molecular mobility, i.e., topological and rheo-
logical phenomena, rather than surface heterogeneity as generally assumed. Finally, he
addresses the issue of the transition between the interfacial film region and the bulk water.
The presumption is made that this interfacial zone is of the same order of thickness as the
structured, adsorbed water layer at the surface.
Finally, it is worthwhile to consider a recent statistical thermodynamic model proposed
by Etzler for water structure near solid interfaces.50 The model is based upon the hypothesis
that the hydrogen bonding probability for water increases the closer the water molecule is
to the surface. This model has been shown to be consistent with experimental observations
reported by a number of authors in the literature. In a recent experimental study, a density
of water of 0.966 g/cm3 was found for water in the narrow pores of a silica gel (a hydrogel-
like surface?).51 This value compares favorably to a value of 0.970 g/cm3 calculated by
Etzler based upon his “ bond percolation” model. However, the experimentally determined
values for the density of water within narrow pores is also not without controversy.51
In conclusion, we believe there is compelling evidence to suggest that water reorganization
occurs at all surfaces and more specifically, with regard to this article, at hydrogel surfaces.
However, for hydrogel surfaces, there is as yet little unambiguous experimental support for
this surface water reorganization. New and more refined experimental tools for investigating
structure at these surfaces are clearly needed.
ACK N O W LED G M EN T
Support was received from NIH grant RR 01296 during the preparation of this manuscript.
REFERENCES
1. Samal, R. K., Iwata, H., and Ikada, Y., Introduction of reactive groups onto polymer surfaces, in
Physicochemical Aspects of Polymer Surfaces, Vol. 2, Mittal, K. L., Ed., Plenum Press, New York, 1981,
801.
2. Epton, R., Chemical and biochemical reactions in the gel phase, in Reactions on Polymers, Moore, J. A.,
Ed., Reidel Publishing, Boston, 1973, 286.
3. Ratner, B. D. and Hoffman, A. S., The effect of cupric ion on the radiation grafting of /V-vinyl-2-
pyrrolidone and other hydrophilic monomers onto silicone rubber, J. Appl. Polym. Sci., 18, 3183, 1974.
4. Ratner, B. D. and Hoffman, A. S., Synthetic hydrogels for biomedical applications, in Hydrogels for
Medical and Related Applications, Vol. 31, ACS Symposium Series, Andrade, J. D., Ed., American
Chemical Society, Washington, D. C., 1976, 1.
5. Jansen, B. and Ellinghorst, G., Modification of polyetherurethane for biomedical application by radiation
induced grafting. II. Water sorption, surface properties, and protein adsorption of grafted films, J. Biomed.
Mater.Res., 18, 655, 1984.
6 . Tazuke, S. and Kimura, H., Surface photografting. II. Modification of polypropylene film surface by
graft polymerization of acrylamide, Makromol. Chem., 179, 2603, 1978.
7. Matsunaga, T. and Ikada, Y., Surface modifications of cellulose and polyvinyl alcohol, and determination
of the surface density of the hydroxyl group, in Modification of Polymers, Vol. 121, ACS Symposium
Series, Carraher, C. E. and Tsuda, M., Eds., American Chemical Society, Washington, D. C., 1980, 391.
8 . Freed, K. F., Excluded volume effects in polymer attached to surfaces: chain conformational renormalization
group, J. Chem. Phys., 79, 3121, 1983.
9. Silberberg, A., The gel-solvent interface, Polym. Prepr. Am. Chem. Soc. Div. Polym. Chem., 11, 1289,
1970.
10. Krindel, P. and Silberberg, A., Flow through gel-walled tubes, J. Colloid Interface Sci., 71, 39, 1979.
11. Silberberg, A., The role of membrane-bound macromolecules and macromolecules in solution in cell/cell
encounters in flowing blood, Ann. N. Y. Acad. Sci., 416, 83, 1983.
12. Ratner, B. D., Biomedical applications of hydrogels: review and critical appraisal, in Biocompatibility' of
Clinical Implant Materials, Vol. 2, Williams, D. F., Ed., CRC Press, Boca Raton, Fla., 1981, 145.
13. Holly, F. J. and Refojo, M. F., Wettability of hydrogels. I. Poly(2 -hydroxyethyl methacrylate), J. Biomed.
Mater. Res., 9, 315, 1975.
14. Holly, F. J. and Refojo, M. F., Water wettability of hydrogels, in Hydrogels for Medical and Related
Applications, Vol. 31, ACS Symposium Series, Andrade, J. D., Ed., American Chemical Society, Wash-
ington, D. C., 1976, 252.
15. Ko, Y. C., Ratner, B. D., and Hoffman, A. S., Characterization of hydrophilic-hydrophobic polymeric
surfaces by contact angle measurements, J. Colloid Interface Sci., 82, 25, 1981.
16. Ratner, B. D., Surface characterization of biomaterials by electron spectroscopy for chemical analysis,
Ann. Biomed. Eng., 11, 313, 1983.
17 Ratner, B. D., Weathersby, P. K., Hoffman, A. S., Kelly, M. A., and Scharpen, L. H., Radiation
grafted hydrogels for biomaterial applications as studied by the ESCA technique, J. Appl. Polym. Sci., 22,
643, 1978.
18. Baszkin, A., Boissonnade, M. M., Proust, J. E., Tchaliovska, S., TerMinassian-Saraga, L., and
Wajs, G., Silicone grafted with poly(vinyl pyrrolidone) for contact lenses. Surface properties and stability
of thin tear film, J. Bioeng., 2, 527, 1978.
19. Yasuda, H., Sharma, A. K., and Yasuda, T., Effect of orientation and mobility of polymer molecules
at surfaces on contact angle and its hysteresis, J. Polym. Sci. Polym. Phys. Ed., 19, 1285, 1981.
20. Yasuda, H., Sherry, B., El-Nokaly, M. A., and Friberg, S. E., Preparation of cationic polymer surfaces
by grafting polymerization, J. Appl. Polym. Sci., 27, 1735, 1982.
21. Bigelow, R. W., Bailey, F. C., Salaneck, W. R., Pochan, J. M., Pochan, D. F., Thomas, H. R., and
Gibson, H. W., Structure variation of sulfonated polystyrene surfaces, in Advances in Chemistry Series,
Ions in Polymers, Vol. 187, Eisenberg, A., Ed., American Chemical Society, Washington, D. C., 1980,
295.
22. Miller, D. R. and Peppas, N. A., Surface analysis of poly(vinyl-alcohol-co-/?-vinyl-2-pyrrolidone) by
ESCA, Polym. Prepr. Am. Chem. Soc. Div. Polym. Sci., 24, 12, 1983.
23. Burkstrand, J. M., Copper-polyvinyl alcohol interface: a study with XPS, J. Vac. Sci. Technol. A, 16,
363, 1979.
24. Gilding, D. K., Paynter, R. W., and Castle, J. E., Quantitative evaluation of water structuring and
protein adsorption on the surface of hydrophilic polymers by ESCA, Biomaterials, 1, 163, 1980.
25. Ratner, B. D. and Hoffman, A. S., Surface characterization of hydrophilic-hydrophobic copolymer model
systems. I. A preliminary study, in Adhesion and Adsorption of Polymers, Part B, Lee, L. H., Ed., Plenum
Press, New York, 1980, 691.
26. Franks, F., The role of water structure in disperse systems, Chem. Ind., 7, 560, 1968.
27. Franks, F., Water, A Comprehensive Treatise, Vol. 1, Plenum Press, New York, 1972.
28. Andrade, J. D., Lee, H. B., Jhon, M. S., Kim, S. W., and Hibbs, J. B., Jr., Water as a biomaterial,
Trans. Am. Soc. Artif. Int. Organs, 19, 1, 1973.
29. Jhon, M. S. and Andrade, J. D., Water and hydrogels, J. Biomed. Mater. Res., 7, 509, 1973.
30. Klotz, I. M., Water: its fitness as a molecular environment, in Membranes and Ion Transport, Vol. 1,
Bittar, E. E., Ed., John Wiley & Sons, New York, 1970, 93.
31. Bruck, S. D., Blood Compatible Synthetic Polymers, Charles C Thomas, Springfield, 111. 1974.
32. Cooke, R. and Kuntz, I. D., The properties of water in biological systems, Annu. Rev. Biophys. Bioeng.,
3, 95, 1974.
33. Tait, M. J. and Franks, F., Water in biological systems, Nature (London), 230, 91, 1971.
34. Drost-Hansen, W., The effects on biologic systems of higher-order phase transitions in water, Ann. N. Y.
Acad. Sci., 125, 471, 1965.
35. Ling, G. N., The physical state of water in living cells and its physiological significance. Intern. J.
Neurosci., 1, 129, 1970.
36. Garda, C., Anderson, J. M., and Barenberg, S. A., Hemocompatibility: effect of structured water,
37. Lee, H. B., Jhon, M. S., and Andrade, J. D., Nature of water in synthetic hydrogels. I. Dilatometry,
specific conductivity, and differential scanning calorimetry of polyhydroxyethyl methacrylate, J. Colloid
Interface Sci., 51, 225, 1975.
38. Choi, S., Jhon, M. S., and Andrade, J. D., Nature of water in synthetic hydrogels. III. Dilatometry,
specific conductivity, and dielectric relaxation of poly(2,3-dihydroxypropyl methacrylate), J. Colloid In-
terface Sci., 61, 1, 1977.
39. Taniguchi, Y. and Horigome, S., The states of water in cellulose acetate membranes, J. Appl. Polym.
Sci., 19, 2743, 1975.
40. MacKenzie, A. P. and Rasmussen, D. H., Interactions in the water-polyvinylpyrrolidone system at low
temperatures, in Water Structure at the Water-Polymer Interface, Jellinek, H. H. G., Ed., Plenum Press,
New York, 1976, 146.
41. Fleming, W. W., Fornes, R. E., and Memory, J. D., Evidence of distinct water species in cellulosic
environments from broad-line NMR, J. Polym. Sci. Polym. Phys. Ed., 17, 199, 1979.
42. Carles, J. E. and Scallan, A. M., The determination of the amount of bound water within cellulosic gels
by NMR spectroscopy, J. Appl. Polym. Sci., 17, 1855, 1973.
43. Hoeve, C. A. J., The structure of water in polymers, in Water in Polymers, Vol. 127, ACS Symposium
Series, Rowland, S. P., Ed., American Chemical Society, Washington, D. C., 1980, 136.
44. Adamson, A. W., The water-polymer interface, in Water in Polymers, Vol. 127, ACS Symposium Series,
Rowland, S. P., Ed., American Chemical Society, Washington, D. C., 1980, 87.
45. Klier, K. and Zettlemoyer, A. C., Water at interfaces: molecular structure and dynamics, in Colloid and
Interface Science, Vol. 1, Kerker, M., Zettlemoyer, A. C., and Rowell, R. L., Eds., Academic Press,
New York, 1977, 231.
46. Ibach, H. and Lehwald, S., The bonding of water molecules to platinum surfaces, Surf. Sci., 91, 187,
1980.
47. Drost-Hansen, W., Effects of vicinal water on colloid stability and sedimentation processes, in Colloid
and Interface Science, Vol. 1, Kerker, M., Zettlemoyer, A. C., and Rowell, R. L., Eds., Academic Press,
New York, 1977, 267.
48. Lyklema, J., Water at interfaces: a colloid-chemical approach, in Colloid and Interface Science, Vol. 1,
Kerker, M., Zettlemoyer, A. C., and Rowell, R. L., Eds., Academic Press, New York, 1977, 257.
49. Stigter, D., Mobility of water near charged interfaces, Adv. Colloid Interface Sci., 16, 253, 1982.
50. Etzler, F. M., A statistical thermodynamic model for water near solid interfaces, J. Colloid Interface Sci.,
92, 43, 1983.
51. Etzler, F. M. and Fagundus, D. M., The density of water and some other solvents in narrow pores, J.
Colloid Interface Sci., 93, 585, 1983.
Chapter 5
IM M O B ILIZA TIO N OF BIOM OLECULES AND CELLS ON AND W ITHIN

SY N TH ETIC POLYM ERIC HYDROGELS
Wayne R. Gombotz and Allan S. Hoffman
TA BLE OF CONTENTS
I. Introduction.......................................................................................................................96
II. Preparation of Hydrogels ...............................................................................................96

A. Bulk Polymerization........................................................................................... 96
B. Grafting to a Support..........................................................................................96
C. Conversion of Polymers...................................................................................100
III. Surface Characterization Techniques.......................................................................... 100
IV. Types of Molecules and CellsUsed in Hydrogel Immobilizations......................... 101
V. Immobilization Techniques..........................................................................................102
A. Entrapment......................................................................................................... 102
1. Bulk Polym ers....................................................................................... 102
2. Encapsulation........................................................................................108
B. A dsorption......................................................................................................... 108
C. Covalent Attachm ents...................................................................................... 109
D. Combined Techniques...................................................................................... I l l
VI. Microenvironmental Effects..........................................................................................112
VII. Applications................................................................................................................... 113

A. Analytical and Diagnostic............................................................................... 114
1. Column Techniques.............................................................................. 114
2. B iosensors............................................................................................. 115
3. Im m unoassays...................................................................................... 116
B. Therapeutic Bioreactors and Drug Delivery System s..................................117
C. Industrial Applications .....................................................................................118
VIII. Conclusion.......................................................................................................................119
Acknowledgment.......................................................................................................................120
References 120
I. IN TROD UCTION
The study and application of immobilized biologically active molecules and cells have
become increasingly important endeavors in both medicine and industry. Biomolecules and
cells can be immobilized on and within many different supports using a variety of techniques.
Hydrogels have some properties which make them particularly suitable for this purpose.
A hydrogel can be defined as a polymeric material which exhibits the ability to swell in
water and retain a significant fraction of water within its structure without dissolving. Small
hydrophilic molecules can readily diffuse through hydrogels. Hydrogels exhibit good tissue
biocompatibility and may interact less strongly with immobilized species than more hydro-
phobic materials. Thus, molecules and cells immobilized on or within hydrogels may be
more likely to retain their biological activity for longer periods of time. Hydrogels may also
have a large number of polar reactive sites on which biomolecules and cells can be im-
mobilized by relatively simple chemistries.
This chapter will, first, briefly review the preparation and surface characterization of
hydrogels already discussed in Chapters 1 and 4, in order to stress the important characteristics
of hydrogels in relation to immobilization procedures. The wide variety of biomolecules
and cells suitable for hydrogel immobilization will be considered next, along with the different
immobilization techniques. Finally, applications of hydrogels containing immobilized biom-
olecules and cells will be discussed.
II. PREPA R A TIO N OF HYDROGELS
There are three ways to prepare hydrogels for the immobilization of biomolecules and
cells: (1) bulk polymerization, (2) grafting to a support, and (3) conversion of existing
polymers.
A. Bulk Polymerization
Many vinyl monomers exist that can potentially be used to synthesize a hydrogel. Table
1 shows some of the commercially available synthetic monomers. Bulk hydrogels can be
formed with one or more monomers. This wide variety of compounds enables one to
specifically adjust the physical properties of a hydrogel for a given application, (see Chapter
1). Usually a small amount of cross-linking monomer is included in any hydrogel formulation.
The polymerization reaction is normally initiated with radiation, ultraviolet, or chemical
catalysts. The choice of a suitable initiator will depend upon the particular monomers and
solvents being employed. Radiation polymerization does not require catalysts and can be
used with any monomer. Each initiaion process yields free radicals which initiate a chain
reaction with the monomer molecules resulting in a polymerized gel.
Chemical initiators include heat labile peroxides or room temperature redox systems with
reducing agents such as ferrous salts, sodium metabisulfite, or tetramethylethylenediamine
(TEMED) plus oxidizing agents such as ammonium persulfate or hydrogen peroxide. Sources
for radiation polymerization include Cobalt 60, Cesium 137, and electron beam accelerators.
The polymerized hydrogel may be produced in a wide variety of forms including films and
membranes, rods, particles, and emulsions.
Some hydrogels such as polyesters and polyamides are formed by condensation or step-
addition polymerization. In these reactions a small molecule may be split out resulting in
the formation of an ester or amide bond. Gilding and Reed2 5 synthesized copolymers of
polyethylene oxide (PEO)/polyethylene terephthalate using condensation polymerization.
B. Grafting to a Support
Bulk polymerized hydrogels have an inherently weak structure. To improve the mechanical
Table 1
SYNTHETIC MONOMERS USED IN HYDROGELS1
M O N O M E R S U S E D IN H Y D R O G E L S
N EU TRA L = = = = = = = = = = a c id ic o r a n io n ic
/CH3 ACRYLIC A CID, CH2 = C - C 0 2 H

h y d r o x y m et h y l c h 2 = cv
M ETHACRYLATE ^C C ^C Ô H D E R IV A T IV E S
(R = —H ,- C H 3 )
c h 3
G LY C ER Y L M ETHA C RY LATE CH2 = CRO TON IC ACID CH3 - C = C H - C 0 2 H (w ith V A c )
c o 2ch 2 c h - c h 2o h
OH
SODIUM S T Y R E N E /= \ 0 ®
/C H 3 CH 2 = CH - < w > - S 0 3 No
SULFO NA TE
PROPYLENEGLYCOL
c h 2 " C \ C 0 2 CH2 C H - C H 3
M ETHACRYLATE
OH
B A SIC O R C A T IO N IC
2 ,4 P E N T A D IE N E -I-O L C H 2 = CH —CH = CH — CH 2 OH
fjt A M IN O E T H Y L METHACRYLATE, CH2 = C '

D ER IV ATIV ES c o 2- c 2h 4 n - r
A CR Y L A M ID E CH2 = C - C O - N - R
D ER IV A T IV E S R" ( r , r ' , r " = - h - c h 3 , - C4 H g ) ^
(r = - h , - c h 3 ) P Y R ID E N E c h 2 =c h ^ n
(r , r "= h , - c h 3 - c 2 h 5 , — c h 2 c h o h c h 3
... w
N -V IN Y L P Y R R O L ID O N E CH2 = C H - N < f ~ ~ | C R O S S L IN K E R S
E T H Y LE N EG L Y C O L c h 2 =c"R
R Dl M E T H A C R Y L A T E
A C R Y L IC S CHp = C ^ D ER IV A T IV E S <?0 R
* x c o 2 r
0
0
- 0
£
-
(R = -H ,- C H 3 ) ( c h 2 c h 2 O+j-CO
( r '= — c h 3 , — c 4 h 9 )
M E T H Y L E N E -B IS - C H 2 = CH CH = C H 2
A C R Y L A M ID E
CO CO
NH NH
^CHg/
properties of a hydrogel, it can be grafted or surface coated onto a stronger support. The
most efficient technique for preparing grafted hydrogels involves generating free radicals
on the support surface and then polymerizing monomers directly onto that surface. The
result is a chain of monomers covalently bonded to the support. Figure 1 shows the variety
of techniques used to generate radicals on surfaces. Grafting of hydrogels to a wide variety
of polymeric supports has been investigated.1-6 12
Surface grafting of hydrogels with ionizing radiation can be performed by three techniques.
The mutual or direct irradiation technique involves exposure of the surface to radiation in
the presence of monomers (see Chapter 1). This method of grafting must be carried out in
the absence of oxygen which would inhibit the polymerization reaction. Pre-irradiation
grafting is done by first irradiating the polymer in the presence of oxygen in which case
peroxides are formed on the surface, or in vacuo, in which no peroxides are formed but
surface and trapped radicals are generated. Once a polymer contains peroxides, it can be
exposed to polar monomers and then heated and/or contacted with a reducing agent (redox
couple) which causes polymerization to occur. A surface containing trapped radicals will
polymerize vinyl monomers without an initiator present. In all cases, the polymerization
reactions must be carried out in the absence of oxygen or other free radical traps.
The type of radiation source used can influence the character of the grafted hydrogel.
Cobalt 60 produces y-rays which are deeply penetrating and the subsequent hydrogel graft
may be located both on the surface as well as within the matrix of the substrate where the
monomer has penetrated. Electron beam accelerator initiated reactions form grafts primarily
on the surface since these forms of irradiation are both higher in intensity and less penetrating.
Radiation dose and dose rate also have an influence on the nature of the resulting graft. A
higher dose will cause more polymerization to occur thus increasing the amount of graft on
the surface. A larger dose rate will produce more of a surface graft while a smaller dose
Methods of surface activation
Peroxide formation
CH, CH, CH-j

AAAAA „ C CH 3CH 3 CH 3 CH 3
I I I Oxidizing__ Fe +3
kOH * Fe
/ / } / / / Conditions //////'
Ceric ions
OH OH OH Ce 0 OH OH
I I I 1 I I
CH? CH? CHp C H ? CHp CHp
I I I +H
/ / / / / / /
"Active Voporllor radical transfer

CH, CH, CH, CHp CH, CH 3
I I I I I I RH
) / / / / / / Plasma discharge / / / / / / / '
atoms or chemical
catalyst radicals
Ionizing radiation
CH* CH3 c h 3 C H p CH 3 CH 3
wwv-
1 lonizing radiation
U. V.
c h 3 c h 3 c h 3 (PS)
1 , JL U.V.+ J v J ' / i ' Z + (PSlH
photosensitizer
FIGURE 1. Methods of surface activation for grafting of hydrogels.
rate will cause the graft to be distributed both on and within the substrate. Temperature,
solvent composition, monomer, concentration, pH, and the presence of various metal ions
are additional factors that must be considered when grafting hydrogels with ionizing radiation.
One of the earliest applications of the graft polymerization technique to the preparation
of hydrogel coated materials for biomedical applications was reported in 1964 by Yasuda
and Refojo.11 They grafted /V-vinyl-2-pyrrolidone (NVP) to silicone rubber in an effort to
increase the hydrophilicity of the rubber surface. A list of other papers describing grafted
hydrogels designed for biomedical applications can be found in the work by Ratner and
Hoffman.1
We have been studying the grafting of methacrylic acid (MAAc) onto various polymer
substrates using both the mutual and pre-irradiation techniques. Figures 2 and 3 are from
mutual irradiation systems. Figure 2 shows how varying the MAAc concentration in an
ethanol/water solvent can influence the weight percent graft on polypropylene (PP) film.
The dose used was 0.25 Mrad from a Cobalt 60 source. As the MAAc concentration increases,
more monomer is available to diffuse into and graft to the film and an increase in the weight
percent of the graft is observed.
Figure 3 shows how the amount of ethanol in the solvent affects the percent graft in the
same system described above while holding the monomer concentration constant at 5%. The
maximum graft occurs at an ethanol concentration of about 75%. Figure 4 is a pre-irradiation
system in which the ethanol concentration in the solvent was varied. The monomer is again
MAAc, while the substrate is a PP-polyethylene (PE) copolymer. The radiation dose was
2.60 Mrads with a 10% MAAc concentration. In this case, the maximum graft occurs at an
ethanol concentration of about 10%. The different shapes of these two curves show us that
MAAC CONC. VS PERCENT GRAFT BY WEIGHT

ONTO POROUS PP FILM
5I!
!0
X
~
.00 X
Jll
:1;
X
X
G 10
A X
A X
F
T
~
iE
X )(
iE
IC
a • lO u U! l3 l. lll
MAAC CONC. IN SOLVENT (VOL %)
FIGURE 2. Mutual irradiation grafting of MAAc onto PP film in a 90% ethanol solvent.
PERCENT GRAFT vs. EtOH CONCENTRATION IN SOLVENT
J
E
A
ll
c ll
E l•
N
T l
G lO
A •
A
F
T •
2
0 I I I I I I I I I I I I +-I I I I I
0 lll ~ ~ 10 Jll .oo ~ !0 !Ill ~ ~ ro ~ 10 ~ ~ ~ ~
Et OH CONC. (Vo 1 %)
FIGURE 3. Mutual irradiation grafting of MAAc to PP with a monomer con-

centration of 5%.
the mutual irradiation and pre-irradiation methods of grafting proceed by different mecha-
nisms. There are many other substrates that can be used in the grafting of hydrogels for
biomedical and industrial applications including nylon, cellulose, and polyurethanes to men-
tion a few.
\ Craft
300r-~,--r-r~~r-r-,--r-r~-,.-.-.--r-,
/(y- 't.>
\
\
\
\ 0
\
\
0 \
\
\
\
\
\
\
\ .......
00~~~~--~~~~--~~~~--~~~~~
10 15 20 25 30 35 40 45 50 55 60 65 70 75 80
Ethanol \
FIGURE 4. Pre-irradiation grafting of MAAc to a PP/PE copolymer film with a

monomer concentration of l 0%.
Radio frequency (RF)-plasma deposited surface grafts occur when the monomer molecules
are dissociated in a plasma environment to a variety of charged and free radical species.
These species will graft to any substrate that is placed in the RF-plasma reactor. The nature
of the graft is influenced by several factors including the monomer concentration, monomer
type, the reaction time, the presence of other chemicals, and the energy level of the plasma.
C. Conversion of Polymers
Synthetic hydrogels may also be prepared from the conversion of existing polymers.
Natural polymers such as proteins can be used to form hydrogels but will not be discussed
in this review. Figure 5 shows some converted polymers used as hydrogels. Sefton and
Merril 13 subjected triblock copolymer films of styrene-butadiene-styrene to surface hy-
droxylation. This treatment resulted in a material composed of a soft hydrogel layer bound
to a tough elastomeric substrate. Kudela et al. 14 hydrolyzed polyacrylonitrile gels with
hydrochloric acid to produce hydrogel membranes containing amide and carboxyl groups.
Graham et al. 15 cross-linked PEO with 4,4'-diphenylmethane diisocyanate to make a po-
lyurethane hydrogel network. Aladesulu and Graham 16 used this network to prepare interstitial
hydrogels with poly(methylacrylate) and poly(MAAc).
Ill. SURFACE CHARACTERIZATION TECHNIQUES
Once a hydrogel has been made, it is important to characterize what type and how many
functional groups are present on its surface (see Chapter 4). This is particularly important
CONVERTED POLYMERS USED AS HYDROGELS
POLYELECTROLYTE
COMPLEX
t CH2- rH +---f CH 2 - <fH~CH 2 - rH+-

CN y=O C02 H
NH2
+CH2- C H + - MIXTURE OF 2
~ POLYMERS
o~o
NATURAL t H N - C H J j - - RECONSTITUTED
1 )(
-f HN-7H-~ 3;
Ry Rz
FIGURE 5. Converted polymers used as hydrogels.
if the hydrogel is to be employed for subsequent biomolecule or cellular immobilizations.

Many techniques are available to characterize hydrogel surfaces, including gel water content
determination, electron spectroscopy for chemical analysis (ESCA), Fourier transform in-
frared spectroscopy (FTIR), surface energy analysis with a contact angle goniometer, mi-
croscopy (light, SEM, TEM), and chemical methods including titrations and dyes. A detailed
discussion of surface analysis of hydrogels can be found in Chapter 4 by B. D. Ratner in
this volume.
IV. TYPES OF MOLECULES AND CELLS USED IN HYDROGEL

IMMOBILIZATIONS
There are numerous biomolecules and cell types that have been immobilized on and within
hydrogels. Table 2 lists some biologically active species which may be used along with
some of their potential applications and a general reference list.
Table 2
BIOLOGICALLY ACTIVE SPECIES THAT CAN
BE IMMOBILIZED ON AND WITHIN
HYDROGELS AND THEIR APPLICATIONS
Species Applications Reference
Enzymes Therapeutic agents, biosen- 1, 17, 18

sors, artificial organs, red
blood cell substitutes, thera-
peutic reactors, industrial
reactors,
Antibodies and Immunoassays, therapeutics 19
antigens and diagnostics, biosensors
Antithrombogenic Artificial organs, blood-com- 13, 20, 21
agents patible surfaces
Drugs Drug delivery systems, drug 22, 23
mechanism studies, (receptor
research)
Hormones Biosensors 22
Neurotransmitters Biosensors 22
Cells and organelles Bioreactors, artificial organs, 17, 24, 25

biosensors, ’
Amino acids Peptide synthesis 26
DNA DNA probe assays
RNA Peptide synthesis
V. M ETHODS OF IM M OBILIZATIO N
There are several techniques for the immobilization of biomolecules and cells on and
within hydrogels. They include physical entrapment, electrostatic attraction, physical ad-
sorption with or without cross-linking, and chemical bonding. Combination techniques have
also been used such as entrapment with chemical cross-linking. This section will discuss
each of these methods of immobilization giving specific examples. Physical adsorption and
electrostatic attraction will be included in the same discussion since it is often difficult to
distinguish between the two methods.
A. Entrapment
The entrapment method of immobilization is based on the occlusion of biomolecules or
cells within a constraining hydrogel which has a tight enough structure to prevent the species
from diffusing rapidly into the surrounding medium. Entrapment is not necessarily a per-
manent method of immobilization and slow release of the biomolecules or cells may still
occur. This release may be desired for an entrapped drug-hydrogel system. Entrapment may
be employed with bulk and grafted hydrogels. Bulk hydrogels may be used in the form of
microspheres, fibers, tubes, discs, and membranes. The advantage of entrapment over other
methods of immobilization is that it is a simple process and many types of biomolecules
and cells can be immobilized in this way. Entrapment usually does not cause a change in
molecular structure of the immobilized species, although microenvironmental effects can
influence its bioactivity significantly. High levels of biomolecule activity may be retained
in many systems.
1 . Bulk Polymers
Entrapment may be carried out by the polymerization of the hydrogel in a solution
containing the biomolecules or cells to be immobilized. Both chemical and radiation initiation
of the polymerization reaction can be used. If radiation is the method of choice, it can be
done at a low temperature so the species to be entrapped will retain greater activity.27 When
choosing the monomer type as well as the polymerization method, one may also control the
degree of cross-linking within the hydrogel. Tightly linked gels will immobilize biomolecules
or cells well but may inhibit product and substrate diffusion into and out of the hydrogel
matrix. The monomer type is also important since different functional groups may effect
the microenvironment of the immobilized species leading to an increase or decrease in its
biological activity.28
Numerous enzymes have been entrapped within hydrogels. Several systems are described
in the review by Goldman et al.29 Enzyme entrapment within hydrogels using chemical
initiation of the polymerization reaction will be discussed first. Early studies involved the
polymerization of poly(acrylamide) (PAAm) gel blocks containing the entrapped enzyme
followed by fragmentation.30 31 This procedure results in a wide range of sizes and shapes
of particles which have less mechanical stability than the parent blocks.
An alternative to the fragmentation technique is the emulsion or suspension polymerization
method of entrapment which produces microbeads of well-defined specific size and good
mechanical stability. Nilsson et al.32 described the simultaneous entrapment of trypsin in an
AAm bead polymerization reaction in which ammonium persulfate and TEMED were used
as a catalyst system. The enzymes entrapped in these beads showed a decrease in leakage
when compared to the mechanical fragmentation technique. Bernfeld et al.33 used a similar
method to entrap aldolase in beads polymerized from A,V'-methylenebisacrylamide (BIS).
Other monomers have been investigated for the entrapment of enzymes with the goal of
controlling the mechanical strength and water content the hydrogel. O’Driscoll et al.34
entrapped trypsin in a comonomer gel of 2-hydroxyethylmethacrylate (HEMA) and PNVP
using di(secbutyl)peroxidicarbonate as an initiator.
Epton et al.35 described a method for the entrapment of (3-D-glucosidase in a variety of
hydrogel beads using suspension polymerization. BIS was copolymerized with several dif-
ferent monomers including l-acrylopiperidine-4-spiro-2'-(r,3'-dioxacrylopentane), 1-ac-
ryloy1-4-piperidine, and aeryloylmethoxyamine. Potassium persulfate was used as the chemical
initiator. The enzyme activity in the beads was comparable to the activity of the free enzyme.
The entrapped enzyme also retained appreciable activity when heated, in comparison to the
free enzyme which decreased in activity.
Radiation initiation is an alternative to using chemicals to catalyze the polymerization of
the hydrogel. According to Kawashima and Umeda,36 the radiation induced method of
polymerization has several advantages over chemical initiation. These include:
1. There is no enzyme inactivation due to chemical catalysts.

2. Supercooled solutions can be polymerized thus eliminating the problem of heat of
polymerization.
3. Immobilization can be accomplished quickly.
4. Various spongy and textured structures having a large surface area can be prepared.
Descriptions of methods which have been used to vary the surface character of grafted
hydrogels are described in several papers.37 38
There are also certain undesirable side reactions which can occur with the hydrogel
polymerizations initiated by radiation. These include:
1. Polymer degradation of the substrate and for the graft itself

2. Extensive cross-linking
3. Formation of unwanted chemical species (peroxides, acids)
4. Inactivation of the enzyme (only when enzyme is present during irradiation)
.
,.:
>-
.....
>
,_
u 50
<(
u
.....
<(
::r
>-
N
z
w
0
-150 -100 -so 0 so
IRRADIAT 10 N TEMPERATURE < ·c
FIGURE 6. Effect of irradiation temperature on the enzymatic activity of Strepto-

myces phaerochromogenes. Irradiation dose: 10 Mrad. HEMA concentration: 30%.
(From Kaetsu, I., Biotechnol. Bioeng., 21, 679, 1979. With permission.)
Degradation and cross-linking can be minimized by using low radiation doses. Formation
of most undesired functional groups can be avoided by excluding oxygen and reactive solvents
from the polymerization system. Kumakura, Yoshida, and Kaetsu 39 have employed low
temperatures to protect the enzyme from inactivation during the irradiation. Figure 6 shows
the effect of irradiation temperature on the enzyme activity of Streptomyces phaerochrom-
ogenes entrapped in a HEMA gel.
Kawashima 36 immobilized several enzymes by radiation polymerization of AAm including
glucose oxidase, invertase, o-amino acid oxidase, aminoacylase, and alkaline protease. A
dose of 0. 5 to 0. 7 Mrad was used from a Cobalt 60 source at a temperature of - l5oC.
Maeda et al. 40 ·41 used a similar technique to entrap glucoamylase, invertase, and ~-galac
tosidase in a bulk PAAm gel. The temperature was also - l5°C while the dose was 2 Mrad.
Maeda et al. 42 .4 3 also employed radiation polymerization to entrap these same enzymes in
hydrogels made from several different monomers. The monomers used included, acrylic
acid (AAc), sodium acrylate, BIS, 2-hydroxyethylacrylate (HEA), and NVP. They found
that in the case of HEA, the monomer-enzyme solution gelled at an irradiation dose of less
than l Mrad but it was difficult to eliminate enzyme leakage from the gel. When leakage
was eliminated by an increased irradiation dose, the enzyme activities became very low.
They concluded that this high dose formed polymer chains containing greater amounts of
cross-linking than found in the low dose polymerized gels. These differences in the hydrogel
structures influenced enzyme leakage and activity.
Yoshida et al.44 found that by using a comonomer system of both hydrophobic and
hydrophilic monomers in a radiation polymerized system, the activity of the enzyme glu-
coamylase was enhanced in the more hydrophobic matrix. The hydrophilic matrices showed
more enzyme leakage over time than the hydrophobic matrix. Hydrophilic monomers in-
cluding HEM A, HE A, A Am, an NVP and hydrophobic monomers such as hexanediol
monomethacrylate, diethylene glycol dimethacrylate, and methylmethacrylate were used.
Kumakawa and Kaetsu45 entrapped papain in hydrogels prepared from various hydrophilic
monomers including HEM A, HE A, and polyoxyethylene dimethacrylate. Again, low tem-
perature irradiation was used to initiate the polymerization. The hydrogel made from HE A,
the most hydrophilic monomer used in the experiment, showed the highest enzyme activity.
The results seem to contradict the results discussed above in which the more hydrophobic
gel showed higher enzyme activity. The differences may be attributed to the different enzymes
used. They also found that the enzyme activity reached a maximum at a certain monomer
concentration and then decreased if the concentration was increased further. They attributed
this maximum activity to the formation of an optimal hydrogel pore size at a particular
monomer concentration, which allowed for good product and substrate diffusion. Similar
results were found using a system in which a-glucosidase and glucose oxidase were entrapped
in HEMA gels with different monomer concentrations.27 Kaetsu and Kumakura46 also en-
trapped enzymes within hydrogels that were radiation grafted to a support.
Photoinitiation has also been used as a polymerization initiation method for the entrapment
of enzymes in hydrogels. Tanaka et al.47 entrapped yeast invertase in a polyethylene glycol
(PEG) dimethacrylate matrix grafted to a polyester film. The oligomer has photosensitive
methacryloyl groups at both ends of the PEG chain. Benzoin ethyl ether was added as an
initiator to the monomer enzyme solution. This mixture was layered on the polyester film
and illuminated at a wavelength between 300 to 400 nm for 3 min. The entrapped invertase
was more stable at high temperature than the free enzyme and the immobilized enzyme was
stable for over 30 batch reactions with no loss of activity. Yeast invertase has also been
entrapped in hydrophilic urethane prepolymers by Fukushima et al.48
Entrapment of antibodies and antigens in bulk hydrogels has been done primarily with
PA Am gels which were then used as immunoabsorbents.49’50 Chemical initiation is often
used with this method of entrapment.
As with biomolecule entrapment, A Am has been commonly used as a monomer for the
entrapment of cells and organelles. An organelle is defined as a functionally and structurally
discrete part of a cell which is often bound by a membrane. Examples include the nucleus,
mitochondria, ribosomes, endoplasmic reticulum, and chloroplasts. Organelles often contain
many biomolecules which are organized into supramolecular systems capable of catalyzing
complex reactions. These organelles can be immobilized and act as biocatalysts for reactions
that would otherwise be very difficult to perform. Polymerization can again be initiated
either chemically or by radiation. Eukaryotic cells have been difficult to entrap within a
hydrogel matrix since they may multiply and break the support. Also, there is a problem in
supplying cells deep within the matrix with an adequate amount of oxygen. Prokaryotic cells
and eukaryotic organelles, on the other hand, have been successfully entrapped within
hydrogels using both chemical and radiation initiated polymerizations. Chemically initiated
systems will be discussed first.
Mattiasson et al.51 entrapped Saccharomyces cerevisiae, bakers yeast, in an AAm/BIS
gel using ammonium persulfate and TEMED to initiate the polymerization. This system was
used as a microbe thermistor. Schimizu et al.52 entrapped Brevibacterium ammoniagenes in
a PA Am gel in an attempt to make a coenzyme A synthesizing system. Chibata et al.53
entrapped Escherichia coli in a PA Am gel lattice using potassium persulfate and (3-dime-
Table 3
CELLS ENTRAPPED IN HYDROGELS AND THEIR REACTION
PRODUCTS55
Hydrogel Cells Reaction (substrate/product)
Polyacrylamide Achromobacter aceris NAD-kinase activity

Achromobacter butyri Production of D-glucose 6 -phosphate
Actinomycetes sp. Biodegradation of steroids
Arthrobacter oxydans Synthesis of FAD (1,3-propanediol/lac-
tic acid)
Arthrobacter simplex (Cortisol/prednisolone)
Bacillus sp. Production of bacitracin
Bacillus subtilis Production of a-amylase
Brevibacterium Production of NADP, synthesis of CoA
ammoniagenes
Candida lipolytica Production of citric acid
Candida tropicalis Degradation of phenol
Clostridium butylicum Production of hydrogen
Corynebacterium (D-Glucose/L-glutamine)
glutamicum
Escherichia coli Production of L-tryptophane
Gluconobacter melanogenes (L-Sorbose/2-ketogulonic acid)
+ Pseudomonas sp.
Gluconobacter oxydans (Glycerol/dihydroxyacetone)
Hansenula polymorpha (Methanol/formaldehyde)
Hyphomicrobium sp. Nitrate degradation
Lactobacillus bulgaricus + (Milk/yogurt)
Streptococcus
thermophilus
Lactobacillus casei Malic acid degradation
Lactobacillus delbrueckii (D-Glucose/lactic acid)
Methanogenic population (Waste water/methane)
Penicillium chrysogenum Production of penicillin
Penicillium cyaneofulvum Production of erythorbic acid
Pseudomonas aeruginosa Nitrate degradation
Pseudomonas denitrificans Biological denitrification
Pseudomonas fluorescens Synthesis of pyridoxal 5 '-phosphate
Pseudomonas putida Benzene metabolism
Pseudomonas tesosteroni A'-Dehydrogenation of Reichstein com-
pound S
Saccharomyces cerevisiae (AMP/ATP)
(CMP/CDP)
(D-Glucose/ethanol)
(Grape juice/water)
(Molasses/ethanol)
Production of glutathione
Streptococcus faecalis L-Arginine catabolism
Tetrahymena pyriformis Evidence of survival
Thiobacillus ferrooxydans Ferrous ion oxidation
Copolymer of acrylamide B. ammoniagenes NAD-kinase activity
and acrylate
Polymethacrylamide C. tropicalis Oxidative degradation of phenol
E. coli (Penicillin G/6 -amino penicillanic acid)
Polystyrene C. tropicalis Degradation of phenol
Poly(vinyl alcohol) B. ammoniagenes Production of CoA
2-Hydroxyethyl Streptomyces (D-Glucose/D-fructose)
methacrylate phaerochromogenes
Polyurethane Kloekera yeast
Nocardia rhodocrous Bioconversion of steroids
Table 3 (continued)
CELLS ENTRAPPED IN HYDROGELS AND THEIR REACTION
PRODUCTS55
Hydrogel Cells Reaction (substrate/product)
Maleic polybutadiene A. simplex (Hydrocortisone/prednisolone)

(PBM) N. rhodocrous Bioconversion of steroids
Polyethylene glycol) A. simplex (Hydrocortisone/prednisolone)
N. rhodocrous Bioconversion of steroids
Polypropylene glycol) A. simplex (Hydrocortisone/prednisolone)
N. rhodocrous Bioconversion of steroids
Polycondensation of E. coli (Penicillin G/6 -aminopenicillanic acid)
epoxides
Silica hydrogel S. cerevisiae Evidence of metabolic activity
thylaminopropionitrile as an initiating system. The heat of stability of these cells was higher
than that of free cells. Omatas54 found that as the hydrophobic content of the gel increased,
the activity of entrapped Nocardia rhodocrous cells also increased. Table 3 lists some of
the other prokaryotic cell types that have been entrapped in various hydrogels along with
their reaction products.
Chemical initiation has been used to successfully entrap organelles in hydrogels. Attempts
have also been made to entrap chloroplasts isolated from plant cells in both PAAm and PVA
gels using chemically initiated polymerizations.56 59 Similar work was done with bacterial
chromatophore entrapment in PAAm.60 Successful entrapment of yeast peroxisomes has also
been done in PAAm gels.61
The polymerization of AAm monomers in the presence of viable cells and organelles often
results in considerable losses of biological activity due to the toxicity of the monomers and
the heat evolved during the polymerization. Freeman and Aharonowitz62 describe the en-
trapment of Streptomyces clavuligerus in PAAm using preformed linear water-soluble chains
partially substituted with acyl hydrazide groups. The gel was cross-linked with a dialdehyde
and cell damage caused by the AAm monomer was avoided. The cells were able to synthesize
the antibiotic cephalosporin with yields similar to those of free resting cells.
Much work has been recently done using radiation polymerization at low temperatures as
a method to entrap cells and organelles within hydrogels. Hydrogels are particularly suitable
for this type of immobilization because they remain stable when supercooled and are easily
polymerized at low temperatures. Kumakura et al.39 entrapped Streptomyces phaerochrom-
ogenes in hydrogels of HEM A, irradiating the solution at —24°C, with a dose of less than
0.5 Mrad. They found that the porosity of the polymer matrix increases with increasing
water content in the monomer solution and could thus be varied. The water freezes during
the polymerization reaction and when warmed, it melts, leaving pores behind.
Hayashi et al.63 used an interesting technique to immobilize Brevibacterium ammonia
genes in four different monomer solutions composed of various ratios of AAm, BIS, sodium
acrylate, magnesium acrylate, and potassium acrylate. The cell monomer solution was
injected into precooled petroleum ether. The solution was immediately frozen in a super-
cooled state in the shape of small beads and subjected to a y-radiation dose of 0.5 Mrad.
After irradiation, the petroleum ether was discarded and the beads were gradually thawed
in ice water. Beads of 0.5 to 1.0 mm in diameter were obtained. The immobilized cells
were subjected to a 5-hr reaction and then washed with water. This was repeated up to 20
times without any loss of activity.
Fujimura and Kaetsu64 entrapped Saccharomyces formosensis cells by radiation induced
polymerization of methoxyPEG dimethacrylate or PEG dimethacrylate at temperatures of
0°C and 25°C. Ethanol production by these cells was found to increase threefold over yeast
cells growing in suspension. Kaetsu65 also immobilized E. coli cells in HEM A using a
radiation dose of 1.0 Mrad at a temperature of —78°C. Immobilization was shown to have
no effect on the ability of the cells to bind antibodies when compared to cells cultured in
liquid medium.
Organelles, primarily chloroplasts, have also been entrapped using radiation as the
polymerization initiator.66 68 Monomers and polymers used included HEA, PEG diacrylate,
PEG dimethacrylate, methoxyPEG methacrylate, glycidyl methacrylate, and glycidyl acrylate.
2. Encapsulation
Encapsulation within a hydrogel membrane is another method used to entrap biomolecules
and cells. In these systems, the species are not held within the hydrogel matrix as in bulk
polymer entrapment, but are suspended within small aqueous compartments contained by
the membrane. The membrane prevents the escape of the encapsulated molecules or cells
but is permeable to the smaller molecules of their substrates and products.
Chang69 71 was one of the first to report the use of semipermeable microcapsules for the
entrapment of enzymes. More recently eukaryotic cells have been contained within a mi-
crocapsule called Encapcel® made by Damon Biotech Inc. The capsule is made of a gel
coated with a porous polymer. The diameter of the capsule is critical for high cell growth
rates, since this controls the diffusion of substrates and products to and from the cells.
Encapcel® can also contain entrapped prokaryotic cells, organelles, enzymes, antibodies,
drugs, and other proteins.
Klomp and Ronel72 73 used hydrogels for the encapsulation of pancreatic islet cells in an
attempt to make an artificial pancreas. This system did not employ microcapsules but rather
a porous poly(HEMA) membrane chamber. The membranes were first synthesized using the
redox initiators of ammonium persulfate and sodium metabisulfite with ethylene glycol
dimethacrylate as a cross-linking agent. Cells were placed in membrane chambers which
were fused with heat or cyanoacrylate.
The papers mentioned in this section are by no means an exhaustive list. For further
readings concerning immobilized cells and organelles the reader is referred to the following
reviews.24’25,74
B. Adsorption
Physical adsorption via secondary molecular forces is another way to immobilize a bio-
logical species on a hydrogel. Adsorption may sometimes be less attractive than other methods
when immobilizing molecules because it is reversible. The species may desorb from the
hydrogel especially after the hydrogel is exposed to a protein solution such as blood. In this
case, the proteins may compete for sites on the hydrogel surface, thus causing desorption
of the species of interest.75 Another problem with adsorption immobilization of biomolecules
onto hydrogels is the possibility that a residue involved in the active site of the molecule
may interact with the hydrogel, leading to inactivation of the biomolecule.
The two types of bonds encountered most frequently with adsorption of molecules onto
hydrogels are hydrogen bonds and bonds resulting from charge-charge interaction (ionic and
salt bridge). Hydrophobic interactions are also present in many adsorption systems. Poly-
electrolyte complexes are particularly suited for adsorption immobilization because a net
charge (anionic or cationic) can easily be incorporated in the hydrogel system. This is done
by stoichiometrically adding greater or lesser amounts of one of the two polymeric com-
ponents during formulation. Heparin absorbed onto cationic hydrogels is a classical example
of biomolecule adsorption onto a polyelectrolyte complex. Some specific polyelectrolyte
complexes are noted in the review by Ratner and Hoffman.1 Adsorption of enzymes onto a
variety of surfaces is discussed in detail by Mosbach76 although there is little specific mention
of hydrogels.
Adsorption of viable cells onto hydrogels is the most attractive method for cellular im-
mobilization since cell adhesion is fundamental to traditional eukaryotic cell culture tech-
niques. Proliferation of anchorage-dependent cells can only occur after adsorption to a culture
surface. Once bonded the cells continue to adhere and desorb from a surface during the
normal process of growth and division. The cytoskeletal structure of a eukaryotic cell forms
focal adhesions within the cell membrane which contact the surface upon which the cell is
growing. These structures are dynamic and will change depending upon the type of surface
they contact. Hydrogels may prove to be important in producing new surfaces for cell culture
because the rate of cell growth may be greatly affected by the hydrophilicity, surface
topography, local pH, and net ionic charge of the surface. All of these parameters can be
varied by the molecular engineering of appropriate hydrogels. For example, a hydrogel
which is lightly cross-linked and highly water swollen will have flexible polymeric molecular
segments or loops extending into the surrounding medium. These polymer chains could
interfere with protein adsorption as well as adhesion and spreading of cells.
Fujimura and Kaetsu77 absorbed Saccharomyces formosensis on fragmented hydrogels
prepared from the monomers methoxyPEG methacrylate and HE A. Polymerization was
initiated by gamma irradiation. The yeast cells grew over the hydrogel fragments and even-
tually infiltrated the gel by multiplication. The ethanol productivity of the immobilized cells
was found to be 13 times that of free yeast cells.
Compatibility of hydrogels with eukaryotic cells is an important aspect in the design of
artificial organs. Ultimately one would like to develop a material which is completely
accepted by the tissues in the body. Schnarr et al.78 derivatized PAAm gels with A-acetyl-
glucosamine and observed the adhesion and growth of chicken hepatocytes on the gels.
Gilding et al.79 studied the growth of soft tissue ingrowth into PAAm gels prepared by either
bulk polymerization or grafting to silicone rubber and polyurethane supports. The reactions
were initiated by Cobalt 60 irradiation. The hydrogels were implanted subcutaneously in
sheep and rats in the form of discs. They found that the gels with a water content greater
than 98% had a large amount of tissue ingrowth while no ingrowth was seen in gels with
a water content of less than 90%. Yoshi and Kaetsu80 have also studied the adsorption of
eukaryotic cells onto hydrogels prepared by radiation-induced grafting of various monomers.
The cells cultured were rat glial and pituitary tumor cells and Chang liver cells. The hydrogels
were polymerized from methyl methacrylte, HEM A, glycidyl methacrylate, trimethylolpro-
pane trimethacrylate, and tetramethylolethane tetraacrylate. They found the rate of cell
growth to be higher on hydrophobic than on hydrophilic polymers. An interesting application
of eukaryotic cell adsorption to surface modified PAAm beads is described by Truffa-Bachi
and Wafsy.81 They covalently bound haptens to the beads which were placed in a column.
When a mixture of cells was run through the column, only those that produced antibodies
specific to the hapten absorbed to the beads.
C. Covalent Attachments
There are several advantages of covalently attaching a biological species to a hydrogel
when compared to other methods of immobilization. The species will not desorb during use
or when assayed for activity. The composition of the medium to which the species is exposed
may also be changed without desorption. The species will retain its activity provided the
covalent bonds are formed with nonactive sites on the molecule. Disadvantages may include
the inability to recover the species once it is immobilized and the possibility of inactivating
the moleclule during the immobilization process.
There are few applications involving covalent immobilization of viable cells to hydrogels,
particularly for eukaryotic cells. Normal cell growth involves a continual synthesis and
turnover of surface proteins. Covalent attachment of these proteins to a hydrogel would
disrupt this process. If a covalently immobilized cell did divide, the daughter cell would be
free to leave the surface. Another problem is that the coupling agents employed can often
be toxic to the cell.82 Jirku et al.83-84 have done some work with yeast cells. They covalently
attached Zygosaccharomyces lactis and Saccharomyces cerevisiae to hydroxyalkyl meth-
acrylate gels using glutaraldehyde as a coupling agent.
Three methods may be employed to covalently bond a biomolecule to a hydrogel. The
first involves pre-activation of the hydrogel surface for subsequent reaction with groups on
the biomolecules to be immobilized. Small molecules or “ arms” may be attached to the
hydrogel surface before immobilizing the biomolecule. By chosing different size arms, the
distance between the covalently bonded biomolecule and the hydrogel can be varied. The
second method is to use a coupling reagent which links the biomolecule to the gel. In this
case, the molecule to be immobilized and the coupling agent are exposed to the hydrogel
concurrently. The third technique is to preactivate the biomolecule for coupling to the gel.
Figure 7 illustrates three methods using an enzyme as a model.
To covalently bind a biologically active protein to an insoluble hydrogel, functional groups
on the protein which are nonessential for its biological activity must be used. Coupling
conditions are often modified to avoid involvement of the active site residues. Table 4
includes a list of some of the most common functional groups on proteins that are suitable
for covalent binding to hydrogels. There are a large variety of coupling reactions which can
be employed to covalently bond molecules to hydrogels. The one of choice for a particular
system will depend upon both the functional group of the hydrogel as well as the groups
on the protein. The various techniques used will not all be discussed in detail in this chapter
but there are several good reviews which have been published on this material.17’76,85 86
PA Am has been extensively used for covalent immobilization of biomolecules. Com-
mercial PAAm beads are available from several companies. PA Am derivatives bearing useful
functional groups can also be purchased under the tradename of Enzacryl®. A good review
of the coupling reactions used to attach proteins to PAAm is given by Inman.87 Barker et
al.88-89 describe a technique to covalently binding the enzymes alpha and beta amylase to
different Enzacryl® derivates. Enzacryl® AH contains acid hydrazide groups while Enzacryl®
AA contains arylamino groups. Samples of Enzacryl® AA were activated with either thio-
phosgene or by dizaotization while samples of Enzacryl® AH were activated with nitrous
acid. Other papers describe techniques for the covalent immobilization of enzymes to PAAm
gels using ethylene diamine90 and glutaraldehyde91 as coupling agents.
Many other monomers have been employed for the covalent immobilization of enzymes.
They have been prepared in the form of microbeads, bulk polymers, and graft copolymers
on numerous substrates. Turkova92 has written a good review of immobilization of enzymes
on hydroxyalkyl methacrylate gels. Manecke and Vogt93 immobilized trypsin, papain, and
glucose oxidase of a variety of copolymer gels. Monomers used include NVP and A Ac.
Hoffman et al.94 have covalently bound heparin, streptokinase, and albumin to graft co-
polymers of NVP/HEMA on silicone rubber films. They employed cyanogen bromide to
couple the biomolecules to the hydroxyl groups in the graft. When an arm of €-amino caprioc
acid was incorporated, the activity of the biomolecules increased (Figure 8). There are many
additional papers describing radiation grafting of various hydrogels to polymers with the
subsequent covalent immobilization of biomolecules.95102 In our laboratory we have recently
been working on the immobilization of L-asparaginase onto PMAAc which has been radiation
grafted onto porous PP films or hollow fibers. The enzyme is covalently bound to the
carboxyl groups by activating these groups with carbodimide and A-hydroxy succinimide
(Figure 9).
Campbell et al.103 covalently immobilized antigens onto cellulose supports in 1951. Since
this time many antibodies and antigens have been covalently linked to a variety of other
hydrogels.104 Molday, Dreyer, Rembaum, and Yen105 107 have covalently bound antibodies
to radiation synthesized hydrogel microsphere latices for use as cell markers and for other
immunological separations and diagnoses.
l. PRE-ACTIVATION COUPLING TO THE HYDROGEL
I I I
I
0
I I
I -X + y--+ I -X-Y I -X~+ y
I I I
I activated I I
coupling agent
2. DIRECT COUPLING TO HYDROGEL
I
I
j -x-@ + Y-Y
I -X + Y-Y + 8--+ I or
I
I coupling j -X-Y-Y-@
agent
3. PRE-ACTIVATION OF ENZYME
I
I
I
I
-x-Y-@
a) §+Y--+ + I -X --+ I or
I
I
I
I -x-@ Y +
~~
8
R R
b) + y + ~C=CH 2 --t :;-c=CH 2 ___,. ~ ~
X monomer .---<;.NX
~
~bulk copolymer
/ hydrogel
CH2=C,
surface grafted
j~
j N
hydrogel I
FIGURE 7. Methods of covalently binding an enzyme to a hydrogel.
As mentioned earlier, one of the ways to covalently immobilize a biomolecule to a substrate

is to first activate the biomolecule. This method has been used to conjugate an enzyme to
a monomer and then copolymerize the conjugate with free monomer. The result is a hydro-
gel in which the biomolecules are both entrapped and covalently bonded. D' Angiuro
et al. 108 - 110 employed the redox system, Fe 2 + - H20 2 , or UV irradiation to initiate the po-
lymerization of these conjugates. Other papers have been published in which similar methods
of enzyme immobilization have been employed. 111 - 114
D. Combination Techniques
A combination of both physical entrapment and covalent bonding of enzymes to hydrogels
has been investigated by Mosbach. 115 - 116 The enzymes citrate synthetase, hexokinase, trypsin,
and glucose-6-phosphate dehydrogenase were added to solutions containing the monomers
AAm and HEMA along with carbodiimide or cyanogen bromide. These mixtures were
polymerized with ammonium persulfate and TEMED. The result is an enzyme preparation
that is both entrapped and covalently bonded to the polymer matrix. Pollak et al. 117 im-
mobilized hexokinase in a P AAm gel using both covalent bonding and entrapment. The
reaction was carried out in the presence of species which occupied the active sites of the
enzyme thus preventing inactivation of the enzyme during covalent bonding to the matrix.
Table 4
FUNCTIONAL GROUPS ON PROTEINS WHICH ARE SUITABLE FOR
COVALENT BINDING TO HYDROGELS
------ C-AMINO OF LYSINE (L y s ) AND N-TERMINUS AMINO GROUP
— SH SULFHYDRYL OF CYSTEINE (C y s )
— COOH CARBOXYL OF ASPARTATE (A sp ) AND GLUTAMATE (G lu ) AND

C-TERMINUS CARBOXYL GROUP
-S -S — DISULFIDE OF CYSTINE
CH3- S - THIOETHER OF METHIONINE (M et)
— CHÔH HYDROXYL OF SERINE (S er ) AND THREONINE (T h r)
H ^N H
— N -C GUANIDINO OF ARGININE (A r g )
PHENOLIC OF TYROSINE (T y r )
'l— i IMIDAZOLE OF H IST ID IN E (H is)
INDOLE OF TRYPTOPHAN (T r p )
N
VI. MICROENVIRONMENTAL EFFECTS
Whenever a biomolecule or cell is immobilized on or within a hydrogel support, it is

placed in a different environment than it encounters in its natural state. This microenvironment
can have profound effects upon the immobilized species and may result in an increase,
decrease, or complete loss of activity. Katchalski28 was one of the first investigators to point
out the importance that the microenvironment has on the activity of immobilized biomole-
cules. Both the structure and chemical composition of a hydrogel will influence the mi-
croenvironment of an immobilized species. Large pore diameters will allow good diffusion
of product and substrate in the case of an immobilized enzyme or organelle. Smaller pore
diameters can cause decreases in this diffusion thus decreasing the activity of the enzyme.
Fawcett and Morris118 have studied how changes in the pore size of PA Am gels effect the
gels’ permeability to various proteins.
y _
-OH y -o.
y \ y
' c = NH y ;c = N-Biomolecule
BrCN B1 0 -NH9 y
/ — X_
-OH y -o' A
Hydrogel
FIGURE 8 . Covalent binding of a biomolecule to a grafted hydrogel with and without the use of an arm . 94
Different hydrogels, particularly polyelectrolytes, can carry different net charges. Local
changes in pH may also be found in the hydrogel microenvironment of an immobilized
species. These factors will effect the biological activity of these species. As an example of
these effects, we found in our laboratory that the extent of grafting of MAAc onto PP film
significantly influences the activity of the subsequently immobilized L-asparaginase mole-
cule. Enzyme activity increases dramatically up to a 10% graft after which it decreases
(Figure 10). Upon analyzing the total amount of bound enzyme on each of these films, we
found that the amount of covalently bound enzyme increases almost linearly with the increase
in graft level. Thus, the films with high graft levels contained more L-asparaginase than
those with lower graft levels, but the enzyme was somehow being inactivated. There are
several possible explanations for this result. At higher graft levels, the enzyme could be
attached by multiple covalent bonds, thus increasing the chance that its catalytic site is
inactivated.119 Another explanation is that the L-asparaginase is bound within a tighter
hydrogel matrix at higher graft levels which results in poorer diffusion of substrate and
product. Manecke et al.120 describe the effects that several different hydrogels have on
covalently bound enzymes. They state that if an enzyme is attached by many covalent bonds,
a matrix-imposed rigidity may occur and the enzyme may be unable to undergo the con-
formational changes necessary for catalysis to occur. There are several other references
which discuss the importance of the microenvironment on an immobilized biomole-
cule.121124
VII. APPLICATION S
The potential applications for immobilized biomolecules and cells on and within hydrogels
are numerous and range from large scale industrial catalytic processes to the development
of new cancer therapeutic agents.18 25125126
ENZYME IMMOBILIZATION
/ 0
/ II
/ -C-NR1
N-ACYLISOUREA / I
/ c«o
/ I
NR
2
/ N / 0 N
/ II / II II
/ -COOH C —► / -c -o -c
/ II -I- M* / 1
/ N / NH
/ 1 / 1
*2 *2
POLYPROPYLENE GRAFTED
WITH METHACRYLIC ACID CARBODIIMIDE O-ACYLISOUREA
N-HYDROXY + HO-N
SUCCINIMIDE
/ 0
/ II
/ -C-NH-ENZYME
/
/
/
FIGURE 9. Covalent enzyme immobilization onto PP film grafted with MAAc using carbo-
diimide and Af-hydroxy succinimide as coupling agents.
A. Analytical and Diagnostic Applications

One of the advantages of using immobilized biomolecules, particularly enzymes and
antibodies, for chemical analysis and preparation is their high substrate specificity. Other
benefits include increased atability (storage, thermal, and operational) and reusability of the
immobilized species.
1. Column Techniques
Biomolecules and cells can be immobilized on and within hydrogel beads which are then
packed in a column.31 A column of this type can be used as an analytical tool which separates
certain molecular species from a heterogenous sample. In an early study, Hicks and Updike127
investigated the characteristics of an immobilized enzyme packed column. They entrapped
glucose oxidase and lactose dehydrogenase within PAAm gels. The gels were fragmented
and packed into columns. After passing substrates of either glucose or lactic acid through
each column the products of the enzymatic reactions were mixed with a color reagent. These
reaction products could then be quantified using a photometric cell.
L-ASPARAGINASE ACTIVITY vs. PERCENT GRAFT

OF MAAc ON PP FILM
•,
A
A
R
A
•
I
N
A
••
A
c
T
I
V
I
T
,
y
I!
,
R
I
L
•
ri
u PERCENT GRAFT BY WEIGHT OF MAAC ON FILM
FIGURE 10. The effect of percent graft on the activity of covalently bound L-aspara-
ginase. One IU (International Unit) is the amount of enzyme that degrades one micromole
of L-asparagine per min.
2. Biosensors
Biosensors are becoming an increasingly important method for chemical analysis. 128 Clark
and Lyons 129 were one of the first to apply the enzyme electrode concept as a basis for
biochemical-specific sensors. A biosensor consists of an electrochemical sensor that contains
immobilized molecules or cells on its surface. These biospecies are usually contained on or
within a hydrogel membrane which is attached to the physical component of the biosensor.
This membrane serves an important function since it interfaces the biomolecular or cellular
system to the physical component. 130 This physical component converts a chemical signal
into an electrical one. The component may be an optical fiber, 131 a piezo-electric crystal,
or more commonly, an electrode system (Figure 11).
With an enzyme biosensor, the substrate is converted to a product which changes the net
charge on the surface of the sensor. This is read as a change in surface potential on the
electrode. Microbe thermistors may contain immobilized microorganisms placed in proximity
to a thermistor which measures the metabolic heat evolved. Mattiasson et al. 132 entrapped
yeast cells in a PAAm gel attached to a thermistor and were able to measure glucose
concentrations in various solutions.
The potential for biosensors in medical and industrial settings is great. Biosensors could
improve the ability of the physician to rapidly diagnose particular diseases in the office
rather than using a clinical laboratory which often has a slow response time. Routine mon-
itoring may also be done with biosensors, especially with critical care pateints. Sensors
BIOSENSOR COMPONENTS
~--SIGNAL
~--- (ANALYTE)
SIGNAL TRANSMISSION SEMI-PERMEABLE A) USUALLY IMMOBILIZED

MEMBRANE BARRIER(S) ON OR WITHIN
POLYMER COMPONENT
CONDUCTOR
HYDROPHOBIC POLYMER B) BINDS TO OR REACTS
SEMI-CONDUCTOR WITH ANALYTE
HYDROGEL POLYMER
FIBER-OPTIC
ANION OR CATION ANTIBODY
PIEZOELECTRIC PERM-SELECTIVE POLYMER ANTIGEN
ENZYME
NUCLEOTIDE
FIGURE 11. The three major components of a biosensor.
small enough to fit in a catheter could be used to continually monitor blood electrolyte and
oxygen levels. Kumakura et al. 133 describe a biosensor which measures the potassium ion
concentration in blood serum. The sensor membrane is made of HEMA in which valinomycin
is entrapped using radiation polymerization. O'Driscoll et al. 134 monitored blood glucose
levels using gel entrapped glucose oxidase. In industry, biosensors can be valuable in
monitoring fermentation processes in which the concentrations of nutrients, oxygen, and
metabolites are important.
3. Immunoassays
Immunoassays are used to quantify amounts of antigen or antibody present in a solution
and may be used in a wide variety of diagnostic tests. Included are assays for drug monitoring,
viral diseases, sexually transmitted diseases, respiratory diseases, tissue typing, and blood
typing. Most of these assays depend upon one of three types of signals. These include
radioactivity (radioimmunoassay, RIA), fluorescence (fluorescent immunoassay, FIA), and
visible color change (enzyme-linked immunoassay, ELISA). 135 The following section will
discuss solid phase immunoassays in which the antigen or antibody is bound to a hydrogel.
Bemfeld and Won49 were one of the first groups to immobilize antigens in a PAAm gel.
They were able to quantitively assay a specific antibody in a blood serum preparation. Carrel
and Barandun50 investigated the influence the cross-linking of the PAAm gel had on the
subsequent antibody antigen binding reaction. Kardana et al. 136 used PAAm fragments
obtained from an Enzacryl® gel as a solid support for an immunoadsorbent. There are many
other papers discussing the covalent attachment of antigens to PAAm beads for the preparation
of immunoadsorbent. 137- 141 Recently, other hydrogels including HEMA, HEA, octyl meth-
acrylate, hexyl methacrylate, and hydroxypropyl methacrylate have been used to immobilize
antigens on and with beads, discs, and membranes for immunoassays. 142- 146 Radiation was
used to initiate hydrogel polymerization in these experiments.
BLOOD - PLASMA
WHOLE - -.... OUT
BLOOD
IN
1 PLASMA
OUT
FIGURE 12. A porous hollow fiber blood separation device.
Antibodies and antigens have also be used as components in a biosensor. These differ
from enzyme biosensors in that the antibody-antigen binding reaction is usually irreversible.
The identification of cell surface receptors has been made possible through the use of
immobilized antibodies, enzymes, drugs, and amino acids onto hydrogel beads. 107 • 147 . 153
B. Therapeutic Bioreactors and Drug Delivery Systems

Porous hollow fiber blood plasma filters lend themselves well to enzyme immobiliza-
tion. 154155 Figure 12 is a schematic of such a device and Figure 13 is a scanning electron
micrograph of a porous PP hollow fiber in this type of filter. One can see from the micrograph
that there is a large surface area available for biomolecule immobilization.
In our laboratory, we have immobilized L-asparaginase onto a porous hollow fiber plas-
mapheresis device to make a bioreactor for the treatment of acute lymphocytic leukemia.
The PP fibers are radiation grafted with MAAc to which L-asparaginase is covalently bound.
Blood is pumped into the reactor where the plasma flows through the pores of the fibers
while the cells do not. L-asparagine is converted to L-aspartic acid and ammonia by the
enzyme. The L-asparagine free plasma is then returned to the patient. The leukemia cells
which lack L-asparagine synthetase activity, are unable to produce their own L-asparagine
and die while the normal cells are unaffected. This reactor has been tested in sheep and the
activity of the immobilized L-asparaginase is retained even after exposure to the harsh blood
environment (Figure 14). Other blood detoxification devices have been developed in which
antibodies are immobilized on hydrogel beads which are packed in a column. 156 When blood
or plasma is passed through the column the immobilized antibodies will selectively bind
and remove the unwanted antigens from the blood.
A wide variety of drug delivery systems have been developed for achieving a regulated
or sustained release of therapeutic agents over a sustained and predetermined period of time.
Hydrogels have been used for sustained release of a drug by acting as either a diffusion
controlling barrier membrane in resevoir devices or as a matrix for containment and release
of an active agent in monolithic devices. The hydrogel may be designed to degrade in the
body or to resist attack from the body's defense system and be removed at a later time.
Several reviews have been written which discuss the use of hydrogels in controlled drug
delivery systems. 157 . 169
Resevoir systems may be in the form of membranes, capsules, or microcapsules. Matrix
devices may be made from discs, 169 beads, or pellets. 170 Many papers have been written
describing the controlled release of drugs from microbeads. 171 · 173
Some hydrogel membranes can be made to swell in the body due to an increase or decrease
in pH within the gel. The swelling opens pores in the membrane thus allowing the drug to
diffuse into the blood. Insulin entrapped in a resevoir device is the primary drug that has
been used for this type application. 174. 177 Horbett and Ratner 178 . 179 have developed an insulin
delivery system which works on the principles described above. Glucose diffuses into a
ASPARAGINASE ACTIVITY OF
REACTOR 15
A
s
p
A
R
A
G
I
N
A
s ---.-DAY 0
E
--e- DAY 5
_.._ DAY 41
A
c
T
I
V
I
T
y
I B 10 1!1 10 2!1 Ill !Ill .., 411
u
TIME (MIN)
FIGURE I4. L-Asparaginase activity of a bioreactor before (Day 0 and Day 5) and
after (Day 4 I) being tested in a sheep for I hr.
synthetic hydrogel network made from HEMA, N,N' -dimethylaminoethyl methacrylate, and
cross-linker. The glucose reacts with an enzyme, glucose oxidase, that is entrapped within
the hydrogel, and is converted to gluconic acid. The acid can protenate basic functional
groups in the membrane. As a consequence of this reaction, charged sites are created in the
hydrogel. Electrostatic repulsion will occur which causes an expansion of the membrane.
This expansion increases the permeability of the membrane, thus allowing insulin to diffuse
through the pores. In the absence of glucose, the hydrogel membrane returns to the unswollen,
nonpermeable state.
Specific-site (targeted) drug delivery for the treatment of various cancers has recently
stimulated much interest. 180- 182 Most anticancer drugs work by attacking rapidly dividing
cells. They lack, however, selectivity for cancer cells and as a result are toxic to many other
normal cell types. Through hybridoma technology, the production of highly purified mon-
oclonal antibodies specific to antigens found primarily on cancer cells may eventually be
possible. These antibodies can be bound to hydrogel microcarriers which contain a cytotoxic
agent. When introduced into the body, these antibody-coated carriers should bind only to
the cancer cells bearing the appropriate antigen. The carriers would then release the drug
and kill only the cancer cells. (Figure 15).
C. Industrial Applications
Immobilized biomolecules and cells on and within hydrogels can be used in synthetic
reactions for production of bioenergy (ethanol, hydrogen, and methane), organic acids,
amino acids, and drugs. Much work has been devoted to immobilized enzyme reactors which
have several advantages over conventional reactors. As catalysts, enzymes allow industrial
processes to be performed faster, under milder conditions, and with fewer side reactions
than conventional chemical processes. Furthermore, unlike most organic reactions, enzymic
reactions are run primarily in water.
There are four types of bioreactors: ( 1) packed bed, (2) fluidized bed, (3) continuous flow
stirred tank, and (4) continuous flow hollow fiber reactors. Detailed descriptions with dia-
grams of each reactor can be found in several references. 17 •25 • 183 - 187 The first three reactors
MICROSPHERE CONTAINING
DRUG AND COATED WITH
~
~
ANTIBODY
MICROSPHERES BIND TO TUMOR CELLS

AND RELEASE DRUG ON CELL SURFACE
OR WITHIN CELL AFTER ENDOCYTOSIS
OF PARTICLE
FIGURE 15. Targeted drug delivery employing a tumor-cell-specific antibody-coated microsphere which contains
a toxic drug.
may use hydrogel fragments or microspheres onto or within which biomolecules or cells
have been immobilized. In the continuous flow hollow fiber reactor, the species of interest
is immobilized onto the wall of the hollow fiber. Reviews describing some of the applications
of immobilized enzymes can be found in References 188 and 189.
Single enzymes can be immobilized to produce one product from a substrate. Whole cells
and organelles, on the other hand, have several advantages over individual enzyme systems
since their enzymes are often organized into the requisite metabolic pathways required for
a multienzyme reaction. A table with references listing present industrial applications of
immobilized microbial cells in PAAm gels can be found in the work by Linko and Linko. 190
Other papers describe the applications of immobilized microorganisms in PAAm beads which
are used in a packed bed reactor. 31 • 191
Cell culture is important for the production of many biological materials including vac-
cines, enzymes, hormones, antibodies, interferons, and nucleic acids. Hydrogel microcarriers
in suspension or in packed beds are becoming a new method for cell culture. The cells
attach to and grow as monolayers on the particle surface or may be entrapped within a bulk
hydrogel or a hydrogel membrane. The large surface to volume ratio and the use of large
culture volumes (ranging from a few milliliters to several hundred liters) results in increased
production. 24
VIII. CONCLUSION
The physical and chemical properties of a hydrogel will depend upon the monomers or
polymers from which it is made. There is a large variety of monomers and polymers suitable
for making a wide range of hydrogels. Many different biomolecules, cells, and organelles
may be immobilized on and within or otherwise combined with hydrogels. There exists a
variety of ways to immobilize a species of interest and the properties of the system will
depend upon the method of choice. Because of this great diversity, there are many medical
and industrial applications of hydrogel-immobilized biomolecules and cells.
ACK N O W LED G M EN T
The authors would like to thank Cheryl Kruesel for typing the final manuscript and Jan
Gombotz for typing the drafts.
REFERENCES
1. Ratner, B. D. and Hoffman, A. S., Synthetic hyrdogels for biomedical applications in Hydrogels for
Medical and Related Applications, Vol. 31, ACS Symposium Series, Andrade, J. D., Ed., American
Chemical Society, Washington, D. C. 1, 1976.
2. Reed, A. M., Gilding, D. K., and Wilson J., Biodegradable elastomeric biomaterials — polyethylene
oxide/polyethylene terepthalate copolymers, Trans. Am. Soc. Artif. Intern. Organs, 23, 109, 1977.
3. Gilding, D. K. and Reed, A. M., Biodegradable polymers for use in surgery-poly(ethylene oxide)
polyethylene terepthalate) (PEO/PET) copolymers. I, Polymer, 20, 1454, 1979.
4. Gilding, D. K. and Reed, A. M., Biodegradable polymers for use in surgery-polyglycolic/poly(lactic acid)
homo- and copolymers. I, Polymer, 20, 1459, 1979.
5. Reed, A. M. and Gilding, D. K. Biodegradable polymers for use in surgery-poly(glycolic/poly(lactic acid)
homo- and copolymers. II. in vitro degradation, Polymer, 22, 494, 1981.
6. Chapiro, A., High polymers in Radiation Chemistry of Polymeric Systems, Vol. 15, Interscience, New
York, 1962.
7. Kaetsu, J., Okubo, H., and Hayashi, K., Radiation polymerization of glass forming systems. I. Effect
of temperature on the initial polymerization rate, J. Polym. Sci. Part A, 10, 2203, 1972.
8. Haruvy, Y. and Rajbenbach, A., Nylon 6 water permeable membranes prepared by electron bean radiation-
induced graft copolymerization, J. Appl. Polym. Sci., 26, 3065, 1981.
9. Haruvy, Y. and Rajbenbach, Grafting of acrylamide to Nylon-6 by electron beam pre-irradiation technique.
II. Kinetic aspects and film permeability, J. Appl. Polym. Sci., 27, 2711, 1982.
10. Lee, H. B., Shim, H. S., and Andrade, J. D., Radiation grafting of synthetic hydrogels to inert polymer
surfaces. I. Hydroxyethyl methacrylate, Polym. Prepr. Am. Chem. Soc. Div. Polym. Chem., 13, 729, 1972.
11. Yasuda, H. and Refojo, M., Graft copolymerization of vinylpyrolidone onto polydimethylsiloxane, J.
Polym. Sci., 2, 5093, 1964.
12. Stannett, V., Grafting, Radiat. Phys. Chem., 18, 215, 1981.
13. Sefton, M. V., and Merril, E., Surface hydroxylation of styrene-butadiene-styrene block copolymers for
biomaterials, J. Biomed. Mater. Res., 10, 33, 1976.
14. Kudela, V., Stoy, A., and Urbansova, R., Electrochemical properties of hydrophilic weak-acid membranes
from hydrolysed polyacrylonitrile. I, Eur. Polym. J., 10, 905, 1974.
15. Graham, N. B., Nwachuku, N. E., and Walsh, D. J., Interaction of poly(ethylene oxide) with solvents.
16. Aladesulu, I., Graham, N. B., and Richards, R. W., Interstitial polymers based on a polyurethane
network. I. Melting points and fractional crystallinites, Polymer, 24, 279, 1984.
17. Zaborsky, L., Immobilized Enzymes, CRC Press, Boca Raton, Fla., 1973.
18. Wingard, L., Katchalski, E., and Goldstein, L., Applied Biochemistry and Bioengineering, in Analytical
Applications of Immobilized Enzymes and Cells, Vol. 3, Academic Press, New York, 1981.
19. Line, W. F. and Becker, M. J., Solid phase immunoassay, Immobilized Enzymes, Antigens, Antibodies,
and Peptides, Weetall, H. H., Ed., Marcel Dekker, New York, 1975, 497.
20. Goosen, M. F. A., Sefton, M. V., and Hatton, M. W. C., Inactivation of thrombin by antithrombin III
on a heparinized biomaterial, Thromb. Res., 20, 543, 1980.
21. Sefton, M. V. and Goosen, M. F. A., Irreversible immobilization of heparin for biomaterials, Dev.
Biochem., 12, 463, 1981.
22. Venter, J., Immobilized and insolubilized drugs, hormones, and neurotransmitters: properties, mechanisms
of action, and applications, Pharmacol. Rev., 34, 153, 1982.
23. Wingard, L., Immobilized drugs and enzymes in biochemical pharmacology, Biochem. Pharmacol., 32,
2647, 1983.
24. Mattiasson, B., Immobilized Cells and Organelles, CRC Press, Boca Raton, Fla., 1983.
FIGURE 13. A scanning electron micrograph of a porous PP hollow fiber. The bar equals
1 0 0 microns.
25. Chibata, I. and Wingard, L., Applied biochemistry bioengineering in, Immobilized Microbial Cells, Vol.
4, Academic Press, New York, 1983.
26. Geysen, H., Meloen, R., and Barteling, S., Use of peptide synthesis to probe viral antigens for epitopes
to a resolution of a single amino acid, Proc. Natl. Acad. Sci. U.S.A., 81, 3998, 1984.
27. Kaetsu, I., kumakura, M., and Yoshida, M., Enzyme immobilization by radiation-induced polymeri-
zation of 2-hydroxyethyl methacrylate at low temperatures, Biotechnol. Bioeng., 21, 847, 1979.
28. Katchalski, E., Silman, I., and Goldman, R., Effect of the microenvironment on the mode of action of
immobilized enzymes in, Advances in Enzymology, Vol. 34, Nord, F. F., Ed., Interscience, New York,
445, 1971.
29. Goldman, R., Goldstein, L., and Katchalski, E., Water-insoluble enzyme derivatives and artificial enzyme
membranes in, Biochemical Aspects of Reactions on Solid Supports, Stark, G. R., Ed., Academic Press,
30. Hicks, G. P. and Updike, S. J., The preparation and characterization of lypholized polyacrylmamide
enzyme gels for chemical analysis, Anal. Chem., 38, 726, 1966.
31. Mosbach, K. and Mosbach, R., Entrapment of enzymes and microorganisms in synthetic cross-linked
polymers and their applications in column techniques, Acta Chem. Scand., 20, 2807, 1966.
32. Nilsson, H., Mosbach, R., and Mosbach, K., The use of bead polymerization of acrylic monomers for
immobilization of enzymes, Biochem. Biophys. Acta, 268, 253, 1972.
33. Bernfeld, P., Bieber, R. E., and MacDonnell, P. C., Water insoluble enzymes: arrangement of aldolase
within an insoluble carrier, Arch. Biochem. Biophys., 127, 779, 1968.
34. O’Driscoll, K. F., Izu, M., and Korus, R., Gel entrapment of enzymes, Biotechnol. Bioeng., 14, 847,
1972.
35. Epton, R., Marr, G., and Shackley, A., New hydrophilic acrylate monomers suitable for the gel bead
entrapment of enzymes, Polym. Rep., 22, 553, 1981.
36. Kawashima, K. and Umeda, K., Immobilization of enzymes by radiopolymerization of acrylamide,
Biotechnol. Bioeng., 16, 609, 1974.
37. Ratner, B. D. and Hoffman, A. S., Surface grafted polymers for biomedical applications, Synthetic
Polymeric Biomaterials Concepts and Applications, Szycher, M. and Robinson, W., Eds., Technomic
Publishing, Westport, Conn. 1980, 133.
38. Ratner, B. D. and Hoffman, A. S., The effect of cupric ion on the radiation grafting of V-vinyl-2-
pyrrolidone and other hydrophilic monomers to silicone rubber, J. Appl. Polym. Sci., 18, 3183, 1974.
39. Kumakura, M., Yoshida, M., and Kaetsu, I., Immobilization of Streptomyces phaerochromogenes by
radiation-induced polymerization of glass forming monomers, Biotechnol. Bioeng., 21, 679, 1979.
40. Maeda, H., Yamauchi, A., and Suzuki, H., Preparation of immobilized enzymes by y-ray irradiation,
Biochim. Biophys. Acta, 315, 18, 1973.
41. Maeda, H., Yamauchi, A., and Suzuki, H., Preparation of immobilized enzymes by A-vinyl pyrrolidone
and the general properties of the glucoamylase gel., Biotechnol. Bioeng., 16, 1517, 1974.
42. Maeda, H., Suzuki, H., and Yamauchi, A., Preparation of immobilized enzymes using polyvinyl alcohol,
43. Maeda, H., Suzuki, H., Yamauchi, A., and Sakima, A., Preparation of immobilized enzymes from
acrylic monomers under y-ray irradiation, Biotechnol. Bioeng., 17, 119, 1975.
44. Yoshida, M., Kumakura, M., and Kaetsu, I., Immobilization of enzymes by radiation induced co-
polymerization of 2-hydroxyethyl methacrylate and other hydrophilic or hydrophobic comonomers, J.
Macromol. Sci. Chem., 14, 555, 1980.
45. Kumakura, M. and Kaetsu, I., Properties of immobilized papain by radiation polymerization, Polymer
J., 16, 113, 1984.
46. Kaetsu, I., Kamukura, M., Asana, M., Yamada, A., and Suzuki, Y., Immobilization of enzymes for
medical uses on plastic surfaces by radiation induced polymerization at low temperatures, J. Biomed. Mater.
Res., 14, 199, 1980.
47. Tanaka, A., Yasuhara, S., Fukui, S., Iida, T., and Hasegawa, E., Immobilization of invertase by the
use of photocross-linkable resin oligomers and properties of the immobilized enzyme, J. Ferment. Technol.,
55, 71, 1977.
48. Fukushima, S., Nagai, T., Fujita, K., Tonaka, A., and Fukui, S, Hydrophilic urethane prepolymers:
convenient materials for enzyme entrapment, Biotechnol. Bioeng., 20, 1465, 1978.
49. Bernfeld, P. and Won, J., Antigens and enzymes made insoluble by entrapping them into lattices of
synthetic polymers, Science, 142, 678, 1963.
50. Carrel, S. and Barandun, S., Protein-containing polyacrylamide gels: their use as immunoadsorbents of
high capacity, Immunochemistry, 8, 39, 1971.
51. Mattiasson, B., Larsson, P. O., and Mosbach, K., The microbe thermistor, Nature (London), 268, 519,
1977.
52. Shimizu, S., Morioka, H., Iani, Y., and Ogata, K., Synthesis of coenzyme A by immobilized microbial
cells, J. Ferment. Technol., 53, 77, 1975.
53. Chibata, I., Tosa, T., and Tato, T., Immobilized aspartase-containing microbial cells: preparation and
enzymatic properties, Appl. Microbiol., 27, 878, 1974.
54. Omatas, T., Iida, T., Tanaka, A., and Fukui, S., Transformation of steroids by gel-entrapped Nocardia
rhodocrous cells in organic solvents, Eur. J. Appl. Microbiol. Biotechnol., 8, 143, 1979.
55. Kennedy, J. F. and Joaquim, M. S., Immobilized living cells and their applications in, Applied Bio-
chemistry and Bioengineering, Vol. 4, Chibata, I. and Wingard, L. B., Eds., Academic Press, New York,
1983, 189.
56. Ochiai, H., Shibata, H., Matsuo, T., Hashinokuchi, K., and Yukawa, M., Immobilization of chloroplast
photosystems, Agric. Biol. Chem., 41, 721, 1977.
57. Ochiai, H., Shibata, H., Matsuo, T., Hashinokuchi, K., and Inamura, I., Immobilizaion of chloroplast
photosystems with polyvinyl alcohols, Agric. Biol. Chem., 42, 683, 1978.
58. Ochiai, H., Shibata, H., Fuiishima, A., and Honda, K., Photocurrent by immobilized chloroplast film
electrode, Agric. Biol. Chem., 43, 881, 1979.
59. Karube, I., Aizawa, K., Ikeda, S., and Suzuki, S., Carbon dioxide fixation by immobilized chloroplasts,
60. Tanaka, A., Yasuhara, S., Gellf, G., Osumi, M., and Fukui, S., Immobilization of yeast microbodies
and the properties of immobilized microbody enzymes, Eur. J. Appl. Microbiol. Biotechnol., 5, 17, 1978.
61. Yang, H. S., Leung, H. H., and Archer, M. C., Preparation and properties of bacterial chromatophores
entrapped in polyacrylamide, Biotechnol. Bioeng., 18, 1425, 1976.
62. Freeman, A. and Aharonowitz, Y., Immobilization of microbial cells in cross-linked, prepolymerized,
linear polyacrylamide gels: antibiotic production by immobilized Streptomyces clavuligeres cels, Biotechnol.
Bioeng., 23, 2747, 1981.
63. Hayashi, T., Tanaka, Y., and Kawashima, K., Immobilization of microbial cells containing NAD-
Kinase, Biotechnol. Bioeng., 71, 1019, 1979.
64. Fujimura, T. and Kaetsu, I., Immobilization of yeast cells by radiation-induced polymerization, Z.
Naturforsch., 37, 102, 1981.
65. Kaetsu, I., Kumakura, M., Kikuchi, S., Adachi, S., and Suzuki M., Immobilization of E. coli cell as
an antigen by radiation polymerization method. Z. Naturforsch., 38, 812, 1983.
66. Fujimura, T., Yoshii, F., Kaetsu, I., Inoue, Y., and Shibata, K., Effect of irradiation and immobilization
on spinach chloroplast activities, Z. Naturforsch., Teil C, 35, 477. 1980,
67. Yoshii, F., Fujimura, T., and Kaetsu, I., Stabilization of chloroplast by radiation-induced immobilization
with various glass-forming monomers. Biotechnol. Bioeng., 23, 833, 1981.
68. Yoshii, F., Fujimura, T., and Kaetsu, I., Stabilization of chloroplasts by radiation induced-immobilization
with various glass-forming monomers, Biotechnol. Bioeng., 23, 833, 1981.
69. Chang, T. M. S., Semipermeable Microcapsules, Science, 146, 524, 1964.
70. Chang, T. M. S., Macintosh, F. C., and Masen, S. G., Semipermeable aqueous microcapsules. I.
Preparation and properties Can. J. Physiol. Pharmacol., 44, 15, 1966.
71. Chang, T. M. S., Immobilized enzymes and their biomedical applications, in Immobilized Enzymes,
Antigens, Antibodies, and Peptides, Weetall, H. H., Ed., Marcel Dekker, New York, 1975, 245.
72. Klomp, G. F., Ronel, S. H., Hashiguchi, H. D., Andrea, M., and Dobelle, W. H., Hydrogels for
encapsulation of pancreatic islet cells, Trans. Am. Soc. Artif. Int. Organs, 25, 74, 1979.
73. Ronel, S. H., D’andrea, M. J., Hashiguchi, H., Klomp, G. F., and Dobelle, W. H., Macroporous
hydrogel membranes for a hybrid artificial pancreas. I. Synthesis and chamber fabrication, J. Biomed.
Mater. Res., 17, 855, 1983.
74. Jack, T. R. and Zatic, J. E., The immobilization of whole cells, Adv. Biochem. Eng., 5, 125, 1977.
75. Lee, R. G., Adamson, C., and Kim, S. W., Competitive adsorption of plasma proteins onto polymer
surfaces, Thromb. Res., 4, 485, 1974.
76. Messig, R. A., Immobilized enzymes in, Methods in Enzymology, Vol. 44, Mosbach, K., Ed., Academic
Press, New York, 1976, 148.
77. Fujimura, T. and Kaetsu, I., Growth of yeast cells immobilized with porous swelling carriers produced
by radiation polymerization, Appl. Biochem. Biotechnol., 8, 145, 1983.
78. Schnaar, R. L., Weigel, P. H., Kuhlenschmidt, M. S., Lee, Y. C., and Roseman, S., Adhesion of
chicken hepatocytes to polyacrylamide gels derivatized with V-acetylglucosamine, J. Biol. Chem., 253,
7940, 1978.
79. Gilding, D. K., Green, G. F., Annis, D., and Wilson, J. G., Soft tissue ingrowth into hydrogels, Trans.
Am. Soc. Artif. Int. Organs, 24, 411, 1978.
80. Yoshii, F. and Kaetsu, I., Cell culture on polymers prepared by radiation-induced polymerization of
various glass-forming monomers, Appl. Biochem. Biotechnol., 8, 115, 1983.
81. Truffa-Bachi, P. and Wafsy, L., Specific separation of cells or affinity columns, Proc. Natl. Acad. Sci.
U.S.A., 66, 685, 1970.
82. Zabriskie, D., Ollis, D., and Burger, M., Activity and specifically of covalently immobilized wheat germ
agglutinin toward cell surfaces, Biotechnol. Bioeng., 15, 9981, 1973.
83. Jirku, V., Turkova, J., Kuchynkova, A., and Krumphanzyl, V., Modified alkyl methacrylate gel as
support for immobilization of yeast cells, Eur. J. Appl. Microbiol. Biotechnol., 6, 217, 1979.
84. Jirku, V., Turkova, J., and Krumphanzyl, V., Immobilization of yeast cells with retention of cell
division and extracellular production of macromolecules, Biotechnol. Lett., 2, 509, 1980.
85. Goldstein, C. and Manecke, G., The chemistry of enzyme immobilization in, Applied Biochemistry
Bioengineering, Vol. 1, Wingard, L. B., Katchalski, E., and Goldstein, L., Eds., Academic Press, New
York, 16, 1976.
86. Jacoby, W., and Wilchek, M., Eds., Methods in Enzymology, Vol. 34, Academic Press, New York,
1974.
87. Inman,J., Covalent linkage of functional groups, ligands and proteins to polyacrylamide beads in, Methods
in Enzymology, Vol. 34, Jakoby, W. and Wilchek, M., Eds., Academic Press, New York, 1974, 30.
88. Barker, S., Somers, P., Epton, R.,and McLaren, J., Cross-linked polyacrylamide derivatives (Enzacryls)
as water-insoluble carriers of amylolytic enzymes, Carbohydr. Res., 14, 287, 1970.
89. Barker, S., Somers, P., and Epton, R., Recovery and reuse of water-insoluble amylase derivatives,
Carbohydr. Res., 14, 323, 1970.
90. Axen, R., Vletblad, P., and Porath, J., The use of isocyanides for the attachment of biologically active
substance to polymers, Acta Chem. Scand., 25, 1129, 1971.
91. Weston, P. and Avrameas, S., Proteins coupled to polyacrylamide beads using glutaraldehyde, Biochem.
Biophys. Res. Commun., 45, 1574, 1971.
92. Turkova, J., Immobilization of enzymes on hydroxyalkyl methacrylate gels in, Methods in Enzymology,
Vol. 34, Jacoby,W. and Wilchek, M., Eds., Academic Press, New York, 1974, 66.
93. Manecke, G. and Bogt. H., On some reactive polymers and immobilized enzymes, Pure Appl. Chem.,
50, 655, 1978.
94. Hoffman, A. S., Schmer, G., Harris, C., and Kraft, W. G., Covalent binding of biomolecules to
radiation-grafted hydrogels on inert polymer surfaces, Trans. Am. Soc. Artif. Int. Organs, 18, 10, 1972.
95. Mate, P., Horbett, T. A., Hoffman, A. S., and Ratner, B. D., Covalent coupling of small molecules
and proteins to poly(2-hydroxyethyl methacrylate-methacrylic acid) hydrogels, Enzyme Engineering, Vol.
2, Pye, E. K., and Wingard, L. B., Eds., Plenum Press, New York, 1974, 137.
96. Venkatraman, S., Horbett, T. A., and Hoffman, A. S., The kinetic behavior of a-chymotrypsin im-
mobilized inside hydrogel-grafted tubes, J. Mol. Catal., 2, 275, 1977.
97. Venkatraman, S., Studies of the Immobilization of a-chymotrypsin on 2-hydroxyethylmethacrylate-meth-
acrylic Acid Hydrogels and the Kinetics of a Tubular Immobilized a-chymotrypsin Reactor, Ph.D. thesis,
University of Washington, Seattle, 1976.
98. Venkatraman, S., Horbett, T. A., and Hoffman, A. S., The reactivity of a-chymotrypsin immobilized
on radiation-grafted hydrogel surfaces, J. Biomed. Mater. Res. Symp., 8, 111, 1977.
99. Plate, N. A., Valuev, C. I., and Chupov, V. V., Effects of methods of immobilization of proteolytic
enzymes in polymeric hydrogels on blood compatability of modified polymeric materials, Vyssokomool.
Soedin Ser. A, 22, 1963, 1980.
100. Muller-Schulte, D. and Horster, F., Radiation grafted polyethylene as a carrier for protein immobilization,
Polym. B ull, 7, 77, 1982.
101. J e n d r y c h o w s k a - B o n a m o u r , A., Hudrogels prepares par le Greffage Radioch imique pour Yimmobilisation
des enzymes, J. Polym. Sci. Polym. Chem. Ed., 21, 2589, 1983.
102. Abdel-Hay, F., Beddows, C., Gil, M., and Guthrie, J., The use of acrylic acid-co-nylon substrates for
immobilizing enzymes, J. Polym. Sci. Polym. Chem. Ed., 21, 2463, 1983.
103. Campbell, D. H., Luescher, E., and Lerman, L. S., Immunologic adsorbents. I. Isolation of antibody
by means of a cellulose protein antigen, Proc. Natl. Acad. Sci. U.S.A., 31, 515, 1951.
104. Petrov, R., Kabonov, V., Khaitov, R., Nekrasov, A., Alekseeva, N., Aparin, P., and Gelkina, S.,
Immunizing effect of bacterial polysaccharides with synthetic polyelectrolytes, Immunologiya, 5, 40, 1983.
105. Yen, S. P. S., Rembaum, A., and Molday, R. S., Improved hydrophilic polymeric microspheres and
magnetic functional microspheres for immunological research, Polym. Prepr. Am. Chem. Soc. Div. Polym.
Chem., 20, 577, 1979.
106. Rembaum, A., Cell labeling with magnetic and non-magnetic immunomicrospheres for separation and
diagnosis, Pure Appl. Chem., 56, 1305, 1984.
107. Molday, R. S., Dreyer, W. J., Rembaum, A., and Yen, S. P. S., Latex spheres as markers for studies
of cell surface receptors by scanning electron microscopy, Nature (London), 249, 81, 1974.
108. D’Anquiro, L., Mazzola, G., Vecchio, G., Focher, B., and Cremonesi, P., Comparison of characteristics
of immobilized enzymes prepared by graft-copolymerization and support activation, J. Appl. Biochem., 2,
208, 1980.
109. D’Anquiro, L. and Cremonesi, P., Enzyme immobilization by ultra-violet graft copolymerization: influ-
ence of the enzyme on reaction mechanism and on the characteristics of the immobilized enzyme copolymers,
J. Appl. Biochem., 3, 459, 1981.
110. D’Anquiro, L. Vecchio, G., Mazzola, G., Cantafi, R., and Cremonesi, P., Bisacryloylpiperazine as
vinylating agent for enzyme immobilization by graft-copolymerization onto polysaccharides, Angexv. Mak-
romol. Chem., 104, 129, 1982.
111. Martinek, K., Klibanov, A., Goldmacher, U., and Berezin, I., The pmciples of enzyme stabilization.
1. Increase in the thermostability of enzymes covalently bound to a complementary surface of a polymer
support in a multipoint fashion, Biochim. Biophys. Acta, 485, 1, 1977.
112. Torchilin, V., Tischenko, E., Smirnov, V., and Chazov, E., Immobilization of enzymes on slowly
soluble carriers, J. Biomed. Mater. Res., 11, 223, 1977.
113. Mozhaev, V., Siksnis, A., Torchilin, V., and Martinek, K., Operational stability of copolymerized
enzymes at elevated temperatures, Biotechnol. Bioeng., 25, 1937, 1983.
114. Plate, N., Valuev, L., and Chupov, V., The chemistry of macromonomers based on physiologically active
substances and some problems of the synthesis of polymers for biomedical applications, Pure Appl. Chem.,
56, 1351, 1984.
115. Mosbach, K., Matrix bound enzymes. I. The use of different acrylic copolymers as matrices, Acta Chem.
Scand., 24, 2084, 1970.
116. Mosbach, K. and Mattiasson, B., Matrix bound enzymes. II. Studies on a matrix-bound two-enzyme-
system, Acta Chem. Scand., 24, 2093, 1970.
117. Poliak, A., Blumenfeld, H., Wax, M., Baughn, R., and Whitesides, G., Enzyme immobilization by
condensation copolymerization into cross-linked polyacrylamide gels, J. Am. Chem. Soc., 102, 6327, 1980.
118. Fawcett, J., and Morris, C., Molecular-sieve chromotography of proteins on granulated polyacrylamide
gels, Sep. Sci., 1, 9, 1966.
119. Carr, P. W., and Bowers, L. D., Immobilized Enzymes in Alytical and Clinical Chemistry, Vol. 56,
Elvig, P. J., Winfordner, J. D., Kolthaff, I. N., Eds., John Wiley and Sons, New York, 1980.
120. Manecke, G., Gunzel, G., and Forster, H., Enzymes covalently bound to various polymers and the
effects on the properties of these enzymes, J. Polym. Sci., Polymer Symp., 30, 607, 1970.
121. Gestrelius, S., Mattiasson, B., and Mosbach, K., On the regulation of the activity of immobilized
enzymes microenvironmental effects of enzyme-generated pH changes, Eur. J. Biochem., 36, 89, 1973.
122. Samsonov, G., Ponomareva, R., and Melenevsky, A., Thermodynamics and kinetics of the interaction
synthetic and semi-synthetic polymers with enzymes during their immobilization and modification, Pure
Appl. Chem., 51, 1409, 1979.
123. Goldstein, L., Microenvironmental effects on enzyme catalysis: a kinetic study of polycationic derivatives
of chymotrypsin, Biochemistry, 1, 4072, 1972.
124. Goldstein, L., Pecht, M., Blumberg, S., Atlas, D., and Levin, Y., Water insoluble enzymes. Synthesis
of a new carrier and its utilization for preparation of insoluble derivatives of papain, trypsin, and subtilo-
peptidase, Biochemistry, 9, 2372, 1970.
125. Bowers, L. D., Immobilized enzymes in chemical analysis, Trends in Anal. Chem., 1, 191, 1982.
126. Klibanov, A. M., Immobilized enzymes and cells as practical catalysts, Science, 219, 722, 1983.
127. Hicks, G. and Updike, S., The preparation and characterization of lypholized polyacrylamide enzyme gels
for chemical analysis, Anal. Chem., 38, 726, 1966.
128. Cheung, P. W., Fleming, D. G., Ko, W. H., and Neuman, M. R., Theory, Design, and Biomedical
Applications of Solid State Chemical Sensors, CRC Press, Boca Raton, Fla., 1978.
129. Clark, L. and Lyons, C., Electrode systems for continuous monitoring in cardiovascular surgery, Ann.
N. Y. Acad. Sci, 102, 29, 1962.
130. Ross, P., Biosensors a new analytical technology, Biotechniques, 1, 204, 1983.
131. Mansouri, S. and Schultz, J., A miniature optical glucose sensor based on affinity binding, Biotechology,
2, 885, 1984.
132. Mattiasson, B., Larsson, P., and Mosbach, K., The microbe thermisitor,Nature (London), 268, 519,
1977.
133. Kumakura, M., Adachi, S., and Kaetsu, I., A new electrochemical determination of potassium ion
concentration with immobilized valinomycin disc, Electrochem. Sci. Technol., 131, 1272, 1984.
134. Driscoll, K., Kapoulas, A., Albisser, A., and Gander, R., Continuous monitoring of blood glucose
using gel entrapped glucose oxide, in Hydrogels for Medical and Related Applications, Vol. 31, ACS
Symposium Series Andrade, J. D., Ed., American Chemical Society, Washington, D. C., 1, 1976.
135. Guesdon, J. and Avrameas, S., Solid phase enzyme immunoassays m,Applied Biochemistry Bioengi-
neering, Vol. 3 .Wingard, L.,Katchalski, E., Goldstein, L., Eds., Academic Press, New York, 207, 1981.
136. Kardana, A., Squire, M., Adam, T., Goka, G., and Bagshawe, K., Sensitive radioimmunoassays using
partially purified gamma globulins coupled to enzacryl (acrylamide polymer) solid support, J. Immunol.
Methods, 30, 47, 1974.
137. Inman, J. and Dintzis, H., The derivitation of cross-linked polyacrylamide beads. Controlled introduction
of functional groups for the preparation of special-purpose, biochemical adsorbents, Biochemistry, 8, 4074,
1969.
138. Ternynck, T. and Avrameas, S., Polyacrylamide-protein immunoadsorbents prepared with glutaldehyde,
FEBS Lett., 23, 24, 1972.
139. Guesdan, J. and Avrameas, S., Polyacrylamide-agarose beads for the preparation of effective immu-
noadsorbents, J. Immunol. Methods, 11, 129, 1976.
140. Kumakura, M., Suzuki, M., Adachi, S., and Kaetsu, I., Polyacrolein microspheres as immunoreagents,
J. Immunol. Methods, 63, 115, 1983.
141. Kumakura, M., Suzuki, M., ad Kaetsu, I., Properties of functional polymeric microspheres obtained by
radiation polymerization of acrolein, J. Colloid Interface Sci., 94, 157, 1984.
142. Kumakura, M., Kaetsu, I., Suzuki, M., and Adachi, S., Immobilization of antibodies and enzymes
labeled antibodies by radiation polymerization, Appl. Biochem. Biotechnol., 8, 87, 1983.
143. Kumakura, M., Kaetsu, I., and Adachi, S., Preparation by irradiation of a solid support for enzyme
immunoassay, J. Immunol. Methods, 69, 79, 1984.
144. Kumakura, M. and Kaetsu, I., Polymeric microspheres for immunoresearch, Immunol. Commun., 13,
119, 1984.
145. Kumakura, M., Adachi, S., and Kaetsu, I., Flexible porous discs by radiation polymerization method
for enzyme immunoassay of a-fetoprotein, Arch. Immunol. Ther. Exp., 32, 135, 1984.
146. Kumakura M., and Kaetsu, I., Sheets obtained by radiation polymerization for enzyme immunoassay,
Inti. J. Appl. Radiat. I sot., 35, 471, 1984.
147. Stocker, J. and Heusser, C., Methods for binding cells to plastic; applications to a solid phase radioim-
munoassay for cell-surface antigens, J. Immunol. Methods, 26, 87, 1976.
148. Pachmann, K. and Gabrilova, J., Use of antibody-coated polyacrylic plastic beads for semiquantitative
determination of cell surface antigens, J. Immunol. Methods, 64, 383, 1983.
149. Uchida, T., Hosaka, S., and Murao, Y., Polymer microbeads bound to C3 fragments for detecting and
labeling cells with C3 receptors, Immunol. Commun., 1, 467, 1982.
150. Pachmann, K., Thierfeld, S., and Rodt, H., Poly-L-lysine as a coating agent for the attachment of active
antibodies onto polyacrylic plastic beads, Biomedicine, 33, 210, 1980.
151. Tritton, T., Yee, G.,and Wingard, C., Immobilized andriamycin: a tool for separating cellsurface from
intracellular mechanisms. Fed. Proc. Fed. Am. Soc. Exp. Biol., 42, 284, 1983.
152. Geysen, H., Meloen, R., and Barteling, S., Use of peptide synthesis to probe viral antigens for epitopes
to a resolution of a single amino acid, Proc. Natl. Acad. Sci. U.S.A., 81, 3998, 1984.
153. Kronick, P. L. and Rembaum, A., Reactions of human platelets with microspheres of poly(hydroxyethyl
methacrylate) and polyacrylamide, J. Biomed. Mater. Res., 11, 39, 1977.
154. Lewis, W. and Middleman, S., Conversion in a hollow fiber membrane enzyme reactor , A.l.Ch.E.J.,
20, 1012, 1974.
155. Rony, P., Multiphase catalysis. II. Hollow fiber catalysts, Biotechnol. Bioeng., 13, 431, 1971.
156. Pineda, A. A., Ed., Selective Plasma Component Removed, Futura Publishing, Mount Kisco, New York,
1984.
157. Langer, R. S. and Peppas, N. A., Present and future applications of biomaterials in controlled drug
delivery systems, Biomaterials, 2, 201, 1981.
158. Heller, J., Helwing, R., Baker, R., and Tuttle, M., Controlled release of water-soluble macromolecules
from bioerodable hydrogels, Biomaterials, 4, 762, 1983.
159. Buckles, R., Biomaterials for drug delivery systems, J. Biomed. Mater. Res., 17, 109, 1983.
160. Rosen, H., Chang, J., Wnek, G., Linhardt, R., and Langer, R., Bioerodable polyanhydrides for
controlled drug delivery, Biomaterials, 4, 131, 1983.
161. Kopecek, J., Controlled biodegradability of polymers-a key to drug delivery systems, Biomaterials, 5, 19,
1984.
162. Graham, N. and McNeil, M., Hydrogels for controlled drug delivery, Biomaterials, 5, 27, 1984.
163. Majkus, V., Horakova, Z., Vymola, F., and Stol, M., Employment of hydron polymer antibiotic vehicle
in otolaryngology, J. Biomed. Mater. Res., 3, 443, 1969.
164. Drobnik, J., Spacek, P., and Wichterle, O., Diffusion of anti-tumor drugs through membranes from
hydrophilic methacrylate gels, J. Biomed. Mater. Res., 8, 45, 1974.
165. Hosaka, S., Ozawa, H., and Tanzawa, H., Controlled release of drugs from hydrogel matrices, J. Appl.
Polym. Sci., 23, 2089, 1979.
166. Lee, E. S., Kim, S. W ., Kim, S. H., Cardinal, J. R., and Jacobs, H., Drug release from hydrogel
devices with rate controlling barriers, J. Membr. Sci., 7, 293, 1980.
167. Rosiak, J., Burezak, K., and Pekala, W., Polyacrylamide hydrogels as sustained release drug delivery
dressing materials, Radiat. Phys. Chem., 22, 907, 1983.
168. Gyselinck, P., Schacht, E., Van Severen, R., and Braeckman, P., Preparation and characterization of
therapeutic hydrogels as oral dosage forms, Acta Pharm. Technol., 29, 9, 1983.
169. Heller, J. and Trescony, P., Controlled drug release by polymer disolution II. Enzyme-mediated delivery
device, J. Pharm. Sci., 686, 919, 1979.
170. Yoshida, M., Asano, M., Kaetsu, I., Nakai, K., Yamanaka, H., Suzuki, T., Shida, K., and Suzuki,
K., In vivo release of testosterane from vinyl polymer composites prepared by radiation-induced polymer-
ization, Biomaterials, 4, 33, 1984.
171. Goosen, M., Leung, Y., Chou, S., and Sun, A., Insulin-albumin microbeads: an implantable, biode-
gradable system, Biomater. Med. Devices Artif. Organs, 10, 205, 1982.
172. Mueller, K. and Heiber, S. Gradient-IPN-modified hydrogel beads: their synthesis by diffusion-polycon-
densation and function as controlled drug delivery agents, J. Appl. Polym. Sci., 27, 4043, 1982.
173. Edman, P. and Sjotiolm, I., Acrylic microspheres in vivo. VI. Antitumor effect of microparticles with
immobilized I-asparaginase against 6C3HED lymphone, J. Pharm. Sci., 72, 654, 1983.
174. Schubert, W., Baurschmidt, P., Nagel, J., Thull, R., and Schaldach, M., An implantable artificial
pancreas, Med. Biol. Eng. Comp., 18, 527, 1980.
175. Ishihara, K., Kobayashi, M., Ishimaru, N., and Schinohara, I., Glucose induced permeation control
of insulin through a complex membrane consisting of immobilized glucose oxidase in a poly(amine). Polym.
J., 16, 625, 1984.
176. Ishihara, K., Kobayashi, M., and Shinohara, I., Insulin permeation through amphiphilic polymer mem-
branes having 2-hydroxyethyl methacrylate moiety, Polym. J., 16, 647, 1984.
177. Sefton, M. V., and Nishimura, E., Insulin permeability of hydrophilic polyacrylate membranes, J. Pharm.
Sci., 69, 208, 1980.
178. Horbett, T. A., Ratner, B. D., Kost, J., and Singh, M., A bioresponsive membrane for insulin delivery
in, Recent Advances in Drug Delivery Systems Anderson, J. M., Kim, S. W., Eds., Plenum Press, New
York, 1984, 209.
179. Horbett, T. A., Kost, J., and Ratner, B. D., Swelling behavior of glucose sensitive membranes, Polym.
Prepr. Am. Chem. Soc. Div. Polym. Chem., 24, 34, 1983.
180. Raso, V., Antibody mediated delivery of toxic molecules of antigen bearing target cells, Immunol. Rev.,
62, 93, 1982.
181. Poste, G. and Kirsh, R., Site-specific (targeted) drug delivery in cancer therapy, Biotechnology, 1, 869,
1983.
182. Collier, J. and Kaplan, D., Immunotoxins, Sci. Am., 251, 56, 1984.
183. Skinner, K., Enzyme Technology, Chem. Eng. News, 53, 22, 1975.
184. Bowers, L. and Carr, P., Immobilized enzymes in analytical chemistry, Adv. Biochem. Eng., 15, 89,
1982.
185. Margaritis, A. and Wallace J., Novel bioreactor systems and their applications, Biotechnology, 2, 447,
1984.
186. Gaden, E. L., Jr., Production methods in industrial microbiology, Sci. Am., 245, 181, 1981.
187. Cooney, L. L., Bioreactors: design and operation, Science, 219, 728, 1983.
188. Maugh, T., A renewed interest in immobilized enzymes, Science, 223, 474, 1984.
189. Buhak, S., Biotechnological Applications of Proteins and Enzymes, Academic Press, New York, 1977.
190. Linko, P. and Linko, Y., Applications of immobilized microbial cells in, Applied Biochemistry Bioen-
gineering, Vol. 4, Chibata, I., and Wingard, L., Eds., Academic Press, New York, 1984, 54.
191. Sata, T., Mari T., Tosa, T., Chibata, I., Furi, M., Yamashita, K., and Sumi, A., Engineering analysis
of continuous production of I-aspratic acid by immobilized Escherichia coli cells in fixed beds, Biotechnol.
Bioeng., 17, 1797, 1975.
Chapter 6
PROTEIN ADSORPTION TO HYDROGELS
Thomas A. Horbett
TABLE OF CONTENTS
I. Introduction.....................................................................................................................128
II. Adsorption From Single Protein Solutions................................................................ 128
III. Adsorption to Hydrogels From ProteinMixtures........................................................143
IV. Summary and Conclusions........................................................................................... 166
Acknowledgments..................................................................................................................... 168
References...................................................................................................................................168
I. INTRODUCTION
Protein adsorption is thought to be one of the more important initial events which occurs
upon exposure of solid surfaces to the biological environment. Many different types of
observations have contributed to this conclusion, but it derives most fundamentally from
the fact that the rapid formation of a relatively tightly held layer of protein prevents direct
contact of cellular elements with the surface. Thus, all of the situations in which solid
surfaces interact with biological systems actually involve interactions between the adsorbed
protein layer present at the interface and the various other biological elements. The adsorption
of proteins from a mixture to surfaces is selective because of differences in the affinity of
proteins for surfaces of differing chemical properties. Since each type of cell tends to have
highly specific attraction for only certain proteins, the differences in the composition of the
adsorbed layer on different surfaces probably constitute the major mechanism underlying
the infuence of the adsorbed layer on biological interactions with foreign surfaces (see also
Chapter 4).
The adsorption of proteins at interfaces has been the subject of several recent reviews,
but none of these has specifically focused on hydrogels.13 In this chapter, our intention,
therefore, is to present the available information on protein interactions with hydrogels in
some depth and, wherever possible, contrast it with analogous studies on nonhydrogel
materials. The major purpose of this article is to highlight the special or unusual properties
of hydrogels in regards to protein adsorption. This approach reflects our view that the “ state
of the art” in protein adsorption studies does not permit unambiguous information about
mechanisms of adsorption or allow interpretation of the data in terms of a single theory.
Instead, we must largely rely on a comparative, somewhat phenomenological approach. The
specific questions we plan to try to answer for the reader are
1. What types of data are available on protein adsorption to hydrogels?

2. How does protein adsorption to hydrogel and nonhydrogel surfaces differ?
3. What is the significance of the differences in adsorption to hydrogels and nonhydrogels?
As it will be seen, measurements of adsorption isotherms, competitive adsorption studies

from simple protein mixtures, studies of the types and amounts of protein adsorbed from
blood plasma, and characterization of the desorption/exchange reactions all reveal significant
differences between hydrogels and nonhydrogels.
II. ADSORPTION FROM SINGLE PROTEIN SOLUTIONS
Protein adsorption to a variety of hydrogels has been studied with various methods and
purposes, as summarized in Table 1. In most cases, adsorption to the hydrogel differed
significantly from that for the nonpolar, nonhydrogel material typically used as a control.
However, the particular aspect of the adsorption process which was studied and revealed
these differences was not always the same. Furthermore, although these differences may be
due to the weaker protein interactions predicted for hydrogels in some analyses4 6 of their
surface properties, the data available typically do not provide a direct measure of the affinity
of the protein for the surface. The interpretation of the differences in adsorption to hydrogels
in comparison to nonhydrogel material is therefore not entirely straightforward. In this
section, we present summaries of the available work together with critical evaluations of
the work.
Among the first reports of single protein adsorption studies of synthetic hydrogels were
those published by Horbett and his collaborators in 1975 and 1977, respectively,78 The
hydrogels studied were polymers of hydroxyethylmethacrylate (HEMA) and A-vinyl pyr-
Table 1
PROTEIN ADSORPTION FROM SINGLE PROTEIN SOLUTIONS TO HYDROGELS
Proteins Substrates Measurements Methods Comments Ref.
Fibrinogen HEMA Isotherms 1251 MAAc enhanced adsorption 7

Albumin NVP HEMA decreased adsorption
7 -Globulin Silastic® Desorption greater from HEMA
Fibrinogen HEMA (H) Isotherms 125I Less adsorption on HEMA/PE than PE or 8
EMA (E) EMA

H-E copolymers More desorption from HEMA
Polyethylene (PE) Freundlich exponent higher on HEMA
than EMA
Albumin Cuprophane® Desorption 125I Desorption and exchange greater on Cu- 14

Polyethylene Exchange prophane® than on polyethylene
Albumin PEU/PEG Kinetics t25I Greatly reduced adsorption to hydrogel- 17

Fibrinogen PEU/PPG like PEU/PEG compared to hydrophobic
PEU/PPG
Albumin Cuprophane® Isotherms 125I Less adsorption to Cuprophane® than PVC 24

Fibrinogen PVC Exchange for all proteins
Immunoglobulin G Desorption Enhanced exchange of IgG on Cupro-
phane® vs. PVC
Thrombin Cuprophane® Isotherms 1251 Weak and strong sites for thrombin on 25
Antithrombin III PVC Cuprophane®
ATIII adsorption less on Cuprophane®
Thrombin Cuprophane® Isotherms l2T Adsorption to Cuprophane® lower than 26

Activation PVC <100 [xg/m€
PVC
ATIII binding Adsorbed thrombin amidase inactive on
Antibody both surfaces
binding Thrombin on both surfaces induced fibrin
formation so
Table 1 (continued)
PROTEIN ADSORPTION FROM SINGLE PROTEIN SOLUTIONS TO HYDROGELS
Proteins Substrates Measurements Methods Commefits Ref.
Prothrombin Cuprophane® Isotherms l2T Less adsorption to Cuprophane® 27

PVC Activation Venom Partially activatable on all surfaces
PAN Activity Enzyme
■2 5 J
Fibrinogen Cuprophane® Isotherm Antifibrinogen uptake 28
Radioautography Uniform adsorption
Fibrinogen NVP Relative 125 I, NP-40 Albumin increased on NVP 11

Adsorption
Albumin PTFE 125I antibodies Fibrinogen decreased on NVP

IgG Total serum Prothrombin unchanged
proteins
Prothrombin IgG varied

Albumin HEMA Kinetics Elution Free fluorescence label penetrated gels 29
HEMA-MMA Isotherms Fluorescence Preferential adsorption of labeled protein
Albumin HEMA Isotherms Depletion Less adsorption to HEMA 30
Fibrinogen Polyethylene
Cellophane
Carbon
Albumin HEMA/ Distribution TEM Domain forming block copolymers 31

y-Globulin Polystyrene Albumin preferred HEMA domains
Fibrinogen Copolymers
Fibronectin HEMA Relative ELISA Fibronectin adsorption to HEMA much 33

Polyacrylamide adsorption reduced
Polyvinylacetate added to HEMA restored
adsorption
Albumin Cuprophane® Kinetics 125I Less adsorption to Cuprophane® 36
Fibrinogen Silastic® Temperature Temperature effects minimal except for al-
bumin on silicon rubber
Diffusion limited model with saturation fit
most data well
Albumin Polyvinyl alcohol — — Albumin adsorption decreased as water 35

Fibronectin Polyacrylamide content increased
Fibronectin adsorption less on
polyacrylamide
Fibrinogen Dextran Adsorption Ellipsometry Adsorption less on dextran 37

Silane Desorption
Silica
Albumin HEMA Adsorption 125I High adsorption to all 38

HEA Detergent HEA had twice adsorption of others
GMA wash
Polyacrylamide
Albumin HEMA Adsorption FTIR/ATR “ Denaturation” slower on HEMA/MAA 40

HEMA/MAA Spectra HEMA/MAA
Mucin Silicone (S) Adsorption l4C Greater adsorption to NVP/S than S 44

NVP/S kinetics
Lysozyme HEMA Interfacial Contact Adsorption enhanced wetting of HEMA 45

Albumin GMA/MMA tension angles and GMA/MMA contact lenses
y-Globulin
Adsorption
( "g/cm2 )
50
El row
40 0 •-Globulin 6.8 l~x10 5
83 Fibrinogen 5.5 3.4xlo 5
A Albumin 4.8 0. 7xlo 5
30
20
10
%MAAc in HEP'A
FIGURE I. The effect of methacrylic acid on protein adsorption

onto PHEMA/Silastic® at low ionic strength. The solvent was 0.005
M Hepes, pH 7.4. Protein concentration was 0.5 mg/m£. (From
Horbett, T. A. and Hoffman, A. S., Adv. Chem. Ser., 145, 230,
1984. With permission.)
rolidone (NVP), prepared as grafts to Silastic®7 or polyethylene. 8 The adsorption of fibri-

nogen, albumin, and )'-globulin or immunoglobulin G were studied, mainly using 125 I-labeled
forms of each protein added to a solution of the unlabeled form of the same protein. Some
experiments were done with unlabeled proteins, using a ninhydrin assay to detect protein
by measuring the amino acids in a base hydrolysate of adsorbed surface.
The strong effect of methacrylic acid present as a contaminant in most HEMA preparations
on protein adsorption was demonstrated in the initial study. 7 Thus, )'-globulin adsorption to
HEMA/Silastic® grafts from water (12.6 to 23 jJ..g/cm2 , depending on graft) was found to
be much higher than to Silastic® (0.8 jJ..g/cm2 ) unless highly purified HEMA preparations
and physiologic concentrations of salt were present in the buffer (0.11 jJ..g/cm2 observed
under these conditions). The effect of various amounts of methacrylic acid added to purified
HEMA on the adsorption of )'-globulin, fibrinogen, and albumin is illustrated in Figure I.
The protein with the most positive charge at pH 7 .4, i.e., )'-globulin, is adsorbed most
strongly by the negatively charged MAAc groups. The isoelectric pH of )'-globulin, fibri-
nogen, and albumin are about 6.8, 5.5, and 4.8, respectively, so that at the pH of the
adsorption experiments (7 .4), the amount of adsorption of each protein would be expected
to decrease in the same order, as Figure 1 shows. These data also illustrate that relatively
small amounts of MAAc are necessary to cause great increases in )'-globulin adsorption,
e.g., an increase of almost 50-fold between 0.02 and 1.5% MAAc. This finding confirms
calculations which indicate that the presence of MAAc in hydrogels made with unpurified
HEMA could reasonably explain the high )'-globulin adsorption from water observed for
such hydrogels. In the presence of physiologic concentrations of salt, adsorption of these
proteins onto these P(MAAc-co-HEMA) hydrogels was much lower (less than 0.5 j.Lg/cm2 ).
Since increased ionic strength is commonly used to elute proteins bound to ion exchange
resins during purification it is clear that the P(MAAc-co-HEMA) hydrogel behaves as a
typical ion exchange resin in its interactions with proteins. In this regard, although the
10
Amount Adsorbed
(jig/cm 2)
075
050
025
0.0 +---,------,,-------,-----,--.,..----.
0 2 3 4 5 6
GRAFT (mg/cm2)
FIGURE 2. Fibrinogen adsorption onto HEMA/Silastic®: effect of graft level. Films

radiation grafted with HEMA to various degrees were equilibrated at 37°C for 20 hr
in l mg!m€ fibrinogen solution in O.OlM Hepes, O.l47M NaCl, 0.02% azide, pH
7 .4, and then rinsed and soaked in buffer at room temperature overnight. (From
Horbett, T. A. and Hoffman, A. S.,Adv. Chem. Ser., 145, 230, 1984. With permission).
adsorption of these proteins from solutions of physiological ionic strength did not differ
greatly, many other proteins do adsorb readily onto ion exchanges at this ionic strength.
Thus, preferential adsorption of certain proteins was predicted for P(MAAc-co-HEMA)
hydrogels exposed to the in vivo-like environment. The potential for adsorption by hydrogels
made with impure HEMA may well have undesirable effects when the materials are used
as biomaterials in contact with blood or tissue. These expectations were confirmed in later
studies with contact lenses, where lysosyme adsorption to P(HEMA-co-MAAc) gels is very
high even in the presence of saline (see Section Ill).
The grafting of PHEMA to Silastic® was observed to decrease fibrinogen adsorption, as
shown in Figure 2. Many other studies have also revealed lower adsorption to hydrogels,
as will be noted below, but some apparent exceptions exist. As shown in Figure 3, the
adsorption isotherms for fibrinogen on PHEMA/Silastic® again display lower adsorption
than to Silastic®, but the other hydrogel system studied, PNVP/Silastic, ®appeared to adsorb
about the same amount of fibrinogen as Silastic. ® This unexpected divergence in the prop-
erties of the two hydrogels is caused by certain unique properties of the PNVP/Silastic®
hydrogels. Although the PNVP/ Silastic® graft retains about 60 wt% water compared with
about 30% for the PHEMA graft, the surface of the PNVP/Silastic® films does not appear
wet, i.e., the contact angle of water droplets adhering to the film is relatively high. In
contrast, there is an absence of drop formation on the wettable PHEMA/Silastic® graft
surface. Based on this and several other types of observations, the PNVP graft is thought
to be intermingled with the poly(dimethylsiloxane) (PDMS) chains of the silicone rubber
matrix while the PHEMA graft rests above and is separate from the underlying silicone
rubber matrix. 9 • 10 Other studies with PNVP grafted to polytetrafluoroethylene (PTFE) have
shown decreased fibrinogen adsorption with increasing graft level, 11 similar to the results
for PHEMA in Figure 2. The lowered adsorption of fibrinogen to P(NVP-g-TFE) but not
on PNVP/Silastic® presumably occurs because NVP is unable to penetrate the much more
crystalline PTFE substrate so a true surface graft of poly NVP is obtained in this case.
In our initial studies as well as those of most other investigators since, the comparison
of adsorption to the hydrogel and a chemically nonhomologous, nonhydrogel material was
made, e.g., HEMA to Silastic®, but clearly many major differences exist between such
chemically disparate polymers and it is therefore not exactly clear which differences are
Amount Adsorbed
(f1g/cm 2)
1.0
0.75
0.50
0.25
•
OL----L----~----L---~----~----~--~
0 1.0 2.0 3.0
Fibrinogen concentration (mg/ml)
FIGURE 3. Fibrinogen adsorption isotherms on Silastic®, PNVP/Silastic®, and PHEMA/Silastic®. Untreated

Silastic® (squares) and Silastic® grafted with PNVP (2.7 mg/cm 2 ) (triangles) or PHEMA (5.8 mg/cm 2 ) (circles)
were equilibrated at 37°C for 20 hr in fibrinogen solutions in 0.01 M HEPES, 0.147 M NaCI, 0.02% azide, pH
7 .4, and then rinsed and soaked in buffer at room temperature overnight.
ascribable to the hydrogel nature and which are due to other factors. Our second study of
protein adsorption to hydrogels was partly designed to alleviate this difficulty by using a
better "control" surface, namely the chemically homologous ethylmethacrylate (EMA). 8
The HEMA, EMA, and HEMA-EMA based copolymers were grafted to polyethylene in
this study. HEMA grafted polyethylene exhibited reduced fibrinogen adsorption compared
to ungrafted polyethylene, just as observed with HEMA grafts to Silastic®, whereas in-
creasing EMA grafts enhanced fibrinogen adsorption. The latter effect was consistent with
a surface area enhancement because the grafts are relatively rough and because smooth, cast
EMA films had much lower adsorption than grafts. 8 • 12
Fibrinogen adsorption isotherms on the P(HEMA-co-EMA-g-E) copolymer series were
measured after 24 hr, a time shown to give steady-state adsorption values in separate kinetic
studies. In all cases, the isotherms were not of the Langmuir type since they lacked a true
saturation or plateau value even at concentrations as high as 4 mglm€, but instead displayed
a slightly increasing slope. Isotherms for all materials are presented in Figure 4 using log-
log (Freundlich) coordinates. The constants obtained from a least squares fit of the isotherms
to the Freundlich equation (adsorption = K[concentration]b) are plotted in Figure 5 as a
function of graft composition. HEMA-rich materials were found to have a higher concen-
tration dependence, b, but a lower coefficient, K, than the EMA-rich materials. Both para-
meters vary uniformly with composition. The concentration dependence (b) for fibrinogen
adsorption to hydrophilic HEMA-rich surfaces was about double the value found for the
hydrophobic, EMA-rich homologs. Other studies suggest that the value of b increases as
the affinity of a protein for the surface increases. 13 If this relation between b and affinity is
applied to this study, then the affinity of fibrinogen for the hydrophilic HEMA grafts would
seem to be higher than to the hydrophobic EMA grafts.
The desorption and exchange of fibrinogen was also found to differ substantially in the
E
5
.5
.I
• .5
r------.-----.......------r-----t.l
0.001 001 0.1 I 10
Fibrinogen Concentration (mg/mU
FIGURE 4. Fibrinogen adsorption isotherms to P(HEMA-co-EMA) copolymers

grafted to polyethylene. Films were exposed to various concentrations of '"!-
fibrinogen for 24 hr at 3TC and dilution-displacement rinsed. Least squared fitted
lines to log of adsorption vs. log of concentration are drawn. The HEMA/EMA
monomer ratios used in making the polymers were: A, 4/0; B, 3/1; C, 2/2; D, I/
3; E, 0/4. (From Weathersby, P. K., Horbett, T. A., and Hoffman, A. S., J.
Bioeng., I, 395, !984. With permission).
P(HEMA-co-EMA-g-E) series. A portion of the adsorbed protein was desorbed much more
rapidly than the remainder. This rapidly removed "loose fraction" constituted about 25%
of the adsorbed fibrinogen on HEMA-rich grafts, but only 5% was rapidly removed from
surfaces containing at least 50% EMA. The halflives of the "strongly bound" fibrinogen,
as determined from the slope of a plot of the log of adsorbed protein vs. time of desorption,
also varied among the polymers. Fibrinogen desorbed from HEMA-rich surfaces with half-
lives of 30 to 70 days, while on predominantly EMA surfaces the release were nearly twice
as slow. The desorption from predominantly EMA grafts (113 day half-life) was increased
several fold in the presence of dissolved fibrinogen (42 to 45 day half-life). Presumably,
the enhanced desorption in the presence of solution phase fibrinogen is due to exchange
between adsorbed and solution phase fibrinogen. On PHEMA surfaces, fibrinogen exchange
occurred at nearly five times the rate of desorption from buffer alone. This result is similar
to the enhanced rates of removal of albumin from surfaces in the presence of albumin
136 HydrogeL~ in Medicine and Pharmacy
.5
b
.4
.3
k
4
0
4/0 3/1 2/2 1/3 0/4
HEMA/EMA Monomer Ratio
FIGURE 5. Freundlich analysis of fibrinogen adsorption to P(HEMA-

co-EMA) copolymers grafted to polyethylene. The plots show the pa-
rameters k and b of the empirical adsorption isotherm function a = kc".
where a is adsorption (in J,Lg/cm') and c is solution concentration (in mg/
mt). (From Weathersby, P. K., Horbett, T. A., and Hoffman, A. S.,
J. Bioeng .. I, 395, 1984. With permission).
solutions reported by Brash et al. 14 However, the rates of exchange in both studies were
still quite low relative to the adsorption rates and suggest that adsorption is not readily
reversible.
The studies of fibrinogen adsorption to P(HEMA-co-EMA-g-E) graft copolymers therefore
revealed marked differences in fibrinogen interaction with PHEMA and PEMA. However,
the adsorption isotherms seem to reflect higher affinity of fibrinogen for PHEMA than PEMA
because b is higher for HEMA, while the desorption data is suggestive of a lower degree
of interaction because desorption and exchange are more rapid on HEMA. This apparent
contradiction was originally interpreted to mean that the properties of the free solution
fibrinogen molecule, which determine its adsorption behavior, differ from those of the
adsorbed molecule, which control the desorption behavior. However, it was shown that the
protein layer adsorbed at higher concentrations appeared to contain a greater amount of
protein with electrophoretic properties different than native fibrinogen. Brash has recently
reported that more degraded fibrinogen is eluted first from glass bead columns. 15 Fibrinogen
preparations typically contain a series of partially degraded fibrinogen molecules. 16 These
observations suggest that the populations of fibrinogen molecules adsorbed to PHEMA and
PEMA may differ. The HEMA, which is more easily loaded with fibrinogen as indicated
by a higher exponential coefficient b in the Freundlich analysis, may be expected to have
more of the lower affinity fibrinogen variants on its surface than EMA. These lower affinity
variants would also be expected to be more easily desorbed.
The differences in fibrinogen desorption behavior evident in our studies of hydrogel and
nonhydrogel materials were also observed by Brash et al. in studies of albumin desorption
from Cuprophane® and polyethylene.14 Cuprophane® is a commercial hemodialysis mem-
brane made from regenerated cellulose and typically adsorbs substantial water into its struc-
ture (37% reported by Brash et all4). Although Cuprophane® is a processed natural product
rather than a synthetic material like PHEMA and it has never been proposed for use as an
implant, Cuprophane® has the essential properties of a hydrogel and has been the subject
of many protein adsorption studies. On polyethylene, albumin desorption into buffer was
negligible for periods up to 200 hr, whereas approximately 50% desorption occurred from
Cuprophane® in this period. When unlabeled albumin (at 10 mg%) was added to the buffer
to allow exchange with the adsorbed l25I protein, further desorption occurred on Cuprophane®
but no removal from polyethylene was observed unless relatively high concentrations (370
mg%) of albumin were present. Although the amounts of adsorption on Cuprophane® (about
70 ng/cm2 at 10 mg% albumin) were substantially less than on polyethylene (about 200 ng/
cm2) at lower concentrations, adsorption to Cuprophane® appeared to increase linearly
throughout the concentration range studied (10 to 100 mg%) and surpassed adsorption for
polyethylene at 100 mg%. These results might indicate partial entry of albumin or free l25I
iodide into the pores of the Cuprophane®. If this were to occur, the enhanced “ desorption”
from Cuprophane® might instead reflect the wash out of sorbed albumin or iodide rather
than detachment of adsorbed protein.
The adsorption kinetics of albumin and fibrinogen to hydrogel-like polyurethanes have
also been studied by Brash and Uniyal.17 The hydrogel-like polymers were made with
polyoxyethylene glycol of Mn 600 (PEU/PEG 600) or 1540 (PEU/PEG/540) and had water
contents of 17 and 67%, respectively.18 An analogous polyurethane made with polyoxypro-
pylene glycol of Mn 1200 (PEU/PPG 1200) behaved as a hydrophobic polymer, i.e., it did
not swell in water and had a high water contact angle. The adsorption of both albumin and
fibrinogen to the hydrogel-like polyurethanes (0.02 to 0.08 fxg/cm2) were much lower than
to the hydrophobic polyurethane (0.6 to 1.1 |ag/cm2). However, Brash et al. noted that the
lower adsorption to these hydrogels might have reflected greater desorption during the rinse,
i.e ., that adsorption to the PEG hydrogels could have been equivalent to that on PPG surfaces
but was much more loosely held on the PEG surface. Either explanation of the results would
be consistent with other studies showing a lower adsorption and greater exchangeability of
proteins on hydrogels.
A preliminary report of greatly reduced protein adsorption to quartz covalently coated
with polyethylene glycol has been given by Gregonis et al.19 A variety of hydrophilic coating
techniques have been used to reduce protein adsorption to matrices used in size exclusion
chromatography, including polyethylene oxide adsorption to controlled pore glass,20 gly-
cerylpropylsilane bonded to controlled pore glass,21 “ an ether functional organochlorosilane”
bonded to silica particle (“ |ul Bondgel” ), and other proprietary materials designed along the
same lines (“ TSK-Gel SW” ,22 “ Spherogel-TSK Type SW” 23). The adsorptive losses to
hydrogel-like chromatographic matrices such as Sephadex and Bio Gel P are also very low
although direct measurements of the amount bound, e.g., in |xg/cm2,were not found in the
literature. Although chromatography supports are not hydrogels in a biomedically useful
sense and measurements of protein adsorption to them has largely been qualitative, e.g.,
showing reduced losses to controlled pore glass after treatment,21 they do provide perspective
on protein adsorption to the biomedical hydrogels.
Chuang and co-workers have published a series of studies of protein adsorption to Cu-
prophane® and other surfaces.24 28 Adsorption isotherms for fibrinogen, immunoglobulin G,
albumin,24 28 thrombin and antithrombin III,25 and prothrombin27 were obtained and compared
to those on polyvinylchloride (PVC). Except for thrombin,25 all the isotherms displayed a
saturation or plateau adsorption which was substantially lower on Cuprophane® than on
PVC.
In the thrombin studies, the istotherms were analyzed to obtain “ kinetic constants” ,
apparently by analogy to Michaels-Menten analysis of enzyme kinetics which display a
maximal or saturation velocity.25 Although the isotherms are clearly not kinetic measure-
ments, their shape (rectangular hyperbola) is the same observed in plots of enzyme velocity
vs. substrate concentration. Thus, a double reciprocal plot of each variable linearizes both
types of data and allows a slope and intercept to be calculated. Chuang et al. further assigned
the term dissociation constant and maximum surface concentration to denote the parameters
obtained from this analysis of their isotherms, in analogy with similar constants obtained
from enzyme kinetics. This approach is attractive for its simplicity and ease, but it is highly
doubtful that the dissociation constants so calculated reflect the protein-substrate affinity.
The analogy fails because while the enzyme-substrate complex is reversible, the protein-
surface complex is not.
According to this analysis, thrombin binding to Cuprophane® reflects the presence of both
low affinity (dissociation constant, KD = 7190 nM and a high affinity (KD = 300 nM)
binding sites, whereas all the other proteins have a single type of site with KD values in the
range of 300 to 800 nM. The isotherm and “ kinetic analyses” for thrombin on PVC and
Cuprophane® are shown in Figure 6 while corresponding data for antithrombin are shown
in Figure 7; Table 2 summarizes these analyses. While the KD values cannot readily be
assigned their usual meaning, i.e., reciprocal association constants, they are a convenient
summary of this data. The KD values for albumin, fibrinogen, and antithrombin III on both
Cuprophane® and PVC are quite similar whereas IgG has very different KD values on these
two surfaces. Thus, the concentration dependence of adsorption is very similar in all cases
except for IgG on PVC. On the other hand, the maximal or plateau concentrations are higher
on PVC than on Cuprophane.® Chuang et al.24 also observed a somewhat larger fraction of
exchangeable protein on Cuprophane® than on PVC, analogous to the observations of Brash
et al.14 In a more recent study, Chuang et al.26 observed that although the amidase activity
of adsorbed thrombin was undetectable on both Cuprophane® and PVC, antithrombin-III
uptake by thrombin-treated Cuprophane® and PVC could occur and thrombin on each surface
could induce fibrin formation when exposed to a fibrinogen solution. The differences in
thrombin isotherms to the surfaces do not seem to reflect detectable differences in the
functional state of the molecule on the two surfaces.
Prothrombin adsorption isotherms to Cuprophane® and PVC were also reported by Chuang
et al.27 The plateau adsorption to Cuprophane (about 100 ng/cm2) was significantly less than
on PVC (about 160 nm/cm2) but as with albumin, IgG, antithrombin III, and fibrinogen,
no great difference in the initial slope of the prothrombin isotherm on Cuprophane® or PVC
was obvious. The activation of adsorbed prothrombin by a snake venom preparation occurred
to similar extents on both surfaces, whether measured by amidolytic activity (about 26% of
solution value) or the release of the activation peptide (about 16% on both surfaces). Thus,
despite the presumably different environment in the hydrogel and nonhydrogel surfaces, the
prothrombin molecule did not appear to behave differently with respect to its function.
Finally, Chuang et al. have recently reported the application of in situ immunoradiometric
assay of adsorbed fibrinogen on Cuprophane®.28 Comparative studies with PVC were not
reported. The adsorbed fibrinogen readily and specifically reacts with the antifibrinogen
whether adsorbed from pure fibrinogen solutions or from protein mixtures. A radioautograph
of a Cuprophane® disc treated first with fibrinogen and then with 125I antifibrinogen revealed
a uniform distribution of grains, whereas a more nonuniform image was obtained with a
clinically used piece of Cuprophane ®.
The adsorption of fibrinogen, immunoglobin G, prothrombin, and albumin to polytetra-
fluoroethylene (PTFE) and PNVP radiation grafted to PTFE were studied by Boffa et al.11
Although the buffers used were usually close to physiologic pH and ionic strength, the
buffers also contained the nonionic detergent NP-40 at 0.01%, making comparison of their
600
A !......
N
E
B
...
u 10
.....
N
....
E
.....
500
"'Q
c
8
.= 400 .
u
z
0
z
6
0 300 0
(.)
(.) 4
_1&1
1&1
(.) 200 (.)
c(
c( 1&. 2
1&.
a:: a::
::::l 100 ::::l
(/)
(/)
0 20 40 60 80 100
0 0.2 0.4 0.6 0.8 1.0
I (mg/mlr 1
THROMBIN
THROMBIN (mg/ml)
FIGURE 6. Adsorption of Thrombin onto Cuprophane® and polyvinyl chloride. (A) Isotherms for thrombin
adsorption onto Cuprophane (o-o) and PVC (x--x); (B) double reciprocal plot for thrombin adsorption
onto Cuprophane® (o-o) and PVC (x--x). (From Chuang, H. Y. K., Crowther, P. E., Mohammed, S.
F., and Mason, R. G., Thromb. Res., 14, 273, 1984. With permission).
1200 -c..
N" A N
E
u
B
E
u
.....
0
1000 ...c
..... 30
5
..
BOO .Q
0 20
z 600
0 <.J
0 z
0
w 400 0
0 w 10
<
1.1..
0
a: 200 c(
I.L.
::::> a::
en :::>
(/)
0 0.2 0.4 0.6 0.8 1.0
0 5 10 15 20 25
ANTITHROMBIN m (mg/ml) I <motmlr 1

ANTITHROMBIN m
FIGURE 7. Adsorption of Antithrombin Ill onto Cuprophane® and Polyvinyl chloride. (A) Isotherms for AT-Ill
adsorption onto Cuprophane® (o-o) and PVC (x--x); (B) double reciprocal plot for AT-Ill adsorption
onto Cuprophane® (o-o) and PVC (x--x). (From Chuang, H. Y. K., Crowther, P. E., Mohammed, S.
F., and Mason, R. G., Thromb. Res., 14, 273, 1984. With permission).
work to that of others doubtful. The authors did not explain why NP-40 was present in the
buffer. The NP-40 may have helped stabilize the granular polymer suspensions used in some
of this work. The data were not reduced to J.Lg/cm2 but instead presented either as observed
cpm or uptake relative to PTFE. Despite these limitations, the data merit some attention
because the only other available study of adsorption to pNVP is thought to be compromised
by the intermingling of silicone polymer (see Reference 7 and the disoussion of it given
above). Fibrinogen adsorption was observed to be much lower on P(NVP-g-TFE) than PTFE.
Albumin adsorption to P(NVP-g-TFE) was much greater than to PTFE, however. Prothrom-
bin adsorption was the same on PTFE and on P(NVP-g-TFE). The effect of the NVP graft
on immunoglobulin G (IgG) adsorption was complex. At intermediate degrees of NVP graft,
lgG adsorption was lower than PTFE but it was almost the same as PTFE at low or high
percent grafting.
Table 2
COMPARISON OF THE KINETIC CONSTANTS OF THE
INTERACTIONS OF FIVE DIFFERENT PLASMA PROTEINS WITH
CUPROPHANE® AND POLY VINYL CHLORIDE8
Cuprophane® Polyvinyl chloride
Max surface Max surface

Plasma proteins Cone in blood cone cone
(mol wt) (g/ - 0 Kd (nM) (pmol/cm2) Kd (nM) (pmol/cm2)
Albumin 35—45 869 1.77 869 3.14

(69,000)
Fibrinogen 2 -4 .5 397 1.06 326 2 .1 2
(340,000)
Immunoglobulin G 8— 18 394 1.64 35 2.40
(160,000)
Thrombin 0 300 2.26 382 8 .11
(36,600) 7190 12.40
Antithrombin III 0.17—0.30 692 9.32 969 16.71
(65,000)
a Copied from Reference 25 by permission of the copyright owner, Pergamon Press.
From Chuang, H. Y. K., Crowther, P. E., Mohammed, S. F., and Mason, R. G., Thromb. Res.,
14, 273, 1984. With permission.
The decreased fibrinogen adsorption observed for PNVP grafts to PTFE is analogous with
similar effects caused by PHEMA grafts to Silastic® and polyethylene noted above. The
enhanced albumin adsorption on pNVP/PTFE seems to be consistent with observations of
enhanced albumin adsorption to hydrogels made by others.
Brynda and co-workers have studied protein uptake by ungrafted PHEMA coatings.29-30
In the first study, coatings on glass were prepared by pouring HEMA solution in acetone-
methanol into chromic acid cleaned glass tubes and allowing solvent evaporation.29 In the
second study, chromic acid treated polyethylene “ foils” were coated with PHEMA in methyl
cellosolve.30 To avoid the possible uptake of unbound l25I iodide in l25I protein preparations
by hydrogels, these authors attempted to use fluorescently labeled proteins in their first
study.29 However, problems were encountered with the fluorescent labeled proteins. Thus,
the fluorescent label was released by the protein and sorbed by the gel in some cases (e.g.,
dansylated or fluorescein isothiocyanate labeled albumin) and preferential adsorption of the
labeled protein occurred with others (e.g., fluorescamine labeled albumin). Neither the time
course of adsorption nor the adsorption isotherm obtain with fluorescamine labeled albumin
on P(HEMA-co-MMA) copolymers displayed any plateau or saturation level. This study
resulted in far more information about the pitfalls of using fluorescently labeled proteins in
adsorption studies to hydrogels than about the adsorption behavior of proteins on PHEMA.
As discussed below, we have developed what appears to be a reliable method for using l25I
proteins to study adsorption to hydrogels, using extensive dialysis and excess unlabeled
sodium iodide to prevent adsorption of l25I iodide.
In their second study, Brynda et al. have tried to avoid all these problems by using
unlabeled proteins.30 The procedure involved the use of polyethylene or HEMA-coated
polyethylene bags (10 x 20 cm) filled with Ringers solution and then injected with a protein
stock solution. The solution in the bags was “ stirred” by keeping the bags in motion.
Adsorption was determined by measuring the difference between the 280 nm absorbance of
the solutions in the bags and of reference solutions. Adsorption at 30, 60, or 120 min was
found to be the same and the results were therefore reported as equilibrium values. Fibrinogen
adsorption (in (xg/cm2) from 1 mg/m€ human fibrinogen was reported to be 0.73 ± 0.006
on polyethylene, 0.4 ± 0.14 on cellophane, and 0.23 ± 0.01 on PHEMA-coated pol-
yethylene. Therefore, the results are consistent with those of others discussed above in
demonstrating lower protein uptake by the hydrogel surface.
The spatial distribution of albumin, y-globulin, and fibrinogen adsorbed on domain form-
ing HEMA-styrene copolymers has been examined with transmission electron microscopy
(TEM) by Sakurai et al.31 These polymeric systems appear to segregate into PHEMA or
polystyrene-rich domains since osmium stained samples have alternately light and dark areas
upon TEM examination. Osmium is thought to be taken up only by PHEMA. Some samples
were also osmium-stained and examined after first soaking them in single protein solutions
for periods ranging from 30 sec to 6 hr. The microscopic appearance of the surfaces changed
after protein treatment but each protein gave a distinct pattern. The round areas thought to
be polystyrene domains disappeared after fibrinogen or y-globulin treatment but
were still evident after albumin treatment. Control studies with the homopolymers showed
that albumin coverage of PHEMA homopolymer was much more extensive after the same
time of treatment than with the other proteins, while on polystyrene homopolymer, the
reverse was true. The authors concluded that the hydrophilic PHEMA domains preferred
to adsorb albumin, while y-globulin and fibrinogen preferred hydrophobic polystyrene do-
mains. The possibility that such local enrichment of proteins occurs even from complex
protein mixtures on all surfaces due to chemical heterogeneity of the surfaces merits some
attention in view of the frequent observation of the initiation of blood clotting at rather
localized sites on surfaces.32
Fibronectin adsorption to a wide variety of substrates measured with an antibody uptake
method was reported by Klebe.33 Polyacrylamide and PHEMA-coated tubes displayed no
uptake of fibronectin, whereas many hydrophobic polymers, e.g., polystyrene, polyme-
thylmethacrylate, had uptakes which were 50% or more of the amount observed on glass.
The addition of even trace amounts of poly(vinyl acetate) to PHEMA solutions used to coat
glass tubes greatly enhanced fibronectin adsorption, e.g., surfaces coated with 0.001%
poly(vinyl acetate) in PHEMA had 34% of the uptake of surfaces coated with pure poly(vinyl
acetate) solutions. The large effect of trace quantities of poly (vinyl acetate) probably reflects
localization of the poly (vinyl acetate) polymer at the outer surface of the coating. Surface
enrichment is expected with coatings made from polymer mixtures which are dissimilar in
their surface properties. The more hydrophobic species would be expected to prefer the air
interface while the more hydrophilic one should prefer the hydrophilic glass substrate.34
Lower adsorption of fibronectin to PHEMA compared to PEMA has also been observed in
our laboratory. Adsorption from a 0.022 mg/m€ fibronectin solution in phosphate buffer
saline to HEMA coated glass was only 40 ng/cm2 compared to 170 ng/cm2 on EMA coated
glass and 140 ng/cm2 on glass itself. Ikada et al. have also reported that fibronectin adsorption
to acylamide grafted on polyethylene decreases with increasing degree of graft.35
The kinetics of albumin and fibrinogen adsorption to Cuprophane® and Silastic® were the
subject of a detailed, comprehensive study by Bomzin and Miller.36 As observed in many
of the studies previously presented, the steady state or saturation levels of albumin and
fibrinogen were substantially less on Cuprophane® than on Silastic®. Albumin adsorption
was 200 ng/cm2 on Cuprophane® compared to 806 on Silastic. Fibrinogen adsorption was
244 ng/cm2 on Cuprophane® vs. 473 ng/cm2 on Silastic®. The kinetics appeared to be
diffusion controlled, i.e ., independent of surface type, except for the albumin-silicone rubber
system. Temperature effects on saturation values or kinetics were also negligible except in
the case of the albumin-silicone rubber system, where increased temperature increased the
rate of reaction significantly. The difference in the effect of temperature on albumin ad-
sorption to Cuprophane® and Silastic® was attributed to the presence of two types of reaction
sites on Silastic®, one of which is not diffusion controlled, i.e., a slower or weaker site.
The general weakness or absence of temperature effects on adsorption indicates that little
protein denaturation occurs on either hydrophilic Cuprophane® or hydrophobic Silastic®
during adsorption because bulk phase protein denaturation is known to be highly temperature
sensitive. In our studies of fibrinogen adsorption to the HEMA-EMA/PE series of copoly-
mers, we also observed relatively small effects of temperature.8 The coefficient b in the
Freundlich analysis, for HEMA at 4, 20, and 37°C was 0.51, 0.59, and 0.50, respectively,
while on EM A the b values were 0.29, 0.29 and 0.28 at these temperatures.
The adsorption of albumin to poly(vinyl alcohol) (PVA) films has also been reported by
Ikada.35 By increasing the temperature of annealing used in preparing the films, the water
content of the resulting film could be lowered. Decreases in water content were attributed
to increased cystalllinity induced by higher temperature treatments. Albumin adsorption
increased from about 0.01 |xg/cm2 at 60% water to about 0.18 |xg/cm2 at 25% water content
films. The increased adsorption to lower water content films was thought to be due to the
less diffuse, more ordered nature of the interface in lower water content films. That is, the
interfacial excluded volume is presumably lower on lower water content, more ordered films.
However, vinyl alcohol-ethylene copolymer surfaces grafted with dextran were found to
have much enhanced uptake of both albumin and fibrinogen. PVA grafts to this same substrate
had lower adsorption at intermediate graft levels which increased to near the original value
at higher grafts. Thus, while the importance of a diffuse hydrophilic interface on protein
uptake was stressed by Ikada, the available data is at best unclear on this question. Elam et
al. have also recently reported reduced fibrinogen adsorption and greater desorption on
dextran grafted to silicon surfaces in comparison to either oxidized, hydrophilic silicon or
silanized, hydrophobic silicon.37
Adsorption to hydrogels used in contact lens applications have been studied by a number
of groups because protein deposition can cause lens clouding and discomfort. Holly used
125I albumin of unspecified origin to study the adsorption from 1 mg/m€ bovine serum
albumin solutions in 0.9% saline after 15 hr.38 Polymers of 2-hydroxyethyl methacrylate
(HEMA), hydroxyethyl acrylate (HEA), glyceryl methacrylate (GMA), and acrylamide (AAm)
were studied. Albumin uptakes of 15.9 (HEMA), 18.6 (PGMA), 16.1 (PAAm), and 41.7
(PHEA) |xg/cm2 were reported. The water content of the polymers was given as 42% for
PHEMA and 84 to 87% for the three other polymers. The albumin adsorption reported is
far higher than monolayer levels and might indicate penetration of albumin as Holly spec-
ulated. However, the permeability of PHEMA to a variety of substances has been studied
and appears to be far too low to permit albumin penetration.39 Our experience has been that
if extensive dialysis is not performed to remove free iodide from 125I protein preps, erro-
neously high protein uptakes by hydrogels appear to occur due to iodide penetration and
retention. Holly’s results probably are an artifact due to failure to remove ,25I iodide from
his preparations.
Castillo et al. studied human serum albumin adsorption to contact lenses made from
HEMA or P(HEMA-co-MAA) copolymers using Fourier transform attenuated total reflec-
tance infrared spectroscopy (see also Volume II, Chapter 3).40 The lenses were exposed to
3.6 mg/m€ albumin solution in an isotonic buffer at 37°C for periods ranging from 1 to 50
hr rinsed for 1 min, and spectra were taken immediately, i.e., without drying, by pressing
the lens directly against an ATR (attenuated total reflectance) crystal. The spectra of the
adsorbed proteins differed from the protein in solution, but the adsorbed spectra also changed
with time of adsorption. The adosrbed protein spectra was more similar to that observed for
heat denatured albumin. The data therefore support the conclusion that adsorbed albumin
becomes denatured, but not completely nor immediately upon adsorption. Since the spectra
of albumin adsorbed to the higher water content P(HEMA-co-MAA) lenses approached that
for the heat denatured proteins more slowly than on HEMA, the denaturation of adsorbed
proteins appears to depend on hydrophilicity. These studies, if correct, provide an extremely
important insight into protein behavior at interfaces, particularly that the long proposed
structural rearrangements of adsorbed protein may occur slowly and at different rates on
different surfaces. Unfortunately, there is as yet no corroboration of these observations by
others, nor do we know how the spectra of proteins on nonhydrogels might change with
time. These workers have performed similar studies with lysozyme,41 mucin,42 and 7-
globulin.43 Structural transitions of irreversibly bound molecules directly in contact with the
polymer were detected, although the reversibly adsorbed protein appeared to be in its native
state.
Mucin adsorption to poly(N-vinyl pyrrolidone) (PNVP) grafted silicone contact lenses has
been measured using 14C-labeled protein and a gas flow chamber to detect radioactivity
continuously in situ.44 This study is entirely unique in the literature in its use of this
methodology, which requires very careful correction for radiation emanating from the bulk
phase beneath the contact lens. Adsorption kinetics revealed steady state was achieved in
approximately 10 hr. Adsorption isotherms displayed steadily rising adsorption throughout
the concentration range studied (0.01 to 0.2 mg/m€) on both surfaces, although less ad-
sorption occurred on silicone compared to PVP silicone surfaces. The lack of saturation or
even a decrease in slope of the isotherms at higher concentrations is inconsistent with most
protein adsorption isotherms in the literature. This may indicate incomplete correction for
bulk phase radioactivty since small errors in this regard are expected to be magnified at
higher protein concentrations. The enhanced adsorption to the PVP lenses was thought to
explain their enhanced wettability by tear solution, although how this could be distinguished
from the intrinsically greater wettability of PVP itself compared to silicon was not made
clear.
An allergic reaction is sometimes observed in contact lens wearers and it is therefore
thought that denaturation of proteins adsorbed to lenses could provide an antigenic stimulus.
The saline contact angles of hydrogels used in contact lenses after exposure to the tear fluid
proteins albumin, lysozyme, and 7-globulin have been reported to be decreased.45 After
boiling lenses exposed to albumin or 7-globulin, the contact angles increased. Adsorption
of proteins undoubtedly caused the initial decrease in contact angles. The increase in contact
angles after boiling was attributed to denaturation of the adsorbed proteins. The data therefore
provide some support for the possible allerginicity of adsorbed proteins.
III. ADSORPTION TO HYDROGELS FROM PROTEIN MIXTURES
The adsorption of proteins from mixtures to hydrogels has been the subject of several
studies, as summarized in Table 3. These studies represent various approaches to two major
questions. First, how does the adsorption from complex mixtures differ from that from single
protein solutions? Second, can understanding this process provide a more directly relevant
insight into the reactions occurring when a foreign material is placed in the biological
environment? As will be seen, these studies show that many different proteins adsorb to
hydrogels from complex protein mixtures, but the type and amount depends on the particular
surface under study. The adsorption properties of hydrogels exposed to protein mixtures
appear to have certain general properties, differing from nonhydrogels, including lower
adsorption and more rapid and complete desorption. In the remainder of this section, the
individual studies listed in Table 3 are summarized, critically reviewed, and discussed with
the purpose of substantiating these conclusions.
The adsorption of proteins from human plasma to Cuprophane® was studied by Limber
and collaborators with a column specially designed to hold particulate material.46,47 After
pre-equilibration of the ground, sieved Cuprophane® with degassed buffer, plasma was
pumped upward into the column, equilibrated for a chosen time (1 to 120 min) and then
buffer was pumped through the column. Samples taken after 150 or 1000 m€ of buffer had
Table 3
PROTEIN ADSORPTION TO HYDROGELS FROM MIXTURES OF PROTEINS
Proteins Surfaces Measurements Methods Comments Ref.
Albumin Cuprophane' Types of Elution Fibrinogen not detected 46

Immunoglobulin G Glass proteinElutability Electrophoresis Nine proteins detected on glass, three on
Factor XII Siliconized Immunodiffusion siliconized glass, one on Cuprophane® in
glass initial wash
One protein detected on glass, none on
siliconized glass, one on Cuprophane® in
final wash
Detergent eluates revealed two additional
proteins on siliconized glass
Proteins identified were albumin and IgG
Contact activation products detected on
glass
Albumin Cuprophane® (CP) Types of protein Elution Up to seven proteins detected, depending 47
Immunoglobulin G Albuminized Cp Electrophoresis on conditions
Heparinized Albuminization and heparinization in-
creased the number of detected proteins
Fibrinogen Silastic® (SR) Adsorption l25I Plasma fibrinogen adsorption reduced 7

Albumin PHEMA/SR Kinetics more on SR than on HEMA/SR
7 -Globulin NVP/SR Competition Alb/Fib and 7 -G/Fib competition similar
on all surfaces
Many HEMA Types of In situ Albumin, IgG, fibrinogen, other proteins 48

proteins radiolabeled observed
Electrophoresis Albumin appeared first (30-sec exposure)
Many HEMA-EMA/PE Types of In situ Variable uptake into protein, depending 49

copolymers proteins radiolabeled on polymer
Iodine uptake Elution
Electrophoresis
Many Silastic® (SR) Types of Elution Different proteins on AAm/Silastic® and 50
p-Acrylamide/SR proteins Electrophoresis Silastic®
Silver staining
Albumin Polyethylene (PE) Adsorption ,251 Hemoglobin greatly enriched over bulk 51
7-Globulin PHEMA/PE from plasma phase
Fibrinogen PEMA/PE Variable uptake into proteins, depending
Hemoglobin on polymer
Albumin preferred over fibrinogen on
PHEMA, reverse on PEMA
Fibrinogen HEMA-EMA/PE Adsorption ,25I Variable uptake for each protein, depend- 52
Immunoglobulin G from plasma ing on polymer
Albumin
Hemoglobin
Albumin PHEMA/PE Adsorption ,251 Decreasing adsorption kinetics for fibrino- 57, 58
IgG PEMA/PE kinetics from gen on PE and PEMA/PE
Fibrinogen Polyethylene plasma Increasing adsorption kinetics for fibrino-
Hemoglobin gen on PHEMA/PE
Fibrinogen Many Adsorption 125I Increasing kinetics on PHEMA, other po- 59

kinetics from lar materials
plasma
Albumin PEU/PEG Adsorption ,25I Decreasing kinetics on PEMA, other non-
Immunoglobulin G PEU/PPG kinetics from polar materials
Fibrinogen plasma Albumin and fibrinogen adsorption was 63
absent on hydrogel-like PEG/PEU
Lysozyme HEMA , 25]

Adsorption MAAc enhances lysozyme uptake greatly 66, 67
Albumin MMA from tear fluid
IgG NVP
MAAc
in
passed through the column were designated “ initial” and “ final” , respectively. These
samples were taken by collecting the effluent resulting from slight additional buffer passage
(buffer wash) or by switching to a buffer containing 0.1% Titron® X-100 detergent (detergent
eluate). The buffer washes and detergent eluates were analyzed by polyacrylamide gel
electrophoresis and immunodiffusion to detect the number and type of proteins present.
Fibrinogen was not detected in any of the washes or eluates, but it was also stated that the
detergent in the eluate prevented detection of fibrinogen even when fibrinogen was purposely
added to a sample. Furthermore, the only protein positively identified in the washes and
eluates was albumin. Other proteins sometimes observed could not be positively identified
because of technical limitations of the techniques used. Finally, no information was obtained
on the completeness of the elution so the authors could not rule out the presence of other
proteins that remained bound to the surfaces.
Despite these limitations, the studies of Limber et al. did reveal some relevant and
interesting facts about Cuprophane® interactions with plasma. Thus, while nine proteins
were present in the “ initial” buffer wash from glass and three from siliconized glass, only
one was detectable in the Cuprophane® buffer wash. Similarly, the initial detergent eluates
contained nine proteins from glass, five from siliconized glass, but again only one on
Cuprophane.® The final buffer wash from glass and siliconized glass had only one protein
and none were observed in the final Cuprophane® wash. The final detergent eluate from
Cuprophane® did reveal one protein, however, The detection of primarily albumin in the
washes and elutes from plasma-treated Cuprophane® compared to several other proteins
observed in the initial rinses from glass or siliconized glass suggests an affinity of albumin
for Cuprophane® which is similar to the enhanced affinity observed on HEMA and other
hydrogels in other studies (see below).
In their second study, Limber et al. showed that albumin or heparin pretreatment of
Cuprophane® increased the number of proteins detectable in washes and eluates following
the plasma equilibration step.47 Thus, while three proteins were detected in the initial wash
from Cuprophane® after plasma equilibration, up to six were detected if the Cuprophane®
had been pretreated with albumin or heparin. The additional proteins were not identified nor
was any explanation offered for the apparently enhanced protein binding of these treated
surfaces.
Several studies of protein adsorption from plasma to hydrogels have been performed in
my laboratory.748 52 We have also studied protein uptake from simulated tear fluid to hy-
drogels intended for use as contact lenses (see below). The plasma adsorption studies have
largely been concerned with defining the composition of the adsorbed protein layer which
forms on surfaces exposed to plasma and determining the influence of the surface chemistry
of the polymer on the composition of this layer. In pursuit of the latter information, a series
of polymers and copolymers made from HEMA and EM A have been used. The major
conclusions of these studies are summarized in Table 4. The basis for these conclusions will
now be elucidated by presentation of the studies in some detail.
The amount of fibrinogen adsorption onto Silastic from plasma (0.06 to 0.1 fig/cm2) was
found to be significantly depressed below its saturation value measured in buffer (0.65 fxg/
cm2) presumably because of competition from other components of the plasma.7 The ad-
sorption of fibrinogen onto PHEMA/Silastic® from plasma (0.05 to 0.16 (xg/cm2) was not
as greatly depressed relative to adsorption from buffer (0.3 |xg/cm2), however. Thus, fibri-
nogen adsorption onto Silastic® from plasma was found to be about the same as onto PHEMA/
Silastic®. To determine the basis for these differences, the competitive adsorption of fibri-
nogen against albumin and 7-globulin from binary mixtures was studied (see Figures 8 and
9). These experiments suggest that competition from these proteins may be a quantitatively
dominant factor in modifying surface adsorption of fibrinogen from plasma at normal con-
Table 4
PROTEIN ADSORPTION FROM PLASMA TO HEMA-EMA COPOLYMERS
Adsorption from plasma is markedly different than from solutions of single protein.
The adsorbed protein layer differs in composition from the bulk plasma composition.
The composition of the adsorbed protein layer is different on each polymer.
The composition of the adsorbed layer changes drastically during the early stages of the adsorption process.
Fibrinogen adsorption kinetics are markedly different on HEMA and EMA as well as on other nonpolar materials.
100
%of Pure Fibrinogen o Silastic
Adsorption 0 polyHEMA/Silastic
75 t:. polyNVP/Silastic
50
25
0;------r-----r-----r--- -.------.----,
0.01 0.1 1.0 10 100 1000 Weight/Weight
0.05 0.5 5.0 50.0 500 5000 Mole/Mole
AI bumin I Fibrinogen
FIGURE 8. Fibrinogen-albumin competitive adsorption onto Silastic®, PNVP/Silastic"', and

PHEMA/Silastic®. Untreated Silastic@ and films grafted with PNVP (2.7 mg/cm') or PHEMA
(5.8 mg/cm') were equilibrated in solutions containing 0.01 mg fibrinogen/m€ plus the concen-
tration of albumin corresponding to the ratio shown. The solvent was 0.01 M Hepes, 0.147 M
NaCI, 0.02% azide, pH 7.4. Equilibration was carried out at 37°C for 20 hr and was ended by
dilution displacement rinse followed by a 20-hr soak rinse with buffer. (From Horbett, T. A. and
Hoffman, A. S., Adv. Chem. Ser., 145, 230, 1984. With permission).
centrations of these proteins. The albumin/fibrinogen and -y-globulin/fibrinogen molar con-

centration ratios in plasma are approximately I 00 and 6, respectively. The competition
experiments show that about 60 and 30% reductions in plasma fibrinogen adsorption would
be expected at these ratios of albumin and -y-globulin, respectively. The 90% reduction in
fibrinogen adsorption expected from the combined competitive effect of these two proteins
is apparently reflected in the roughly 90% reduction in fibrinogen adsorption onto Silatic®
from plasma in comparison with its adsorption from buffer. However, neither protein com-
peted so differently to Silastic® and PHEMA/Silastic® as to account for the very different
behavior of these surfaces toward fibrinogen in plasma and in buffer. Thus, the plasma
fibrinogen experiments as well as the albumin and -y-globulin competition experiments
indicate the importance of other factors in significantly modifying fibrinogen adsorption onto
surfaces (see below).
The composition of the adsorbed protein layer which is formed on surfaces exposed to
plasma was therefore thought to be markedly different on PHEMA/Silastic® and Silastic.®
To investigate this possibility, a method for analyzing the eluates which overcame the
00 t:. Silastic
%of Pure Fibrinogen o polyHEMA/Silastic
Adsorption o polyNVP/Silastic
75
50
25
01-----.-----~----.-------~~
0.01 0.1 1.0 10 00 1000 Weight/Weight
0.02 0.2 2 20 200 2000 Mole/Mole
;r Globulin/Fibrinogen
FIGURE 9. Fibrinogen--y-globulin competitive adsorption onto Silastic®. PNVP/Silastic®, and

PHEMA/Silastic®. Untreated Silastic® and films grafted with PNVP (2.7 mg/cm') or PHEMA
(5.8 mg/cm') were equilibrated in solutions containing 0.01 mg fibrinogen/m€ plus the concen-
tration of -y-globulin corresponding to the ratio shown. The solvent was 0.01 M Hepes, 0.147 M
NaCI, 0.02% azide, pH 7.4. Equilibration was carried out at 37°C for 20 hr and was ended by
dilution displacement rinse followed by a 20-hr soak rinse with buffer. (From Horbett, T. A. and
Hoffman, A. S., Adv. Chem. Ser., 145, 230, 1984. With permission).
difficulties encountered by Limber and his eo-workers had to be developed. We, therefore,
began a series of studies using a new method in which adsorbed proteins are radiolabeled
in situ, eluated with sodium dodecyl sulfate (SDS), and separated by electrophoresis on
polyacrylamide gel electrophoresis in SDS (SDS-PAGE). The initial study showed that the
method was capable of detecting many more proteins adsorbed to surfaces than was possible
previously, e.g., nine on Teflon or PHEMA/Silastic® exposed to plasma. 48 We, therefore,
performed this type of analysis on three P(HEMA-co-EMA-g-E) copolymers as well as
PHEMA and P(EMA-g-E), with the results shown in Figure 10. 49
The iodograms obtained from the five different graft copolymers after both 0.5 min and
150 min plasma exposure had the major protein peaks seen and tentatively identified in
other studies with this method. 13 .48 The high molecular peak seen at the origin of the gel
(over 140,000 mol wt) is assigned to fibrinogen (peak 1); the next peak (#2, about 120,000
mol wt) to IgG; and the 13,000 mol wt peak (#9) to hemoglobin. These assignments are
based on the molecular weight observed for purified preparations of these proteins when run
nonreduced on SDS gels. 13.48 Reduction of the eluates with DTT followed by electrophoresis
showed patterns which are also consistent with these assignments. To quantitate the distri-
bution of 1251 into the various proteins adsorbed to the copolymers, the fraction of total
125 1 in each of nine molecular weight ranges was calculated. The observed protein peaks
always fell within these molecular weight ranges, as inspection of Figure 10 will show. Some
of the results obtained with films exposed to plasma for 150 min are summarized in
Figure 11.
Large and systematic differences in iodine uptake by the same protein adsorbed to different
surfaces were observed in this study. For example, the fraction of 1251 taken up by fibrinogen
(adsorbed from plasma after 150 min) was 50% on PHEMA, 65% on 3/1 P(HEMA-co-
EMA), and less than 5% on pEMA. The variation in iodine uptake probably reflects dif-
ferences in the composition of the adsorbed protein layer on the series of polymers. The
differences in 1251 distribution appear to be due partly to compositional differences in the
5
cpm -3 4
A.
sFexiO
3
0
4
3 B.
2
8 c.
6
0
10
8
3 E.
00 150 lOO 504030 20 10 5

MOLECULAR WEIGHT (xi0- 3 )
FIGURE 10. Iodograms of P(HEMA-co-EMA) copolymer

grafts on polyethylene exposed to plasma for 150 min. Letter
codes refer to the following monomer volume ratios of HEMA/
EMA used in preparing the films: A, 4/0; B, 3/1; C, 2/2; D,
113; and E, 0/4. (From Horbett, T. A., and Weathersby, P. K.,
J. Biomed. Mater. Res., 15, 403, 1984. With permission).
adsorbed layer, while others are due to differences in reactivity of the protein to the 1251
label. Differences in the composition of the adsorbed layer on Silastic® and polyacrylamide/
Silastic® have also been observed in a study performed by silver staining of the SDS-PAGE
gels of the eluates. 50 This technique is nominally sensitive enough to be routinely used in
place of the iodination technique but in practice it was found to be difficult to reproducibly
obtain the required sensitivity.
The in situ radioiodination method does not provide quantitative measurements of the
Peak I
00 M.W. = 140,000
%of
Total
125 I 60
Protein
40
20
0 25 50 75 100 +
PE
A
Peak 2
35 MW.= 120,000±20,000
%of 30
Total
125 I
25
Protein
20
15
10
0 25 50 75 100 PE
% EMA
FIGURE 11. Distribution of "'I in adsorbed protein after !50 min plasma exposure. The percent of total
"'I in four of the nine protein peaks observed is plotted against P(HEMA-co-EMA) copolymer film com-
position. The data for polyethylene are presented at the far right. (From Horbett, T. A. and Weathersby,
P. K., J. Biomed. Mater. Res., 15, 403, 1984. With permission).
Peak 4
M. W. = 72,000i8000
20
% of
Total
125I 15
Protein
10
0 25 50 75 100 t
PE
% EMA
FIGURE IIC.
Peak 9
M. W. = I 3,000 ±. 2500
25
"'o of
Total
125J 20
Protein
15
10
0 25 50 75 100 f
PE
"'o EMA
FIGURE IlD.
Amount 0.20
•• A
Adsorbed
()Jg/cm 2 l •• _,...·----;.::.-.
0.15 •• I :
••
,\. •..
. ••
010
005
,.·•
.•
"·~.
.......
I
..
I .•
• I
F
H
00
0 25 50 75
% EMA in Monomer Mixture
100 t
PE
FIGURE 12. Protein adsorption to HEMA-EMA-PE graft copolymers. Fibrinogen

( - - ) , immunoglobulin G (-·-)albumin (e), and hemoglobin (--)average adsorption
curves are plotted. Data for PE are also shown to the right (fibrinogen = F, immunoglobulin
= I, albumin = A, hemoglobin = H). (From Horbett, T. A., J. Biomed. Mater. Res .. 15,
673, 1984. With permission).
composition of the adsorbed layer because of nonuniform incorporation of iodine into pro-
teins. Albumin in particular was found to be poorly iodinated in the adsorbed layer. To
determine more accurately the magnitude of the difference in protein distribution on hydrogels
and nonhydrogels, a study of plasma protein adsorption to the P(HEMA-EMA)/PE copol-
ymers was done with the prelabeled protein technique. 51 •52
The adsorption of 1251 labeled proteins from plasma to the P(HEMA-co-EMA-g-E) series
of graft copolymers was measured after 2 hr of equilibration at 37°C. Fibrinogen, immu-
noglobulin G, albumin, and hemoglobin adsorption were measured in each set of experi-
ments, using the same plasma and polymer preparations. Four sets of such experiments were
done, each with different plasma and copolymer preparations, using freshly iodinated pro-
teins. The multiple adsorption data for each protein were averaged to obtain the data of
Figure 12. A characteristic trend in adsorption on the various copolymers is evident for each
protein, indicating substantial difference in the affinity of the proteins for each copolymer.
The enhanced adsorption of albumin to PHEMA evident in Figure 12 is similar to the
observations of Limber et al with Cuprophane® noted above. It appears that each of the
proteins has a different propensity to adsorb on each of the polymers. The result is a complex
variation in the composition of the adsorbed protein layer as a function of polymer com-
position. The adsorption of proteins from plasma to surfaces is therefore a competitive
process in which all of the plasma proteins probably participate to some degree. While
surface activity is a general property of all proteins, the surface activity of each protein
differs, so that the various proteins in plasma have a wide range in affinity and rate of
adsorption, Furthermore, the plasma concentration of each protein varies, so the influence
of mass action on adsorption is also important. As a result, the protein layer that adsorbs
to a polymer surface from plasma is a mixture, whose composition reflects many competitive
factors.
The P(HEMA-co-EMA) copolymer series used in this study varies in surface hydropho-
bicity, as demonstrated by the variable uptake of substantial amounts of water (30% or more
on PHEMA, 5% on PEMA). Since differences in the affinity of proteins for hydrophobic
moieties are used practically in the separation of proteins with hydrophobic chromatogra-
phy,53 the variable adsorption of plasma proteins to the P(HEMA-co-EMA) copolymers is
not surprising. The ability of polymers to fractionate the plasma proteins and concentrate
them at their surface is a key factor in the complex processes which determine the com-
patibility of polymers implanted in vivo or as supports for cells in vitro. Further research
is needed however, to delineate the role of specific plasma proteins in specific cellular
reactions with polymers.
The preceding studies were largely done with a fixed period of contact of the surface with
the protein mixture and provided little information on dynamic aspects of the adsorption
process from complex mixtures. However, the steady state or chronic interaction of foreign
materials with the biological environment may be strongly influenced by events occurring
very early in the process, i.e., the transient phase events may dictate results occurring much
later. Examples indicating the reasonableness of this hypothesis include the initial activation
of Factor XII by surfaces leading to the formation of a fibrin clot on glass surfaces,54 the
requirement for fibrinogen adsorption to achieve platelet adhesion from blood to some
surfaces,55 and the enhanced thrombogencity of surfaces treated with various proteins.56 The
kinetics of protein adsorption from plasma to the hydrogel and other polymers have therefore
been the subject of several studies in my laboratory.57'59
In our initial kinetic studies, adsorption from plasma was measured after ten different
exposure periods under static conditions, from 0.5 to 240 min.57 58 An equal volume of full
strength ACD plasma was added to the polymers immersed in CPBSz buffer and mixed by
repeated pipetting. The final plasma concentration was thus 50% of full strength. Hemo-
globin, immunoglobin G, fibrinogen, and albumin adsorption were measured. The adsorption
of each protein was measured on four polymers: PE, PHEMA/PE, P(HEMA-co-EMA)/PE,
and PEMA/PE. Some of the results are presented in Figures 13 to 16. Hemoglobin adsorption
to each of the polymers is similar in that it increases with adsorption time. However, the
actual amount adsorbed at any given time depends on the surface. Thus, the amount adsorbed
to the PE and PEMA/PE (0.04 and 0.09 fxg/cm2) is much higher than to the more hydrophilic
PHEMA/PE and P(HEMA-co-EMA)/PE (0.02 and 0.02 |jig/cm2). Hemoglobin adsorption
to PE appears to reach a steady level state at about 64 min. Adsorption to the grafted
polymers appears to increase slowly at longer times and thus steady-state adsorption levels
cannot be assigned for these polymers. On all polymers, hemoglobin adsorption is far higher
than one might expect for a protein present at such low levels (0.03 mg/m€ in plasma out
of 61 mg/m€ total protein concentration). This is due to the highly preferential adsorption
of this protein on surfaces, as previously reported.13 51 On all four polymers, however, the
amount present at the shortest equilibration period (0.5 min) is very low (0.01 |xg/cm2 or
less) and only gradually increases to the unexpectedly high value. Thus, the rate of hemo-
globin adsorption to surfaces from plasma is far below what diffusion would allow, suggesting
other phenomena are more important.
Immunoglobulin G adsorption to all polymers gradually increases with time. However,
the initial (0.5 min) adsorption to the more hydrophobic polymers PE and PEMA/PE (about
0.05 |xg/cm2) is much higher than to the more hydrophilic polymers PHEMA/PE and P(HEMA-
co-EMA)/PE (about 0.01 |xg/cm2). A steady-state adsorption level is observed for PE, but,
as with hemoglobin, the adsorption to the grafted copolymers continues to slowly increase
at longer times. For all polymers, the adsorption of IgG from plasma does not appear to be
diffusion limited.
Fibrinogen adsorption kinetics strongly depend on the polymer studied. The adsorption
to both the hydrophobic polymers PE and PEMA/PE is initially much higher than at longer
times and a steady-state adsorption level is reached after about 64 min. On PHEMA/PE,
however, fibrinogen adsorption is initially quite low but gradually and continuously increases
with time without reaching a steady-state value. The rapid initial decline in fibrinogen
HEMOGLDBIN ADSORPTION TO PE
AMOUNT
RDSDRBED
[ ~5 ./CM .2 J
~.IS
~.12
FIGURE 13. Hemoglobin adsorption from bovine plasma to PE (A). PEMA/PE (B), PEMA-PHEMA/PE (C),
and PHEMA/PE (D). Upper curves: after initial rinse; lower curves: after an additional overnight soak rinse.
HEM05LOBIN ADSORPTION TO P[HEMR-EMAJ/PE

AMOUNT
RDSORBED
[ ~5 ./CM .2 J
~.IS
~.12
~.~9
~.mi
:t
+ t
~.ID
+
~.~~
0 I~ 2~ 30 LJ0 S0 60 120 2LJ0
ADSORPTION TIME [MIN. J
FIGURE 13C.
HEM05LOBIN ADSORPTION TO PHEMA/PE

AMOUNT
RDSORBED
[ ~5./CM .2 J
H.IS
H.l2
H.H9
H.mi :1:
.
+
+
:1:
:t
H.ID
H.lm
0 m 20 30 lJH S0 ~ 120 2LJH
FIGURE 130.
IMMUNOGLOBULIN ADSORPTION TO PE
AMOUNT
AD50Rf3ED
[~5 ./CM .2 J
~-~~
~-Y~
+
~-]~
0.20
++ +
t i
+ :t:
~-~~ *
~-~~
~ I~ 2~ 3~ ~ m
~1!1 I~ 2Y0
A
IMMUND5LD8ULIN ADSORPTION TO PEMA/PE

AMOUNT
ADSDRf3ED
[ ~5 ./CM .2 J
~-~
+
+
+
+
*
~.lk:1
~-~ +----~---+---------+---+----+-----t
~ 10 2~ ]~ ~ ~0 50 120 2Y0
B
FIGURE 14. Immunoglobulin G adsorption from bovine plasma to PE (A), PEMA/PE (B), PEMA-PHEMA/PE
(C). and PHEMA/PE (D). Upper curves: after initial rinse: lower curves: after an additional overnight soak rinse.
IMMUNOGLOBULIN ADSORPTION TO P[HEMA-EMAJ/PE

AMOUNT
ADSDRBED
[ ~6 ./CM .2 J
~.~
+
+
m 20 30 40 s0 m !31
FIGURE 14C.
IMMUND5LD6ULIN ADSORPTION TO PHEMR/PE

AMOUNT
ADSORBED
[ ~6 ./CM .2 J +
~.sa +
PJ. y~
+
~.3~
±
~.2~ +
+
~.I~
0.~~
LJ0 ~ m 12~ 2~~
ADSORPTION TIME [M! N. J

FIGURE 140.
FIBRINOGEN ADSORPTION TO PE
AMOUNT
ADSORSED
[~G./CM .2 J
~-~~
+
+
* +
~- ~0 .f.--------------------~-----------<
0 11£1 ]~ 40 ~1£1 61£1 121£1 240
ADSORPTION T!ME [MIN. J
A
FIBRINOGEN ADSORPTION TO PEMR/PE

AMOUNT
ADSORSED
[~G./CM .2 J
~-~~
H.3H
H.2~ +
+ +
+ i
H. I~ +
i
H.~0 0+------------------------~
11£1 ~ ~ ~ ~ I~

B
FIGURE 15. Fibrinogen adsorption from bovine plasma to PE (A), PEMA/PE (B), PEMA-PHEMA (C). and
PHEMA/PE (D). Upper curves: after initial rinse; lower curves: after an additional overnight soak rinse.
FIBRINOGEN R D 50RPTI0N TO PEHEMR-EMR1/PE

RMOUNT
ADSORBED
FIBRINOGEN RD5QRPTI ON TO PHEHR/PE

RMOUNT
RD50RBED +• +
FIGURE 15D.
RLBUMIN RDSDRPTIDN TO PE
AMOUNT
ADSORBED
[~G./CM .2 J
1.~
B.EIB
B.ba
+
8.L!B +
+ + +
t +
i
8.28
8.~
~ 18 2~ ]~ ~ ~ ~ 128 2Y8
A
RL6UMIN RD5nRPTIDN TO PEMR ON PE

AMOUNT
ADSORBED
[~6 ./CM .2 J
1.0~
8.8B
+
0.~ t
+
+
+
+ +
+ + +
8.Y~
8.2~
8.~
~~------
I~ 3~ ~~ ~ -------
YB ~~~~--~
----~-- 128 ----1
2Y8
B
FIGURE 16. Albumin adsorption from bovine plasma to PE (A), PEMA/PE (B), PEMA-PHEMA/PE (C),
PHEMA/PE (0), Upper curves: after initial rinse; lower curves: after an additional overnight soak rinse.
ALBUMIN ADSORPTION TO PHEMA-EMR ON P[

AMOUNT
ADSDRBED
[~G./CM .2 J
1.~
+
~.m
~.fll
+
~.ttl
+
+ +
+
~.2H i
*
~.~
~ I~ 2~ ]~ t.t Slii Ei I~ 2~~
ADSORPTION TIME [MIN.J

FIGURE I6C.
ALBUMIN ADSORPTION TO PHEMA ON PE.

AMOUNT +
ADSDRBED +
+
[ ~n./ CM .2 J
1.~~
+
~.Em +
~.6~
+ t
~.tm
~.2~
~.~~ I
~ I~ 2~ ]~ ~ s~ ~ 12~ 2~~

FIGURE 160.
Table 5
FIBRINOGEN ADSORPTION TO POLYMERS FROM
BABOON PLASMA IN VITRO
Adsorption (p,g/cm2) at
Line Polymer 10 Sec 360 Sec
1 PEMA/PE 0.38 ± 0.003 0.11 0.006

2 PHEMA/PE 0.014 ± 0 . 0 0 1 1 0.032 ± 0.0018
3 PEMA,-HEMA,/PE 0.036 ± 0.0015 0.044 ± 0 . 0 0 1 1
4 PEMA.-HEMAyPE 0 . 0 1 1 ± 0.0009 0.024 ± 0.0008
5 PEMA/glass 0.085 0.013 0.086
6 PHEMA/glass 0.0091 0 .0 0 0 2 0 .0 2 0
7 PEMA,-HEMA,/glass 0 . 0 2 1 ± 0.0018 0.023
8 Polyethylene 0.088 ± 0.0025 0.044 ± 0.0024
9 Glass 0.0073 ± 0.0003 0 .0 1 2
10 Teflon® 0.052 ± 0.0006 0.028 ± 0.0062
11 Pellethane® 0.073 ± 0.0018 0.060 ± 0.0025
12 Superthane® 0.081 ± 0.0078 0.083 ± 0 .0 1 0
13 Renathane® 0.014 0.0008 0.051 ± 0.0016
14 Erythrothane® 0.061 ± 0.0045 0.094
15 PAAM/SR 0.068 ± 0 . 0 0 0 0 0.80 ± 0.034u
16 Silastic® 0.029 0.005 0.0050 ± 0 .0 0 1 0
a This measurement was made after 320 sec of adsorption instead of 360 sec.
adsorption on PE and PEMA/PE is reminiscent of the rapid loss of reactivity to antifibrinogen

by plasma exposed surface reported by Vroman and Adams.60
Albumin adsorption kinetics also strongly depend on the polymer. Thus, on PE and PEMA/
PE albumin adsorption appears to reach a peak of adsorption at 1 min, and declines rapidly
to a much lower steady state value at longer times. On PHEMA/PE and P(HEMA-co-EMA)/
PE, however albumin adsorption is initially quite low and gradually and continuously in-
creases, with no clear steady-state adsorption being reached.
In summary, the adsorption of hemoglobin and immunoglobulin G to each polymer
increased with time, although the absolute amounts depended on the polymer types. Fibri-
nogen and albumin adsorption on PHEMA/PE and P(HEMA-co-EMA)/PE also increased
with time, but on both PE and PEMA/PE there was a decrease with time following the
initial adsorption of these proteins. It appears that although fibrinogen and albumin react
more rapidly than others on these surfaces, they are eventually displaced to some degree by
the later deposition of the more slowly reacting proteins, hemoglobin, and immunoglobulin
G. The kinetics of adsorption of proteins from plasma are complex and appear to be strongly
dependent on the hydrophilicity of the adsorbing surface. It seems likely that the complexity
of the kinetics reflects differences in the nature of the protein-surface complex which is
formed.
In order to understand the relevance of these differences in kinetics of adsorption of plasma
proteins on hydrogels and nonhydrogel surfaces, we have also performed a comparative
study of in vitro and in vivo adsorption of baboon fibrinogen to a variety of surfaces for
which the thrombogencity measured by platelet consumptivity had already been determined.59
Fibrinogen adsorption kinetics from plasma in vitro are summarized in Table 5. For the
series of HEMA-EMA copolymers, the adsorption kinetics differed in a consistent way as
the composition changed. Fibrinogen adsorption was higher initially on PEMA/PE than at
later times. On PHEMA/PE, adsorption was low initially and increased to a final or plateau
level with time. On the intermediate copolymers HEMA,-EMA1/PE and HEMA3-EMAj/PE,
the adsorption kinetics were intermediate between the behavior of either EM A or HEM A,
and tended to be almost the same at all times. Fibrinogen adsorption to the substrate material
polyethylene (PE) begins to decrease within 20 sec and then approaches a plateau. Thus,
adsorption kinetics to the more hydrophobic polymers differed markedly from the adsorption
kinetics on the more hydrophilic surfaces of this series of polymers. The kinetics of fibrinogen
adsorption from plasma to HEMA-EMA/PE with baboon fibrinogen are thus essentially the
same as found originally with bovine fibrinogen, as already described.
The amounts of baboon fibrinogen adsorbed to other tubings after plasma-surface contact
for 10 or 360 sec from baboon plasma are also listed in Table 5. Whether the adsorption
increased or decreased as the surface-plasma contact time increased depended on the surface
studied. The amount of fibrinogen adsorption from plasma decreased with contact time for
hydrophobic, nonpolar surfaces, such as polyEMA, polyethylene, and Teflon. But for the
hydrophilic, more polar surfaces, the adsorption of fibrinogen from plasma increased as
contact time increased, as seen for poly HEM A, glass, Erythrothane, and Renathane.
The deposition of 125I-fibrinogen on a series of polymers from baboom blood was studied
at times ranging from 1 to 60 min using the in vivo arteriovenous shunt. The time course
of deposition to the polymers was basically of two types. Fibrinogen deposition on P(HEMA3-
EMA,)/PE is an example of the first type (Figure 17). There were two phases in the time
course of fibrinogen deposition on P(HEMA3-EMA1)/PE from nonheparinized animals. Within
the initial 6 min, the adsorption was very low and was the same as observed after heparinizing
the animal. The deposition in this initial time interval was lower than in vitro adsorption
from plasma. As blood contact time increased, the deposition proceeded into a second phase
in which surface retention of 125I fibrinogen increased rapidly. Red blood clots could be
seen in this stage. This second phase of high deposition was not observed or was greatly
depressed when the experiment was repeated after heparinizing the animal. Materials for
which two phases of fibrinogen deposition were observed from nonheparinized animals are
considered to be highly reactive or moderately reactive materials. Renathane and poly aery-
lamide/Silastic® were highly reactive materials. The HEMA-EMA-PE series of copolymers
were moderate to highly reactive materials, depending on the polymer composition.
Fibrinogen deposition on polyethylene in vivo is an example of the second type (see
Figure 17). There is only one phase in the time course of deposition, in both nonheparinized
and heparinized animals. The adsorption was low throughout the time course. Materials
with this type of fibrinogen deposition behavior are considered to be low reactivity materials.
Polyethylene, Pellethane,® Superthane,® and Silastic® exhibited low reactivity in this study.
The thrombogenicity of the polymers used in this study vary widely, as indicated by the
enhancement of the rate of platelet destruction (cannula platelet consumption) measured in
an in vivo arteriovenous shunt on baboons.6162 Polymers which were reactive with respect
to acute fibrinogen adsorption, i.e., fibrinogen adsorption underwent a large increase in later
times of adsorption, a process largely inhibited by heparin, also had relatively high chronic
platelet consumption. Conversely, materials inducing low platelet consumption such as
polyethylene had only a single phase of fibrinogen adsorption kinetics resulting in low
adsorption which was not markedly altered by heparin administration. Thus, even though
the time frame of fibrinogen adsorption events studied is far shorter than the more clinically
relevant periods used in platelet consumption studies, the initial events appear to be related
to the longer term processes.
The combined study of in vitro and in vivo fibrinogen adsorption to a series of polymers
varying greatly in thrombogenicity revealed clear differences in the kinetics of the process
which are generally well correlated with one another. The less polar, more hydrophobic
polymers, e.g., polyethylene, exhibited decreasing fibrinogen adsorption kinetics from plasma
in vitro and also had only a single phase of in vivo fibrinogen adsorption which was not
affected by heparin. More polar, more hydrophilic mateials, e.g., P(HEMA3-co-EMA,)/PE,
had increasing in vitro fibrinogen adsorption kinetics from plasma and a two phase in vivo
2.6
2 2.2
l'g/cm
1.8
14
1.0
0.6
0.2
•
0.06
t
l
004
!I f
0.02
10 20 30 40 50 60
TIME (min)
Amount
Deposited
(JJg/cm2)
0.06
0.04
0.02
10 20 30 40 50 60
Time (min.)
FIGURE 17. Kinetics of baboon fibrinogen deposition from blood in vivo to P(HEMA,-PEMAJPE (A) and PE
(B).
fibrinogen adsorption time course which was strongly affected by heparin. The ability of
polymers to influence the adsorption kinetics of fibrinogen therefore appears to be an im-
portant correlate of thrombogenicity. Whether the different kinetics indicate acutal differences
in the state of adsorbed fibrinogen which somehow cause the difference in thrombogencity
remains to be explored.
Brash and Uniyal have studied the kinetics of adsorption of albumin, immunoglobulin G,
and fibrinogen from human plasma to polyurethanes containing polyoxyethylene (PEG)
or polyoxypropylene (PPG) as well as polyethylene, glass, siliconized glass, and polysty-
rene.63 Initially high fibrinogen adsorption followed by a decrease to a lower, steady-state
value was observed on polyethylene in agreement with analogous observations made in our
studies, as noted above. However, adsorption to the hydrogel-like PEG containing polymers
had no parallel in our studies in terms of polymer type or results. The adsorption of albumin
and fibrinogen to these polymers was below the detection limit of the experiment, i.e.,
essentially zero, while values of 0.10 |xg/cm2 were observed for immunoglobulin G on both
PEG 600 (water content 17%) and PEG 1540 (water content 67%). On a hydrophobic
polyurethane containing polyoxypropylene, albumin and immunoglobulin G adsorption were
both found to be about 0.10 (Jig/cm2 but no fibrinogen adsorption was detected. The shortest
time used in these static adsorption experiments was 2 min, however. The flow studies done
in my laboratories allowed much shorter times (10 to 160 sec) to be studied and revealed
substantial changes in adsorption to some surfaces during this time interval. Thus, as noted
by Brash and Uniyal, it is quite possible that the adsence of fibrinogen on some surfaces
reflected a more rapid decrease in adsorption than their experiment could detect.
It appears that the results of Brash and Uniyal63 are generally consistent with those made
in my laboratories in that the patterns of adsorption kinetics depends strongly on the surface.
However, the effects of hydrophilic vs. hydrophobic surfaces on the kinetics differ somewhat
in the separate studies. For example, Brash and Uniyal observed decreasing fibrinogen
adsorption kinetics (or the implication of this when no fibrinogen was observed at all) for
all surfaces, whereas our data showed differences in the kinetics for hydrophilic and hy-
drophobic materials. However, since Brash and Uniyal used diluted plasma (20% of initial)
whereas we used undiluted plasma and both labs have subsequently observed strong effects
of plasma concentration on both the amounts and kinetics of adsorption,64 65 the differences
in the studies are probably partly due to this “ mass action” effect. Further studies will
therefore have to be done before the effect of hydrogel character on adsorption kinetics is
fully understood.
Two other studies of protein adsorption to hydrogels from protein mixtures have also been
done in our labs. These studies employed a simulated tear fluid for measurement of protein
adsorption to hydrogels used in contact lens applications. The three protein mixture (3.88
mg/m€ albumin, 1.6 mg/m€ immunoglobulin G, and 1.20 mg/m€ lysozyme in 0.01 M
citrate, 0.01 M phosphate, 0.12 M NaCl, 0.02% sodium azide, pH 7.4) is based on the
concentrations of these proteins found in normal tear fluid. The first study concerned ad-
sorption to polymers and copolymers of HEMA and MMA (methyl-methacrylate) intended
as a model for a new soft/hard lens with minimal protein uptake.66 More recently, we have
used an improved version of the same techniques to examine protein uptake by commercially
available contact lenses as well as chemical homologues of the commercial lenses67 (see
also Volume II, Chapter 3).
The P(HEMA-co-MMA) copolymers series were found to vary in the uptake of the three
proteins studied, but considerably more total protein was apparently taken up by HEMA-
rich polymers than by MMA-rich polymers. However, as was mentioned in the article, this
could have been an artifact due to “ penetration of free radioactive iodide disassociated from
the protein” . The studies were repeated with extensively dialyzed proteins and equilibration
of the surfaces in buffer containing 0.05 M Nal, both changes intended to eliminate iodide
Table 6
PROTEIN ADSORPTION TO CONTACT LENS MATERIAL3
Amount adsorbed (jjig/cm2)
Material11 Albumin Immunoglobulin G Lysosyme Total
PMMA 0.278 0 .101 0.144 0.523

P(60HEMA/4OMMA) 0.133 0.058 0.024 0.215
P(HEMA) 0.032 0.026 0 .0 1 0 0.068
P(80HEMA/20NVP) 0.027 0.016 0.023 0.066
P(60HEMA/40NVP) 0.028 0.0 2 1 0 .1 0 0 0.149
P(40HEMA/60NVP) 0.029 0 .021 0.227 0.277
P(90HEMA/10MAAc) 0.036 0.040 390 390
P(70HEMA/30MAAc) 0.038 0.106 374 374
P(80HEMA/20AAm) 0.016 0 .0 1 0 0.153 0.179
1 0 .1 2 2 0.0043 0.0065 0.133
2 0.108 0.018 0.013 0.139
3 0.069 0.0082 8 .2 0 8.277
4 0 .1 1 0 0.018 15.5 15.628
5 0.458 0.032 0.106 0.569
6 0.0104 0.0032 9.78 9.794
7 0.005 0.0015 12.7 12.706
8 0.017 0.0031 6 .2 2 6.240
a After exposure to simulated tear fluid for 2 hr followed by an initial rinse and a further overnight
soak rinse.
b Materials 1 to 8 are commercial contact lenses. The exact formulation are proprietary but the
major chemical components were as follows: 1 and 2 PHEMA; 3, HEMA-dimethyl-3-oxobutyl
acrylamide copolymers; 4, HEMA-MAAC-NVP terpolymer; 5, unknown; 6 to 8 , HEMA-meth-
acrylic acid copolymers.
artifacts. In this system, total protein uptake was less on HEM A than on MM A, a finding
consistent with other observations showing lower adsorption on hydrogels than on hydro-
phobe materials as discussed previously. However, the iodide used in this study appears to
have enhanced albumin adsorption to MMA for reasons that are not clear. These technical
factors therefore limit the usefulness of these data.68
More recent studies have also employed dialyzed proteins, but the buffer contained only
0.01 M Nal, a level found to have no effect on albumin adsorption.67 The major results of
these studies, shown in Table 6, are the tremendous range in total protein uptake largely
due to enormous uptake of lysozyme by some of the materials. Polymers containing meth-
acrylic acid (MAAc) adsorb great amounts of lysozyme, probably because of an ion-exchange
process. Lysozyme is a cationic protein with a high isoelectric pH (10.5 to 11.0) and therefore
carries considerable positive charge at the pH of the experiment (7.4), causing it to be
strongly adsorbed by MAAc. In polymers not containing MAAc and commercial lenses with
low total protein uptake, the relative distribution of albumin, immunoglobulin G, and ly-
sozyme still varies substantially. Thus, these hydrogel systems also appear to be capable of
fractionating the tear fluid protein, but by at least two distinctly different mechanisms, i.e.,
ion exchange vs. affinity differences to neutral polymers.
IV. SUMMARY AND CONCLUSIONS
Protein adsorption to hydrogels has been studied with a variety of proteins and hydrogels
and many aspects of the process have been characterized. These studies do not provide a
completely coherent or uniform body of information about the behavior of proteins on
hydrogels. However, certain important observations do stand out from the diverse sets of
data available which provide some answers to the questions posed in the introduction.
The types of data available on protein adsorption to hydrogels basically amount to quan-
titative measurements of the amount of a particular protein adsorbed under a variety of
conditions, including exposure to single protein solutions, to simple or complex mixtures,
and after various adsorption or desorption times. Little or no data on other important aspects
of the adsorption process to hydrogels are available, e .g ., the structure of an adsorbed protein
or the activity of an adsorbed enzyme. Studies directed at the mechanisms of adsorption
and the influence of the adsorbed protein on the reactions of the hydrogel with cellular
elements involved in the reactions of tissue to foreign implants are also lacking. The latter
is an especially important gap in the data base because studies of adsorption alone cannot
provide us with the necessary guidelines to productive, directed research on protein ad-
sorption. In this sense, the types of protein adsorption data available on hydrogels represent
too narrow an approach and frequently do not contribute information which is as relevant
as it could be. This field of research needs to move on to protein adsorption studies of
hydrogels done in the context of the influence of adsorptive events on cellular interactions
with hydrogels.
The studies I have reviewed provide a somewhat more complete answer to the question
of how adsorption to hydrogel and nonhydrogel surfaces differ. The most frequent observation
is a quantitative reduction in the amount of a particular protein adsorbed to the hydrogel in
comparison to a nonhydrogel, control (typically hydrophobic) surface. Inspection of the
comments on the studies summarized in Table 1 substantiate this conclusion, as does a
listing of the relatively few exceptions to it, e.g., thrombin on Cuprophane®,25 albumin on
NVP/PTFE.11 It also appears that the entire field of gel chromatography of proteins relies
on the low adsorptivity of the hydrogel-like matrices used, although direct quantitative
demonstration of how low the adsorption to such matrices is (in terms of p,g/cm2) is not
available.
Secondary to the reduced adsorption, qualitative differences in the nature of protein
adsorption to hydrogels compared to nonhydrogels appear to exist. Thus, desorption and
exchange studies indicate proteins are less tightly bound to hydrogels than to nonhydrogels.
This observation is tied closely to quantitative reductions in adsorption because almost all
studies cited used some kind of a rinse procedure to remove bulk protein prior to the adsorption
measurement, a step in which loosely bound protein is probably largely removed. Thus, as
Brash has speculated,17 hydrogels in equilibrium with the protein solution may have as much
protein held near the interface as do nonhydrogels but it is removed during the rinse. The
in situ adsorption measurements with the intrinsic fluorescence technique employed by
Gregonis et al.19 appear capable of resolving this question.
A further qualitative difference in protein adsorption to hydrogel and nonhydrogel surfaces
is the apparent preference of certain proteins, especially albumin, for hydrogel surfaces.
Thus, albumin appears to prefer HEMA domains in HEMA-styrene block copolymers,31
albumin adsorption to NVP/PTFE was higher than to PTFE,11 only albumin was detected
in eluates of Cuprophane® exposed plasma,46 and the amount of albumin adsorbed to
polyHEMA from plasma was much greater than fibrinogen adsorption while the reverse was
true on pEMA.52 The results for albumin adsorption to hydrogels suggest that although
adsorption is quantitatively less on hydrogels for some proteins, this is not true for all
proteins. Apparently, differences in the affinity of various proteins for hydrogel and non-
hydrogel surfaces results in shifts in the composition of the adsorbed layer. In this context,
it is especially important to consider that when adsorption occurs from a mixture of proteins
such as plasma, the reduced adsorption to hydrogels for certain proteins necessarily provides
an opportunity for the adsorption of other proteins which may normally not compete well.
That is, adsorption sites normally occupied by these proteins are now “ open” . Since the
intrinsic surface activity of various proteins differs substantially, the reduction in the amount
of adsorption of many of the proteins is accompanied by a qualitative change in the nature
of the substrate which may be more adsorptive for other proteins.
These quantitative and qualitative differences apparent in protein adsorption to hydrogels

raise again the final question posed initially, namely, what is the significance of these
differences? Perhaps the most interesting answer to this question is that the data indicate
that it may well be possible to produce clinically useful biomaterials with very low protein
adsorption and potentially much enhanced biocompatibility. Thus, while I have concluded
that the quantitative reductions in adsorption of some proteins on certain hydrogels apparently
leaves the door open to adsorption of still others and causes a qualitative change in the
adsorbed layer formed from complex mixtures, it may be possible to go beyond this case
to one in which adsorption of all or most proteins are greatly reduced. Therefore, I suggest
the general trends evident in the studies of adsorption to hydrogels I have reviewed are
correct but can be carried much further than generally realized. The basis for this suggestion
is the partly reduced adsorption evident in hydrogels such as HEMA which are relatively
low in water content and cross-linked significantly, compared to the great reductions in
adsorption to polyethylene glycol containing polymers19 63 and the high water content gel
matrices used in chromatography.
Little comparative, quantitative data on protein adsorption to polymers which exhibit very
low adsorption are available. As a result, the factors in a polymer such as polyethylene
glycol which might result in lower adsorption have not yet been elucidated. In a general
way, however, it seems likely that the combination of low affinity for any of the sites in
the polymer with high chain mobility would result in generally low adsorption. Chain mobility
minimizes the chances for multipoint attachment thought to be required for protein adsorption.
The highly hydrated but neutral nature of the polyethylene glycol molecule would be expected
to greatly reduce the strength of its binding to proteins, while the presence of only single
bonds in the chain and lack of sterically inhibitory side chains presumably favor a high
degree of chain mobility. Other polymers with these properties would also be expected to
be essentially nonadsorbing, e.g., polyacrylamide. The preparation and testing of polymers
of this type in the form of useful biomaterials, perhaps as surface coatings, thus represents
an interesting approach to improved biocompatibility. The examination of protein adsorption
to polymers of this type provides a convenient means to evaluate new materials and should
therefore prompt much more extensive and careful measurements of adsorption to polyeth-
ylene glycol like systems.
A C KNO W LEDGM EN TS
The financial support of the N. H. L. B. I. through grant HL19419 made the preparation
of this article possible and is gratefully acknowledged. Discussions with my colleagues
Buddy D. Ratner, John Brash, and Don Gregonis were also helpful. Hanson Y. K. Chuang
deserves special thanks for allowing me to reproduce some of his work and even providing
useable proofs of the necessary figures.
REFERENCES
1. Horbett, T. A., Protein adsorption on biomaterials, Biomaterials: Interfacial Phenomena and Applications,
Vol. 199, ACS Symposium Series, Cooper, S. L. and Peppas, N. A., Eds., American Chemical Society,
Washington, D.C., 1982, 233.
2. Cooper, S. L., Young, B. R., and Lelah, M. D., The physics and chemistry of protein-surface interactions,
in Interaction of the Blood with Natural and Artificial Surfaces, Salzman, E. W., Ed., Marcel Dekker,
New York, 1981, 1.
3. Brash, J. L., Protein interactions with artificial surfaces, in Interaction of the Blood with Natural and
Artificial Surfaces, Salzman, E. W., Ed., Marcel Dekker, New York, 1981, 37.
4. Hoffman, A. S., Principles governing biomolecule interactions at foreign interfaces, J. Biomed. Mater.
Res. Symp., 5, 77, 1974.
5. Andrade, J. D., Lee, H. B., Jhon, M. S., Kim, S. W., and Hibbs, J. B., Jr., Water as a biomaterial,
6 . Porath, J., Molecular sieving and non-ionic adsorption in polysaccharide gels, Biochem. Soc. Trans., 7,
1197, 1979.
7. Horbett, T. A. and Hoffman, A. S., Bovine plasma protein adsorption on radiation-grafted hydrogels
based on hydroxyethyl methacrylate and A-vinyl-pyrrolidone, in Applied Chemistry at Protein Interfacesf
Advances in Chemistry Series, Vol. 145, Baier, R. E., Ed., American Chemical Society, Washington,
D.C., 1975, 230.
8. Weathersby, P. K., Horbett, T. A., and Hoffman, A. S., Fibrinogen adsorption to surfaces of varying
hydrophilicity, J. Bioeng., 1, 395, 1977.
9. Hoffman, A. S. and Harris, C., Radiation-grafted hydrogels on silicone rubber surfaces — a new biom-
aterial, Polym. Prepr. Am. Chem. Soc. Div. Polym. Chem., 13, 740, 1972.
10. Ratner, B. D. and Hoffman, A. S., Radiation grafted hudrogels on silicone rubber as new biomaterials,
Am. Chem. Soc. Org. Coat. Plast. Chem. Prepr., 33, 386, 1973.
11. Boffa, G. A., Lucien, N., Faure, A., Boffa, M. C., Jozefonviz, J., Szubarga, A., and Larrieu, M.
J., Polytetrafluoroethylene-A-vinylpyrrolidone graft copolymers: affinity with plasma proteins, J. Biomed.
Mater. Res., 11, 317, 1977.
12. Weathersby, P. K., Horbett, T. A., Hoffman, A. S., and Kelly, M. A., Surface analysis of methacrylate
graft copolymers varying in hydrophilicity, J. Bioeng., 1, 381, 1977.
13. Horbett, T. A., Weathersby, P. K., and Hoffman, A. S., The preferential adsorption of hemoglobin to
polyethylene, J. Bioeng., 1, 61, 1977.
14. Brash, J. L., Uniyal, S., and Samak, Q., Exchange of albumin adsorbed on polymer surfaces, Trans.
Am. Soc. Artif. Int. Organs, 20, 69, 1974.
15. Brash, J. L., Szota, P., and Thibodeau, J. A., Interactions of plasma with glass: identification of adsorbed
proteins, Trans. Soc. Biomat., 7(Abstr.), 264, 1984.
16. Mosesson, M. W., Galankis, D. K., and Finlayson, J. S., Comparison of human plasma fibrinogen
subfractions and early plasmic fibrinogen derivatives, J. Biol. Chem., 249, 4656, 1974.
17. Brash, J. L. and Uniyal, S., Dependance of albumin-fibrinogen simple and competitive adsorption on
surface properties of biomaterials, J. Polym. Sci. Polym. Symp., 6 6 , 377, 1979.
18. Whicher, S. J. and Brash, J. L., Platelet-foreign surface interactions: release of granule constituents from
adherent platelets, J. Biomed. Mater. Res., 12, 181, 1978.
19. Gregonis, D. E., Buerger, D. E., Van Wagenen, R. A., Hunter, S. K., and Andrade, J. D., Polyethylene
glycol) surfaces to minimize protein adsorption, Trans. Soc. Biomat., 10 (abstr.) 266, 1984.
20. Hiatt, C. W., Shelokov, A., Rosenthal, E. J., and Galimore, J. M., Treatment of controlled pore glass
with polyethylene oxide) to prevent adsorption of rabies virus, J. Chromatogr., 56, 362, 1971.
21. Regnier, F. E. and Noel, R., Glycerolpropylsilane bonded phases in the steric exclusion chromatography
of biological macromolecules, J. Chromatogr. Sci., 14, 316, 1976.
22. Fukano, K., Komiya, H., Sasaki, H., and Hashimoto, T., Evaluation of new supports for high-pressure
aqueous gel permeation chromatography: TSK-Gel SW type columns, J. Chromatogr., 166, 47, 1978.
23. Somack, R., McKay, V. S., and Giles, J. W., Biological applications on Spherogel TSK-SW-Type Gel,
in Size Exclusion Chromatography, Vol. 138, ACS Symposium Series, Provder, T ., Ed., American Chemical
Society, Washington, D.C., 1980, 285.
24. Chuang, H. Y. K., King, W. F., and Mason, R. G., Interaction of plasma proteins with artificial surfaces:
protein adsorption isotherms, J. Lab. Clin. Med., 92, 483, 1978.
25. Chuang, H. Y. K., Crowther, S. F., Mohammed, S. F., and Mason, R. G., Interactions of thrombin
and antithrombin III with artificial surfaces, Thromb. Res., 14, 273, 1979.
26. Chuang, H. Y. K., Mohammad, S. F., Scharma, N. C., and Mason, R. G., Interaction of human
alpha-thrombin with artificial surfaces and reactivity of adsorbed alpha-thrombin, J. Biomed. Mater. Res.,
14, 467, 1980.
27. Chuang H. Y. K. and Mitra, S. P., Adsorption and reactivity of human prothrombin to artificial surfaces,
J. Biomed. Mater. Res., 18, 695, 1984.
28. Chuang, H. Y. K., In situ immunoradiometric assay of fibrinogen adsorbed to artificial surfaces, J. Biomed.
Mater. Res., 18, 547, 1984.
29. Brynda, E., Drobnik, J., Vacik, J., and Kalal, J., Protein sorption on polymer surfaces measured by
fluorescence labels, J. Biomed. Mater. Res., 12, 55, 1978.
30. Brynda, E., Cepalova, N. A., and Stol, M., Equilibrium adsorption of human serum albumin and human
fibrinogen on hydrophobic and hydrophilic surfaces, J. Biomed. Mater. Res., 18, 685, 1984.
31. Sakurai, Y., Akaike, T., Kataoka, K., and Okano, T., Interfacial phenomena in biomaterials chemistry,
in Biomedical Polymers, Polymeric Materials and Pharmaceuticals for Biomedical Use, Goldberg, E. P.
and Nakajima, A., Eds., Harcourt Brace Jovanovich, New York, 1980, 335.
32. Dutton, R. C., Baier, R. E., Dedrick, R. L., and Bowman, R. L., Initial thrombus formation on foreign
surfaces, Trans. Am. Soc. Artif. Int. Organs, 14, 57, 1968.
33. Klebe, R. J., Bentley, K. L., and Schoen, R. C., Adhesive substrates for fibronectin, J. Cell. Physiol.,
109.481.1981.
34. Ratner, B. D., Graft copolymer and block copolymer surfaces, in Surface and Interfacial Aspects of
Biomedical Polymers, Vol. 1, Andrade, J. D., Ed., Plenum Press, New York, 1985, 373.
35. Ikada Y., Blood-compatible polymers, in Advances in Polymer Science, Vol. 57, Dusek, K., Ed., Springer-
Verlag, Berlin, 1984, 103.
36. Bornzin, G. A. and Miller, I. F., The kinetics of protein adsorption on synthetic and modified natural
surfaces, J. Colloid Interface Sci., 8 6 , 539, 1982.
37. Elam, J. H., Nygren, H., and Stenberg, M., Covalent coupling of polysaccharides to silicon and silicon
rubber surfaces, J. Biomed. Mater. Res., 18, 953, 1984.
38. Holly, F. J., Protein and lipid adsorption by acrylic hydrogels and their relation to water wettability, J.
Polym. Sci. Polym. Symp., 6 6 , 409, 1979.
39. Ratner, B. D. and Miller, I. F., Transport through crosslinked poly(2-hydroxyethyl methacrylate) hydrogel
membranes, J. Biomed. Mater. Res., 7, 353, 1973.
40. Castillo, E. J., Koenig, J. L., Anderson, J. M., and Lo, J., Characterization of protein adsorption on
soft contact lenses, I. Conformational changes of adsorbed human serum albumin, Biomaterials, 5, 319,
1984.
41. Castillo, E. J., Koenig, J. L., Anderson, J. M., and Lo, J., Protein adsorption on hydrogels. II. Reversible
and irreversible interactions between lysozyme and soft contact lens surfaces, Biomaterials, 6 , 346, 1985.
42. Castillo, E. J., Koenig, J. L., Anderson, J. M., and Jentoft, N., Protein adsorption on soft contact
lenses. III. Mucin, Biomaterials, 7, 9, 1986.
43. Castillo, E. J., Koenig, J. L., and Anderson, J. M., Characterization of protein adsorption on soft
contact lenses. IV. Comparison of in-vivo spoilage with the in-vitro adsorption of tear proteins, Biomaterials,
7, 89, 1986.
44. Baszkin, A., Proust, J. E., and Boissonnade, M. M., Adsorption of bovine submaxillary mucin on
silicone contact lenses grafted with poly(vinyl pyrrolidone), Biomaterials, 5, 175, 1984.
45. Refojo, M. F., and Holly, F. J., Tear protein adsorption on hydrogels: a possible cause of contact lens
allergy, Contact Intraocul. Lens Med. J., 3, 23, 1977.
46. Limber, G. K., Glenn, C. H., and Mason, R. G., Studies of the proteins elutable from certain artificial
surfaces exposed to human plasma, Thromb. Res., 5, 735, 1974.
47. Limber, G. K. and Mason, R. G., Studies of proteins elutable from cuprophane exposed to human plasma,
Thromb. Res., 6 , 421, 1975.
48. Weathersby, P. K., Horbett, T. A., and Hoffman, A. S., A new method for analysis of the adsorbed
plasma protein layer on biomaterial surfaces, Trans. Am. Soc. Artif. Int. Organs, 22, 242, 1976.
49. Horbett, T. A. and Weathersby, P. K., Adsorption of proteins from plasma to a series of hydrophilic-
hydrophobic copolymers. I. Analysis with the in situ radioiodination technique, J. Biomed. Mater. Res.,
15.403.1981.
50. Hoffman, A. S., Horbett, T. A., Ratner, B. D., Hanson, S. R., Harker, L. A., and Reynolds, L. O.,
Thrombotic events on grafted polyacrylamide-Silastic surfaces as studied in a baboon, in Biomaterials:
Interfacial Phenomena and Applications, Vol. 199, ACS Advances in Chemistry Series, Cooper, S. L. and
Peppas, N. A., Eds., American Chemical Society, Washington, D.C., 1982, 59.
51. Horbett, T. A., Weathersby, P. K., and Hoffman, A. S., Hemoglobin adsorption to three polymer
surfaces, Thromb. Res., 12, 319, 1978.
52. Horbett, T. A., Adsorption of proteins from plasma to a series of hydrophilic- hydrophobic copolymers.
II. Compositional analysis with the prelabeled protein technique, J. Biomed. Mater. Res., 15, 673, 1981.
53. Kissing, W. and Reiner, R. H., Chromatographic investigations into the interaction between proteins and
polymer matrices. II. Chromatography of serum proteins on substituted agarose gels, Chromatographia,
11, 83, 1978.
54. Griffin, J. H., Surface-dependent activation of blood coagulation, in Interaction of the Blood with Natural
and Artificial Surfaces, Salzman, E. W., Ed., Marcel Dekker, New York, 1981, 139.
55. Zucker, M. B. and Vroman, L., Platelet adhesion induced by fibrinogen adsorbed onto glass, Proc. Soc.
Exp. Biol. Med., 13, 318, 1969.
56. Young, B. R., Lambrecht, L. K., Cooper, S. L., and Mosher, D. F., Plasma proteins: their role in
initiating platelet and fibrin deposition on biomaterials, in Biomaterials: Interfacial Phenomena and Ap-
plications, Vol. 199, ACS Advances in Chemistry Series, Cooper, S. L. and Peppas, N. A., Eds., American
Chemical Society, Washington, D.C., 1982, 317.
57. Horbett, T. A., The kinetics of adsorption of plasma proteins to a series of hydrophilic-hydrophobic
copolymers, ACS Org. Coat. Plast. Chem. Prepr., 40, 642, 1979.
58. Horbett, T. A., The kinetics of adsorption of plasma proteins to a series of hydrophilic-hydrophobic
copolymers, in Adhesion and Adsorption of Polymers, Part B, Lee, L. H., Ed., Plenum Press, New York,
1980, 677.
59. Horbett, T. A., Cheng, C. M., Ratner, B. D., Hanson, S. R., and Hoffman, A. S., The kinetics of
baboon fibrinogen adsorption to polymers: in vitro and in vivo studies, J. Biomed. Mater. Res., in press,
1986
60. Vroman, L. and Adams, A. L., Identification of rapid changes at plasma-solid interfaces, J. Biomed.
Mater. Res., 3, 43, 1969.
61. Hanson, S. R., Harker, L. A., and Ratner, B. D., and Hoffman, A. S., Evaluation of artificial surfaces
using baboon arteriovenous shunt model, in Biomaterials 1980, Advances in Biomaterials, Vol. 3, Winter,
G. D., Gibbons, D. F., and Plenk, H., Eds., John Wiley & Sons, New York, 1982, 519.
62. Hanson, S. R., Harker, L. A., Ratner, B. D., and Hoffman, A. S., In vivo evaluation of artificial
surfaces with a nonhuman primate model of arterial thrombosis, J. Lab. Clin. Med., 95, 289, 1980.
63. Uniyal, S. and Brash, J. L., Patterns of adsorption of proteins from human plasma onto foreign surfaces,
Thromb. Haemostas. (Stuttg.), 47, 285, 1982.
64. Horbett, T. A., Mass action effects on competitive adsorption of fibrinogen from hemoglobin solutions
and from plasma, Thromb. Hemostas., 51, 174, 1984.
65. Brash, J. L. and ten Hove, P., Effect of plasma dilution on adsorption of fibrinogen to solid surfaces,
Thrombos. Haemostas. (Stuttg.), 51, 326, 1984.
66. Royce, F. H., Jr., Ratner, B. D., and Horbett, T. A., Adsorption of proteins from artificial tear solutions
to poly(methyl methacrylate-2-hydroxyethyl methacrylate) copolymers, in Biomaterials: Interfacial Phe-
nomena and Applications, Vol. 199, ACS Advances in Chemistry Series, Cooper, S. L. and Peppas, N.
A., Eds., American Chemical Society, Washington, D.C., 1982, 453.
67. Ratner, B. D. and Horbett, T. A., Protein adsorption to contact lens materials, Trans. Soc. Biomat.,
7(Abstr.), 76, 1984. (Abstract)
6 8 . Royce, F. H., Adsorption of proteins from artificial tear solution and oxygen transport studies on copolymer
contact lens materials, M. S. thesis, University of Washington, Seattle, 1981.

INDEX
A Antithrombin III, adsorption of

to Cuprophane, 137— 140
to poly (vinyl chloride), 129, 137— 140
AAm, see Acrylamide
Antithrombogenic agents, 102
Achromobacter
Applications of hydrogels, see also specific applica-
aceris, 106
tions, 2, 113— 119
butyri, 106
Archimedes’ buoyancy principle, 45
Acrylamide (AAm), 97
Arthrobacter
Acrylic acid, 97
oxydans, 106
Acrylics, 97
simplex, 106— 107
Actinomycetes, 106
Artificial organs, 102, 108
Activation energy, diffusional, 64— 65
l -Asparaginase, 110, 113, 115, 117— 118
Activation entropy, 6 6
Association, 20
Adsorption isotherms, 134— 136
Auger spectroscopy, 91
Aggregation, 20
Average number of cross-links per chain, 11
AIBN, see Azobisisobutyronitrile
Avisco wet gel, 75
Albumin, 167
Avrami equation, 23
adsorption of
Azobisisobutyronitrile (AIBN), 6
to Cuprophane, 129, 131, 137— 142, 144, 146
to PEU/polyethylene glycol, 129, 145
to polyacrylamide, 131, 142 B
to polyethylene, 129, 137, 145, 153, 160—
161, 163
to poly (glyceryl methacrylate), 131, 142 Bacillus, 106
to poly(hydroxyethyl acrylate), 131, 142 Benzoyl peroxide, 6
Binary multicomponent diffusion coefficient, 59
to poly(2 -hydroxyethyl methacrylate), 128—
Biocompatibility, 2, 96
132, 141— 142, 152, 165— 166
to poly(2 -hydroxyethyl methacrylate-co-ethyl Bio Gel P, 137
Bioreactor, 102, 117— 118
methacrylate-g-E), 152
to poly(2 -hydroxyethyl methacrylate-co-ethyl Biosensor, 102, 115— 117
methacrylate)/poly ethylene, 145, 153, 160— Blob model, 70, 75
161, 163 Blocking factor, 78
to poly(2 -hydroxyethyl methacrylate-co-methyl Bond percolation model, 92
methacrylate), 130, 142— 143, 165— 166 Bovine serum albumin, diffusion through poly(vinyl
to poly(2 -hydroxyethyl methacrylate)/polyethyl- alcohol), 75
ene, 153, 160— 161, 163 Branching, degree of, 23
to poly(2-hydroxyethyl methacrylate)/Silastic, Brevibacterium ammoniagenes, 105— 107
144, 146— 147 Bulk method, for hydrogel preparation, 7
to poly (methyl methacrylate), 145, 165— 166
to polystyrene, 141
to polytetrafluoroethylene, 130, 138— 139
c
to polyurethane, 137
to poly(vinyl alcohol), 131, 142 Candida
to poly (vinyl chloride), 137— 140 lipolytica, 106
to poly(A-vinyl pyrrolidone), 129, 138— 139 tropicalis, 106
to poly(A-vinyl pyrrolidone)/Silastic, 146— 147 Cell culture, 119
to Silastic, 141— 142, 146— 147 Cellophane, protein adsorption to, 141
denatured, 142— 143 Cells, immobilized, see also specific organisms,
immobilized, 1 1 0 1 0 2 , 106— 110
Alkaline protease, 104 Cellulose, irradiation of, 14

d - Amino acid oxidase, 104 Ceric ions, 98
Amino acids, 102 Chain extension, 36
Aminoacylase, 104 Chain mobility, 65
Aminoethyl methacrylate, 97 at surface, 86—89
Amylase, 110 Chain scission, 18— 19
Analytical applications, of immobilized molecules Chain transfer, 6 , 8
and cells, 11A— 117 Chapman and Enskog analysis, 58
Antibodies, immobilized, 102, 105, 116— 118 Chemical potential 30, 59—60, 62
Antigens, immobilized, 102, 105, 116 Citrate synthetase, 111
Classification of hydrogels, 2 irreversible thermodynamic equations for, 60—61

Clostridium butylicum, 106 nonisothermal, 59
Clustering, 20 through nonporous hydrogel, 63—78
Concentrated hydrogel, 77 polymer structure and, 76—78
Contact angle goniometer, 101 through porous hydrogel, 78—80
Contact angle hysteresis, 88—89, 92 Diffusion coefficient, 11, 61
Contact lens, 142, 165 determination of
allergic reaction to, 143 by free-volume theories, 65—72
Converted polymers, 100 by scaling theories, 72—76
Copolymerization/cross-linking reactions, 6 — 12 by statistical mechanics, 64— 65
kinetic mechanism of, 8 —9 Diffusion equations, analysis of mass transport by,
molecular weight distribution of copolymer, 9— 58—60
10 Dilatometry, 23
Correlation length of mesh size, 54 Dilute hydrogel, 77
Corynebacterium glutamicum, 106 Dissipation factor, 60—61
Covalent attachment to hydrogels, 109— 113 Drug delivery systems, 102, 117— 119
Creep experiments, 28 Dynamic mechanical tests, 28
Critical mesh size, 68— 69
Cross-linking, 2
by chain fracture, 4 E
copolymerization/cross-linking reactions, 6 — 1 2
degree of, 23, 77 Effective diffusion coefficient, 78—79
intermolecular, 18— 19 Effective number of cross-linked subunits, 28
intramolecular, 18— 19 EGDMA, see Ethylene glycol dimethacrylate
irradiation-induced, 15— 17 Elastic free energy, 29—35
by side-chain fracture, 5 Elastic modulus tests, 28
Cross-linking agents, 4 Electron beam accelerator, 13, 96—97
Cross-linking density, 28—29, 77 Electron beam irradiation, 12— 13
of poly(2-hydroxyethyl methacrylate), 50 Electron diffraction, 91
of poly(vinyl alcohol), 47, 49 Electron microscopy, 23
Cross-linking ratio, 4 Electron spectroscopy for chemical analysis
of poly(2-hydroxyethyl methacrylate), 50—52 (ESCA), 89—90, 101
of poly(vinyl alcohol), 47—50 Electrostatic potential, 59
Cross-linking reactions, 2— 8 Elongation, 53
Crotonic acid, 97 Emulsion method, for hydrogel preparation, 7
Crystalline phase, diffusion through, 77—78 Encapcel, 108
Crystallinity, degree of, 23 Encapsulation, 108
Crystallites, 20—21, 48 End-group analysis, 23
Crystallization, 21—23 End-linking, 4, 15— 16
Cumyl peroxide, 6 End-to-end distance, 28, 34— 36, 42, 54
Cuprophane Enskog theory, 63
pretreatment of, 146 Entanglements, 5— 6 , 20—21
protein adsorption to, 129— 131, 137— 146 Enthalpic parameter, 46
Cyclization, 16 Entrapment in hydrogels, 102— 108
Entropic parameter, 46
Entropy, elastic contribution to, 31— 33
D Enzacryl, 110, 116
Enzyme biosensor, 115— 116
Decomposition of polymers, 15 Enzyme-linked immunoassay, 116
Degradation of polymers, 15, 17— 18 Enzymes, see also specific enzymes
Detergent eluate, 146 immobilized, 102— 105, 108— 114, 118— 119
Detour ratio, 78 microenvironmental effects on, 1 1 2
Dextran, protein adsorption to, 131, 142 Equilibrium polymer volume fraction, 30, 36, 38—
Diagnostic applications, of immobilized molecules 40
and cells, 114— 117 of poly(vinyl alcohol), 49
Diethyl hexyl phthalate, 48 Equilibrium swelling theory
Differential scanning calorimetry, 23 Gaussian models in, 28—34
Diffusion non-Gaussian models in, 34— 38
Fickian, 61, 78 Equilibrium volume swelling ratio, 29
generalized Fickian equations for, 60 Equilibrium weight swelling ratio, 29
generalized Stefan-Maxwell equations for, 58—59 Erythrothane, protein adsorption to, 162— 163
ESC A, see Electron spectroscopy for chemical to polyacrylamide, 130— 131, 141
analysis to poly(2-hydroxyethyl methacrylate), 130, 141
Escherichia coli, 105— 108 to poly (methyl methacrylate), 141
Ethylene glycol dimethacrylate (EGDMA), 8 , 50 to polystyrene, 141
structure of, 3, 97 Fickian diffusion, 61, 78
Expanded state, 20 Fickian equations, generalized, 60
Expansion coefficient, 70 Flory-Huggins equations, 36
Expansion factor, 30—33 Flory polymer-solvent interaction parameter, 30—
Extension factor, of poly(vinyl alcohol), 53 31, 36, 38—39, 45^17
Eyring’s rate theory, 63, 65 Fluorescent immunoassay, 116
Fluorinated polyethylene, irradiation of, 14
Forced diffusion, 60
F Fourier transform attenuated total reflectance in-
frared spectroscopy, 142
Factor XII, adsorption to glass, 144 Fourier transform infrared spectroscopy, 101
Ferrous salts, 96 Free energy of mixing, 29—31, 35
Fibrinogen Free radicals, 97—98
adsorption of produced by irradiation, 15— 19
to carbon, 130 Free-volume theory, 11
to cellophane, 130, 141 determination of diffusion coefficient by, 65—72
to Cuprophane, 130, 137— 142, 146 Freezing point depression, 46—47
to dextran, 131, 142 Freundlich coordinates, 134
to Erythrothane, 162— 163 Freundlich equation, 134
to glass, 162— 163 Front factor, 42—44
to Pellethane, 162— 163
to PEU/polypropylene glycol, 129, 145
to polyacrylamide/Silastic, 162— 163
G
to polyethylene, 129— 130, 141, 145, 153,
158— 159, 162— 164 Gaseous method, for hydrogel preparation, 7
to poly(ethyl methacrylate), 129, 134 Gas evolution, during irradiation of polymers, 14,
to poly (ethyl methacrylate)/polyethylene, 145, 17
153, 158— 159, 162— 163 Gaussian models, in equilibrium swelling theory,
to poly(2 -hydroxyethyl methacrylate), 128— 28—34
129, 132, 141, 162— 163 Gelation, 11, 16
to poly(2 -hydroxyethyl methacrylate-co-ethyl Gel fraction, 28
methacrylate), 148— 149 Gel permeation chromatography, 23
to poly(2 -hydroxyethyl methacrylate-co-ethyl Gel point, 11, 16
methacrylate-g-E), 134, 152 Glass, protein adsorption to, 144, 162— 163
to poly(2 -hydroxyethyl methacrylate-co-ethyl Glass transition temperature, 11— 12, 23, 28
methacrylate)/polyethylene, 153, 158— 159, Globular orders, 20
163— 164 y-Globulin
to poly(2 -hydroxyethyl methacrylate)/polyethyl- adsorption of
ene, 153, 158— 159, 162— 163 to poly(2 -hydroxyethyl methacrylate), 128,
to poly(2-hydroxyethyl methacrylate)/Silastic, 132, 141
133— 134, 146— 148 to poly(2 -hydroxyethyl methacrylate)/polyethyl-
to polystyrene, 141 ene, 145
to polytetrafluoroethylene, 138— 139 to poly(2-hydroxyethyl methacrylate)/Silastic,
to polyurethane, 137, 165 147— 148
to poly (vinyl chloride), 129, 137— 140 to polystyrene, 130, 141
to poly(A-vinyl pyrrolidone), 138— 139 to poly (A-vinyl pyrrolidone), 144
to poly(A-vinyl pyrrolidone)/Silastic, 147— 148 to poly(A-vinyl pyrrolidone)/Silastic, 147— 148
to poly(A-vinyl pyrrolidone-g-tetrafluoroethy- to Silastic, 129, 147— 148
lene), 133 denatured, 143
to Renethane, 162— 163 Glucoamylase, 104— 105
to silane, 131 Gluconobacter
to Silastic, 131, 141— 142, 144, 148, 162— melanogenes, 106
163 oxydans, 106
to silica, 131 Glucose oxidase, 104— 105, 110, 114, 118
to Teflon, 162— 163 Glucose-6 -phosphate dehydrogenase, 111
degraded, 136 a-Glucosidase, 105
Fibronectin, adsorption of (3-Glucosidase, 104
Glyceryl methacrylate (GMA), 97 to poly(2-hydroxyethyl methacrylate), 128, 132,

Grafting to support, 96— 100 165— 166
Gravity gradient, 59—60 to poly(2 -hydroxyethyl methacrylate-co-ethyl
G value, 15— 16 methacrylate), 148— 149
to poly(2 -hydroxyethyl methacrylate-co-ethyl
methacrylate-g-E), 152
H to poly(2 -hydroxyethyl methacrylate-co-ethyl
methacrylate)/polyethylene, 145, 153, 156—
Hansenula polymorpha, 106 157, 162— 163
HDEEMA, see Hydroxydiethoxyethyl methacrylate to poly(2 -hydroxyethyl methacrylate-co-methyl
Heat transfer, 60 methacrylate), 165— 166
HEEMA, see Hydroxyethoxyethyl methacrylate to poly(2 -hydroxyethyl methacrylate)/polyethyl-
Helmholtz free energy, 38—44 ene, 153, 156— 157, 163
HEMA, see 2-Hydroxyethyl methacrylate to poly(methacrylic acid), 145
Hemoglobin, adsorption of to poly(methyl methacrylate), 165— 166
to poly(ethyl methacrylate)/polyethylene, 153— to polytetrafluoroethylene, 130, 138— 139
155, 163 to poly(vinyl chloride), 129, 137— 140
to poly(2 -hydroxyethyl methacrylate-co-ethyl to poly(A-vinyl pyrrolidone), 138— 139, 145
methacrylate), 148— 149 Industrial applications, of immobilized molecules
to poly(2 -hydroxyethyl methacrylate-co-ethyl and cells, 118— 119
methacrylate-g-E), 152 Infrared spectroscopy, 23, 91, 101
to poly(2 -hydroxyethyl methacrylate-co-ethyl Inhomogeneities, 20
methacrylate)/polyethylene, 153— 155, 163 Initiators, 6 — 8 , 96
to poly(2 -hydroxyethyl methacrylate)/polyethyl- Insulin, 117— 119
ene, 153— 155, 163 Interaction energy density, 45
to polyethylene, 153— 155, 163 Interpenetrating polymeric networks (IPN), 2
Inverse gas chromatography, 46—47
Heparin, 110
Invertase, 104— 105
Hexokinase, 111
Iodide penetration, 142
Hole distribution function, 6 8
Ionic charge, 59
Hormones, 102
Ionic hydrogels, 2
Hydraulic permeability, 63
Ionic strength, 132
Hydrodynamic radius of solute, 6 6 — 6 8 , 73—74
Ionizing radiation, 12— 16
Hydrodynamic theories, 63
IPN, see Interpenetrating polymeric networks
Hydroxydiethoxyethyl methacrylate (HDEEMA), 3
Irreversible process, theory of, 60—61
Hydroxyethoxyethyl methacrylate (HEEMA), 3
Isotropic extension, 36
2-Hydroxyethyl methacrylate (HEMA), 3, 8 , 50
Hydroxymethyl methacrylate, 97
Hyphomicrobium, 106 j
Hysteresis effects, 20
Jump frequency, 64
I Jump length, 64— 65
Ideal network, 4, 19—20 K

Immobilization in hydrogels
by adsorption, 108— 109 Kedem-Katchalsky theory, 62—63
applications of, 113— 119 Kloekera, 106
by combination techniques, 1 1 1
by covalent attachment, 109— 1E3
by entrapment, 102— 108 L
microenvironmental effects in, 112— 113
molecules and cells used in, 1 0 1 — 1 0 2 Lactobacillus
Immunoassay, 102, 116— 117 bulgaricus, 106
Immunoglobulin G, adsorption of casei, 106
to Cuprophane, 137— 140, 144 delbrueckii, 106
to glass, 144 Lactose dehydrogenase, 114
to PEU/polypropylene glycol, 145 Lamellar single crystals, 21
to polyethylene, 153, 156— 157, 163 Light scattering, 23, 46
to poly (ethyl methacrylate)/polyethylene, 145, Loops, 5— 6 , 20—21
153, 156— 157, 163 Lysozyme
adsorption of diffusion in, 63—78

to poly(2 -hydroxyethyl methacrylate), 131, Nucleic acids, 102
145, 165— 166 Number average degree of polymerization, 10
to poly(methacrylic acid-co-2 -hydroxyethyl Number average molecular weight between cross-
methacrylate), 133, 165— 166 links, 4, 10, 13, 23, 77
to poly (methyl methacrylate), 165— 166 determination of
denatured, 143 rubber elasticity studies, 51—54
swelling studies, 28—31, 36—41, 44— 51
of poly(2-hydroxyethyl methacrylate), 50—52, 55
M of poly (vinyl alcohol), 47—48, 53—54
Number average molecular weight of uncross-linked
MAAc, see Methacrylic acid polymer, 10, 30—31, 69
Macromolecular relaxation, 77 NVP, see A-Vinyl pyrrolidone
Macroporous hydrogel, 63
diffusion through, 78—80
Macrosyneresis, 20 o
Maleic poly butadiene, cells immobilized in, 107
Mass transfer theory Onsager reciprocal relations, 61—63
continuum approach to, 58—61 Organelles, 102, 105, 107— 108
simplifications for ordinary diffusion in hydro- Osmometry, 46
gels, 61—63 Osmotic pressure, 46, 62
MDEEMA, see Methoxydiethoxyethyl methacrylate Oxygen, effect on irradiation-induced cross-linking,
MEEMA, see Methoxyethoxyethyl methacrylate 15
MEMA, see Methoxy ethyl methacrylate
Membrane osmometry, 23
Mesh size, 68—69, 73—77 P
determination of, 54
of poly(2-hydroxyethyl methacrylate), 55 PAAm, see Polyacrylamide
Methacrylic acid (MAAc), 3, 132, 166 PAN, adsorption of proteins to, 130
Methanogens, 106 Pancreas, artificial, 108
Methoxydiethoxyethyl methacrylate (MDEEMA), 3
Papain, 105, 110
Methoxyethoxyethyl methacrylate (MEEMA), 3
Pellethane, protein adsorption to, 162— 163
Methoxyethyl methacrylate (MEMA), 3
Penicillium
Methylene-bis-acrylamide, 97
chrysogenum, 106
Microbead, 103, 117
cyaneofulvum, 106
Microcapsule, 108, 117
2,4-Pentadiene-l-ol, 97
Microporous hydrogel, 63
Peptide synthesis, 102
diffusion through, 78—80
Permeability coefficient, 76
Microscopy, 101
Microsyneresis, 20 Permeability of hydrogels, 62—63
Mobility factor, 11 Permselectivity, 63
Molecular weight, 23 Peroxides, 97—98
Monomers, 3, 7, 96—97 PEU/polyethylene glycol
Mucin, adsorption of adsorption of proteins to, 129
to poly (A-vinyl pyrrolidone), 131, 143 protein adsorption to, 145
to silicone, 131 PEU/polypropylene glycol, adsorption of proteins
Multicomponent diffusion coefficient, 61—62 to, 129
Mutual diffusion coefficient, 71—72 Phase separation, 20
PHEMA, see Poly(2-hydroxyethyl methacrylate)
Phenomenological coefficients, 61
N Photoelectron spectroscopy, 91
Planck equation, 6 6
Nemst-Planck equations, 61—62 Plasma method, for hydrogel preparation, 7, 100
Network defects, 5— 6 , 19—20 Plasmapheresis device, 117
Network structure, 19—20 Plasma proteins, see also specific proteins
Neurotransmitters, 102 adsorption of
Neutral hydrogels, 2 to Cuprophane, 143— 146
Nocardia rhodocrous, 106— 107 to poly (ethyl methacrylate), 146
Non-Gaussian models, in equilibrium swelling the- to poly(2-hydroxyethyl methacrylate), 146
ory, 34— 38 to poly(2 -hydroxyethyl methacrylate-co-ethyl
Nonporous hydrogel, 63 methacrylate)/polyethylene, 152— 153
to poly(2-hydroxyethyl methacrylate)/Silastic, grafted to silicone rubber, 89—90

146— 147 protein adsorption to, 128— 132, 141— 147, 152,
to Silastic, 146— 147 162— 166
Platelet destruction, 163 structural characteristics of, 50—52, 55
Polyacrylamide (PAAm), 168 surface characteristics of, 87—90
antibodies and antigens immobilized in, 105, 116 swelling characteristics of, 2 2
cells immobilized in, 104— 107 Poly(2-hydroxyethyl methacrylate-co-ethyl methac-
enzymes immobilized in, 103— 105, 110— 111, rylate), 144— 145, 148— 151, 153— 164
114 Poly(2-hydroxyethyl methacrylate-co-ethyl methac-
grafted to silicone rubber, 90 rylate-g-E), 152— 153
irradiation of, 14 Poly(2-hydroxyethyl methacrylate-co-ethyl methac-
organelles immobilized in, 107 rylate)/polyethylene, 145
protein adsorption to, 130— 131, 141— 142, 145 Poly(2-hydroxyethyl methacrylate-co-methacrylic
proteins immobilized in, 1 1 0 acid), 2 2
surface characteristics of, 89—90 Poly(2-hydroxyethyl methacrylate-co-methoxyethyl
Poly(acrylamide-A,A'-methylenebis acrylamide), methacrylate), 75
cells immobilized in, 105 Poly(2-hydroxyethyl methacrylate-co-methyl
Polyacrylamide/Silastic, 149, 162— 163 methacrylate)
Poly aery late, 14, 104 protein adsorption to, 130, 142— 143, 165— 166
Poly(acrylic acid), 104 swelling charcteristics of, 2 2
Poly [ 1-acrylopiperidine-4-spiro-2'-( 1' ,3 '-dioxyacry- Poly(2-hydroxyethyl methacrylate-co-A-vinyl pyrrol-
lopentane)], 103 idone), 2 2
Poly(acryloylmethoxyamine), 103 Poly (2-hydroxyethy 1 methacry late)/polyethy lene,
Poly( 1-acryloy 1-4-piperidine), 103 153— 163
Polyamide, 14, 96 Poly(2-hydroxyethyl methacrylate)/Silastic, 146—
Poly(diethylene glycol dimethacrylate), 105 148
Polyester, 14, 96 Poly(hydroxypropyl methacrylate), 116
Polyethylene Poly isobutylene, 14
irradiation of, 14 Polymers
protein adsorption to, 129— 130, 137, 141, 153— bulk, entrapment in, 102— 108
164 conversion of, 1 0 0
Polyethylene glycol, 168 Polymer volume fraction, 45, 14— 77
cells immobilized in, 107 of poly(vinyl alcohol), 47
on quartz, 137 Poly (methacrylamide), 14, 106
Polyethylene glycol dimethacrylate, 105, 107— 108 Poly(methacrylate), 14
Polyethylene oxide/polyethylene terephthalate, 96 Poly(methacrylic acid), 100
Poly (ethyl methacrylate), 129, 134, 146— 147, grafted to polypropylene, 98—99
162— 163 grafted to polypropylene-polyethylene copolymer,
Poly(ethyl methacrylate)/polyethylene, 153— 163 98— 100
Poly(glyceryl methacrylate) protein adsorption to, 145
protein adsorption to, 131, 142 Poly(methacrylic acid-co-2-hydroxyethyl methacry-
surface characteristics of, 89 late), 132— 133
Poly(glycidyl methacrylate), 109 Poly(methoxydiethoxyethyl methacrylate), 22
Poly(hexanediol monomethacrylate), 105 Poly(methoxyethoxyethyl methacrylate), 22
Poly(hexyl methacrylate), antigens, 116 Poly(methoxypolyethylene glycol dimethacrylate),
Poly(hydroxyalkyl methacrylate), 6 , 110 107— 109
Poly(hydroxydiethoxyethyl methacrylate), 22 Poly(methylacrylate), 100
Poly(hydroxyethoxyethyl methacrylate), 22 Poly (N, N '-methy lenebisacry lamide), 103— 104
Poly(hydroxyethyl acrylate) Poly(methyl methacrylate)
antigens immobilized in, 116 cells immobilized in, 109
cells immobilized in, 109 enzymes immobilized in, 105
enzymes immobilized in, 104— 105 protein adsorption to, 141, 145, 165— 166
protein adsorption to, 131, 142 swelling characteristics of, 2 2
surface characteristics of, 89 Polynitrile, 100
Poly(2-hydroxyethyl methacrylate) (PHEMA), 2, Poly(octyl methacrylate), 116
128, 140 Polyoxyethylene, 165
antigens immobilized in, 116 Poly(oxyethylene dimethacrylate), 105
cells immobilized in, 104, 106— 109 Polyoxyethylene glycol, 137
diffusion through, 75, 79 Polyoxypropylene, 165
enzymes immobilized in, 103, 105, 111 Polyoxypropylene glycol, 137
grafted to Silastic, 133 Polypropylene, 14
Polypropylene glycol, 107 from protein mixtures, 143— 166

Polysiloxane, 14 from single protein solutions, 128— 143
Polystyrene to nonhydrogel surfaces, 167
cells immobilized in, 106 covalent binding to hydrogels, 1 1 0 — 1 1 2
irradiation of, 14 fluorescent labeled, 140
protein adsorption to, 130, 141 Prothrombin, adsorption of
Polytetrafluoroethylene (PTFE), 133 to Cuprophane, 130, 137— 140
protein adsorption to, 130, 138— 139 to PAN, 130
Poly(tetramethylolethane tetraacrylate), 109 to polytetrafluoroethylene, 130, 138— 139
Poly(trimethylolpropane trimethacrylate), 109 to poly(vinyl chloride), 130, 137— 140
Polyurethane, 100 to poly(/V-vinyl pyrrolidone), 138— 139
cells immobilized in, 106 Pseudobinary diffusion coefficient, 62
protein adsorption to, 137, 165 Pseudomonas, 106
Poly (vinyl acetate), 141 PTFE, see Polytetrafluoroethylene
Poly(vinyl alcohol) (PVA), 23 PVA, see Poly(vinyl alcohol)
cells immobilized in, 106 PVC, see Poly(vinyl chloride)
diffusion of bovine serum albumin through, 75
diffusion of theophylline through, 75
extension factor of, 53 R
heparinized, 75
irradiation of, 14 Radiation, ionizing
organelles immobilized in, 107 effects on macromolecules, 13— 14
protein adsorption to, 131, 142 types of, 12— 13
structural characteristics of, 47—50, 53—54 y-Radiation, 12— 13
swelling characteristics of, 2 2 Radiation grafting, 97—98
tensile strength of, 53 Radiation polymerization, 12— 19, 96, 103— 104,
Poly(vinyl chloride) (PVC), 48 107
adsorption of proteins to, 129— 130 in solution, 17— 19
irradiation of, 14 Radioimmunoassay, 116
protein adsorption to, 137— 140 Red blood cell substitute, 102
Poly(vinylidene chloride), 14 Reflection coefficient, 63
Poly(A-vinyl pyrrolidone), 132 Relaxed polymer volume fraction, 31
adsorption of proteins to, 130— 131 of poly(2-hydroxyethyl methacrylate), 50
enzymes immobilized in, 103— 105 Renethane, 162— 163
grafted to polytetrafluoroethylene, 133 Renkin equation, 79
grafted to Silastic, 133— 134 Retractive equilibrium force, 34
grafted to silicone rubber, 98 Rigidity factor, 54
irradiation of, 14 Rubber elasticity studies, 23, 51—54
protein adsorption to, 129— 130, 138— 139, Rubber elasticity theory, 38—44
143— 145
Poly(A-vinyl pyrrolidone-co-2-hydroxyethyl methac-
rylate), 1 1 0 s
Poly(A-vinyl pyrrolidone)/Silastic, 146— 148
Pore wall partition coefficient, 78 Saccharomyces
Porosity, 78 cerevisiae, 105— 107, 110
Porous hydrogel, diffusion through, 78—80 formosensis, 107— 109
Potassium persulfate, 103 Salicylic acid, 75
PP fiber, hollow, 117— 118 Scaling theory, determination of diffusion coeffi-
Pre-existing order, 20 cient by, 72—76
Preparation of hydrogels, 1—23, 96— 100 Sedimentation, 23
by bulk polymerization, 96 Self-diffusion coefficient, 70—71
by chemical cross-linking, 2— 13 Semicrystalline hydrogel, 20—23, 48
grafting to support, 96— 100 diffusion through, 77—78
by radiation, 12— 19 Semidilute hydrogel, 73, 77
Pressure diffusion, 60 Sephadex, 137
Probability of activation, 6 6 Sieving mechanism, 68—69, 73—74
Progesterone, 75 Silane, protein adsorption to, 131
Propagation, 8 Silastic, protein adsorption to, 129, 131, 141— 148,
Propylene glycol methacrylate, 97 162— 163
Proteins Silica hydrogel
adsorption of cells immobilized in, 107
protein adsorption to, 131 Teflon, 162— 163

Silicone, 131 TEMED, see Tetramethylethylenediamine
Sodium metabisulfite, 96 Tensile strength, of poly(vinyl alcohol), 53
Sodium styrene sulfonate, 97 Tensile tester, 51—54
Sol fraction, 28 Termination of polymerization, 8
Solute diffusion, 57—80 Tetrahymena pyriformis, immobilized, 106
generalized mass transfer theory, 58—63 Tetramethylethylenediamine (TEMED), 96
nonporous hydrogels, 63—78 Theophylline, 75
porous hydrogels, 78—80 Thermal diffusion coefficient, 59—60
Solute partition coefficient, 76—78 Thermal expansion coefficient, 70, 72
Solute permeability, 63 Thermal gradient, 60
Solution method, for hydrogel preparation, 7 Thermoelectricity, 60
Solution viscometry, 46 Thermogravimetric analysis, 23
Solvent, for cross-linking reactions, 4, 6 —7 Thermomechanical analysis, 23
Specific hole free-volume, 70—71 Theta hydrogel, 77
Specific interstitial free-volume, 70 Theta-temperature, 46
Spheres of influence, 70 Thiobacillus ferrooxydans, 106
Spherulites, 21 Thrombin, adsorption of
Statistical mechanics, determination of diffusion to Cuprophane, 129, 137— 140
coefficient by, 64— 65 to poly(vinyl chloride), 129, 137— 140
Stefan-Max well equations, generalized, 58—59 Thrombogenicity, 163— 165
Strain at break, 52—53 Tortuosity, 78
Streptococcus Transport coefficient, 58
faecalis, 106 Triton X-100 detergent, 146
thermophilus, 106 Trypsin, 110— 111
Streptokinase, 110
Streptomyces
clavuligeris, 107 u
phaerochromogenes, 104, 106— 107
Structure of hydrogels, 2, 19—20 Ultraviolet light, 98
determination of, 23 Unreacted functionalities, 20—21
Styrene-butadiene-styrene copolymer, 100 Unstrained state, 20
Supercooling, 20
Superheating, 20
Superthane, 162— 163 V
Support, grafting to, 96— 100
Supramolecular order, 20 VAc, see Vinyl acetate
Surface characteristics, 85—92, 100— 101 Van de Graaff accelerator, 13
chain mobility, 86—89 Vinyl acetate (VAc), 3
of polyacrylamide, 89—90 A-Vinyl pyrrolidone (NVP), 3, 97
of poly(glyceryl methacrylate), 89 Viscometry, 23
of poly(hydroxyethyl acrylate), 89 Viscous flow, 60
of poly(2-hydroxyethyl methacrylate), 87—90 Vitrification, 11— 12
water structuring, 89—92 Void space, 65—72
Suspension method, for hydrogel preparation, 7
Swelling
characteristics of selected hydrogels, 2 2 w
degree of, 28—34, 41, 6 8 , 73—74, 77
of poly(2-hydroxyethyl methacrylate), 50
Water content, 142, 168
isotropic, 30
Water structuring, 86—87, 89—92
Swelling methods, 28
Swelling ratio, 43—44
of poly(vinyl alcohol), 47, 50 X
Swelling studies, 23, 44— 51
Swelling theory, see Equilibrium swelling theory
Swollen polymer volume fraction, 31 X-ray diffraction, 23
of poly(2-hydroxyethyl methacrylate), 50 X-ray scattering, low angle, 23
T z
Tear fluid proteins, 146, 165— 166 Zygosaccharomyces lactis, 110

Nikolaos A. Peppas (Editor) - Hydrogels in Medicine and Pharmacy-Fundamentals-CRC Press (1986)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Nikolaos A. Peppas (Editor) - Hydrogels in Medicine and Pharmacy-Fundamentals-CRC Press (1986)

Uploaded by

Copyright:

Available Formats

Hydrogels

Boca Raton London New York

CRC Press is an imprint of the

Reissued 2019 by CRC Press

© 1986 by Taylor & Francis Group, LLC

No claim to original U.S. Government works

A Library of Congress record exists under LC control number:

ISBN 13: 978-0-367-24923-6 (hbk)

Hydrogels are cross-linked macromolecular networks swollen in water or biological fluids.

Nikolaos A. Pep pas, Sc.D., is a Professor of Chemical Engineering at Purdue University.

Barbara D. Barr-Howell, M.S. Steven R. Lustig, M.S.

PREPARATION METHODS AND STRUCTURE OF HYDROGELS

Nikolaos A. Peppas and Antonios G. Mikos

II. Hydrogel Preparation by Chemical Cross-Linking...................................................... 2

III. Hydrogel Preparation by Radiation ..............................................................................12

IV. Network Structure and Defects.......................................................................................19

Hydrogels are water-swollen networks (cross-linked structures) of hydrophilic homopol-

1. Cross-linking of a homopolymer or copolymer in solution or in the solid state with

II. HYDROGEL PREPARATION BY CHEMICAL CROSS-LINKING

Monomer Name Abbreviation Chemical structure

Hydroxyethyl methacrylate HEMA CH,=C(CH 3)COOCH,CH,OH

fC H 2-C H }n + PY -> fC H 2-C H }k + XY

Cross-linking by side-chain fracture

End-linking by main-chain fracture

Figure 1 presents a schematic representation of a hydrogel produced by this technique.

Polymerization Problems related to polymer

Initiation (six such reactions)

Propagation and cross-linking (ten such reactions)

Termination by combination (three such reactions)

Termination by disproportionation (three such reactions)

Chain transfer to monomer (ten such reactions)

Here I is the initiator, A is a molecule with initiated radical, M, is the monomethacryl

monomethacryl and dimethacryl monomer terminal groups, respectively, and Mp q r is a dead

3. Molecular Weight Distribution

rA = 2fkdI - [k„M, + ki2(2M2 + M3)]A (3)

rM, = [kj,A + (kpn + kf|l)P + (kp2l + kf2l)Q]M, (4)

rM2 = _ 2[ki2A + (kpi2 + kf|2)P + (kP22 + kf22)Q]M2 (5)

Total living polymer, P

rP = [kSlA + (kp21 + kf2l)Q]M, - [(kpi2 + kfl,)(2M2 + M3) ( 6)

Total living polymer, Q

rQ = [ki2A + (kpi2 + kf|2)P](2M2 + M3) - [(kP2l + kf21)M, (7)

Here, M3 is the concentration of pendent methacryl groups, defined as:

q(Pp.q.r + Qp.q.r + Mp,q,r) (8)

»k.j.k = 2 S X P'q'rk(Pp,q,r + Qp.q.r + Mpqr) (11)

r,ooo = [kuM, + ki2(2M2 + M3)] A - l- ktcllP2 - ktcl2 PQ - ktC22Q2

+ [kfllM, + kfl2(2M2 + M3)]P + [kf21M, + kf22(2M2 + M3)]Q (13)

V o .o - IKA + (kpu + k fi.)p + ( k P2. + kf2|)Q]M, (14)

r*oio = [ki2A + (kpi2 + kfl2)P + (kp22 + kf22)Q](2M2 - M3) (15)

Finally, the rate equation for the concentration of cross-links is

M'O.O.I 2[k12A + (kpi2 + kfl2)P + (kp22 + kf22)Q]M3 (16)

w i^i,o,o W 2(i|iQj Q + l/2l[/001)

polymerization rate of cross-linked chains

^gp___^gp —. ____ 1_______ /^)o\

III. HYDROGEL PREPARATION BY RADIATION

A. Effect of Ionizing Radiation on Polymers

2. Effects of Ionizing Radiation on Macromolecules

-(C H 2-C )-n

cross-link, while polymers with the structure

-(C H 2-C )-n

Reaction with oxygen

Decomposition and rearrangement

FIGURE 3. Schematic representation of cross-linking and end-linking during irradiation of

B. Irradiation of Polymers in Solution

^gp_^gp —. 1_____ /^)o\