POLYMERIC FOAMS SERIES series editor S.T.

Lee

POLYMERIC
FOAMS
Mechanisms and Materials

Edited by
S.T. Lee
N. S. Ramesh

CRC PR E S S
Boca Raton London New York Washington, D.C.

© 2004 by CRC Press LLC

www.ketabdownload.com
 Library of Congress Cataloging-in-Publication Data

Polymeric foams : mechanisms and materials / [edited by] S.T. Lee, N.S. Ramesh.
p. cm. — (polymeric foams series)
Includes bibliographical references and index.
ISBN 0-8493-1728-2 (alk. paper)
1. Plastic foams. I. Lee, S.-T. (Shau-Tarng), 1956- II. Ramesh, N. S. (Natarajan S.)
III. Title. IV. Series.

TP183 .F6P63 2004

668.4'93—dc22 2003068613

This book contains information obtained from authentic and highly regarded sources. Reprinted material
is quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable
efforts have been made to publish reliable data and information, but the author and the publisher cannot
assume responsibility for the validity of all materials or for the consequences of their use.

Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic
or mechanical, including photocopying, microfilming, and recording, or by any information storage or
retrieval system, without prior permission in writing from the publisher.

All rights reserved. Authorization to photocopy items for internal or personal use, or the personal or
internal use of specific clients, may be granted by CRC Press LLC, provided that $1.50 per page
photocopied is paid directly to Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923
USA. The fee code for users of the Transactional Reporting Service is ISBN 0-8493-1728-2/04/
$0.00+$1.50. The fee is subject to change without notice. For organizations that have been granted a
photocopy license by the CCC, a separate system of payment has been arranged.

The consent of CRC Press LLC does not extend to copying for general distribution, for promotion, for
creating new works, or for resale. Specific permission must be obtained in writing from CRC Press LLC
for such copying.

Direct all inquiries to CRC Press LLC, 2000 N.W. Corporate Blvd., Boca Raton, Florida 33431.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are
used only for identification and explanation, without intent to infringe.

Visit the CRC Press Web site at www.crcpress

© 2004 by CRC Press LLC

No claim to original U.S. Government works

International Standard Book Number 0-8493-1728-2
Library of Congress Card Number 2003068613
Printed in the United States of America 1 2 3 4 5 6 7 8 9 0
Printed on acid-free paper

© 2004 by CRC Press LLC

www.ketabdownload.com
 Dedication

To the Lord, whose breath brings forth wisdom.

© 2004 by CRC Press LLC

www.ketabdownload.com
 Preface

Since World War II, polymeric foam has become a very important part of
the polymer industry, influencing almost every aspect of our daily lives. The
industry has grown rapidly due to the beneficial characteristics of polymeric
foam, including its light weight, sound absorption properties, and positive
role in material savings, and it is used widely in protective packaging,
thermal insulation, and seat cushioning. Unique properties, processes, and
technologies have been developed to satisfy and explore new application
opportunities. As a result, the polymeric foam industry has become quite
diversified.
Foaming itself is a dynamic and complex phenomenon that involves sci-
entific principles and engineering parameters that govern the manufacturing
process. One of the main objectives of this volume is to provide a clear
understanding of the fundamental mechanisms and material characteristics
of polymeric foam. Chapter 1 offers an introduction to mechanisms and
materials of polymeric foams. These fundamental mechanisms appear to be
common to all foams, which are made through foaming phenomena consist-
ing of basic mechanisms—cell nucleation, growth, and stability. Chapters 3
and 5 are dedicated to these mechanisms.
Although foaming is an unstable separation process with dynamic com-
plexity, material strength plays a key role in determining a foam’s expansion
boundary and cellular structure. Chapter 2 is devoted solely to discussing
material strength, which is crucial to thermoplastic foam extrusion. Foam
processes are thoroughly discussed in Chapters 4, 6, and 7, which cover
thermoplastic foam and flexible and rigid thermoset polyurethane foams. It
is our hope that this volume brings forth a clear comprehension of mecha-
nisms and material features and their interactions with polymeric foams.
We acknowledge that as understanding expands, the number of unknowns
also grows. We hope this book will help industrial and academic scientists,
chemists, and engineers who are performing fundamental and applied
research attain lucid understanding of foaming mechanisms and materials
that will lead to great achievements. The volume can also be used as a
textbook at the graduate level and a reference book for undergraduate
courses on foams. We believe this book will generate a tremendous interest
in developing new processes, technologies, and applications in the area of
manufacturing foams.

S.T. Lee
N.S. Ramesh

© 2004 by CRC Press LLC

www.ketabdownload.com
 Acknowledgments

The editors wish to thank all contributing authors and reviewers for their
concentrated efforts in the process of preparing and revising the manuscripts.
The team spirit developed through work on this volume is greatly valued.
Nelson Malwitz, retired from Sealed Air Corporation in 2003, Richard Gend-
ron of the Canadian Research Council, and Professor Masahiro Ohshima of
Kyoto University offered prudent comments in review to make this book
more professional. Input from Dr. Michael (Chien-Yuen) Huang of the New
Jersey Institute of Technology is also appreciated. We also thank Sealed Air
Corporation for its support, and Kevin Lee and Jenny Jun in particular for
faithfully working extra long hours at their computers. Our gratitude to Mrs.
Mjau-Lin Lee and Mrs. Malathi Ramesh for their invaluable support is
beyond that which words can express. Finally, we thank God for the learning
experience and for giving us the strength to continue during difficult times
in the last three years.

© 2004 by CRC Press LLC

www.ketabdownload.com
 Editors

S.T. Lee received his B.E. from the Chemical Engineering Department at
National Tsing-Hua University, Taiwan, and his M.S. and Ph.D. from Stevens
Institute of Technology. He has over 20 years experience working with foam-
ing and polymeric foam.
Presently, Dr. Lee is with Sealed Air Corporation and resides with his wife,
Mjau-Lin, and three children in New Jersey. To date, his publications exceed
80 and include 20 U.S. patents. Dr. Lee is the editor of Foam Extrusion, pub-
lished by CRC Press. He is also a fellow of the Society of Plastics Engineers.

N.S. Ramesh is director of R&D in the Polyolefin Foam Division at Sealed

Air Corporation. He received his BTech. in chemical engineering from the
University of Madras and his M.S. and Ph.D. in chemical engineering from
Clarkson University. He was awarded certificates for attending the summer
rheology program at MIT and the executive management programs at the
University of Michigan, Ann Arbor, and Southern Methodist University Cox
business schools.
Dr. Ramesh has worked in the area of polymeric foams for 15 years and
has received 2 best paper awards for his pioneering work on foams from the
Society of Plastics Engineers. He has more than 50 publications, including
15 U.S. patents and 2 book chapters. Dr. Ramesh served as technical chair
for the Foams 1998 and 2000 international conferences. In 2002, he was
elected a fellow of the Society of Plastic Engineers.

© 2004 by CRC Press LLC

www.ketabdownload.com
 Contributing Authors

P. Berthevas W. Farrissey
The Dow Chemical Company The Dow Chemical Company
Midland, Michigan Midland, Michigan

J. Bicerano J. Fosnaugh
Corporate R&D The Dow Chemical Company
The Dow Chemical Company Midland, Michigan
Midland, Michigan
R. de Genova
R. van den Bosch The Dow Chemical Company
Polyurethanes R&D Midland, Michigan
Dow Benelux B.V.
H.J.M. Grünbauer
Terneuzen, The Netherlands
Polyurethanes R&D
M. Brown Dow Benelux B.V.
The Dow Chemical Company Terneuzen, The Netherlands
Midland, Michigan
R. Herrington
F. Casati The Dow Chemical Company
The Dow Chemical Company Midland, Michigan
Midland, Michigan
J. Hicks
C.P. Christenson The Dow Chemical Company
The Dow Chemical Company Midland, Michigan
Midland, Michigan
K. Hinze
P. Clavel The Dow Chemical Company
Polyurethanes R&D Midland, Michigan
Dow Europe GmbH
Meyrin, Switzerland K. Hock
The Dow Chemical Company
R.D. Daussin Midland, Michigan
Polyurethanes R&D
Dow North America D. Hunter
Freeport, Texas The Dow Chemical Company
Midland, Michigan
M.J. Elwell
Corporate R&D L. Jeng
The Dow Chemical Company The Dow Chemical Company
Midland, Michigan Midland, Michigan

© 2004 by CRC Press LLC

www.ketabdownload.com
 H. Kawabata R. Moore
Polyurethanes R&D The Dow Chemical Company
Dow Chemical Japan, Ltd. Midland, Michigan
Gotemba, Japan
S.E. Moore
H. Kramer Polyurethanes R&D
Polyurethanes R&D Dow North America
Dow Benelux B.V. Freeport, Texas
Terneuzen, The Netherlands
L. Nafziger
D.D. Latham The Dow Chemical Company
Polyurethanes R&D Midland, Michigan
Dow North America
Freeport, Texas
R.A. Neff
D. Laycock BASF Corporation
The Dow Chemical Company Wynadotte, Michigan
Midland, Michigan
M. Norton
S.T. Lee The Dow Chemical Company
Sealed Air Corporation Midland, Michigan
Saddle Brook, New Jersey
B.C. Obi
W. Lidy Polyurethanes R&D
The Dow Chemical Company Dow North America
Midland, Michigan Freeport, Texas

C.W. Macosko V. Parenti

Professor of Chemical Engineering Polyurethanes R&D
and Materials Science Dow Italia S.R.I.
Director of I PRIME, The Industrial Corregio, Italy
Partnership for Research in
Interfacial and Materials D. Parrish
Engineering The Dow Chemical Company
University of Minnesota Midland, Michigan
Minneapolis, Minnesota R. Priester
The Dow Chemical Company
C.A. Martin Midland, Michigan
Polyurethanes R&D
Dow North America N.S. Ramesh
Freeport, Texas Director of R&D
North America Polyolefin Foam
H. Mispreuve Division
The Dow Chemical Company Sealed Air Corporation
Midland, Michigan Grand Prairie, Texas

© 2004 by CRC Press LLC

www.ketabdownload.com
 A.K. Schrock G. Wiltz
Polyurethanes R&D The Dow Chemical Company
Dow North America Midland, Michigan
Freeport, Texas
T. Woods
K. Skaggs
The Dow Chemical Company
The Dow Chemical Company Midland, Michigan
Midland, Michigan
L. Stahler M. Xanthos
The Dow Chemical Company Otto H. York Department of
Midland, Michigan Chemical Engineering and
Polymer Processing Institute
F. Sweet New Jersey Institute of Technology
The Dow Chemical Company Newark, New Jersey
Midland, Michigan
M. Yamaguchi
R. Thomas
TOSOH Corporation
The Dow Chemical Company
Yokkaichi Research Laboratory
Midland, Michigan
Yokkaichi, Japan
R. Turner
The Dow Chemical Company Q. Zhang
Midland, Michigan Center for Biomaterials &
Advanced Techonologies—
H.A. de Vos Medical Devices Group
Polyurethanes R&D A division of Ethicon, Inc.
Dow Benelux B.V. A Johnson & Johnson Company
Terneuzen, The Netherlands Somerset, New Jersey
H.R. van der Wal X.D. Zhang
The Dow Chemical Company Unilever Research
Midland, Michigan Edgewater, New Jersey

© 2004 by CRC Press LLC

www.ketabdownload.com
 Contents

1 Introduction: Polymeric Foams, Mechanisms,

and Materials
S.T. Lee

2 Melt Elasticity of Polyolefins: Impact of Elastic

Properties on Foam Processing
M. Yamaguchi

3 Fundamentals of Bubble Nucleation

and Growth in Polymers
N.S. Ramesh

4 Material Properties Affecting

Extrusion Foaming
Q. Zhang and M. Xanthos

5 Foam Stability in Flexible Polyurethane

Foam Systems
X.D. Zhang, R.A. Neff, and C.W. Macosko

6 Flexible Polyurethane Foams

J. Bicerano, R.D. Daussin, M.J.A. Elwell, H.R. van der Wal,
P. Berthevas, M. Brown, F. Casati, W. Farrissey, J. Fosnaugh,
R. de Genova, R. Herrington, J. Hicks, K. Hinze, K. Hock,
D. Hunter, L. Jeng, D. Laycock, W. Lidy, H. Mispreuve, R. Moore,
L. Nafziger, M. Norton, D. Parrish, R. Priester, K. Skaggs, L. Stahler,
F. Sweet, R. Thomas, R. Turner, G. Wiltz, T. Woods,
C.P. Christenson, and A.K. Schrock

7 Rigid Polyurethane Foams

H.J.M. Grünbauer, J. Bicerano, P. Clavel, R.D. Daussin,
H.A. de Vos, M.J. Elwell, H. Kawabata, H. Kramer, D.D. Latham,
C.A. Martin, S.E. Moore, B.C. Obi, V. Parenti, A.K. Schrock,
and R. van den Bosch

© 2004 by CRC Press LLC

www.ketabdownload.com
 1
Introduction: Polymeric Foams, Mechanisms,
and Materials

S.T. Lee

CONTENTS
1.1 Introduction
1.2 Foaming Technology
1.3 Foaming Mechanisms
1.4 Materials
References

1.1 Introduction
Foam is a material, and foaming is a phenomenon. Both involve the presence
of a gas phase encapsulated by a spherical shell dense phase.
Gas exists throughout the universe and is a very interesting phase, espe-
cially on Earth. It is indispensable to organic life. Despite possessing a mass
and sometimes a shape (e.g., fire), gas is basically formless. The light bonding
of gas makes it almost omnipresent. There may be a large amount of gas
molecules in a limited space, or a small amount in a large space. When gas
is encapsulated in a dense medium, an interesting phenomenon or product
is formed. In nature, cork, wood, and pumice are good examples of this type
of product. In these materials, either there is a hollow space surrounded by
solid, or gaseous voids are dispersed in the solid. The hollow space, as seen
in bamboo and cork, is shaped like hollow sticks bundled together. But foam
is often round in shape.1
The tiny hollow space acts as an amazing transport vehicle for water, from
the root to the top of a tree, a process known as the capillary effect. Foam
impacts almost every aspect of human life. For example, when heated, water
soaked into rice grain expands and renders the grain edible; and gas is

0-8493-1728-2/04/$0.00+$1.50
© 2004 by CRC Press LLC

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 1.1
General Foams in Our Daily Lives
Type Function
Lungs Exchanged gas
Rice and flour Cellular for food
Popped corn, cereal Cellular for food
Meringue Blended air
Soda, beer Froth
Seat cushions Energy absorption
Walls, flooring Insulation

generated in flour through yeast. Both rice and flour possess a cellular struc-
ture that has been compatible with our digestive tissue as food for thousands
of years. Low surface tension surfactant-induced bubbles are another inter-
esting function that removes dirt from textiles. Moreover, natural sponge,
with its open cell structure, is a useful product in absorbing and desorbing
liquids. Table 1.1 shows a brief summary of foams in our lives.2
In most synthesis, foaming is a necessary phenomenon to convert raw
materials to foamed products. Our focus in this volume is on foam as a
product and its processing mechanisms.
Foam contains porous structures perceived as gaseous voids surrounded
by a dense phase or as a gas–solid composite. The drastic differences in
nature between gases and solids make foams unique combinations that have
special properties for applications. For instance, the presence of a cellular
structure can regulate flow velocity, dissipate disturbance, and enlarge mass
transfer area. Metal foam, polymeric foam, paper foam, and ceramic foam
have been developed for unique applications to enrich our lives or to explore
the mysterious universe.
Since the beginning of the twentieth century, synthetic polymers have
become increasingly important materials in industry. As plastication tech-
nology advanced and allowed for more sophisticated operations, the addi-
tion of a blowing agent became a natural extension of polymer processing.
Combining chemistry and engineering principles, foaming methodology
continued to expand in Europe and North America during and after World
War II. Japan joined the developmental list in the 1960s. Presently, foam
extrusion, injection molding, molded bead, x-linked foam, reactive foaming,
and gelation are well-known methods of making polymeric foams, and poly-
meric foams have become a well-developed branch in the polymer industry.
A brief developmental history of polymeric foams is presented in Table
1.2.3–10
Polymers, especially thermoplastic polymers, are characterized by their
viscoelastic nature, which possesses processing and material uniqueness.
When fit into processing and foaming criteria, their expanded cellular struc-
ture shows interesting properties. However, not every polymer is a good
candidate for foam. Considering compatibility with gas, processing window,

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 1.2
A Brief Development History of Polymeric Foams
Year Foam Inventor/Author Reference
1931 Polystyrene foam Munters and Tandberg 3
1937 Polyurethane foam Bayer 1
1941 Polyethylene foam Johnson 4
1966 Polyisocyanurate foam Ashida and Yagi 5
1967 ABS foam Woollard 6
1967 PVC foam Assigned to ICI 8
1968 X-PE foam Asaaka et al. 7
1971 PET foam Dijkstra et al. 9
1972 Polypropylene foam Parrish 10

capability to hold dynamic foaming, and stability during gas replacement

by air, few polymers are left on the list. Surprisingly, these polymeric foams
developed solid and strong applications due to their unique properties.
The gaseous voids dispersed in the polymeric matrix evidently alter the
polymer’s structure, morphology, and properties. Its lighter density has lent
itself to commercial flotation since the 1940s. Also, its cellular structure is a
natural fit in insulation, which is currently a substantial business. Annual
consumption of polymeric foams is forecast to rise to 7.8 billion lb in 2005,
an amount valued at $17.6 billion.11 When a cell wall possesses enough
elasticity, it can be used repeatedly in the athletic and furniture industries.
Other industries, such as automotive, medical, and packaging, are also
enhanced by the usage of polymeric foams. The growth in diversity of
applications and intensity of usage continues. In essence, it is performance
oriented. A simplified performance dependency summary chart is presented
in Figure 1.1. Materials and mechanisms are the primary variables that
directly affect structure formation. When they are combined with residual
gas, mechanical and thermal characteristics are determined.
Polymeric foam can be viewed from different practical perspectives, such
as products, technology, and components. Products can be categorized by
dimension, density, cell size, cell density, morphology, and property; for
instance, block versus film, high density versus low density, microcellular
versus cellular, open cell versus closed cell, and rigid versus soft. As for
technology, it is basically classified as soluble foaming and reactive foaming,
or physical foaming and chemical foaming. Foam can also be classified in
accordance with polymer materials such as thermoplastic and thermoset
foam. Table 1.3 summarizes the perspectives, for which details can be found
in References 12–14. Product variation has been addressed before, and new
developments keep coming to the market, making updates necessary. How-
ever, this book focuses on technology and materials—more precisely, foam-
ing mechanisms and materials. It is appropriate to highlight foaming
technologies prior to discussing foaming and material issues.

© 2004 by CRC Press LLC

www.ketabdownload.com
 Processing
Formula
Foaming

Aging/ Structure/
Curing Morphology

Property/
Application

FIGURE 1.1
Polymeric foam performance dependency summary chart.

TABLE 1.3
Foaming Perspectives
Perspective Terminology
Material Thermoplastic and thermoset
Mechanism Soluble foaming and reactive foaming
Nature Flexible and rigid
Structure Closed cell and open cell
Cell size Microcellular and cellular
Density High density and low density
Dimension Board and thin sheet

1.2 Foaming Technology

The method for making foam is straightforward: generating bubbles and
stabilizing them within a polymeric matrix. Bubble formation, in general, is
a consequence of unstable phenomena or a way to dissipate a “disturbance”

© 2004 by CRC Press LLC

www.ketabdownload.com
 to resume a stable state. Boiling is a good example that is often described as
a transient phenomenon. When a system is excited into an unstable foaming
state, a stabilization mechanism must be established in time to transform the
foaming into a stable foamed product. Again, the fundamental concept
sounds very simple, but there are a great variety of methods for making
useful polymeric foams.
As pointed out, there are two major foaming methodologies in the poly-
meric foam industry: soluble foaming and reactive foaming, or physical
foaming and chemical foaming The former involves physical variation in
polymer states, and the latter is solely dependent on chemical reaction.
The two methodologies can also be described as follows: blowing an agent
into the polymer by mixing within a chamber to take advantage of the
vaporization of gas while reducing pressure; or blending reactants to reactive
conditions for gas evolution within a dense medium. In both methods, the
same three steps—gas implementation, gas expansion, and foam stabiliza-
tion—are involved. It is reasonable to view foaming from the perspective of
material strength. Figure 1.2 illustrates the different paths for common sol-
uble foaming and reactive foaming. Many interesting techniques have been
developed around each methodology in the last half century.
Plasticators, such as screw extrusion and injection molding, are very pop-
ular in making processing foamed polymers, owing to their capabilities in
energy transfer to change the state of thermoplastic polymers to accommo-
date blowing agents in order to form a melting solution for subsequent
bubbling while adequate reduction of pressure occurs. Temperature reduc-
tion is necessary to facilitate the stabilization process. Another common
method is to saturate thermoplastic pellets with a blowing agent. When

Material Strength of PU vs. Extrusion Foaming

PU Foaming
Material Strength

Extrusion
Foaming

Implementation Expansion Stabilization

Foaming Stages

FIGURE 1.2
Material strength variation during PU reactive foaming and extrusion foaming.

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 1.4
General Technologies Used to Make Polymeric Foams
Technology Polymers Applications
Reactive foaming PU, Phenolics Construction, automotive, sports, toy,
furniture, packaging
Extrusion PS, PVC, PE, PP, PET Food, construction, decoration,
packaging, medical
Injection Mold PS, PP, PE Automotive
Mold Bead PS, PP, X-PE Food, packaging, thermoforming
X-PE PE Sports, thermoforming

pressure reduction and temperature increase are applied for enough super-
saturation, controlled foaming makes foamed products.
Reactive foaming generates foaming through chemical reaction. Gas can
evolve out of simple reactions between reactants, or simply by thermally
induced decomposition. Polyurethane, polyisocyanurate, and phenolics are
good examples of foams made by reactive foaming. Basically, the amount of
gas is governed by reactant quantity and ratio, reaction rate by catalyst and
thermal condition, and foam stability by additives. Thermoset polymeriza-
tion amid a gas reaction makes it very unique. The exothermic reaction
enhances foaming by enthalpy. Also, changing formula and reaction condi-
tions creates a great product spectrum for reactive foaming. Slightly over
half of the foam market is designated for reactive foaming products.15,16
Another common practice is to preblend chemicals, which is to decompose
into gaseous components under a certain temperature with the thermoplastic
to cause decomposition inside the processor for foaming. Normally, this type
of reaction generates volatile gases, which makes stabilization of the polymer
very critical. It is very popular for low expansion products. To increase
expansion, polymer strength has to be improved. A good practice is to
generate x-links among the polymer backbone prior to gas generation. Great
success in polyolefin foam has been established. A summary on foaming
technology is included in Table 1.4.17

1.3 Foaming Mechanisms

Production of foamed polymers includes three stages within the polymer:
implementation of gas, expansion of gas, and stabilization of the polymer.
Different mechanisms and parameters occur and sometimes overlap in dif-
ferent stages. When surrounding conditions change, competing mechanisms
appear to make the kinetic process more complex. Sizable research and
development efforts have been devoted to exploring the mechanisms, their
individual roles, and their combined effects.18,19 Selected processing param-
eters and their effects are included in Table 1.5.

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 1.5
Processing Parameters and Their Effects at Different Stages
Implementation Expansion Stabilization
Temperature Solubility ≠ Volatility ≠ Solidification Ø
Viscosity Ø Surface tension Ø Permeability ≠
Reactivity ≠ Viscosity Ø
Diffusivity ≠
Interaction parameter Ø
Pressure Solubility ≠ Shear heat ≠ Solidification –
Viscosity — Surface tension –
Homogenization ≠ Nucleation ≠
Shear Solubility — Nucleation ≠ Solidification Ø
Dissolution ≠ Growth ≠
Dispersion ≠ Cell distribution ≠

The solubility limit is the overriding parameter in the implementation

stage, which is strongly dependent on surrounding pressure, tempera-
ture, and interaction with the polymer. The amount of gas blending—or
generated via reaction—into the polymer is a controllable processing
parameter, which affects melt solution homogenization, foaming dynam-
ics, and stablization. The Flory–Huggins equation is a good guideline for
determining how much gas can be implemented in the polymer; it is
expressed as

DFm = kT(nglnfg + nplnfp + cngfp) (1.1)

where DFm, k, T, and c represent, in order, the free energy of mixing, the
Boltzmann constant, the absolute temperature, and the interaction parame-
ter. Respectively, n, f, and the subscripts p and g denote the molar fraction,
the volume fraction, the polymer, and the gas. The interaction parameter, c,
consists of two parts, entropy and enthalpy. For most polymer/gas systems,
the former has a mean around 0.3, and the latter is determined by solubility
parameters.20 It is expressed as

c = 0.3 + Vg/RT (dp - dg)2 (1.2)

where Vg, R, and d denote the molar volume of gas, the ideal gas constant,
and the solubility parameter, respectively. Common polymer/gas interaction
parameters are tabulated in Table 1.6.21,22
Free energy of mixing for various polymer/gas systems can be prepared
as illustrated in Figure 1.3, with polystyrene and butane at two different
temperatures (processing and application temperatures). The lower the
mixing energy, the more stable it is. The curve sometimes shows two dips
(e.g., PS/butane at 30˚C), which suggests stable gas/polymer coexistence.
One is gas rich and the other polymer rich, corresponding to gas in the cell

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 1.6
Solubility Parameter of Selected Common Polymers and Blowing Agents
B/A
(cal/cm)1/2 Polymer
(cal/cm)1/2
n-Pentane 7.0 Polyethylene 7.9
n-Butane 6.8 Polystyrene 8.6
Propane 6.4 Polypropylene 9.2
Ethane 6.0 Polyvinylchloride 9.4
CCl2F2 11.3 Polyvinylacetate 8.9

and gas in the cell wall. Pressure-dependent parameters are included in

the Sanchez–Lacombe model,23 and additional temperature-dependent
parameters in Simha–Somcynsky.24 Since gas is so pressure sensitive, both
models are better than Flory–Huggins for a gas-charged polymeric system.
The gas molecule is in the Anstrom (10–8 cm) domain, and the “holes” in
the polymeric liquid are 10–6 to 10–7 cm.25 Visible bubbles are of the same
magnitude as cosmic dust, 10–5 cm. How free gas molecules transform into
“clusters” to account for phase separation is an interesting subject. Due to
the limits of analytical instrumentation, some theories have been developed
to explain the transformation.
When the gas solubility limit is met, the gas phase tends to separate from
the polymer phase. A typical phase diagram is illustrated in Figure 1.4, which
suggests a stable single phase, then a metastable single phase, in the unstable
two-phase region. It is believed that the dominant mechanism in the meta-
stable state is nucleation (Nu), and that the dominant mechanism in the
unstable state is spinodal decomposition (SD), or the domains of rearrange-
ment and growth. Both are controlled by diffusion, although the directions

PS/Butane System Enlargement of Figure

Inside Box
0.50 0.00005
0.00000
0.00 0.99984 0.99989 0.99994 0.99999
Free Energy of Mixing (J)

0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00 −0.00005
−0.50 −0.00010
−0.00015
−1.00
−0.00020
−1.50

−2.00

−2.50

−3.00

−3.50

−4.00
Volume Fraction of Butane
FM at 30C (J)
FM at 150C (J)

FIGURE 1.3
Free energy of mixing vs. volume fraction of butane at two different temperatures for PS/
butane system.

© 2004 by CRC Press LLC

www.ketabdownload.com
 Pressure vs. Volume Fraction of a Binary System

U
M
Pressure P

∆P
δP
δφ

Volume Fraction of Hydrocarbons φ

coexistence
classical
spinodal

FIGURE 1.4
Bimodal and spinodal curves; M and U indicate metastable and unstable paths.

of diffusion are opposite, as presented in Figure 1.5.26,27 Figure 1.6 demon-

strates the two different phase separation structures for polystyrene and
partially brominated polystyrene at two different volume fractions.28
SD, counterintuitively, exhibits negative diffusion; namely, diffusion
moves from low concentration to high concentration. However, in the molec-
ular domain, when driving force is high, a molecular cluster could grow
above a certain concentration intensity (critical wavelength) to absorb the
nearby molecules. Depending on the separation path, the metastable state
could be dominant or negligible to cause SD or Nu separation. Table 1.7

TABLE 1.7
Phase Separation Mechanisms: Nucleation and Spinodal Decomposition
Nucleation Spinodal Decomposition

Initial state Stable phase Stable phase

T≠ PØ, or TØ Nucleating sites form Composition fluctuation increases
Activation energy Yes No
Diffusion Positive Negative
Metastable state Yes Minimum
Growth Separate growth Interconnection
Late stage coarsening No Yes

© 2004 by CRC Press LLC

www.ketabdownload.com
 Concentration Profile of Nucleation and Growth during
Phase Separation; R > Rcr
Co positive diffusion coefficient

t1
Concentration
Profile

Time
t2

t3

Distance

Concentration Profile of Spinodal Decomposition during

Phase Separation;

Co t1
negative diffusion
Concentration
Profile

Time
t2

t3

Distance

FIGURE 1.5
Phase separation comparison—nucleation [R > Rcr]; spinodal decomposition (SD) [l > lcr]; R,
l, and subscript cr denote size, concentration wave length, and critical.

FIGURE 1.6
Structure patterns emerging during phase separation in PS/Br-PS blends. Left: nucleation with
PS at 80 vol%. Right: spinodal decomposition (SD) with PS at 50 vol%.

© 2004 by CRC Press LLC

www.ketabdownload.com
 tabulates the two different kinds of separation.29 Clearly, the two different
mechanisms, Nu and SD, are accountable for the separation in the early stage.
It is known that phase separation ranges from single phase (i.e., boiling
and crystallization) to multiphase (i.e., polymer blend processing, foaming).
Although SD has described certain separations in polymer blends, when we
talk about foaming, it is extremely difficult to investigate the dynamic molec-
ular movements to confirm its original mechanism: energy barrier into crit-
ical size (Nu) or energy fluctuation into growth mode (SD). When the gas
phase exceeds the solubility limit, such as with gas evolution in PU foam, it
tends to form a least-energy spherical shape to separate from the surrounding
polymer. It dissipates energy by absorbing the upcoming gases to result in
expansion to a stable state. The energy barrier does not exist. Another exam-
ple is foam extrusion, in which a sharp change into a high degree of super-
saturation occurs, with high temperature or low pressure. The volatile phase
becomes too energetic to coexist with the surrounding polymer phase. Unsta-
ble separation occurs immediately.
Both Nu and SD are plausible mechanisms to account for phase separation
in polymeric foaming, as illustrated in Figure 1.7.30 SD seems prone to
develop more cell coalescence and the corresponding open cell and cell size
distribution. Foaming induced by a sharp energy gradient tends to show
more cell connections and ruptures. For a modest energy gradient, external
aid, such as motion and deformation, is necessary to facilitate nucleation
into existence; phase separation will not occur within process time. Diffusion
to vapor phase above the polymeric melt may become the main mechanism
to reduce the supersaturation and to re-establish the equilibrium.
Whether the process is Nu or SD, enough gas molecules need to gather
together to be sustainable and to overcome surrounding confinement. Spher-
ical shape contains the minimum surface area, or surface energy. According
to the critical bubble radius:

Rcr = 2s/(Pb – P) (1.3)

where s, Pb, and P represent gas surface tension, bubble pressure, and sur-
rounding pressure, respectively. The critical bubble radius is unstable in
nature, according to the mass balance equation.31 Any perturbation can cause
it to move into unstable growth, during which size expansion lowers its
concentration to encourage diffusion and to cause a reciprocal response until
the driving force disappears.
When the surrounding pressure is greater than the gas pressure, gas mol-
ecules are suppressed into the polymer matrix to form a homogenized melt.
In foam extrusion, high pressure is needed to keep phase separation from
happening before the free expansion at the exit.32 Enough pressure reduction
is necessary for gas molecules to gather together to overcome the curvature
confinement. Experiments have determined that modest pressure reduction
seems inadequate to drive the bubble phase into existence.33 A laser beam

© 2004 by CRC Press LLC

www.ketabdownload.com
 Polymer Lean
Phase
Polymer Rich
Phase

Homogeneous
Solution

FIGURE 1.7
Phase separation illustration. Left: spinodal decomposition (SD). Right: nucleation (Nu).

© 2004 by CRC Press LLC

www.ketabdownload.com
 was used to detect the nucleation site for foam extrusion, and extrapolation
was applied to suggest the critical size.34
Kinetics has been applied to the energy barrier nucleation phenomenon;
this has led to the Arrhenius expression for nucleation rate:
J µ exp (–Ea/kT) (1.4)
where Ea represents activation energy, which is determined by interfacial
holding energy and degree of supersaturation. A detailed expression was
presented by Blander and Katz:35
J = N(2s/(pm))1/2exp(-16ps3/(3kTDP2) (1.5)
where N, s, m, and DP denote number of molecules per unit volume, surface
tension, mass of a molecule, and pressure difference (or superheat), respec-
tively.
More energy contribution, as with surface energy or flow energy, leads to
further modification of Equation 1.5. Since nucleation is an unstable phe-
nomenon, growth is in imminent convolution, in which expansion and gas
diffusion occur simultaneously.
Another interesting mechanism is cell opening, especially with high expan-
sion or a weak cell wall. According to the packing theory, a spherical bubble
begins to share a cell wall after expansion over four times. As a result of
sharing a wall, thinning and rupture appear. Figure 1.8 shows five high-
speed photographs of a polydimethylsiloxane membrane stretching to rup-
ture under pressure-induced expansion. A black hole, weak spot, or low
surface tension area forms on top before the membrane bursts out. It is
common for a high degree of expansion to occur in low-density thermoplastic

0.7325 sec 1.2750 sec

1.2755 sec 1.2805 sec 1.400 sec

FIGURE 1.8
Polymethylsiloxane in a box with a hole to allow air to expand the membrane until it bursts;
five high-speed camera photos showing black hole formation before rupture.

© 2004 by CRC Press LLC

www.ketabdownload.com
 foam formation. As a result, an open cell is formed. This area needs further
investigation to correlate it with other properties.
As for thermoset reactive foaming, cell stability involves drainage of the
membrane. It tends to recede to a strut, forming an open cell skeleton. (A
detailed description is presented in Chapter 5.) In fact, when the strut retains
enough strength, the open-cell structure presents unique performance char-
acteristics and is especially good for sound and liquid absorption.

1.4 Materials
Polymeric foam is good for a wide variety of applications. Polymers and
gases are the fundamental components. They are basically two different
materials, both in nature and in general properties. The dissimilar materials
stay together homogeneously under processing conditions, separate into a
cellular structure, and stabilize within a reasonable time. There is no doubt
that material and mechanism need to complement each other for a successful
process. Polymers and gases have to satisfy every mechanism, such as the
conditions of solubilization, homogenization, phase separation, and stabili-
zation. Only certain polymeric foams survive these screening tests.
Different mechanisms involve different competing parameters. As a result,
each step has a window, over or under which foaming efficiency is reduced.
Moreover, the list of conditions is shortened when mass production and the
economics of competition come into consideration.
For soluble foaming, gas solubility in the polymer under processing con-
ditions is a top concern. Solubility is usually dictated by surrounding pres-
sure; a pressure that is too high causes processor concerns, and a pressure
that is too low may not generate enough superheat for proper foaming.
Halogenated hydrocarbons demonstrate an excellent soluble nature in some
thermoplastic melts, but they are too stable to decompose in the low atmo-
sphere and become problematic in regard to ozone depletion and formation
in the high atmosphere. Nowadays, a blowing agent has to comply with not
only technical and economic requirements, but environmental and safety
requirements as well. As for the polymer, it has to be processible with and
without a blowing agent and must possess enough strength to maintain the
dynamic expansion to create a unique cellular structure. The key system
parameters for successful soluble foaming are shown in Table 1.8. Common

TABLE 1.8
Key Parameters for Soluble Foaming
Stage Gas into polymer Gas bubbles in Air bubbles in
polymer polymer
Mechanism Dissolution/homogenization Foaming Curing
Parameters Solubility, diffusivity Volatility, surface Permeability
tension

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 1.9
General Thermoplastic Foams
Polystyrene Polyvinylchloride Polyolefin
Polymer Amorphous Semi-crystalline

Processing Extrusion Extrusion Extrusion

Injection molding Injection molding Injection molding
Mold bead Mold bead
X-linking
Product Rigid Flexible and rigid Soft to rigid
Insulation Fire retardancy Chemical resistance
Low temp stability Low temp stability

foam materials that correspond to soluble foaming principles are polysty-

rene, polyvinylchloride, and polyolefin. Each has unique processing and
product features, which are summarized in Table 1.9.
In reactive foaming, reaction kinetics characterizes materials. Reactions
to generate gases and to form long chain polymers must occur at the right
stage; transient bubbling or limited expansion may otherwise occur. Appro-
priate functional groups and their ratios with catalysts and surfactants
should be controlled such that reaction time and desired foam property
can be obtained. Reaction time dictates the reactor (or mixer) design. The
exothermic nature can easily develop extra heat for additional expansion
of formed gas to acquire a very high expansion, but it also makes stability
hard to control. The right surfactant can certainly widen the favorable
expansion window.36
Polyurethane (PU) foam is a well-established foam technology that
accounts for over half of overall polymeric foam production. It basically
contains flexible and rigid PU foam. The two types of foam are based on a
similar chemistry, and the type and degree of polymerization make the
difference in product and property. A brief summary is presented in Table
1.10. Chapters 6 and 7 are devoted to flexible and rigid PU foam, respectively.

TABLE 1.10
PU Foams
Flexible PU Foam Rigid PU Foam
Formula Polyol + Isocyanate Polyol + Isocyanate
Long polyol with less –OH Short polyol with more –OH
Density, kg/m3 5–200 15–800
Properties Absorption Insulation
Soft Hard
Markets Furniture Construction
Appliance

© 2004 by CRC Press LLC

www.ketabdownload.com
 References
1. Lee, S.T. and Park, C.B., Polymeric Foams: Understanding Foaming Technology,
CRC Press, Boca Raton, FL, in press, 2005.
2. Weaire, D. and Hutzler, S., The Physics of Foams, Oxford University Press, 2000.
3. Munters, G. and Tandberg, J.G., U.S. Patent 2,023,204, 1935.
4. Johnson, F.L., U.S. Patent 2,256,483, 1941.
5. Ashida, K. and Yagi, T., French Patent 1,511,865, 1966.
6. Woollard, D.C., J. Cell. Plast., 4, 1, 16–21, 1968.
7. Assigned to ICI, French Patent 1,487,545, 1967.
8. Asaaka, M., Torii, K., and Yamanaka, E., Jpn. Kokai Tokyo Koho, 43–22 674, 1968.
9. Dijkstra, A.J., Goodman, I., and Reid, J.A.W., U.S. Patent 3,553,157, 1971.
10. Parrish, R.G., U.S. Patent 3,637,458, 1972.
11. Foamed Plastics, Freedonia Group, Inc., Cleveland, OH, 1997.
12. Klempner, D. and Frisch, K.C., eds., Handbook of Polymeric Foams and Foam
Technology, Hanser, Munich, 1991.
13. Lee, S.T., ed., Foam Extrusion: Principles and Practice, CRC Press, Boca Raton,
FL, 2000.
14. Frisch, K.C., Historical developments of polyurethanes, in 60 Years of Polyure-
thanes, eds. J.E. Kresta and E.W. Eldred, Technomic, Lancaster, PA, 1998.
15. Herrington, R., Dow Polyurethanes; Flexible Foams, Dow Chemical, 1997.
16. Woods, G., The ICI Polyurethanes Book, ICI Polyurethanes, 1990.
17. Park, C.B. and Lee, S.T., Foam Seminar, sponsored by Society of Plastics Engi-
neers, Nashville, TN, 2003.
18. Burt, J., The elements of expansion of thermoplastics. II. J. Cell. Plast., Nov./
Dec. 1978.
19. Naguib, H.E., Park, C.B., Yoon, E., and Reichelt, N., Fundamental foaming
mechanisms governing volume expansion of extruded PP foams, Foams 2002,
seminar sponsored by Society of Plastics Engineers, Nashville, TN, 2002.
20. Merrill, E.W., Thermodynamic aspects of devolatilization of polymers, in Polymer
Devolatilization, ed. R.J. Albalak, Marcel Dekker Inc., New York, 1996, chap. 2.
21. Van Krevelen, D.W., Properties of Polymers, 3rd ed., Elsevier Science, Amsterdam,
1990.
22. Grulke, E.A., Solubility Parameter Values, VII/675, Polymer Handbook, 4th ed.,
eds. J. Brandrup, E.H. Immergut, and E.A. Grulke, John Wiley & Sons, New
York, 1999.
23. Sanchez, I.C. and Lacombe, R.H., Statistical thermodynamics of polymer solu-
tions, Macromolecules, 11, 6, 1145–1156, 1978.
24. Simha, R. and Moulinie, P., Statistical thermodynamics of gas solubility in
polymers, in Foam Extrusion, ed. S.T. Lee, CRC Press, Boca Raton, FL, 2000,
chap. 2.
25. Bueche, F., Physical Properties of Polymers, Interscience, New York, 1962.
26. Sperling, L.H., Introduction to Physical Polymer Science, John Wiley & Sons, New
York, 2001.
27. Debenedetti, P.G., Phase separation by nucleation and by spinodal decompo-
sition: fundamentals, in Supercritical Fluids, eds. E. Kiran, P.G. Debenedetti, and
C.J. Peters, Kluwer, Dordrecht, the Netherlands, 2000.

© 2004 by CRC Press LLC

www.ketabdownload.com
 28. Strobl, G.R., The Physics of Polymers: Concepts for Understanding Their Structures
and Behavior, 2nd ed., Springer-Verlag, Heidelberg, 1997.
29. Kiran, E., Polymer miscibility and kinetics of pressure-induced phase separa-
tion in near-critical and supercritical fluids, in Supercritical Fluids, eds. E. Kiran,
P.G. Debenedetti, and C.J. Peters, Kluwer, Dordrecht, the Netherlands, 2000.
30. Gunton, J.D., San Miguel, M., and Sahni, P.S., The dynamics of first-order phase
transitions, in Phase Transitions 8, Academic Press, New York, 1983, chap. 3.
31. Amon, M. and Denson, C.D., A study of the dynamics of foam growth: analysis
of the growth of closely spaced spherical bubbles, Polym. Eng. Sci., 24, 13,
1026–1034, 1984.
32. Thiele, W.C., Foam extrusion machinery features, in Foam Extrusion, ed. S.T.
Lee, CRC Press, Boca Raton, FL, 2000, chap. 8.
33. Biesenberger, J.A. and Lee, S.T., A fundamental study of polymer melt devol-
atilization part three: more experiments on foam-enhanced DV, Polym. Eng.
Sci., 27, 1987.
34. Han, J.H. and Han, C.D., Bubble nucleation in polymeric liquids. I. Bubble
nucleation in concentrated polymer solutions, J. Polym. Sci. B: Polym. Phys., 28,
711, 1990.
35. Blander, M. and Katz, J. L., Bubble nucleation in liquids, AIChE J., 21, 833, 1975.
36. Hilyard, N.C. and Cunningham, A., Low Density Cellular Plastics: Physical Basis
of Behaviour, Chapman & Hall, London, 1994.

© 2004 by CRC Press LLC

www.ketabdownload.com
2
Melt Elasticity of Polyolefins: Impact of Elastic
Properties on Foam Processing

M. Yamaguchi

CONTENTS
2.1 Introduction
2.2 Melt Elasticity
2.2.1 Elastic Properties
2.2.2 Measurements of Elastic Response
2.2.3 Role of Long-Chain Branches
2.2.3.1 Model Polymer
2.2.3.2 Low-Density Polyethylene
2.2.4 Melt Elasticity of New Polyolefins
2.3 Modification of Melt Elasticity
2.3.1 Effect of Processing History
2.3.1.1 Origin of Shear Modification
2.3.1.2 Processing History by Conventional Processing
Machines
2.3.2 Modification by Polymer Blends
2.3.2.1 Background
2.3.2.2 Blends with Weak Gel
2.3.2.3 Processability of Foaming
2.4 Conclusion
References

2.1 Introduction
It is generally accepted that low-density polyethylene (LDPE) produced by
radical polymerization under high pressure is preferable to linear low-den-
sity polyethylene (LLDPE) for foaming. Similarly, high-density polyethylene
(HDPE) with a chromium catalyst is often used for blow molding rather than
that with a Zieglar–Natta catalyst. The reason in both cases is strong melt

0-8493-1728-2/04/$0.00+$1.50
© 2004 by CRC Press LLC

© 2004 by CRC Press LLC

www.ketabdownload.com
elasticity due to long-chain branches. Therefore, a comprehensive knowledge
of the primary elastic properties peculiar to polymer melts and their relation
to foaming is required in order to decide the operating condition and the
recipe for various foaming processes, such as extrusion gas-foaming,
crosslinked foaming by a chemical blowing agent, and bead foaming.
It is, however, difficult to establish the precise relation between rheological
properties, including melt elasticity, and processability. One of the reasons
is that a foamable material in an extruder is not a simple polymer melt. The
effect of solving gas on the rheological properties is dependent on the mis-
cibility, which also affects the melting point.1 Further, the concentration of
the solving gas in a polymer melt decreases during expansion. Likewise,
distribution of a chemical blowing agent in a polymer melt plays a decisive
role on the cell morphology for crosslinked foaming. Another reason for the
difficulty is that foaming is carried out under non-isothermal conditions. In
particular, the processability of extrusion foaming for a crystalline polymer,
such as LDPE, which is often performed near the crystallization temperature
Tc , may be affected by the crystallization behavior, because even a small
amount of crystallites can change the rheological properties drastically. Fur-
ther, the crystallization temperature of LLDPE or LDPE rises rapidly when
blended with a small amount of PE with a higher Tc because the PE with a
higher Tc acts as a crystalline nuclei for the rest.2 As a result, the blend shows
marked melt elasticity around Tc since crystallites behave like crosslink
points, in which molecular weight and polydispersity also play an important
role. Hence, considerable attention has to be paid to the crystallization tem-
perature for extrusion foaming of polyethylene blends. It should also be kept
in mind that adding talc, as the nucleating agent of bubbles often enhances
the crystallization temperature of polyolefins. As for crosslinked batch-foam-
ing, temperature as well as pressure falls off sharply during foaming, since
the gas in the cell is immediately cooled down by adiabatic expansion after
ejection from a mold. Consequently, crystallization takes place rapidly, which
prevents the expanded cells from shrinking.3
This chapter will not cover thoroughly the role that elastic properties play
during the entire foaming process. Nor will it closely examine how this elas-
ticity should be tailored for each of the different phases, such as nucleation,
bubble growth, and stabilization. However, information about the basic rheo-
logical properties of simple polymer melts under isothermal conditions is still
highly relevant for foam practitioners. In the following section, the relation
between molecular structure and melt elasticity will be described, as that
relation pertains to branched polymers and new polyolefins that are being
developed by advanced polymerization technology. This follows a brief review
of melt elasticity and experimental measuring methods. Further, the following
recent experimental studies on the mechanical and material modifications of
melt elasticity will be summarized in Section 2.3: (1) the effect of applied
processing history on the melt elasticity for a branched polymer, which
should be controlled in various foaming processes for a better and more
steady production; and (2) the polymer blend technique, used to enhance
© 2004 by CRC Press LLC
www.ketabdownload.com
melt elasticity, and its effect on foaming. Especially important in the polymer
blend technique are the new types of blends with crosslinked polymers
characterized as weak gels; these are explained in detail, since this blend
technology may make various linear polymers with less melt elasticity avail-
able for foaming.

2.2 Melt Elasticity

2.2.1 Elastic Properties

A polymer melt, as a viscoelastic material, shows various elastic behaviors,

such as strain recovery after cessation and the Weissenberg effect (produced
by normal stress difference), as well as viscous features.4 Further, the Barus
effect, also referred to as extrudate swell and generally ascribed to the Weis-
senberg effect, is quite familiar in actual processing. In general, these elastic
properties are attributed to the recovery force from the low entropy state of
polymer chains, such as anisotropy of molecular orientation and the stretch-
ing of a chain, which are given via entanglement couplings in a similar
fashion to that of permanent, chemical crosslink points in a rubber molecule.
Since entanglement couplings act as crosslink points only temporarily, the
stress generated by deformation in the distant past is of less importance and
fades significantly, as shown in Figure 2.1.
The fading rate, represented by the inverse of relaxation time, dominates
the viscoelastic characteristics of a polymer melt.4 When the relaxation time
is longer than the characteristic time for flow, i.e., the inverse of the strain
rate, the polymer melt responds like an elastic solid. On the other hand,
the viscous properties overcome the elastic ones when the relaxation time
is shorter than the flow characteristic time. Discussions on the mechanical
response of a polymer melt are always carried out employing the dimen-
sionless parameter known as the Deborah number, which is defined as the
relaxation time divided by the inverse of the strain rate. Apparently, elastic
properties are pronounced at a large Deborah number.
The molecular weight dependence of relaxation time is interpreted from the
molecular point of view, which is known as the Doi–Edwards tube theory.5
Doi and Edwards proposed that large-scale motion of macromolecules is
limited in a tube-like region composed of neighbor chains,5 as shown in Figure
2.2A. Since lateral motion is restricted owing to the topological constraints of
the temporary network, a polymer chain diffuses along the backbone chain
by changing its conformation; i.e., it moves within the tube like a snake. This
motion is called reptation. In a simple polymer melt, reptation is the terminal
relaxation mechanism. Through reptation, molecular orientation (or the
memory of the applied strain) decays with time. In the process of reptation,
chain ends relax more quickly than the central part as the result of back-and-
forth movement in the tube (Figures 2.2B, C, and D) in accordance with the
© 2004 by CRC Press LLC
www.ketabdownload.com
 γ0
strain

time

σ0 σ(t ) = σ0 exp (−t /τ)

stress

t=0
time

FIGURE 2.1
Schematic representation of stress relaxation process after a step strain g0 applied at t = 0 for a
viscoelastic body having a characteristic relaxation time t. The initial stress s0 at t = 0 decays
with time following s(t) = s0 exp(–t/t). In the case of a simple polymer melt, relaxation time
in the terminal mode is determined by topological interaction between neighbor chains (entan-
glement couplings). Further, the molecular orientation also decays with time proportional to
stress.

stress.5 (Doi and Edwards derive the relaxation modulus employing the
concept of reptation.) Eventually, anisotropy of molecular orientation, one
of the origins of melt elasticity, decreases with time. Apparently, a longer
chain takes a longer time to relax. The diffusion time is given by the square
of the contour length of the tube divided by the diffusion coefficient.
Considering that the contour length is proportional to molecular weight M
and the diffusion coefficient is proportional to M -1, relaxation time of
reptation t 1 is proportional to M 3, which is in close agreement with the
experimental result t 1 µ M 3.4.

© 2004 by CRC Press LLC

www.ketabdownload.com
 A.
t<0

B.
t=0

C.
t = t1

D.
t = t2

FIGURE 2.2
Schematic illustration following the tube model described in detail in Doi and Edwards. A
polymer chain, a solid line, is confined within a tube, white region, generated by the topo-
logical constraints of neighbor chains. (A) At the equilibrium state, the orientation of chain
segments is isotropic (t < 0). (B) Applied step strain deforms a chain and then orientation
takes place (t = 0). (C), (D) The chain escapes from the tube by back-and-forth movement, or
reptation. Consequently, the degree of orientation, or memory of the applied strain, decreases
with time (t = t1, t2) in accordance with stress. Because of the back-and-forth movement, chain
ends relax more quickly than the central part. In the figures, the dark rectangular regions
represent the average orientation of the polymer chain.

© 2004 by CRC Press LLC

www.ketabdownload.com
 A. B. C.

FIGURE 2.3
Relaxation process of a star-shaped polymer following the tube model. Since a branch prohibits
simple reptation along the tube, the polymer changes its conformation by the contour length
fluctuation. As shown in Figure 2.3B, this process needs the end segment of a branch down to
the branch point, represented by the symbol C, which requires a low entropy state. (From Doi,
M. and Edwards, S.F., The Theory of Polymer Dynamics, Clarendon Press, Oxford, 1986. With
permission.)

Besides molecular weight and distribution, long-chain branches play a

significant role in elastic properties. This is also explained by the
Doi–Edwards theory.5 According to the theory, a branch point is assumed to
be fixed, as shown in Figure 2.3, and thus simple reptation is prohibited.
Consequently, conformation change of long-chain branches has to take place
by utilizing contour length fluctuation, or retraction to the branch point
within the tube. This retraction process requires exposing a polymer chain
to a low enthalpy state, due to distorted conformation around the branch
point. Hence, relaxation time increases exponentially with the length of
branches.5
Since actual polymer melts have numerous relaxation times due to poly-
dispersity and long-chain branches, the distribution as well as the strength
of each relaxation mode is responsible for the viscoelastic properties. As a
rule of thumb, elastic properties are strongly affected by terminal relaxation.
For example, shear storage modulus G G¢¢ representing elastic response to
oscillatory deformation is more sensitive to the long tail of relaxation time
t than loss modulus G ( )
G¢¢¢¢ since G ¢ is proportional to t 2 / 1 + w 2 t 2 , instead
( )
of t / 1 + w t for G ¢¢ , as shown in Equations 2.1 and 2.2:
2 2

•
w2t2
G¢ =
Ú
-•
H ( t)
1 + w2t2
d ln t (2.1)

•
wt
G ¢¢ =
Ú H ( t) 1 + w t
-•
2 2
d ln t (2.2)

where w is the angular frequency and H ( t) is the relaxation spectrum, which

decides the linear viscoelastic properties.
Further, steady-state compliance J e0 , defined by Equation 2.3, is often dis-
cussed as an elastic parameter of the terminal zone in the linear viscoelastic
© 2004 by CRC Press LLC
www.ketabdownload.com
region, which represents average distortion of polymer coils during flow at
w Æ 0 or ġ Æ 0 .4

G¢
Ú H(t) t d ln t 2

J e0 = lim = -• (2.3)
w Æ0 G ¢¢ 2 Ê • ˆ
2

Ë -•
Ú
Á H ( t) t d ln t˜
Á ˜
¯

The Rouse model predicted that steady-state compliance of a linear polymer

relates to molecular weight distribution as follows:6

M z M z +1
J e0 µ (2.4)
Mw 2

Similar to Equation 2.4, Milles has proposed the following relation, which is
in good agreement with the data obtained for various polymers:7

3.4 - 3.7
ÊM ˆ
J e0 µ Á z ˜ (2.5)
Ë Mw ¯

The equations demonstrate that polymers with broad molecular weight

distribution exhibit pronounced melt elasticity, whereas zero-shear viscosity
h0 , the viscous parameter in the terminal zone, is proportional to Mw 3.4 .
Further, the product of J e0 and h0 gives the weight-average terminal relax-
ation time t w , defined as follows:

Ú t H(t)d ln t
2

tw = -•
• = h0 J e0 (2.6)

Ú tH(t)d ln t
-•

Beyond the linear viscoelastic region—at a high strain rate or with a large
strain—rheological properties are more complicated. However, some empir-
ical rules have been proposed to relate steady-state nonlinear viscoelastic
properties to the rheological parameters in the linear region, as follows: 8,9

N 1 ( g˙ )
0.7
G ¢ ÔÏ Ê G ¢ ˆ Ô¸
2
q( g˙ ) ∫ @ 2 Ì1 + Á ˜ ý (2.7)
g˙ 2 w ÔÓ Ë G ¢¢ ¯ Ôþ
w Æ g˙

© 2004 by CRC Press LLC

www.ketabdownload.com
 s( g˙ ) G ¢¢ ÏÔ Ê G ¢ ˆ 2 ¸Ô
0.5

h( g˙ ) ∫ @ Ì1 + Á ˜ ý (2.8)
g˙ w ÔÓ Ë G ¢¢ ¯ Ôþ
w Æ g˙

where q( ġ ) is the primary normal stress coefficient, N 1 ( ġ ) is the primary

normal stress difference, h( ġ ) is the shear viscosity, and s( ġ ) is the shear
stress.
Equation 2.8 is the well-known Cox–Merz relation.9 Apparently, the pri-
mary normal stress difference, one of the elastic parameters, is affected by
G ¢ to a greater degree than by G ¢¢ , indicating that q( ġ ) is more sensitive to
terminal relaxation than h( ġ ) . In addition, the steady-state recoverable strain
g rs ( g˙ ), which relates directly to the degree of molecular orientation during
flow in a simple polymer melt, is given by the following relation:8

1.5
G ¢ ÔÏ Ê G ¢ ˆ Ô¸
2
g rs ( g˙ ) @ Ì1 + Á ˜ ý (2.9)
G ¢¢ ÔÓ Ë G ¢¢ ¯ Ôþ
w Æ g˙

The steady-state recoverable strain is also an important elastic property

for actual processing because it is responsible for the residual strain in prod-
ucts. Equations 2.7–2.9 are quite convenient for predicting nonlinear prop-
erties from linear viscoelastic properties, although they should not be
employed for complex fluids such as polymer blends and composites. Fur-
ther, they make it easier to comprehend the contribution of elastic and vis-
cous properties to each steady-state property.
The elastic properties discussed above are ascribed mainly to the anisotropy
of molecular orientation. The relation with stress is widely understood as the
stress-optical rule. Furthermore, besides the molecular orientation, chain
stretching within the tube, which also lowers the entropy of a molecular chain,
can be the origin of melt elasticity when the strain rate is higher than the
inverse of the characteristic time for chain contraction.10 This phenomenon is
similar to the stress growth of rubber due to the extended chain effect.
Pearson et al. successfully explained the overshooting of the primary nor-
mal stress difference after startup of shear flow, considering chain stretching,10
whereas the overshooting of shear stress is predicted by molecular orienta-
tion.5 Since chain contraction from stretched state to equilibrium is a more
rapid process than the relaxation of molecular orientation or reptation, the
overshooting of the primary normal stress difference takes place at a higher
shear rate than that of shear stress. The degree of chain contraction after
sudden deformation is associated with the damping function, defined as the
depression of the relaxation modulus at a large step strain from a linear one.
Osaki demonstrated that the damping function of a comb-shape polymer and
a random branched polymer like LDPE is larger than that of a linear poly-
mer.11 As for a branched polymer, especially a polymer having two or more

© 2004 by CRC Press LLC

www.ketabdownload.com
 A.

B.

C.

FIGURE 2.4
Schematic illustration of the recovery process from stretched condition for a branched polymer:
(A) equilibrium state, (B) stretched state after sudden deformation, (C) prior to retraction
process of branches (t < ta ; ta is the relaxation time for chain retraction). The chain segments
between branch points are difficult to contract to the initial equilibrium condition because
branches prohibit simple contraction along the tube, whereas the branch part can contract as
shown in Figure 2.4C. Comparison of chain contraction among linear, branched, and fully
crosslinked polymers has been discussed in detail by Laun.12

branch points in a chain, the segments between branch points cannot contract
to the equilibrium state, owing to the topological interaction associated with
long-chain branches as illustrated in Figure 2.4.12 This leads to viscous
enhancement in elongational flow. In order to contract the stretched seg-
ments, the retraction process of the branches must be completed. On the
other hand, a linear polymer can contract within the tube to the equilibrium
immediately.
Thus, once the strain rate is higher than the inverse of the characteristic
time for a chain contraction, start-up curves of elongational viscosity at a
constant strain rate exhibit nonlinear behavior, i.e., upturn departure from
the low strain rate asymptote; this is often referred to as “strain hardening”
or “viscous enhancement.” This elastic behavior affects the processability of
various processes, including foaming. Because of the difficulty of chain
contraction, strain hardening often takes place for a branched polymer. This
feature was well described in recent advanced molecular theories based on
the tube model, which can predict the nonlinear behavior under both shear

© 2004 by CRC Press LLC

www.ketabdownload.com
and elongational flow quantitatively.13–15 For example, the pom-pom model
proposed by McLeish and Larson has been applied to various deformation
modes and polymers.14 Moreover, the molecular stress function (MSF) theory
developed by Wagner et al. has made it possible to predict the upturn behavior
in elongational viscosity, considering the stress compression of branches.13
Further improvement of the MSF theory has recently provided accurate
predictions for various branched polymers.15
Strain hardening plays an important role in foaming because of the reduc-
tion of localized deformation in a thinner part owing to stress increase, which
depresses the coalescence of cells and results in uniform distribution of cell
size. Therefore, rheological properties under elongational flow should be
taken into consideration in order to predict the processability of foaming.
Strictly speaking, discussions of the processability of foaming should be
carried out based on equi-biaxial elongational flow, since biaxial rather than
uniaxial deformation takes place during the growth of bubbles. Nevertheless,
evaluation of uniaxial viscosity is still recommended because a measurement
of the growth curve of biaxial viscosity is difficult, and the results will be
less reliable. Moreover, some experimental results have clarified that material
that shows marked strain hardening in uniaxial elongational viscosity also
exhibits strain hardening in biaxial viscosity.16
In conclusion, the main origins of melt elasticity for a simple polymer melt
are as follows: (1) anisotropy of molecular orientation, which is mostly
responsible for the Weissenberg effect, the Barus effect, and strain recovery;
and (2) chain stretching of contour length, leading to strain hardening in
elongational viscosity and overshooting of normal stress difference in a
steady-shear startup curve.

2.2.2 Measurements of Elastic Response

Experimental techniques for measuring melt elasticity are briefly summa-
rized in this section. Details are well described by Macosko.17
Because of its accuracy, a cone-and-plate, drag-flow-type rheometer is
preferable for the measurement of rheological properties under shear flow,
such as oscillatory shear modulus, primary normal stress difference, recov-
erable strain, and so on. There are two systems for a cone-and-plate rheom-
eter; one is strain control and the other is stress control. The former system,
which is more widely available, is used to measure stress-relaxation and
startup behavior under shear flow, whereas the latter gives recoverable strain
after cessation of shearing. Further, steady-state properties and oscillatory
shear modulus are evaluated by both systems.
In addition to measuring shear viscosity, a capillary, pressure-flow-type
rheometer can also be used to examine extrudate swell, end-pressure drop,
extensibility of a strand, and drawdown force. Further, this rheometer is also
available to obtain the information on the occurrence of surface and/or gloss
volumetric melt fracture because of a similarity to actual extrusion process-
ing. The drawdown force, which is often employed as an elastic parameter

© 2004 by CRC Press LLC

www.ketabdownload.com
 (A)
(G)

(C) (B)

(F)
(E)
(D)

FIGURE 2.5
Typical experimental setup for drawdown force, extensibility, and extrudate swell ratio: (A)
capillary rheometer; (B) laser beam emitter; (C) detector of laser beam; (D) transducer; (E)
pulley; (F) rotating clamps; and (G) computer. (From Yamaguchi, M., J. Polym. Sci. Polym. Phys.
Ed., 39, 228, 2001. With permission.)

in industries, is defined as the force needed for stretching an extrudate strand

at a certain draw ratio, and corresponds well to the elongational viscos-
ity.16,18,19 Varying the draw ratio, the information on the elongational viscosity
at various strain rates is collected. The drawdown force at the maximum
stretching rate is called “melt strength.” A typical experimental setup for
drawdown force, extensibility, and melt strength is shown in Figure 2.5.19
As denoted in the figure, drawdown force is measured by pulling the strand
extruded vertically downward from the die set in a capillary rheometer (A) at
a certain rate by two rotating clamps (F). Furthermore, this machine enables
evaluation of extensibility—the ultimate draw ratio at the breaking point of
the strand—by increasing the pulling rate constantly. Some capillary rheome-
ters also have an attachment to measure the extrudate swell online. The diam-
eter of the extrudates is measured using a laser beam emitted from (B) to (C).
Then the extrudate swell ratio B (= D/D0), the ratio of the diameter of an
extrudate strand D to that of the capillary die D0, is calculated. The strands
are not pulled in case of extrudate swell measurements. Extrudate swell is
affected by the length of a capillary die when t 1 >> tdie , where t 1 is the longest
relaxation time and tdie is the duration in a capillary die, because the memory
of molecular conformation prior to die entrance depends on the length; i.e.,
on tdie . When a die length is long enough to “forget” the memory, i.e., when
t 1 << tdie , then B is described by the normal stress difference at steady-state
as follows:20

(
N 1 2 = 8s 2 B 6 - 1) (2.10)

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 2.1
Experimental Methods for Elongational Viscosity

Mode Method Principle Reference

Uniaxial Meissner- A rectangular or rod-shaped specimen, 21

Elongational type floating by buoyancy or air flow, is
Viscosity stretched by two rotating clamps.
Fiber A strand extruded from a die is pulled 23
spinning downward.
Converging A melt flows from a wide cross-sectional 24
flow area into a narrow one; i.e., converging.
Tubeless A molten fiber appearing from the bath 25
siphon rises upward with free surface to nozzle;
i.e., the opposite of fiber spinning.
Planar or Biaxial Meissner- A polymer sheet is stretched by multiple 26
Elongational type rotating clamps with scissors when
Viscosity needed. The deformation mode is
determined by the ratio of rotating speed
of each clamp.
Lubricant A short rod specimen with lubricant is 27
squeezing compressed. Biaxial flow is attained by a
simple one-axis compression and planar
elongation is given by the compression
of the specimen embedded in a
rectangular channel.
Stagnant Elongational flow field, together with a 28
flow stagnation point, is generated by
impinging two steady-flows that flow
the opposite direction in a line.
Sheet A polymer sheet is inflated by air. Inflation 29
inflation of a sheet over a circular hole attains
biaxial elongation and that from a
polymer tube provides planar
elongation.

In actual extrusion processing, however, the swell ratio is larger than the
value predicted by the equation because molecular chains still “memorize”
the conformation prior to entering the die.
The rheological properties under uniaxial extensional flow for a foamable
material have often been measured by means of the well-known Meissner-
type elongational rheometer,17,21,22 although other methods23–25 have been
proposed, as listed in Table 2.1. Two types of machines are widely available
for Meissner-type uniaxial extension. One of them is an air-supported
system. A rectangular or rod-shaped sample is held by an upward air flow
against gravity and gripped by two rotating clamps fixed in place. The
clamps at both ends rotate at a constant rate, thereby pulling the sample
at a constant strain rate.21 The actual strain rate is usually determined by

© 2004 by CRC Press LLC

www.ketabdownload.com
the time variation of the diameter or width of the sample, which is moni-
tored by a video camera located over the sample. The other system is
composed of the same attachment but uses an oil bath to float the sample
by buoyancy. Both systems give reliable experimental results, with the
exception of samples with low viscosity. In contrast to the development in
uniaxial elongational measurements, the apparatus for equi-biaxial and
planar elongational viscosities, especially their transient response, is less
widely available. Therefore, only a few data have been published, although
various methods have been proposed, as shown in Table 2.1.26–29

2.2.3 Role of Long-Chain Branches

2.2.3.1 Model Polymer
Molecular structure, especially that of long-chain branches, predominantly
affects melt elasticity. Therefore, linear viscoelastic properties and solution
properties of a branched polymer have been well investigated employing
model polymers such as star- and comb-shaped polymers.30–33 Steady-state
compliance increases linearly with branch length for a star-shape polymer,
irrespective of the number of branches, while compliance for a linear polymer
is independent of molecular weight above a critical molecular weight.4,30,31
Broadening of the terminal relaxation mode due to chain retraction of
branches is responsible for the enhancement of steady-state compliance.30,31
Since zero-shear viscosity increases exponentially with branch length, the
relaxation time, i.e., t = J e0 h0 , also greatly increases. As for a comb-shape
polymer, the terminal relaxation mode is the backbone relaxation that takes
place after the chain retraction process for branches as discussed above. The
characteristic time is proportional to Mb 2 t a , where Mb is the molecular
weight of the backbone chain, i.e., the chain segments between branch points,
and t a is the relaxation time for chain retraction.30,32
Raju et al. revealed that apparent flow activation energy increases linearly
with the length of branches employing hydrogenated polybutadiene, the
model polymer of polyethylene.33 Since a larger shift factor is required only
at the terminal zone—i.e., at a lower modulus—for the polymers with long-
chain branches, time-temperature superposition is not applicable to them.34
Thermo-rheological complexity is also connected with the chain retraction
process, since a gauche-rich, compact conformation, which is dominant
during the retraction of branches, has a higher energy state than a trans-
rich, equilibrium conformation.34 Consequently, the existence of long-chain
branches can easily be detected by flow activation energy, at least for PE,
although detailed characterization of branch structure should not be car-
ried out only by flow activation energy for commercial LDPE with com-
plicated structures.16 In the cases of polybutadiene, polystyrene, and
polyisoprene, however, the effect of branch structure on the activation
energy is obscure because the energy difference between both conforma-
tions is not large enough.34

© 2004 by CRC Press LLC

www.ketabdownload.com
2.2.3.2 Low-Density Polyethylene
The branch structure of commercial LDPE has generally been studied accord-
ing to dilute solution properties, based on the pioneering work performed
by Zimm and Stockmeyer.35 Recently, combination measurements of GPC
and light scattering as well as intrinsic viscosity [h] have become powerful
tools for characterizing the branch structure of commercial LDPE in detail.
2
The mean square radius of gyration · Rg Ò of a branched polymer is lower
than that of a linear polymer having the same molecular weight. Their ratio,
defined by Equation 2.11, shows the number of branches:35

Rg 2
g= 2
branched
(2.11)
Rg
linear

Further, since the intrinsic viscosity [h] of a branched polymer is also smaller
than that of a linear polymer, their ratio g’ is smaller than unity:35

g¢ =
[ h] branched (2.12)
[ h] linear
Finally, the parameter b, given by Equation 2.13, gives information on the
branch structure:

g¢ = gb (2.13)

Tackx and Tacx have investigated the branch structure of commercial

LDPE and have found that LDPE produced by the autoclave process, having
lower g and g’, has more long-chain branches than tubular LDPE. Moreover,
they report that the parameter b decreases with an increase in molecular
weight, demonstrating that the molecular shape is changing from star-like
to comb-like with molecular weight for both tubular and autoclave LDPE.36
In addition to dilute solution properties, rheological properties in a molten
state have been studied for commercial LDPE. Yamaguchi and Takahashi
compared various rheological properties for autoclave and tubular LDPE
melts having similar linear viscoelastic properties.16 As seen in Table 2.2, the
autoclave LDPE shows a higher level of drawdown force and a greater extru-
date swell ratio, even though the melt flow ratio (MFR) is larger than that of
tubular LDPE, demonstrating that autoclave LDPE exhibits higher melt elas-
ticity. In particular, the drawdown force of autoclave LDPE is twice as high
as that of tubular LDPE. This result can be attributed to the complicated,
hierarchical branch structure of autoclave LDPE, which is in accordance with
its dilute solution properties.16 Because of its high drawdown force, autoclave
LDPE shows poor extensibility compared with tubular LDPE.
Further, the damping functions of autoclave and tubular LDPE are quite
different. Autoclave LDPE shows larger values (less damping) than tubular

© 2004 by CRC Press LLC

www.ketabdownload.com
TABLE 2.2
Comparison of the Rheological Parameters Obtained by Capillary Rheometer for
Autoclave LDPE and Tubular LDPE
Drawdown
MFR (g/10 min) Force (mN) Swell Ratio
190˚C, 2.16 kg draw ratio = 5 g = 10.8 s–1 Extensibility
Autoclave 1.6 295 1.91 <15
LDPE
Tubular 1.0 135 1.55 50–60
LDPE
* Drawdown force, swell ratio, and extensibility were measured by the die with L/D = 8/2.095
at 160˚C.

LDPE,16 suggesting that the branch structure of autoclave LDPE suppresses

the chain contraction. Chain stretching due to multi-branch points causes
autoclave LDPE to exhibit marked strain hardening in elongational viscosity.
As shown in Figure 2.6, autoclave LDPE shows more pronounced upturn
behavior at all strain rates.16 On the other hand, tubular LDPE shows less
viscous enhancement. Further, the upturn behavior becomes weak at a lower
strain rate, i.e., at a smaller Deborah number.

2.2.4 Melt Elasticity of New Polyolefins

Metallocene catalyst technology is a recent development in the field of poly-

olefins.37 LLDPE (ethylene/a-olefin copolymer) produced by a metallocene
catalyst has excellent mechanical properties in the solid state, such as high
tensile and impact strength, compared with conventional Zieglar–Natta
LLDPE, since the number of short-chain branches is almost independent of
the molecular weight. Because of narrow molecular weight distribution,
however, poor processability due to less melt elasticity often becomes a
serious problem. Therefore, blending with LDPE has been carried out exten-
sively in order to improve the melt elasticity. Further, peroxide modification,
which generates long-chain branches without microgel to enhance the melt
elasticity, has also been investigated.38 Irradiation by electron beam has also
been found to generate long-chain branches in polyethylene. Horii et al.
studied the branch structure of radiation-modified polyethylene and
revealed that both Y-type and X-type branch points exist.39
The processability of foaming for fully crosslinked metallocene LLDPE
by peroxide modification has been investigated by Abe and Yamaguchi.3
According to their study, a foam with a high expansion ratio is obtained
when the average molecular weight between neighbor crosslinked points
is around 105. Further, the rheological properties of the crosslinked LLDPE
are found to be similar to those of a critical gel. Moreover, the expansion
ratio of a crosslinked LLDPE foam is determined not only by the rheological
properties in the molten state but also by the crystallization temperature,

© 2004 by CRC Press LLC

www.ketabdownload.com
 A.
7

160°C
6
log [ηE+ (t, ε) (Pa s)]

5
.

0.019
3η+ (t) 0.039
h+
4 0.077
0.150
tubular LDPE 0.350
0.691
3
−1 0 1 2 3
log [t (s)]

B.
7

160??
160°C
6
log [ηE+ (t, ε) (Pa s)]

5
.

3η+ (t)

0.038
4 h+
0.078
0.158
autoclave LDPE
0.358
0.720
3
−1 0 1 2 3
log [t (s)]

FIGURE 2.6
Growth curves of uniaxial elongational viscosity hE+ (t , e˙ ) at various strain rates ė for (A)
tubular LDPE and (B) autoclave LDPE at 160˚C. The solid line represents 3h+(t), where h+(t) is
transient shear viscosity at the low strain rate asymptote calculated from linear viscoelastic
properties. Numbers in the figure denote the strain rate. (From Yamaguchi, M. and Takahashi
M., Polymer, 42, 8663, 2001. With permission of Elsevier Science.)

© 2004 by CRC Press LLC

www.ketabdownload.com
i.e., the number of short-chain branches. This result reveals that shrinkage
after expansion is prominent for crosslinked PE in spite of rapid cooling
due to adiabatic expansion of gas in cells. As a result, LLDPE with a low
crystallization temperature is allowed to shrink more until crystallization
prohibits large-scale molecular motion completely.3 Further, Abe clarified
that the crosslinked foam obtained from LLDPE exhibits a homogeneous
distribution of cell size, which is quite different from the distribution pro-
duced by conventional LDPE.40 The homogeneous distribution of crosslink
points in crosslinked LLDPE results in an excellent micro-cellular structure.
Further effort has been expended in metallocene technology on improving
processability. A metallocene catalyst makes it possible to incorporate long-
chain branches sparsely into a backbone chain,41,42 which enhances flow acti-
vation energy41 and strain hardening in elongational viscosity to some
degree.42 Malmberg et al. obtained HDPE with long-chain branches by means
of a metallocene catalyst, which shows higher flow activation energy than
conventional HDPE.43 There are many more long-chain branches in this ver-
sion of HDPE than in HDPE produced by means of a chromium catalyst.44
The level of melt elasticity of LLDPE and HDPE with long-chain branches
produced by means of a metallocene catalyst is, however, still considerably
lower than that of conventional LDPE, due to the presence of fewer and/or
shorter long-chain branches.41,42 In the near future, the number and length of
long-chain branches will be controlled by new technologies, such as the
Brookhart catalyst,45 or by copolymerization with a macromonomer obtained
by a metallocene46 or phenoxyimine-based catalyst.47 These advanced tech-
nologies in the polyolefin industry may enable the synthesizing of a polymer
with the appropriate melt elasticity for foaming. They also make it possible
to produce a block or graft copolymer that can be employed for foaming
because they may show unique mechanical properties in the solid state and
widen the application of PE foam. Moreover, copolymers with a polar mono-
mer, which may have attractive mechanical or environmental properties, will
be available in the future. Currently, an ethylene/styrene copolymer48 is avail-
able, and its foam properties have been studied.49
The effort to incorporate long-chain branches has been made not only for
polyethylene but also for isotactic polypropylene (PP), since the poor melt
elasticity of conventional linear PP prohibits using it for foaming. In the past
decade, PP with long-chain branches has been developed by radiation or
peroxide modification and has become commercially available.1,50–53 One of
the targets for the long-chain branch PP is, of course, a foam. Marked strain
hardening in elongational viscosity, which has been described in detail by
Hingmann and Marczinke and by Sugimoto et al.,51 results in less coalescence
of cells and leads to a high expansion ratio with fine cells.50,52 Because of
resistance to heat and organic solvents as well as excellent mechanical prop-
erties, PP foams are expected to replace some polystyrene foams. Its use in
trays for oily foods is one proposed application.53

© 2004 by CRC Press LLC

www.ketabdownload.com
 Although extensive studies have already been carried out on the molecular
architecture of polyolefins, further development bringing about new, tailor-
made polyolefins is also expected in the next decade. This will widen the
applications of polyolefin foams and improve their processability.

2.3 Modification of Melt Elasticity

2.3.1 Effect of Processing History

2.3.1.1 Origin of Shear Modification

Applied processing history has a significant effect on the rheological prop-
erties of a branched polymer, especially on the melt elasticity, including strain
hardening in elongational viscosity, even though it has no effect on the
primary molecular characteristics, such as molecular weight, polydispersity,
and the number and length of long-chain branches.54–68 The phenomenon
called “shear modification” or “shear working” is a reversible, mechanical
modification of the rheological properties, since exposure to annealing or
solvent treatment enables the rheological properties to recover to those of
the original, unprocessed material. Further, the mechanical modification is
barely observable for linear polymers,60,63 indicating that branch structure is
primarily relevant to the phenomenon.
Table 2.3 shows normalized shear storage and loss moduli, i.e., the ratio
of values for the sample processed by an internal batch mixer to those for
the unprocessed sample at various frequencies. The processing history in an
internal mixer noticeably depresses oscillatory shear moduli, especially G ¢
rather than G ¢¢, at lower frequencies, although there is no change in the
molecular weight and the polydispersity.66
The result suggests that applied processing history weakens the effect of the
terminal relaxation mechanism, which is typical of mechanical modification.
The elastic properties, such as primary normal stress difference, extrudate
swell, recoverable strain, and strain-hardening in elongational viscosity, are
thereby more susceptible to processing history than are the viscous proper-

TABLE 2.3
Normalized Oscillatory Shear Moduli for the Processed LDPE
0.01 Hz 0.03 Hz 0.1 Hz 0.3 Hz 1 Hz

G’/G’eq 0.394 0.593 0.696 0.790 0.899

G”/G”eq 0.628 0.782 0.886 0.959 0.994

¢ and Geq
* Geq ¢¢ represent G ¢ and G ¢¢ of the original, unprocessed sample, respectively.

* The sample, commercial LDPE produced by the autoclave process, was processed by a
60-cc internal batch mixer at 30 rpm for 120 min at 160˚C.

© 2004 by CRC Press LLC

www.ketabdownload.com
FIGURE 2.7
Growth curves of uniaxial elongational viscosity hE+ (t , e˙ ) at various strain rates ė at 160˚C for
(o) unprocessed, original LDPE and (•) LDPE processed by a 60 cc internal batch mixer at 30
rpm for 15 min at 160˚C. The solid line represents 3h+(t) at the low strain rate asymptote
calculated from linear viscoelastic properties of the unprocessed sample. (From Yamaguchi et
al. Proceedings of the Annual Meeting of the Society of Polymer Processing, PPS-18. With
permission of the Polymer Processing Society.)

ties.58–68 Figure 2.7 exemplifies the growth curves of uniaxial elongational vis-
cosity at various strain rates for LDPE having different processing histories.67
Unprocessed LDPE, represented in the figure by open symbols, deviates
upward from 3h+ (t) (strain-hardening behavior) beyond a certain strain level,
whereas the sample processed by an internal mixer shows less upturn. In
contrast to this, the linear region of the transient elongational viscosity, in
which hE+ (t) is independent of ė , is hardly changed by the applied process-
ing history. The growth curves for the processed sample, however, may
include some experimental error, because thermal history prior to and/or
during extension has a strong influence on the transient elongational viscos-
ity. Drawdown force is also found to be quite sensitive to applied processing
history16,66–68 and is preferably employed in order to evaluate the magnitude
of the mechanical modification, since the experimental error arising from
applied thermal history after processing can be minimized, owing to the
simplicity of the measuring method. In Figure 2.8, normalized drawdown
force—i.e., the ratio of the drawdown force of the processed sample to that
of the original measured at the same temperature—is plotted against the
processing time in an internal mixer.
© 2004 by CRC Press LLC
www.ketabdownload.com
 1.2

1.0 160°C

0.8
normalized drawdown force

0.6
tubular LDPE

0.4

0.2
autoclave LDPE

0
0 30 60 90 120 150
processing time (min)

FIGURE 2.8
Normalized drawdown force at 160˚C plotted against processing time in an internal mixer at
30 rpm at 160˚C for (o) autoclave LDPE and (•) tubular LDPE. The drawdown force, measured
at a draw ratio of 5, was collected without exposure to annealing in a capillary rheometer.

It has been generally accepted that intense processing history—high strain

rate and/or long residence time in a processing machine—depresses the
drawdown force notably,16,66 which corresponds to the result in Figure 2.8.
Furthermore, the mechanical modification of elastic properties is more
apparent for a branched polymer with well-developed branch structure, as
described below. Thus, LDPE produced by the autoclave process and having
a more hierarchical and complicated branch structure is more susceptible to
applied processing history than is tubular LDPE.16 Taking notice of the phe-
nomenon, Rokudai proposed a characterization method of branch structure
in LDPE,57 which is quite useful because obtained characteristics directly
relate to processability.
Exposure to annealing or solvent treatment makes the rheological proper-
ties recover to the original values. Figure 2.9 shows the recovery curves of
oscillatory shear moduli, such as storage modulus G ¢ and loss modulus G ¢¢ ,
as a function of annealing time in a rheometer for a sample processed in an
internal mixer.66
Oscillatory shear moduli, especially shear storage modulus G ¢ at a lower
frequency, increases more rapidly with annealing time. This is plausible
because applied processing weakens, especially in terminal relaxation mode,
© 2004 by CRC Press LLC
www.ketabdownload.com
as shown in Table 2.3. Therefore, the increase in G ¢ at low frequencies is
prominent, owing to the recovery of terminal relaxation from the mechanical
modification. Similar to G ¢ , normalized drawdown force, the drawdown
force divided by the plateau value at each temperature, also grows rapidly
with annealing time and approaches the original value at the corresponding
temperature, i.e., 1, as shown in Figure 2.10. The figure indicates that the
normalized drawdown force at t = 0 (the value without exposure to anneal-
ing) is independent of the measurement temperature, suggesting that the
same ratio of the drawdown force is lost at each temperature. This result
implies that the normalized drawdown force without annealing, regardless
of the measuring temperature, is the appropriate elastic parameter for dis-
cussing the magnitude of the applied processing history. The figure also
demonstrates that quite a long time is required to recover, especially at a
lower temperature; this becomes more obvious for the sample with a more
intense processing history.16,66

160 C

G′′
log [G′ (Pa)], log [G′′(Pa)]

4 1Hz
G′

3 G′′

0.01Hz

G′
2
0 10 20 30
annealing time (min)

FIGURE 2.9
Growth curves of shear storage modulus G¢ (closed symbols) and loss modulus G≤ (open
symbols) at 1 Hz (circles) and 0.01 Hz (diamonds) at 160˚C plotted against annealing time in
a cone-and-plate rheometer at 160˚C for LDPE processed in an internal mixer at 30 rpm at 160˚C
for 120 min. (From Yamaguchi, M. and Gogos, C.G., Adv. Polym. Technol., 20, 261, 2001. With
permission of John Wiley & Sons, Inc.)

© 2004 by CRC Press LLC

www.ketabdownload.com
 1.0

160°C
145°C
0.8
normalized drawdown force

130°C

0.6

0.4

0.2

0
0 10 20 30
annealing time (min)
FIGURE 2.10
Growth curves of normalized drawdown force at a draw ratio of 5 with annealing time, i.e.,
residence time in the capillary rheometer at the measuring temperature, after processing history
at various temperatures; (•) 130˚C, (o) 145˚C, and (®) 160˚C. The sample was processed three
times by a 30-mm co-rotating twin-screw extruder having conventional conveying screws at
151˚C at 50 rpm.

The quantitative relation between applied shear history and its resultant
rheological properties has not been studied for a long time except by Han-
son,55 who concluded that the applied total strain determines the rheological
properties. Nevertheless, recent studies employing a cone-and-plate rheom-
eter as a processing machine made it clear that applied shear stress, not
strain, determines the rheological properties of a processed sample.66,67
Figure 2.11 shows the recovery curves of the storage modulus plotted against
the annealing time for a sample sheared by a cone-and-plate rheometer.67,68
The figure demonstrates that the recovery curves are determined not by
the total shear strain, i.e., the product of the applied shear rate and the
duration of shearing as shown in Figure 2.11A, but by the product of the
applied shear stress and the duration of shearing as shown in Figure 2.11B.
Further, the result suggests that torque and residence time in an extruder or
a batch mixer are fundamental parameters for controlling rheological prop-
erties, and thus processability. Figure 2.11 also indicates that processing at
a lower temperature, i.e., at a higher stress because of high viscosity,
depresses the melt elasticity more, as demonstrated by Maxwell.62,63 The

© 2004 by CRC Press LLC

www.ketabdownload.com
 A.
500

0.01 Hz 160°C

G′(t) (Pa) 400

300

.
γ ts = const.
200
10 s−1, 30 min
20 s−1, 15 min
30 s−1, 10 min
100
0 20 40 60
annealing time (min)
B.
500

0.01 Hz 160°C

400
G′(t) (Pa)

300

200 σ ts = const.
10 s−1, 30 min
20 s−1, 23 min
30 s−1, 19 min
100
0 20 40 60
annealing time (min)

FIGURE 2.11
Growth curves of shear storage modulus G’ at 0.01 Hz at 160˚C for the samples sheared in a
cone-and-plate rheometer at 160˚C. (A) Applied total strain, i.e., the product of the shear
rate ġ and the duration of shearing ts, is the same for all samples: (o)ġ = 10 s–1 for 30 min,
(‡) ġ = 20 s–1 for 15 min, and (
)ġ = 30 s–1 for 10 min. (From Yamaguchi, M., and Gogos, C.G.,
Adv. Polym. Technol., 20, 261, 2001. With permission of John Wiley & Sons, Inc.) (B) All samples
have the same value of the product of the applied shear stress s and the duration of shearing
ts; (o) ġ = 10 s–1 for 30 min, (‡)ġ = 20 s–1 for 23 min, and (
)ġ = 30 s–1 for 19 min. (From
Yamaguchi, M. et al. Proceedings of the Annual Meeting of the Society of Polymer Processing,
PPS-18, 2002. With permission of the Polymer Processing Society.)

© 2004 by CRC Press LLC

www.ketabdownload.com
results in Figure 2.11 will be explained as regards chain alignment to the
flow direction. Molecular orientation, which is proportional to the stress, is
responsible for disentanglement of temporary couplings associated with
long-chain branches.60,65,66 Consequently, the polymer behaves like a linear
polymer.65 The schematic model of molecular conformation, based on the
tube model, is illustrated in Figure 2.12, in which dots denote neighbor
chains.
Exposure to flow and deformation in a processing machine drags the
branches into the same tube of a backbone chain. Consequently, contraction
of backbone chains between branch points takes place easily, as with a linear
polymer, which results in less strain hardening in elongational viscosity, along
with a low magnitude of the damping function. Moreover, the terminal relax-
ation mechanism, i.e., backbone relaxation after contour length fluctuation of
branches, disappears, which allows the polymer chain to move along the tube
by simple reptation, as with a linear polymer. The recovery of rheological
properties by exposure to annealing can be ascribed to the relaxation of
molecular conformation from a low entropy state, especially around branch
points, to the equilibrium state by micro-Brownian motion. Therefore, higher
temperature enables the rheological properties to recover more quickly, as
shown in Figure 2.10. In contrast, keeping the processed sample below the
crystallization temperature, or glass-rubber transition temperature in case of

flow annealing

FIGURE 2.12
Schematic illustration of molecular conformation (top) before and (bottom) after mechanical
modification, based on the tube model. Molecular motion of the polymer chain represented by
a solid line is confined within the “tube” composed of neighbor chains, denoted by dots. Top:
conformation at equilibrium state. Completion of retraction process of branches is required for
terminal backbone relaxation. Further, chain contraction after sudden deformation is reduced
by the topological constraint. Bottom: Conformation after processing. Because of molecular
orientation resulting from applied flow, branches are aligned to the main chain, which makes
it possible for the branched polymer to use simple reptation along the tube as well as chain
contraction like a linear polymer.
© 2004 by CRC Press LLC
www.ketabdownload.com
 1.2

1.0

tubular LDPE
normalized drawdown force

0.8

0.6

autoclave
LDPE
0.4

0.2 190°C

0
0 10 20 30
annealing time (min)

FIGURE 2.13
Growth curves of normalized drawdown force at a draw ratio of 5 at 190˚C plotted against
annealing time, i.e., residence time in the capillary rheometer at 190˚C, for (o) tubular LDPE
and (•) autoclave LDPE. The samples were processed by an internal mixer at 30 rpm for 30
min at 160˚C.

amorphous polymers,59 retains the rheological properties of the melt because

molecular motion is frozen. Furthermore, it takes a longer time in the molten
state to recover the rheological properties for a sample having a more intense
applied shear history, indicating that intense flow and deformation cause
longer branches to align to a greater degree.
Figure 2.13 compares the recovery curves of normalized drawdown force
for tubular and autoclave LDPE having the same processing history. Auto-
clave LDPE shows a significantly lower value than tubular LDPE, indicating
that applied flow and deformation efficiently lose the feature of the branched
polymer having well-developed branch structure; this is due to the confor-
mation change. Further, the figure demonstrates that autoclave LDPE needs
a longer time to recover from mechanical modification.
In the recovery curve, the low value at t = 0 indicates that more long-chain
branches exist in autoclave LDPE. Moreover, taking a longer time to recover
indicates the existence of longer branches and/or a more hierarchical branch
structure.
As we have observed, the rheological properties of a branched polymer are
determined by mechanical and thermal histories. Under a low shear rate,
© 2004 by CRC Press LLC
www.ketabdownload.com
 7

6 original pellet
η+(t) (kPa s)

processed sample

4
0.4 s−1

190°C

3
0 150 300 450 600
t (s)

FIGURE 2.14
Growth curves of shear viscosity h+(t) at 0.4 s–1 at 190˚C for (o) the original pellet and (•) the
sample processed by an internal mixer at 30 rpm for 150 min at 150˚C. (From Yamaguchi, M.
et al. Proceedings of the Annual Meeting of the Society of Polymer Processing, PPS-18, 2002.
With permission of the Polymer Processing Society.)

micro-Brownian motion may overcome the hydrodynamic force and then

enhance the interaction due to entanglement couplings. Figure 2.14 shows the
growth curve of shear viscosity at a constant shear rate for the sample pro-
cessed by an internal mixer, as compared with the original, unprocessed sam-
ple.67 The shear viscosity of the processed sample increases monotonically even
under shear flow, indicating micro-Brownian force is dominant in this condi-
tion. Further, it is of great interest that stress overshoot, which is ascribed to
overshooting of molecular orientation,5 disappears in the processed sample,
whereas it is clearly detected in the growth curve of the original sample.
The recovery curve from mechanical modification, as described here,
implies information on branch structure. Further research on the quantitative
characterization of the recovery curve employing a model branched polymer
will provide a better understanding of the unique rheological properties as
well as the processability of a branched polymer.

2.3.1.2 Processing History by Conventional Processing Machines

It has been reported that inflation-blown film from a mechanically modified
sample is more transparent than that from the original pellet, which is due
to the low level of external haze, or smooth surface.55,58,60,64 The depression
of surface roughness relates to the weak melt elasticity according to the
processing history. Flow instability in extrusion is also improved by
mechanical modification.54,56,67,68 Due to low melt elasticity, gross, volumet-
© 2004 by CRC Press LLC
www.ketabdownload.com
ric, and chaotic melt fractures, which are caused by long time relaxation
(marked melt elasticity)69 are reduced or take place at a higher shear rate
by means of applied processing history. Moreover, mechanical modification
enhances extensibility in a molten state 60,64 because the depression of strain
hardening in elongational viscosity reduces the stress needed for stretching.
To the best of our knowledge, the direct relation between mechanical mod-
ification and the processability of foaming has unfortunately not been
reported. However, it is expected that reduction of strain hardening for a
processed sample presumably leads to coarse cell morphology, owing to
the coalescence of cells.
The effect of processing history on rheological properties is, of course,
dependent on the processing machine, as shown in Figure 2.15. Although it
is difficult to comprehend quantitatively the difference among conventional
processing machines because of the complicated flow pattern, Figure 2.15
demonstrates that two-roll mill processing has a different influence on the

1.2

1.0
normalized drawdown force

0.8

0.6

0.4

two-roll mill
twin-screw ext.
0.2
conveying screws
160°C
kneading blocks

0
0 10 20 30
annealing time (min)

FIGURE 2.15
Growth curves of normalized drawdown force at a draw ratio of 5 at 160˚C plotted against
annealing time, i.e., residence time in the capillary rheometer at 160˚C, for samples (®) processed
by two-roll mill, and processed three times by co-rotating twin-screw extruder with (•) con-
veying screw elements and (o) forward kneading blocks. (From Yamaguchi, M. et al. Proceed-
ings of the Annual Meeting of the Society of Polymer Processing, PPS-18, 2002. With permission
of the Polymer Processing Society.)

© 2004 by CRC Press LLC

www.ketabdownload.com
rheological properties than processing by other machines. The result suggests
that understanding the processing history from the viewpoint of mechanical
modification is required in order to choose a processing machine as well as
to control processability. According to the flow simulation in a two-roll mill
carried out according to Gaskell’s model,70 the magnitude of the product of
the applied shear stress and the duration of shearing for each pass through
the nip region is too low to depress the melt elasticity.66–68 Moreover, after
passing through the nip region, molecular orientation can relax on the roll
because no deformation takes place except in the nip region. This is why
two-roll mill processing barely depresses melt elasticity.
Compared to the intermittent stress history in two-roll mill processing, an
internal mixer, as shown in Figure 2.7, and extruders, which apply relentless
stress history, depress the melt elasticity drastically. Furthermore, Figure 2.15
shows that screw configuration in a co-rotating twin-screw extruder affects
the melt elasticity to different degrees. The samples in the figure were
extruded three times by a twin-screw extruder with (a) 11.5 of 20-mm pitch
forward conveying screw elements (the total length of the melt conveying
region is 210 mm) or (b) 5 forward kneading blocks KB45/5/42 (each unit
is 42-mm long with 5 discs staggered at 45˚; total length of melt conveying
and/or kneading region is 210 mm), followed by two reverse elements R20/
10 (a unit 10-mm long with a pitch of 20 mm) at fully filled condition.68 These
two types of screw configuration provide the same level of extrusion torque,
average residence time, and resin temperature at the same rotation rate. As
seen in Figure 2.15, processing with the conveying screws depresses the melt
elasticity more than with kneading blocks as long as the torque and the
residence time are the same. Further, the appearance of extrudate strands is
quite different, as shown in Figure 2.16.
The extrusion by conveying screw elements diminishes gross melt fracture,
whereas it is still observed at the strand extruded three times by kneading
blocks. Considering that gross melt fracture is ascribed to terminal relax-
ation,69 the result also supports that conveying screws depress the melt
elasticity more than kneading blocks. The flow direction in kneading blocks
changes largely, abruptly, and frequently with time and location because of
the compression–expansion deformation of a molten polymer.71 On the other
hand, a polymer melt flows along a figure-eight pattern and thus gradually
changes its flow direction along conveying screws.72 The large, abrupt, and
frequent change of the flow direction prohibits a high degree of molecular
orientation, which leads to a low level of melt elasticity. Further, more energy
is consumed to change the flow direction in kneading blocks. Consequently,
the degree of molecular orientation is less than that in the conveying screws
at the same torque level. It is, however, interesting to note that the relentless
change of flow direction, together with elongational deformation, leads to
very effective mixing for polymer blends and composites.71–74
From the experimental results shown in this section, the processing
machine as well as the operating condition should be carefully chosen for
foaming in view of the mechanical modification of rheological properties.
© 2004 by CRC Press LLC
www.ketabdownload.com
 A. 1st 2nd 3rd

B. 1st 2nd 3rd

FIGURE 2.16
Appearance of the strands extruded by the co-rotating twin-screw extruder with (A) conveying
screw elements and (B) forward kneading blocks. (From Yamaguchi M. et al. Proceedings of
the Annual Meeting of the Society of Polymer Processing, PPS-18, 2002. With permission of the
Polymer Processing Society.)

Since the bubble growth process in foaming requires relatively high melt
elasticity, mechanical modification will not be preferable in many cases.
When premixing with filler, such as a flame retardant, wooden powder,
nucleating agent, or pigment, a foaming material must be composed of a
branched polymer. The following should be kept in mind irrespective of the
amount of the filler: Mixing at a high stress condition should be avoided,
and employing a mixing screw is highly desirable for effective mixing with-
out reduction of melt elasticity.

2.3.2 Modification by Polymer Blends

2.3.2.1 Background
Polymer blends and composite systems exhibit complicated rheological
properties.22 As for immiscible polymer blends and polymer composites, there
appears to be an additional relaxation mechanism, which is ascribed to phase
separated morphology. Further, the characteristic time of the relaxation
mechanism is, in general, longer than the reptation time of a matrix polymer.
© 2004 by CRC Press LLC
www.ketabdownload.com
Nevertheless, the relaxation mechanism is of less importance at least for strain
hardening in elongational viscosity, because it is not responsible for the chain
stretching. In cases where the dispersed particles can be regarded as a rigid
body or have a higher viscosity than the matrix, the system is in so-called
suspension. Shikata and Pearson revealed that Brownian motion, the origin
for restoring the equilibrium state from flow-distorted particle distribution, is
responsible for melt elasticity in a suspension system, although it is weak.22,75
Furthermore, Mewis and Metzner have found that suspension with fiber
particles having a large aspect ratio exhibits a high level of steady-state
elongational viscosity because of the microscopic distortion of the flow field
by rigid fibers.17,22,76,77 However, the system does not show strain hardening,77,78
whereas a marked overshoot of shear stress and primary normal stress
difference is observed under shear flow, owing to the orientation of fibers.76
According to Laun, elastic strain under elongational flow for a composite with
long fibers decreases with increasing filler content, since rigid fibers impede
the elongation of polymer melt.77 Moreover, it has been generally accepted
that extrudate swell is often depressed by blending fillers, regardless of the
existence of long relaxation time in the system.79 The result demonstrates that
some elastic properties of polymer composites and polymer blends, especially
the properties ascribed to the interaction due to entanglement couplings, are
less than those of the unfilled polymer melt. Furthermore, the discontinuity
of viscosity or modulus at the interface may cause localized deformations,
which leads to poor extensibility.
When the viscosity of a matrix is higher than that of a dispersion, the
interfacial tension contributes to the terminal relaxation because additional
force is generated by deformation of the dispersion, owing to the increase
in the interfacial area.17,22,80,81 Palierne focused on the phenomenon and pro-
posed a constitutive equation, called the emulsion model, of linear viscoelas-
tic properties for immiscible blend systems in which the matrix has higher
viscosity.81 According to the model, a shoulder in the G ¢ curve in a low
frequency region for an immiscible blend is attributed to the relaxation from
the deformed state of the dispersion to a sphere in order to minimize the
interfacial tension.81 Further, Doi and Ohta studied the effect of interfacial
tension on nonlinear viscoelastic properties and constructed a constitutive
equation.82 Their theory agrees well with experimental results, demonstrat-
ing that primary normal stress difference as well as shear viscosity are
enhanced by interfacial tension in a polymer blend. Strain hardening is,
however, not significantly enhanced by blending an immiscible component.
The most outstanding elastic property observed in this system is pronounced
strain recovery after cessation of flow.83 Although the exact relation with
foaming has not been studied, shrinkage of the cells may take place after
processing, especially in a foam produced from a polyolefin with a low
degree of crystallinity.
Stated briefly, micro-heterogeneous morphology does not directly relate to
strain hardening, one of the most important elastic properties for the bubble
growth process in foaming, although it provides long time relaxation.
© 2004 by CRC Press LLC
www.ketabdownload.com
 In the case of a miscible blend system, the terminal relaxation mechanisms
are entanglement couplings, which bring about strain hardening. The rheo-
logical properties of a blend with a polymer with long-chain branches have
been extensively studied, since long-chain branches are responsible for ter-
minal relaxation as well as chain stretching. The blend enhances the melt
elasticity, including the strain hardening, of a linear polymer to some degree.
Further, it is quite easy to control processability because the rheological
properties, or at least the linear viscoelastic properties, of the blends are
between those of the individual components. Therefore, some blends, such
as LLDPE/LDPE4,84,85 and LDPE/Ionomer,86 are preferably employed for
various kinds of processing, including foaming. Polyethylene blends are one
of the most popular blends employed commercially. Although there is still
controversy over this, most polyethylene blends, as far as the difference in
the number of short-chain branches, are within 40 per 1000 backbone carbon
atoms, and thus are considered miscible in the molten state from the rheo-
logical point of view.87 As a result, strain hardening of LLDPE and HDPE is
improved by blending LDPE, owing to the chain stretching of LDPE.63,84,88
Another important miscible blend system is a blend with an ultrahigh
molecular weight fraction of the same polymer species. It has been demon-
strated by Münstedt89 and Linster and Meissner90 that broadening the mole-
cular weight distribution, i.e., broadening the relaxation spectrum, is also
responsible for strain hardening, at least when the Deborah number associ-
ated with terminal relaxation is beyond 0.5:

1
t 1e˙ > (2.14)
2

Under this condition, coil-stretch transition takes place owing to the drag
flow, and then the molecular chain is stretched. As shown by Dumoulin et
al., however, mechanical mixing may not work for homogenizing the blend.91
This is reasonable because it takes a long time to dissolve each molecular
chain, owing to the low diffusion constant of a high molecular weight frac-
tion. Consequently, mixing has to be carried out intensively to obtain the
thinner striation thickness, i.e., diffusion distance, of an ultrahigh molecular
weight fraction. Nevertheless, the intense mixing often induces molecular
scission due to high stress.

2.3.2.2 Blends with Weak Gel

Recently, Yamaguchi and Miyata found that blending a small amount of a

crosslinked polymer characterized as a weak gel near the sol-gel transition
point enhances melt elasticity of a linear polymer drastically, whereas it has
less or no influence on steady-state shear viscosity.92,93 This is an interesting

© 2004 by CRC Press LLC

www.ketabdownload.com
and anomalous phenomenon since most polymer blends and composites,
except for the system having a branched polymer as a matrix, do not exhibit
strong melt elasticity.
Figure 2.17 shows the growth curves of uniaxial elongational viscosity at
various strain rates for the blends of linear PP and gEHDM, the gel fraction
of a crosslinked terpolymer composed of ethylene, 1-hexene, and ethylidene
norbornene, produced by a metallocene catalyst system.19,92 The precursor
of gEHDM, an uncrosslinked terpolymer (EHDM), is miscible with PP in a
molten state92 because of its high hexene content, 67.5 mol%.87
Strain hardening, which barely takes place for linear PP, is clearly detected
in the curves for the blends at any strain rate in the experimental range, even
though the amount of gEHDM is only 3000 ppm. Further, the upturn behav-
ior becomes pronounced with gEHDM content, whereas the linear region
prior to viscous enhancement is the same as that for pure linear PP. Moreover,
deviation from the linear region is marked even at a low strain rate. The
maximum value of the elongational viscosity for the blend with 3 wt% of
gEHDM is almost 30 times as high as that for pure PP at a strain rate of
0.013–0.016. The upturn behavior is characterized in Figure 2.18, in which
the abscissa axis represents the ratio of measured elongational viscosity to

190°C

7
0.013−0016
log [ηE+ (t, ε) (Pa s)]

0.07−0.08

6
.

0.31−0.34

PP/gEHDM

4
−1 0 1 2 3
log [t (s)]

FIGURE 2.17
Growth curves of uniaxial elongational viscosity hE+ (t , e˙ ) at 190˚C for PP and blends with
gEHDM. (o) PP, (•) PP/gEHDM = 99.7/0.3, (‡) PP/gEHDM = 99/1, and (®) PP/gEHDM =
97/3. Numerals in the figure represent strain rates ė . (From Yamaguchi, M. J. Polym. Sci. Polym.
Phys. Ed., 39, 228, 2001. With permission of John Wiley & Sons, Inc.)

© 2004 by CRC Press LLC

www.ketabdownload.com
 15

3 wt%
190°C

10
+ (t,. ε)/3η+(t)
ηE

1 wt%
5

0.3 wt%

PP
0
0 1 2 3
strain

FIGURE 2.18
Nonlinear parameter, hE+ (t , e˙ ) /3h+(t) where h+(t) is transient shear viscosity at the low strain
rate asymptote calculated from linear viscoelastic properties, plotted against the elongational
strain at 190˚C for PP and blends with gEHDM. (From Yamaguchi, M., SPE ANTEC, 1149, 2001.
With permission of the Society of Plastic Engineers.)

the ratio calculated from linear viscoelastic properties, i.e., the level of non-
linearity, at a strain rate of 0.07–0.08.
All curves remain around unity until an elongational strain of 1, regardless
of the blend ratio. Then the blends show rapid growth with the strain, which
becomes evident with the gEHDM content.
The blend shows not only marked strain hardening but also a high level
of extrudate swell ratio and drawdown force, as shown in Figure 2.19. The
magnitude of extrudate swell ratio and drawdown force increases monoton-
ically with the content of gEHDM, whereas extensibility is depressed because
of the high level of extensional stress.19 The figure shows that the drawdown
force of a blend with only 1% of gEHDM is about twice as high as that of
pure PP. Further, the level of swell ratio and drawdown force of the blend
with 3 wt% of gEHDM is higher than that of a branched PP produced by
peroxide or radiation modification.51

© 2004 by CRC Press LLC

www.ketabdownload.com
 2.0 500

190°C

1.9 400

drawdown force (mN)

1.8 300
swell ratio

1.7 200

1.6 100

1.5 0
0 1 2 3 4 5
gEHDM content (wt%)

FIGURE 2.19
Relation between gEHDM content and (•) extrudate swell ratio at ġ = 10.8 s–1 and (o) draw-
down force at a draw ratio of 5 at 190˚C for PP/gEHDM blends. The dimension of the die
employed was L/D = 8/2.095.

The steady-state shear viscosity of the blend is, on the other hand, similar
to that of pure PP, as shown in Figure 2.20. Moreover, the blend shows a
higher primary normal stress coefficient. The enhancement of primary nor-
mal stress difference is connected with the marked extrudate swell ratio.20
These results prove that the gEHDM acts as a modifier of PP to enhance
various kinds of elastic properties while retaining the viscous properties.
Further, it has been found that gEHDM does not improve the melt elasticity
of polystyrene, which is immiscible with EHDM,92 indicating that miscibility
with network chains is required for the anomalous rheological properties.
Besides miscibility, the molecular structure of a network has a great influence
on the improvement of melt elasticity. The frequency dependence of oscilla-
tory shear modulus of the gEHDM is shown in Figure 2.21.92
Both G ¢ and G ¢¢ increase with frequency, and there seems to be a plateau
in the G ¢ curve in the very low frequency region. Furthermore, tan d

© 2004 by CRC Press LLC

www.ketabdownload.com
 8 6

190°C
7 5

6 4
log [η(γ) (Pa s)]

log [θ(γ) (Pa s 2 )]

.
. 5 3

4 2

η
3 1

2 0
−2 −1 0 1 2 3
.
log [γ (s−1)]

FIGURE 2.20
Steady-state shear viscosity h( ġ ) and primary normal stress coefficient q( ġ ) at 190˚C as a function
of shear rate ġ for (•) PP and (o) PP/gEHDM (97/3) blend. The measurement was performed
by means of a capillary rheometer ( ġ >10 s–1) and a cone-and-plate rheometer ( ġ <10 s–1).

remains almost a constant value irrespective of frequency. The curves dem-

onstrate that gEHDM, having a strong ability to enhance melt elasticity, is
characterized as a gel just beyond the sol-gel transition point at which the
following scaling law governs linear viscoelastic properties:94

G ¢ µ G ¢¢ µ w n (0 < n < 1)
(2.15)

tan d = tan(np / 2) (2.16)

G(t) = St - n (2.17)

where S is gel strength.

© 2004 by CRC Press LLC

www.ketabdownload.com
 7 5

6 190°C 4
log [G′ (Pa)], log [G′′ (Pa)]

5 3

G′
4 2

log [tan δ]
3 1
G′′
2 0

1 tan δ −1

0 −2
−3 −2 −1 0 1 2 3
log [ω (s−1)]
FIGURE 2.21
Frequency dependence of oscillatory shear moduli, such as shear storage modulus G¢, loss
modulus G≤, and loss tangent tan d, for gEHDM at 190˚C. (From Yamaguchi, M. and Miyata,
H. Polym. J. 32, 164, 2000. With permission of the Society of Polymer Science, Japan.)

The average molecular weight between crosslink points of gEHDM is

estimated to be 190,000, which is significantly larger than that between
entanglement couplings of EHDM,92 suggesting that the crosslinked polymer
has sparse crosslink points. According to our recent study,95 homogeneous
distribution of crosslink points is required for the crosslinked polymer to
improve the melt elasticity of a linear polymer efficiently. Furthermore, the
sol fraction does not play an important role in the improvement of melt
elasticity, compared to that of the gel fraction.19,92 A blend with a sol fraction
having many branches will have a similar influence to that with a branched
polymer; e.g., a blend of LLDPE with LDPE. Moreover, a crosslinked poly-
mer having many crosslink points also does not enhance melt elasticity, but
acts as filler because of the high modulus. Consequently, the extrudate swell
ratio of the PP/gEHDM blend is lower than that of pure PP.19,95 Further,
such a crosslinked polymer with high modulus prohibits homogeneous

© 2004 by CRC Press LLC

www.ketabdownload.com
deformation of the blend due to the localized stress concentration, and thus
results in poor extensibility.19,93,95
The morphology of the blend also affects the rheological properties. Figure
2.22 shows transmission electron microscopy for the PP/gEHDM (90/10)
blend stained by ruthenium tetraoxide. The sample was cut out of the strand
extrudated from a capillary rheometer. Since gEHDM is an amorphous poly-
mer (due to less ethylene sequence), the dark area in the picture represents
a gEHDM-rich region, and the white lines with 10–20 nm thickness denote
PP lamellae.
The picture provides evidence that PP lamellae exist even in gEHDM-rich
dark regions, suggesting that mutual dissolution of PP and gEHDM takes
place in the molten state. Further, most lamellae are extended in a horizontal
direction, i.e., perpendicular to the flow direction. Considering that the PP
chains orient perpendicular to the lamellae, the picture demonstrates that
most PP chains orient to the flow direction. Dynamic mechanical properties
in the solid state also support the interdiffusion of both polymer chains. As
seen in Figure 2.23, the blend shows a single peak, ascribed to glass transition
in the E ¢¢ curve of the blend, which is located between those of the individual
pure components (represented by diamonds). The result proves that network
chains of gEHDM are miscible with amorphous PP chains. Further, it also
indicates that network chains are distributed in a wide spatial area of the
blend in spite of the small amount of the gEHDM. Swelling of the network
by PP molecules will be responsible for the wide distribution of network
chains.

100 nm Flow direction

FIGURE 2.22
Transmission electron microscopy of PP/gEHDM (90/10) extrudated from a capillary rheo-
meter at 10.8 s–1 at 190˚C. The sample was stained by ruthenium tetraoxide. (From Yamaguchi,
M. SPE ANTEC, 1149, 2001. With permission of the Society of Plastic Engineers.)

© 2004 by CRC Press LLC

www.ketabdownload.com
 ′

′′

′′

′′

− 100
°C

FIGURE 2.23
Temperature dependence of (o) tensile storage modulus E¢ and (•) loss modulus E≤ for PP/
gEHDM (90/10) at 10 Hz. In the figure, E≤ peaks for (®) pure PP and (‡) gEHDM, shifted
vertically, are also shown. (From Yamaguchi, M. SPE ANTEC, 1149, 2001. With permission of
the Society of Plastic Engineers.)

Although the mechanism of the anomalous rheological properties for this

blend system has not been established precisely, the morphology indicates
that entanglement couplings, especially trapped entanglements associated
with the network, have a substantial effect on strain hardening.4 The trapped
entanglements whose characteristic time is considerably long will lead to
the chain stretching of the network.4 During extension, chain contraction of
the network, maybe with some linear polymers, is prohibited by trapped
entanglements. This is at least one of the origins of the strain hardening. Too
many crosslink points in a crosslinked polymer inhibit swelling of the net-
work, which results in less strain hardening. Further, sol will not play an
important role in enhancing the melt elasticity because there is no network
structure.19,95
Trapped entanglement also leads to marked overshooting in the growth
curves of shear stress and primary normal stress difference. Figure 2.24
shows the startup behavior of shear stress and primary normal stress differ-
ence for a blend with 10 wt% of gEHDM.

© 2004 by CRC Press LLC

www.ketabdownload.com
 A.
20

190°C

15
4.0 s−1
σ+(γ, γ) (kPa)

10
.

2.0 s−1

1.0 s−1
5

0.5 s−1

0
0 50 100 150 200
γ

B.
150

190°C

4.0 s−1
100
N1+ (γ, γ) (kPa)

2.0 s−1

50
1.0 s−1

0.5 s−1
0
0 50 100 150 200
γ

FIGURE 2.24
Startup behavior of (A) shear stress s + ( g , g˙ ) and (B) primary normal stress difference
N 1+ ( g , g˙ ) plotted against shear strain g at various shear rates 05–4.0 s–1 at 190˚C for PP/
gEHDM (90/10).

© 2004 by CRC Press LLC

www.ketabdownload.com
 The blend exhibits prominent overshooting in the curves of both shear
stress and primary normal stress difference. Furthermore, the strain at the
maximum point increases slightly with the strain rate, which is more obvious
in the curve of primary normal stress difference. The level of the maximum
stress rapidly increases with shear strain. It should be noted that the over-
shooting in Figure 2.24 appears at a much larger strain than that for a simple
polymer melt, as shown in Figure 2.14. According to the Doi–Edwards theory,
the strain at the maximum shear stress is around 2, irrespective of strain rate,
which corresponds with numerous experimental results, including PP, with
a minor departure.5 Moreover, it has been clarified by the theory that the
maximum shear stress is ascribed to the overshooting of molecular orienta-
tion, whereas the stretching of contour length also contributes to the over-
shooting of primary normal stress difference. The large strain at the maximum
of primary normal stress difference indicates that chain stretching keeps
occurring to a larger strain via trapped entanglements, which will be con-
nected with the upturn behavior in elongational viscosity. Beyond the max-
imum point, primary normal stress difference decreases with strain. Chain
stretching will be reduced, presumably, by losing interaction due to trapped
entanglements. In other words, prolonged shearing weakens the topological
interaction generated between PP molecules and network chains of gEHDM.
Generally speaking, mismatching of viscosities, if present, will accelerate
the rotation of the dispersion in shear flow, which gives rise to further
disentanglement. On the other hand, under elongational flow, the disentan-
glement between PP molecules and network chains of gEHDM will be
avoided for the following reasons. First, elongational flow is irrotational.
Therefore, rotation of the crosslinked polymer, causing further disentangle-
ment with the linear molecules, barely takes place. Second, deformation
increases exponentially in elongational flow at a constant strain rate. Chain
stretching occurs under the exponential growth of deformation, which has
been confirmed also in shear flow for a branched polymer.96 As a result, the
blend exhibits marked stress enhancement with strain only under extension.
The mechanism of the overshooting discussed here suggests that applied
shear flow changes the morphology in the melt and thus the rheological
properties of the blend, as discussed below.
The mixing method is very important for the blend system because a
crosslinked polymer is difficult to deform in the matrix of a linear polymer
with lower viscosity. To begin, shear-dominant flow should be precluded as
a possibility because it depresses melt elasticity due to the disentanglement
associated with network chains. Further, the diffusion of linear polymers
into the network is required in order to exhibit marked melt elasticity. Hence,
chaotic mixing, in which strain increases exponentially with time, is prefer-
ably employed to decrease the striation thickness of a crosslinked polymer.
Thinner striation enables linear polymers to penetrate into the network
within a certain processing time. Therefore, a twin-screw extruder with
kneading blocks, a rotor-type twin-screw extruder, and a two-roll mill should
be employed for mixing the blend. A single-screw extruder with conven-
© 2004 by CRC Press LLC
www.ketabdownload.com
 A.
7

170°C

6
log [ηE+ (t, ε) (Pa s)]

3η+(t)
. 0.074
5 0.14
0.38

LLDPE

3
−1 0 1 2 3
log [t (s)]

B.
7
170°C
3η+(t)
0.0138
6 0.037
0.074
log [ηE+ (t, ε) (Pa s)]

0.14
. 0.38
5

LLDPE/xLLDPE (97/3)

3
−1 0 1 2 3
log [t (s)]

FIGURE 2.25
Growth curves of uniaxial elongational viscosity hE+ (t , e˙ ) at various strain rates at 170˚C for (A)
LLDPE and (B) LLDPE/xLLDPE (97/3) blend. The solid line represents 3h+(t) at the low strain
rate asymptote calculated from linear viscoelastic properties. (From Yamaguchi, M. and Suzuki,
K., J. Polym. Sci. Polym. Phys. Ed., 39, 2159, 2001. With permission of John Wiley & Sons, Inc.)

tional full-flight screws does not work for mixing, because shear flow dom-
inates in the extruder.
This blend technique—improvement of melt elasticity by blending an
appropriate crosslinked polymer—is available for various polymers. Figure
2.25 shows the growth curves of uniaxial elongational viscosity for LLDPE
© 2004 by CRC Press LLC
www.ketabdownload.com
produced by metallocene catalyst and the blend with 3 wt% of crosslinked
LLDPE (xLLDPE), characterized as the gel just beyond the sol-gel transition
point. The precursor of the crosslinked polymer is, of course, miscible with
LLDPE, although it is difficult to detect the morphology of the blend. It has
also been noted that the gel fraction of xLLDPE is 75 wt% and that the
average molecular weight between crosslink points is 380,000 by swelling
measurement.97
Having no long-chain branch, pure LLDPE does not show upturn behavior
at any strain rate. The viscosity coincides with that of the low strain rate
asymptote calculated from linear viscoelastic properties. On the other hand,
it is obvious that the blend exhibits stronger upturn behavior than conven-
tional LDPE, as shown in Figure 2.6. At a high strain rate, the blend sample
exhibits brittle behavior due to high stress. The maximum stress of the blend
is almost 30 times as high as that of pure LLDPE. In contrast, the steady-
state shear viscosity of the blend is similar to that of pure LLDPE. Further,
blending xLLDPE has no effect on the onset of shark-skin failure,97 which is
often observed for linear polyethylenes. This is reasonable because shark-
skin failure is generally dependent on the apparent shear stress. Figure 2.25B
also indicates that strain hardening is observed even below a strain of 1. This
may be due to the morphology of the blend; i.e., the conformation of the
network chain and/or mutual dissolution between xLLDPE and LLDPE are
not at an equilibrium state.
The experimental results shown here demonstrate that a crosslinked poly-
mer having appropriate structure will be applicable as a modifier for pro-
cessing. Since most linear polymers have less melt elasticity and less strain
hardening in elongational viscosity, they will become available for various
kinds of processing, including foaming, by means of the blend technique.
Of course, the blend technique is also applicable to a branched polymer.

2.3.2.3 Processability of Foaming

The relation between strain hardening in elongational viscosity and the

processability of various processes has been studied intensively, especially
for fiber spinning, film processing, blow molding, and thermoforming, in
which deformation with a free surface takes place most frequently.97–100
According to previous studies, enhanced strain hardening suppresses heat-
sag and thus results in homogeneous thickness of the product. Also, for
foaming, it is expected that elongational viscosity plays an important role
during the bubble growth process in a similar manner.97,100 When the material
does not show the strain-hardening behavior in elongational viscosity, the
thinner part of a cell wall is liable to deform by an inner pressure. Conse-
quently, the thinner part will become much thinner and the bubble may
burst. In cases where the material has marked strain hardening, however,
the deformation in the thinner part will be reduced because of the viscous
enhancement with strain.97–100 As a result, the foam obtained will have uni-
form wall thickness, and thus a large expansion ratio is expected. This is the
© 2004 by CRC Press LLC
www.ketabdownload.com
 0.25

LLDPE/xLLDPE
LLDPE
0.20
volume fraction

0.15

0.10

0.05

0
0.1 0.2 0.3 0.4
diameter of cell (mm)

FIGURE 2.26
Cell size distribution of the foams from (
) original LLDPE and () LLDPE/xLLDPE (97/3)
blend. (From Yamaguchi, M. and Suzuki, K., J. Polym. Sci. Polym. Phys. Ed. 39, 2159, 2001. With
permission of John Wiley & Sons, Inc.)

primary reason why LDPE is preferably employed for foaming instead of

LLDPE.
Yamaguchi and Suzuki have proved that the blends with a crosslinked
polymer exhibit excellent processability for various processes, including
foaming.97,99,100 According to their foaming experiments, conducted in a hot
oven, for a compound containing a chemical blowing agent, the expansion
ratio of the blend with 0.5–3% of the xLLDPE is larger than that of the original
LLDPE.97,100 Further, cell-size distribution of the foams is significantly
improved by blending the xLLDPE, as shown in Figure 2.26. In particular,
the foam from the blend has no huge cell would lead to a rough surface.
The existence of huge cells in the foam from the original LLDPE indicates
that some cells burst and coalesce. The results support that viscous enhance-
ment under elongational flow dominates cell morphology as well as the
expansion ratio. Although it should be based on biaxial elongational viscosity
instead of uniaxial viscosity, as mentioned above, the conclusion will not
change, because the sample showing marked strain hardening in uniaxial
elongational viscosity is apt to exhibit larger strain hardening also in biaxial
viscosity.16 Finally, it is found that the blend exhibits a much wider processing
window (the temperature and the duration in an oven) than the original
LLDPE and commercial LDPE,100 because the melt elasticity of the blend is
less sensitive to temperature than that of LLDPE and LDPE. Consequently,
various blowing agents, such as ADCA and OBSH, are available for foaming,
regardless of the decomposition temperature.100 Similarly, blending gEHDM
© 2004 by CRC Press LLC
www.ketabdownload.com
enhances the processability of foaming for a linear PP. It should be empha-
sized that the blends that are not fully crosslinked exhibit excellent process-
ability for foaming. Moreover, the foaming capability without a fully
crosslinking process implies that the blend is available for gas extrusion
foaming.
Processability of the blend is dependent not only on the mixing method
but also on the processing machine, since the level of the mutual dissolution,
i.e., the degree of swelling of the network, is affected by the applied
processing history. Figure 2.27 shows the effect of processing history on the
drawdown force.
Even though the initial mixing is carried out by a two-roll mill, the following
extrusion history by a single-screw extruder with full-flight conveying screws,
in which shear flow dominates, depresses the drawdown force to a great
degree because of less strain hardening in elongational viscosity.99 On the other
hand, extrusion history by a counter-rotating twin-screw extruder with con-
veying screws has less of an effect on the drawdown force than does the single-
screw extruder. This is a mechanical modification because reprocessing by a

300

two-roll mill
160°C

200
drawdown force (mN)

twin-screw ext.

100

single-screw ext.

0
0 2 4 6 8
xLLDPE content (wt%)

FIGURE 2.27
Drawdown force at 160˚C at a draw ratio of 5 plotted against xLLDPE content for LLDPE/
xLLDPE processed by (•) two-roll mill, (o) single-screw extruder with a conveying screw after
mixing by two-roll mill, and (‡) counter-rotating twin-screw extruder with conveying screws
after mixing by two-roll mill.

© 2004 by CRC Press LLC

www.ketabdownload.com
 6

5 160°C

4
log [G′ (Pa)]

LLDPE
1 LLDPE/xLLDPE
(97/3)
two-roll mill
0 TSE
SSE

−1
−3 −2 −1 0 1 2 3
log [ω (s−1)]

FIGURE 2.28
Frequency dependence of shear storage modulus G¢ at 160 ˚C for (o) LLDPE and (•) LLDPE/
xLLDPE (97/3) blend processed by two-roll mill. Solid and dotted lines represent the blends
reprocessed by single-screw extruder (SSE) and by counter-rotating twin-screw extruder (TSE)
respectively, after mixing by a two-roll mill.

two-roll mill enhances the melt elasticity again, and molecular weight is
unchanged by the processing.99 The phenomenon—mechanical modification
of the rheological properties of the blend system—is very similar to shear
modification of a branched polymer, described in the previous section. It is
interesting that shear dominant flow depresses melt elasticity also for the blend
with a weak gel.
The linear viscoelastic properties for the samples with various processing
histories for the LLDPE/xLLDPE blend are shown in Figure 2.28. It is clear
that terminal relaxation, which may be ascribed to the entanglement cou-
plings associated with network chains, appears in the blend processed by
a two-roll mill because an apparent shoulder is observed in the G ¢ curve in
the low frequency region. Further, the magnitude of the shoulder is reduced
by the processing history, especially that of a single-screw extruder pro-
cessing with a conventional conveying screw, suggesting that the terminal
relaxation mechanism—i.e., the interaction between LLDPE and xLL-
DPE—is weakened by the applied shear flow in the extruder. The magni-
tude of G ¢ for the sample processed by a single-screw extruder is, however,
still higher than that for pure LLDPE. The effect of the processing history
© 2004 by CRC Press LLC
www.ketabdownload.com
 0.10

tan δ 10 Hz

two-roll mill
single-screw ext.

0.01
−80 −60 −40 −20 0 20 40
Temp. (°C)
FIGURE 2.29
Temperature dependence of tan d at 10 Hz for PP/gEHDM (90/10) blends (o) processed by
two-roll mill and (•) reprocessed by single-screw extruder with a conveying screw after mixing
by a two-roll mill.

by a counter-rotating twin-screw extruder falls between that of the two-roll

mill and that of the single-screw extruder, which is in accordance with the
result of the drawdown force shown in Figure 2.27. The experimental result
suggests that elongational flow, as in a two-roll mill, leads to mutual dis-
solution, whereas shear flow, which is dominant in a single-screw extruder,
accelerates the disentanglements between them. The fact that the degree of
mutual dissolution is affected by the applied processing history is also
confirmed by the morphological study of the blend composed of linear PP
and gEHDM. Figure 2.29 shows the temperature dependence of tan d for
the blend in the solid state. The sample processed by a two-roll mill exhibits
a single peak, suggesting that the amorphous chains dissolve each other.
On the other hand, processing history by a single-screw extruder with a
conveying screw leads to double peaks, ascribed to Tg of PP-rich and
gEHDM-rich phases. This result indicates that the structure of the blend
changes from a homogeneous to a heterogeneous morphology composed
of PP-rich and gEHDM-rich phases by shear-dominant flow, although the
level of melt elasticity of the blend having the shearing history is still
considerably higher than that of pure PP.

© 2004 by CRC Press LLC

www.ketabdownload.com
 Another example of the effect of applied processing history is exhibited
in Figure 2.30. Processing history with a single-screw extruder having a
conveying screw makes the magnitude of strain-hardening weaken for a
blend of HDPE and xHDPE, i.e., crosslinked HDPE characterized as a weak
gel, mixed by a two-roll mill.99 It is, however, quite interesting that the blend
having a single-screw extrusion history shows early upturn behavior.
The mechanical modification, i.e., the depression of melt elasticity by shear
flow, is dependent on the molecular structure of both the crosslinked and
the linear polymers. The crosslinked polymer having heterogeneous distri-
bution of crosslink points and/or high density of crosslink points is liable
to depress the melt elasticity by shear history. Further, shear flow has less
effect on the depression of melt elasticity for the blend composed of the
linear polymer with high molecular weight and narrow molecular weight
distribution. Currently, the blend with a crosslinked polymer, which is not
significantly affected by applied processing history, is being developed.100
In conclusion, the blend with a crosslinked polymer, characterized as a
weak gel just beyond the critical point, has great potential to make various
polymers available for foaming. The processing method and mixing proce-
dure, however, have to be taken into consideration.

7
0.013−0.015

0.069−0.071

6
0.35−0.36
log [η+E (t, ε) (Pa s)]

. 5

HDPE/xHDPE
160°C (97/3)

3
−1 0 1 2 3
log [t (s)]
FIGURE 2.30
Growth curves of uniaxial elongational viscosity hE (t , e˙ ) at various strain rates at 160˚C for
+

HDPE/xHDPE (97/3) blend. Open symbols denote the sample mixed by two-roll mill, and
closed symbols denote the sample reprocessed by single-screw extruder after mixing by a two-
roll mill. Numbers in the figure represent the strain rate.

© 2004 by CRC Press LLC

www.ketabdownload.com
2.4 Conclusion
Melt elasticity and its effect on foaming for polyolefins have been described
in this chapter. Since a linear polymer such as LLDPE and PP lacks the
elasticity required for foaming, various approaches to enhance the melt
elasticity have been studied. One of them is a blend technique with a
crosslinked polymer having the appropriate network structure. Catalyst
technology is also being developed extensively in order to incorporate the
appropriate number and length of long-chain branches as well as to generate
new types of polyolefins, such as block or graft copolymers and copolymers
with a polar monomer. Further, applied processing history depresses the
melt elasticity for a branched polymer to a great degree, which is owing to
reduced interactions due to entanglement couplings associated with long
chain branches. This phenomenon has to be considered seriously when a
branched polymer is employed for foaming because the degree of mecha-
nical modification is dependent on the processing temperature, screw con-
figuration, screw rpm, and so on.
It has been shown that strain hardening in elongational viscosity has a
strong effect on the bubble growth process of foaming. However, the quan-
titative relation between the degree of strain hardening and processability
of foaming has not been established yet, since the reported experimental
results are few and restricted. More time should be spent studying elonga-
tional flow, especially biaxial elongational flow, as well as the rheological
features under non-isothermal conditions, in order to derive a more complete
comprehension of foaming processability.

References
1. Gendron, R. and Daigneault, L.E., Rheology of thermoplastic foam extrusion
process, in Foam Extrusion, ed. S.T. Lee, CRC Press, Boca Raton, FL, 2000, chap 3.
2. Yamaguchi, M. and Abe, S., LLDPE/LDPE blends. I. Rheological, thermal, and
mechanical properties, J. Appl. Polym. Sci., 74, 3153, 1999.
3. Abe, S. and Yamaguchi, M., Study on the foaming of crosslinked polyethylene,
J. Appl. Polym. Sci., 79, 2146, 2001.
4. Ferry, J.D., Viscoelastic Properties of Polymers, 3rd ed., John Wiley & Sons, New
York, 1980.
5. Doi, M. and Edwards, S.F., The Theory of Polymer Dynamics, Clarendon Press,
Oxford, 1986.
6. Rouse, P.E., A theory of the linear viscoelastic properties of dilute solutions of
coiling polymers, J. Chem. Phys., 21, 1272, 1953.
7. Mills, N.J., The rheological properties and molecular weight distribution of
polydimethylsiloxane, Eur. Polym. J., 5, 675, 1969; Mills, N.J. and Nevin, A.,
Oscillatory shear measurements on polystyrene melts in the terminal region,
J. Polym. Sci., Part A-2, 9, 267, 1971.

© 2004 by CRC Press LLC

www.ketabdownload.com
 8. Laun, H.M., Prediction of elastic strains of polymer melts in shear and elon-
gation, J. Rheol., 30, 459, 1986.
9. Cox, W.P. and Merz, E.H., Correlation of dynamic and steady flow viscosities,
J. Polym. Sci., 28, 619, 1958.
10. Pearson, D.S. et al., Flow-induced birefringence of concentrated polyisoprene
solutions, J. Rheol., 33, 517, 1989.
11. Osaki, K., On the damping function of shear relaxation modulus for entangled
polymers, Rheol. Acta, 32, 429, 1993.
12. Larson, R.G., A constitutive equation for polymer melts based on partially
extending strand convection, J. Rheol., 28, 545, 1984.
13. Wagner, M.H. and Schaeffer, J., Nonlinear strain measures for general biaxial
extension of polymer melts, J. Rheol., 36, 1, 1992; Wagner, M.H., Bastian, H.,
Hachmann, P., Meissner, J., Kurzbeck, S., Münstedt, H., and Langouche, F. The
strain-hardening behaviour of linear and long-chain-branched polyolefin melts
in extensional flows, Rheol. Acta, 39, 97, 2000.
14. McLeish, T.C.B. and Larson, R.G., Molecular constitutive equations for a class
of branched polymers; the pom-pom polymer, J. Rheol., 42, 81, 1998; Inkson,
N.J. and McLeish, T.C.B., Predicting low density polyethylene melt rheology
in elongational and shear flows with “pom-pom” constitutive equations, J.
Rheol., 43, 873, 1999.
15. Wagner, M.H., Yamaguchi, M., and Takahashi, M., Quantitative assessment of
MSF theory for LDPE, J. Rheol., 47, 779, 2003.
16. Yamaguchi, M. and Takahashi, M., Rheological properties of low density poly-
ethylenes produced by tubular and vessel processes, Polymer, 42, 8663, 2001.
17. Macosko, C.W., Rheology: Principles, Measurements, and Applications, Wiley, New
York, 1994.
18. Wagner, W.H., Bernnat, A., and Schulze, V., The rheology of the rheotens test,
J. Rheol., 42, 917, 1998.
19. Yamaguchi, M., Rheological properties of linear and crosslinked polymer
blends: Relation between crosslink density and enhancement of elongational
viscosity, J. Polym. Sci. Polym. Phys. Ed., 39, 228, 2001.
20. Tanner, R.I., Engineering Rheology, Oxford Science Publications, New York, 1985.
21. Meissner, J., Rheometer zur undersuchung der deformationsmechanischen
eigenschaften von kunststoffe-schmelzen unter definierter zugbeanspru-
chung, Rheol. Acta, 8, 78, 1969; Meissner, J., Development of a uniaxial exten-
sion of polymer melts, Trans. Soc. Rheol., 16, 405, 1972.
22. Larson, R.G., The Structure and Rheology of Complex Fluids, Oxford University
Press, New York, 1999.
23. Ballman, R.I., Extensional flow of polystyrene melt, Rheol. Acta, 4, 137, 1965;
Chang, J.C. and Denn, M.M., An experimental study of isothermal melt spin-
ning of a Newtonian and a viscoelastic liquid, J. Non-Newtonian Fluid Mech., 5,
39, 1979; Jones, D.M., Walters, K., and Williams, P.R., On the extensional vis-
cosity of mobile polymer solutions, Rheol. Acta, 26, 20, 1987.
24. Metzner, A.B. and Metzner, A.P., Stress levels in rapid extensional flows of
polymeric fluids, Rheol. Acta, 9, 174, 1970; Binding, D.M., An appropriate anal-
ysis for contraction and converging flows, J. Non-Newtonian Fluid Mech., 27,
173, 1988; Trembly, B., Estimation of the elongational viscosity of polyethylene
blends at high deformation rates, J. Non-Newtonian Fluid Mech., 33, 137, 1989.
25. Moan, M. and Maguer, A., Transient extensional viscosity of dilute flexible
polymer solutions, J. Non-Newtonian Fluid Mech., 30, 343, 1988.

© 2004 by CRC Press LLC

www.ketabdownload.com
 26. Meissner, J., Stephenson, S.E., Demarmels, A., and Portman, P., Multiaxial
elongational flows of polymer melts—classification and experimental realiza-
tion, J. Non-Newtonian Fluid Mech., 11, 221, 1982
27. Chatraei, S., Macosko, C.W., and Winter, H.H., Lubricated squeezing flow: A
new biaxial extensional rheometer, J. Rheol., 25, 433, 1981; Khan, S.H. and
Larson, R.G., Step planar extension of polymer melts using a lubricated chan-
nel, Rheol. Acta, 30, 1, 1991.
28. Winter, H.H., Macosko, C.W., and Bennett, K.E., Orthogonal stagnation flow:
A framework for steady extensional flow experiments, Rheol. Acta, 18, 323, 1979;
Soskey, P.R. and Winter, H.H., Equibiaxial extension of two polymer melts:
Polystyrene and low density polyethylene, J. Rheol., 29, 493, 1985.
29. Denson, C.D. and Gallo, R.J., Measurements on the biaxial extensional viscosity
of bulk polymers: The inflation of a thin polymer sheet, Polym. Eng. Sci., 11,
174, 1971; Denson, C.D. and Crady, D.L., Measurements on planar extensional
viscosity of bulk polymers: Inflation of a thin, rectangular polymer sheet, J.
Appl. Polym. Sci., 18, 1611, 1974; Denson, C.D. and Hylton, D.L., A rheometer
for measuring the viscoelastic response of polymer melts in arbitrary planar
and biaxial extensional flow fields, Polym. Eng. Sci., 20, 535, 1980.
30. Watanabe, H., Viscoelasticity and dynamics of entangled polymers, Prog. Polym.
Sci., 24, 1253, 1999.
31. Pearson, D.S. and Helfand, E., Viscoelastic properties of star-shaped polymers,
Macromolecules, 17, 888, 1984.
32. Roovers, J. and Toporowski, P.M., Relaxation by constraint release in combs
and star-combs, Macromolecules, 20, 2300, 1987.
33. Raju, V.R., Rachapudy, H., and Graessley, W.W., Properties of amorphous and
crystallizable hydrocarbon polymers. Part 4: Melt rheology of linear and star-
branched hydrogenated polybutadiene, J. Polym. Sci. Polym. Phys. Ed., 17, 1223,
1979.
34. Carella, J.M., Gotro, J.T., and Graessley, W.W., Thermorheological effects of
long-chain branching in entangled polymer melts, Macromolecules, 19, 659, 1986.
35. Zimm, B.H. and Stockmayer, W.H., Dynamics of branched polymer molecules
in dilute solution, J. Chem. Phys., 17, 1301, 1949.
36. Tackx, P. and Tacx, J.C.J.F., Chain architecture of LDPE as a function of molar
mass using size exclusion chromatography and multi-angle laser light scatter-
ing, Polymer, 39, 3109, 1998.
37. Scheirs, S. and Kaminsky, W., eds., Metallocene-Based Polyolefins: Preparation,
Properties, and Technology, Wiley, New York, 2000.
38. Lachtermacher, M.G. and Rudin, A., Reactive processing of LLDPEs in co-
rotating intermeshing twin-screw extruder. 1. Effect of peroxide treatment on
polymer molecular structure, J. Appl. Polym. Sci., 58, 2077, 1995.
39. Horii, F., Zhu, Q., and Kitamaru, R., Carbon-13 NMR study of radiation-in-
duced crosslinking of linear polyethylene, Macromolecules, 23, 977, 1990.
40. Abe, S., personal communication, 2000.
41. Lai, S.Y., Wilson, J.R, Knight, G.W., Stevens, J.C., and Chum, P.W.S., Elastic
substantially linear olefin polymers, U.S. Patent 5,272,236, 1993; Vega, J.F.San-
tamaria, A., Munoz-Escalone, A., and Lafuente, P., Small-amplitude oscillatory
shear flow measurements as a tool to detect very low amounts of long chain
branching in polyethylenes, Macromolecules, 31, 3639, 1998.
42. Wadud, S.E.B. and Baird, D.G., Shear and extensional rheology of sparsely
branched metallocene-catalyzed polyethylenes, J. Rheol., 44, 1151, 2000.

© 2004 by CRC Press LLC

www.ketabdownload.com
 43. Malmberg, A., Kokko, E., Lehmus, P., Löfgren, B., and Seppälä, J. V., Long-
chain branched polyethene polymerized by metallocene catalysts
Et[Ind]2ZrCl2/MAO and Et[IndH4]2ZrCl2/MAO, Macromolecules, 31, 8448, 1998.
44. Hogan, J.P., Levett, C.T., and Werkman, R.T., Melt elasticity in linear PE con-
taining long branches, SPE. J., 23, 87, 1967.
45. Johnson, L.K., Killian, R.M., and Brookhart, M., New Pd(II)- and Ni(II)-based
catalysts for polymerization of ethylene and a-olefins, J. Am. Chem. Soc., 117,
6414, 1995; Johnson, L.K., Mecking, S., and Brookhart, M., Copolymerization
of ethylene and propylene with functionalized vinyl monomers by Paradium
(II) catalysts, J. Am. Chem. Soc., 118, 267, 1996.
46. Hauptman, E., Waymouth, R.M., and Ziller, J.W., Stereoblock polypropylene:
Ligand effects on the stereospecificity of 2-arylindene Zirconocene catalysts, J.
Am. Chem. Soc., 117, 11586, 1995.
47. Kojoh, S., Matsugi, T., Saito, J., Mitani, M., Fujita, T., and Kashiwa, N., New
monodisperse ethylene-propylene copolymers and a block copolymer created
by a titanium complex having fluorine-containing phenoxy-imine chelate
ligands, Chem. Lett., 822, 2001.
48. Schmidt, G.F.Timmers, F.J., Knight, G.W., Lai, S., Nickias, P.N., Rosen, R.K.,
Stevens, J.C., and Wilson, D.R., Constrained geometry addition polymerization
catalysts, processes for their preparation, precursors therefore, methods of use,
and novel polymers formed therewith, Eur. Patent 0416815, 1991.
49. Yang, C.T. and Lee, S.T., Dimensional stability analysis of foams based on LDPE
and ethylene-styrene interpolymer blends, SPE ANTEC, 1782, 2002; Ankrah,
S., Verdejo, R., and Mills, N.J., The mechanical properties of ESI/LDPE foam
blends and sport applications, Cell. Polym., 21, 237, 2002.
50. Bradley, M.B. and Phillips, E.M., Novel polypropylenes for foaming on con-
ventional equipment, Plast. Eng., 82, 1991.
51. Hingmann, R. and Marczinke, B.L., Shear and elongational flow properties of
polypropylene melts, J. Rheol., 38, 573, 1994; Sugimoto, S., Tanaka, T., Masubu-
chi, Y., Takimoto, J., and Koyama, K., Effect of chain structure on the melt
rheology of modified polypropylene. J. Appl. Polym. Sci., 73, 1493, 1999.
52. Cheung, L.K., Park, C.B., and Behravesh, A.H., Effect of branched structure on
the cell morphology of extruded polypropylene foams, SPE ANTEC, 1941, 1996.
53. Lagendijk, R.P., Hogt, A.H., Buijtenhuijs, A., and Gotsis, A.D., Peroxydicarbon-
ate modification of polypropylene and extensional flow properties, Polymer,
42, 10035, 2001.
54. Howells, E.R. and Benbow, J.J., Flow defects in polymer melts, Trans. J. Plast.
Inst., 30, 240, 1962.
55. Hanson, D.E., Shear modification of polythene, Polym. Eng. Sci., 9, 405, 1969.
56. Fujiki, T., Concept of secondary heterogeneous structure of long-chain
branched polyethylene, J. Appl. Polym. Sci., 15, 47, 1971.
57. Rokudai, M., Influence of shearing history on the rheological properties and
processability of branched polymers, part I, J. Appl. Polym. Sci., 23, 463, 1979.
58. Rokudai, M., Mihara, S., and Fujiki, T., Influence of shearing history on the
rheological properties and processability of branched polymers, part II, J. Appl.
Polym. Sci., 23, 3289, 1979.
59. Rokudai, M. and Fujiki, T., Influence of shearing history on the rheological
properties and processability of branched polymers, part III, J. Appl. Polym.
Sci., 23, 3295, 1979.
© 2004 by CRC Press LLC
www.ketabdownload.com
 60. Münstedt, H., The influence of various deformation histories on elongational
properties of low density polyethylene, Colloid Polym. Sci., 259, 966, 1981.
61. Stehling, F., Speed, C.S., and Westerman, L., Causes of haze of low-density
polyethylene blown films, Macromolecules, 14, 698, 1981.
62. Maxwell, B. and Breckwoldt, A., Controlled shear modification of low-density
polyethylene, J. Rheol., 25, 55, 1981.
63. Maxwell, B., Dormier, E.J., Smith, F.P., and Tong, P.P., Processing properties of high-
and low-density polyethylene and their blends, Polym. Eng. Sci., 22, 280, 1982.
64. Ritzau, G., Shear modification of polyolefins and its integration in polymer
processing, Intern. Polym. Process., 1, 188, 1987.
65. Leblans, P.J.R. and Bastiaansen, C., Shear modification of low-density polyeth-
ylene: Its origin and its effect on the basic rheological functions of the melt,
Macromolecules, 22, 3312, 1989.
66. Yamaguchi, M. and Gogos, C.G., Quantitative relation between shear history
and rheological properties of LDPE, Adv. Polym. Technol., 20, 261, 2001.
67. Yamaguchi, M., Todd, D.B., and Gogos, C.G., Effect of processing history on
rheological properties and processability of LDPE, Proceedings of the Annual
Meeting of the Society of Polymer Processing, PPS-18, Guimaraes, Portugal,
June 2002.
68. Yamaguchi, M., Todd, D.B., and Gogos, C.G., Rheological properties of LDPE
processed by conventional processing machines, Adv. Polym. Technol., 22, 179,
2003.
69. White, J.L., Critique on flow patterns in polymer fluids at the entrance of a die
and instabilities leading to extrudate distortion, Appl. Polym. Symp., 20, 155,
1973; Yamaguchi, M., Flow instability in capillary extrusion of plasticized poly
(vinyl chloride), J. Appl. Polym. Sci., 82, 1277, 2001.
70. Gaskell, R.E., The calendering of plastic materials, J. Appl. Mech., 17, 334, 1950.
71. Todd, D.B., Introduction to compounding, and APV systems, Plastics Com-
pounding, Equipment and Processing, ed. D.B. Todd, Hanser, New York, 1998,
chaps. 1 and 6.
72. Sakai, T., Intermeshing twin-screw extruders, Mixing and Compounding of Poly-
mers, ed. I. Manas-Zloczower and Z. Tadmor, Hanser, New York, 1994, chap. 21.
73. Andersen, P.G., Mixing practices in corotating twin-screw extruders, Mixing
and Compounding of Polymers, ed. I. Manas-Zloczower and Z. Tadmor, Hanser,
New York, 1994, chap. 20; Anderson, P.G., The Warner and Pfleiderer twin-
screw corotating extruder system, Plastics Compounding, Equipment and Process-
ing, ed. D.B. Todd, Hanser, New York, 1998, chap. 4.
74. Cheng, H. and Manas-Zloczower, I., Chaotic features of flow in polymer pro-
cessing equipment, relevance to distributive mixing, Intern. Polym. Process., 12,
83, 1997.
75. Shikata, T. and Pearson, D.S., Viscoelastic behavior of concentrated spherical
suspension, J. Rheol., 38, 601, 1994.
76. Mewis, J. and Metzner, A.B., The rheological properties of suspensions of fibres
in Newtonian fluids subjected to extensional deformations, J. Fluid Mech., 62,
249, 1974.
77. Laun, H.M., Orientation effects and rheology of short glass fiber-reinforced
thermoplastics, Colloid Polym. Sci., 262, 257, 1984.
78. Chan, Y., White, J.L., and Oyanagi, Y., A fundamental study of the rheological
properties of glass-fiber-reinforced polyethylene and polystyrene melts, J.
Rheol., 22, 507, 1978.
© 2004 by CRC Press LLC
www.ketabdownload.com
 79. Lin, L., Viscoelastic properties of particle-filled polymers in the molten state,
Ph.D. thesis, Kyoto University, Kyoto, 1989.
80. Taylor, G.I., The formation of emulsions in definable fields of flow, Proc. R. Soc.
A, 146, 501, 1934.
81. Palierne, J.F., Linear rheology of viscoelastic emulsions with interfacial tension,
Rheol. Acta, 29, 204, 1990.
82. Doi, M. and Ohta, T., Dynamics and rheology of complex interfaces I, J. Chem.
Phys., 95, 1242, 1991.
83. Santamaria, A. and White, J.L., Rheological properties, shrinkage and melt
spinning instability of blends of linear polyolefins with low density polyeth-
ylene, J. Appl. Polym. Sci., 31, 209, 1986.
84. Micic, P., Bhattacharya, S.N., and Field, G., Rheological behavior of LLDPE/
LDPE blends under elongational deformation, Intern. Polym. Processing, 12, 110,
1997; Transient elongational viscosity of LLDPE/LDPE blends and its relevance
to bubble stability in the film blowing process, Polym. Eng. Sci., 38, 1685, 1998.
85. Lee, S.T. and Baker, J.J., Foam comprising polyolefin blend and method for
producing same, U.S. Patent 6,096,793, 1998.
86. O’Brain, J.C. and Ehrenfreund, H.A., Synthetic cork-like material and method
of making same, U.S. Patent 4,091,136, 1976; Lee, S.T., Polyolefin/ionomer
blend for improved properties in extruded foam products, U.S. Patent 6066393,
1998.
87. Lohse, D.J. and Graessley, W.W., Thermodynamics of polyolefin blends, Polymer
Blends, ed. D.R. Paul and C.B. Bucknall, Wiley, New York, 1999, chap. 8.
88. Valenza, A., La Mantia, F.P., and Acierno, D., The rheological behavior of
HDPE/LDPE blends. V. Isothermal elongation at constant stretching rate, J.
Rheol., 30, 1085, 1986; La Mantia, F.P., Valenza, A., and Acierno, D., Elongational
behavior of low density/linear low density polyethylenes, Polym. Eng. Sci., 28,
90, 1988.
89. Münstedt, H., Dependence of the elongational behavior of melts on molecular
weight and molecular weight distribution, J. Rheol., 24, 847, 1980.
90. Linster, J.J. and Meissner, J., Melt elongation and structure of linear polyethyl-
ene, Polym. Bull., 16, 187, 1986.
91. Dumoulin, M.M., Utracki, L.A., and Lara, J., Rheological and mechanical be-
havior of the UHMWPE/MDPE mixtures, Polym. Eng. Sci., 24, 117, 1984.
92. Yamaguchi, M. and Miyata, H., Strain hardening behavior in elongational
viscosity for binary blends of linear polymer and crosslinked polymer, Polym.
J., 32, 164, 2000.
93. Yamaguchi, M., Anomalous strain hardening behavior in elongational viscosity
for blends of linear polymer and crosslinked polymer, SPE ANTEC, 1149, 2001.
94. Winter, H.H. and Chambon, F., Analysis of linear viscoelasticity of a crosslink-
ing polymer at the gel point, J. Rheol., 30, 367, 1986.
95. Yamaguchi, M., Suzuki, K., and Maeda, S., Enhanced strain-hardening in elon-
gational viscosity for HDPE/crosslinked HDPE blend: 1. Characteristics of
crosslinked HDPE, J. Appl. Polym. Sci., 86, 73, 2002.
96. Zülle, B., Linster, J.J., Meissner, J., and Hürliman, H.P.., Deformation hardening
and thinning in both elongational and shear of a low density polyethylene
melt, J. Rheol., 31, 583, 1987: Graham, R.S., McLeish, T.C.B., and Harlen, O.G.,
Using the pom-pom equations to analyze polymer melt in exponential shear,
J. Rheol., 45, 275, 2001.

© 2004 by CRC Press LLC

www.ketabdownload.com
 97. Yamaguchi, M. and Suzuki, K., Rheological properties and foam processability
for blends of linear and crosslinked polyethylenes, J. Polym. Sci. Polym. Phys.
Ed., 39, 2159, 2001.
98. Sebastian, D.H. and Dearborn, J.R., Elongation rheology of polyolefins and its
relation to processability, Polym. Eng. Sci., 23, 572, 1983; Fleissner, M., Elonga-
tional flow of HDPE samples and bubble instability in film blowing, Intern.
Polym. Proc., 2, 229, 1988; Shinohara, M., Uniaxial elongational viscosity of high
molecular weight high density polyethylene melts, J. Soc. Rheol. Jpn., 19, 118,
1991; Kanai, T., Dynamics, heat transfer, and structure development, Film Pro-
cessing, ed. T. Kanai and G.A. Campbell, Hanser, New York, 1999, chap. 3.
99. Yamaguchi, M. and Suzuki, K., Enhanced strain-hardening in elongational
viscosity for HDPE/crosslinked HDPE blend: 2. Processability of thermoform-
ing, J. Appl. Polym. Sci., 86, 79, 2002.
100. Yamaguchi, M., Foaming of novel linear polyethylene with enhanced melt
elasticity, Proceedings of the Annual Meeting of the Society of Polymer Processing,
PPS-19, Melbourne, Australia, July 2003.

© 2004 by CRC Press LLC

www.ketabdownload.com
 3
Fundamentals of Bubble Nucleation and
Growth in Polymers

N.S. Ramesh

CONTENTS
3.1 Introduction
3.2 Foaming Process
3.3 Bubble Nucleation Studies
3.4 Nucleation Models and Experiments
3.4.1 Classical Nucleation Principles
3.5 Homogeneous Foam Nucleation
3.6 Heterogeneous Foam Nucleation
3.7 Shear Effects on Heterogeneous Nucleation
3.8 Microvoid Nucleation
3.8.1 Nucleation Mechanism
3.8.2 Nucleation Model
3.8.3 Results and Discussion
3.9 Effect of Concentration of Particles on Nucleation
3.10 Effect of Rubber Particle Size on Nucleation
3.11 Nucleation in Polyurethane Foams
3.11.1 Bubble Nucleation
3.11.2 Bubble Growth Model for Polyurethane Foam
3.12 Foam Growth in Polymers
3.12.1 Bubble Growth Models
3.12.1.1 Single Bubble Growth Models (1917–1984)
3.12.1.2 Cell Model (1984–1998)
3.13 Foam Extrusion Modeling
3.13.1 Foam Growth Equations
3.13.2 Boundary Conditions
3.13.3 Comparison of Theory with Experiment
3.14 Conclusion
Acknowledgment
Nomenclature
References

0-8493-1728-2/04/$0.00+$1.50
© 2004 by CRC Press LLC

© 2004 by CRC Press LLC

www.ketabdownload.com
 3.1 Introduction

Interest in cellular plastics is growing for a number of reasons. The principal

reason has to do with the favorable economics of manufacturing innovative
products at a lower weight. The banning of chlorofluorocarbons (CFCs) in
the early 1980s, along with strong health, safety, and environmental consid-
erations in the foam industry, have increased interest in developing better
methods of controlling foaming operations. In order to control the foaming
process effectively, two key events must be controlled: bubble nucleation
and bubble growth during foam expansion. In this chapter, these two topics
will be discussed from the basics of thermodynamics and other physical
principles. The use of nucleating systems to produce light density foams also
has a potential impact on the disposable household, electronics packaging,
and recreational sports markets.
Various nucleation mechanisms, such as homogeneous, heterogeneous,
mixed-mode, shear-induced, and void nucleation theories, will be discussed.
Suitable data are presented to support theoretical results. Extrusion of foam
involves dissolving the foaming agent in the polymer under high pressure
and then lowering the pressure suddenly by exiting through a die at a
foaming temperature. The sudden instability instantaneously nucleates a
myriad of bubbles with a subsequent uniform distribution. After cell nucle-
ation, they grow due to the diffusion of excess gas. The viscosity of the
polymer, the gas concentration, the foaming temperature, and the amount
of nucleating agent and its nature are some of the variables that control the
foam growth process. Foam dynamics are very complicated. Useful models
and experimental data are presented to explain the basics of foam growth
mechanisms in polymers.

3.2 Foaming Process

A typical foaming process involves the following four steps:

1. Dissolution of foaming agent

2. Bubble nucleation
3. Bubble growth
4. Stabilization

© 2004 by CRC Press LLC

www.ketabdownload.com
 This chapter focuses primarily on the second and third steps. Thermoplas-
tic polymeric foams are made with nucleating agents to aid in cell formation.
The use of nucleating systems to produce various density foams also has a
potential impact on the packaging and nonpackaging markets. Packaging
foams are used everyday for protecting furniture and other objects during
transportation. Polyurethane foams are used in packaging, carpet backing,
sports, leisure, automobiles, and home applications such as seat cushions,
sound, and thermal insulation. Foams are used everywhere from household
items to industrial cushions. Thermosetting foams such as polyurethane
foams are made by generating bubbles within a liquid through a simple
chemical reaction. Foam nucleation is different for different systems. Foam-
ing mechanisms are quite complex, since it is difficult to monitor cell forma-
tion in polymers and liquids at its birth. The size of potential nuclei is very
small, on the order of nanometers or angstroms in scale. Nucleation occurs
in a fraction of a second, and therefore it is difficult for foam scientists to
capture in real life. The following section describes bubble nucleation studies.

3.3 Bubble Nucleation Studies

The birth of tiny bubbles in a polymer system is called nucleation of bubbles
and is part of the process used to make polymeric foams. If the bubbles are
generated from a single homogeneous phase containing no impurity or dirt,
then the process is called homogeneous nucleation. This process is rare, since
most resins are made with additives for various reasons. If tiny particles are
present in the liquid, and if they assist in the formation of cells, then the
process is called heterogeneous nucleation, since the nucleation took place
at a solid and liquid interface. Table 3.1 shows various nucleation studies.
Because the subject of bubble nucleation is so vast, it is impossible to give
a detailed explanation of each category and only key points are mentioned
to explain the basic principles. Refer to the provided references for further
reading about other mechanisms.
It is interesting to note that the physics of various nucleation mechanisms
depend on energy levels involved in the systems. In other words, the birth
of bubbles in a liquid or solid requires an increase in the free energy of the
system. That increase is used to create new surfaces through the formation
of tiny bubbles. The birth of a gas bubble in a polymer through a reversible
thermodynamic process has an excess free energy associated with it. Math-
ematically, that can be expressed as

DG = –Vb DGv + As (3.1)

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 3.1
Nucleation Types and Mechanisms
Nucleation Influencing Nucleating Reference/
Type Nucleation Mechanism Factors Systems Discussion
Homogeneous Bubbles formed due to superheat and Temperature and pressure Boiling water and organic 1–5 (3.4)
pressure reduction liquids
Heterogeneous Bubbles formed with assistance from tiny Geometry and nature of Bubble nucleation using fillers 6 (3.5)
particles particle such as talc in polymers
Mixed-mode Supersaturation of gas in polymers Gas concentration in Polystyrene with zinc stearate, 7 (combination
containing solid particles polymer and its solubility stearic acid, and carbon black of the above 2
limit modes)
Cavity Poor wetting of gas/liquid acts as a potential Geometry of cavity and Boiling of water and organic 8
nucleating site pressure drop in the system liquids in a rough-surfaced
vessel having microcavities.
Shear induced Shear force induces birth of cells Shear rate in foam extrusion Polymeric foam extrusion 9 (3.6)
Microvoid Presence of microvoids due to rapid polymer Cooling rate of polymer melt Bubble nucleation in High 10 (3.8, 3.9, and
cooling and interface cavitations in polymer Impact Polystyrene or its 3.9)
blends blends with polystyrene
Devolatalization Volatile diffusion of solvent within polymer Vacuum level and diffusion Devolatilization of styrene 11
melt rate monomer from polystyrene
Chemical Liquid reaction and generation of gas assisted Temperature, reactant ratio Reaction of polyol and 12 (3.10)
reaction by mixing causes bubble formation and degree of mixing isocyanate to form
polyurethane
Mechanical Pressure change generated by ultrasonic Ultrasonic frequency, LPPE/PE Wax blend and 13–14
energy (mixing wave propagation and mixing of chemicals agitator speed, shear rate, high speed mixing of polyol
and ultrasonic) in polyurethane making temperature and isocyanate

© 2004 by CRC Press LLC

www.ketabdownload.com
 where
Vb = Volume of the bubble nucleus
DGv = Difference between the gas and polymer phases of the standard
Gibbs free energy per unit volume
s = Surface tension of the liquid or polymer
A = Interfacial area
V = Volume of the bubble
subscript b = Bubble

Thus, lowering the surface tension by using surfactants will assist in the
formation of bubbles. Nucleating agents such as talc, diatomaceous earth,
and silica are more effective because they offer voids at the interface.
Before we go into details, it will be useful to review the history of nucle-
ation of bubbles in liquids and polymers. Several investigations involving
the theoretical and experimental analyses of bubble nucleation in fluids and
polymers have been available in the literature since 1900. The historical
development of important models and data is reported in Table 3.2. These
models are applied every day to predict the nucleation rate in liquids, solids,
and molten polymeric systems. The references are given at the end of the
chapter.

3.4 Nucleation Models and Experiments

3.4.1 Classical Nucleation Principles
Classical nucleation theory is based on the following three basic principles
outlined by Kumar:26

• The probability of nucleation is directly proportional to an exponen-

tial function, exp (–W*/kT), where W* is the minimum work required
to make the system unstable and to generate a large number of
bubbles in a continuous phase. Einstein first proposed this relation-
ship in 1930.15
• Fluctuations that create a stable nucleus form and decay by the same
path.
• In 1931, Onsagar developed an idea that applies macroscopic laws
to microscopic entities such as bubble embryos containing clusters
of gas molecules.17

Based on these principles, the steady and transient rates of nucleation are
given as

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 3.2
Historical Development of Bubble Nucleation Studies
Type of Work for Nucleation Theory;
Author(s)/Reference(s) Year Related Work
Einstein (15) 1910 Exponential function on
fluctuation probability
Volmer and Weber (16) 1926 Kinetic approach to homogeneous
nucleation
Onsager (17) 1931 Principles that allow the
application of macroscopic
laws to microscopic entities
such as embryos
Becker and Doring (18) 1935 Kinetic approach to homogeneous
nucleation
Turnbull and Fisher (19) 1949 Rate of nucleation in condensed
systems
Reiss (20); Lothe and Pound (21) 1950; 1962 Statistical mechanical concepts
Gibbs (22) 1961 Phase changes to generate bubbles;
Embryo formation due to
concentration changes
Zettlemoyer (23) 1969 Nucleation studies
Blander et al. (24) 1971 Bubble nucleation in hydrocarbons and
water
Blander and Katz (2) 1975 Homogeneous nucleation theory
Martini (25) 1981 Homogeneous nucleation model
for microcellular foams in
high-impact polystyrene
Colton and Suh (7) 1985 Mixed-mode nucleation
for polystyrene-zinc
stearate system
Kumar (26) 1988 Experiments show the gap
between homogeneous theory
and data of polystyrene-nitrogen
system
Han and Han (11) 1988 Modified free energy
approach to study bubble
nucleation in polystyrene/toluene
solutions
Ramesh et al. (10, 27) 1992, 1994 Microvoid nucleation in high impact
polystyrene
Ruengphrathuengsuka (28) 1992 Nucleation model with
hydrodynamic, heat and
mass transfer effects
Lee (9) 1993 Shear-induced nucleation in LDPE/
CFC-12
Park et al. (29) 1998 Thermal effects on polystyrene/carbon
dioxide system
Ohshima (30) 2002 Experimental studies in
bubble nucleation and
growth rates in batch and
continuous foaming

© 2004 by CRC Press LLC

www.ketabdownload.com
 Ê - DG * ˆ
J s = Zb* N expÁ ˜ (3.2)
Ë kT ¯

( )
J (t) = J s exp -t t (3.3)

where
J = Nucleation rate (nuclei/cm3-s)
s = Steady state
t = Transient state
Z = Zeldovich non-equilibrium factor @ 10 ( -2
)
b* = Rate at which gas molecules are added to the critical nucleus
N = Number of nucleation sites per unit volume
DG* = Gibb’s free energy of forming a critical nucleus
kT = Boltzman’s constant times absolute temperature
t = Induction period for establishing steady-state nucleation conditions

All of these variables change based on the type of nucleating systems and
process conditions. They must be carefully evaluated on an individual basis
to fit real-life applications.

3.5 Homogeneous Foam Nucleation

Homogeneous nucleation represents the formation of a second phase (e.g.,
a gas bubble) in the primary phase (e.g., molten polymer). Homogeneous
nucleation occurs when a sufficient number of dissolved gas molecules form
clusters for a long enough time to make a critical bubble radius to cross over
the resistance path, as shown in Figure 3.1.
Figure 3.2A shows a single phase containing molten polymer saturated
with gas at a certain pressure. Figure 3.2B shows the formation of a second
gas phase when the pressure is reduced from Po to Ps. Thermodynamic
instability is the reason for nucleation of tiny bubbles. Formation of bubbles
involves creation of new surfaces with certain volumes. Usually the embryos
are spherical in size and therefore easier mathematical expressions can be
derived based on thermodynamic principles.
For homogeneous nucleation, Equation 3.1 can be rewritten to express
Gibbs free energy as 7

4 3
DG = - pr ◊ DP + 4 pr 2 s (3.4)
3

© 2004 by CRC Press LLC

www.ketabdownload.com
 ∆G*homo

Nucleus Radius

r * = Critical Bubble Radius

FIGURE 3.1
Homogeneous bubble nucleation.

A. To, Po B. To, Ps (Ps < Po)

0 0 0 0 0 0 0

0 0 0 0 Bubble 0

0 0 0 0 0 0 0

0 0 0 0 0 0 0
Molten Polymer + Gas
0 0 0 0 0 0 0

0 0 0 0 0 0 0
Polymer
(Reduced Pressure)

FIGURE 3.2
Typical nucleation process. To = temperature; Po = initial pressure (higher than surrounding
pressure); Ps = final pressure or surrounding atmosphere pressure.

where r is the bubble radius; DP is the pressure drop (e.g., die pressure drop
in extrusion foaming); and s is the surface tension of the polymer matrix.
The maximum value of DG, denoted as DG*, occurs at a critical size r*, or
when there is a critical number of gas molecules in the embryo, and repre-
sents the free energy of formation of the critical nucleus. If we differentiate
the free energy term with respect to radius and set it equal to zero, the result
becomes

∂DG
=0 (3.5)
∂r

which gives an expression for critical radius, r*:

2s
r* = (3.6)
DP

© 2004 by CRC Press LLC

www.ketabdownload.com
 The spherical shape of the nucleus is assumed to represent minimum resis-
tance for nucleation for a given volume. In general, such an assumption is
reasonable. But in polymeric systems nonspherical geometries might be
encountered. The activation free energy for homogeneous nucleation of a
critical nucleus is derived as

16 ps 3
DG * homo = (3.7)
3 DP 2

where s is the surface tension of the polymer, and DP = Psat – Ps is the

supersaturated pressure. For a batch microcellular process system, Psat is the
gas saturation pressure and Ps is the surrounding pressure at which nucle-
ation is to occur. Usually, Ps is equal to atmospheric pressure.
It should be cautioned that gaseous foaming agents often soften the poly-
mer, and therefore surface tension of the polymer may decrease or increase
based on the foaming agent dissolved in it. Hence, such effects must be taken
into account when predicting real-life examples. Since the activation energy
term has been calculated, the nucleation rate expression can be calculated
based on the classical nucleation theory equation. According to Colton and
Suh,7 the homogeneous nucleation in a gas–polymer system is given by

Ê - DG * homo ˆ
N homo = fo Co expÁ ˜ (3.8)
Ë kT ¯

where fo is a frequency factor for the rate at which gas molecules join a
critical nucleus; and Co is the concentration of gas molecules. We can see
from Equations 3.7 and 3.8 that when the degree of supersaturation is
increased, both critical radius and critical free energy decrease. Physically
this means that a greater amount of gas in the polymer makes it easier for
bubbles to form. Similarly, the higher the pressure drop, the higher the
nucleation rate of bubbles.

3.6 Heterogeneous Foam Nucleation

This is the most common type of nucleation found in polymer systems
containing additives. The efficiency of producing bubbles depends on several
factors, such as the type and shape of nucleating particles and interfacial
tensions of solid and solid–gas interface. Blander and Katz proposed a simple
heterogeneous nucleation model for liquids in 1975.2 The primary benefit
comes from the interface, which acts like a catalyst for nucleation. The pres-
ence of tiny particles and cavities reduces the activation energy required to

© 2004 by CRC Press LLC

www.ketabdownload.com
 Homogeneous
Nucleation

Free Energy
Nucleus Radius
G
∆G*homo ∆G*hetero

Heterogenous
Nucleation

FIGURE 3.3
Heterogenous bubble nucleation. DG*hetero < DG*homo.

Gas
Polymer

Nucleating Particle Surface

FIGURE 3.4
Schematic of nucleating particle interaction with gas and polymer.

achieve a stable nucleus. Figure 3.3 shows the reduction of Gibbs free energy
associated with the heterogeneous nucleation process.
The thermodynamics of heterogeneous nucleation and its mathematical
analysis are given in Uhlmann and Chalmers.4 The heterogeneity factor can
be used to correct the activation energy term derived for homogenous nucle-
ation, as shown in the following:

DG*hetero = DG*homo f(q) (3.9)

For the configuration shown in Figure 3.4, Uhlman and Chalmers derived
an expression for f(q) as

f (q ) =
(2 + cos q)(1 - cos q)2 (3.10)
4

16 ps 3
DG * hetero = f (q ) (3.11)
3 DP 2

© 2004 by CRC Press LLC

www.ketabdownload.com
 A. B.

FIGURE 3.5
Heterogeneous nucleation assisted by rubber particles. (A) Polystyrene with no rubber particles.
(B) Polystyrene with rubber particles. (Reproduced with permission from Ramesh et al., Polym.
Eng. Sci., 34, 1685–1706, 1994.)

where q is the wetting angle, f(q) is the heterogeneity factor, and s represents
the interfacial tensions of a polymer–gas bubble.
According to Colton and Suh, for a typical wetting angle of 20
, f(q) is on
the order of 10–3, which means it is easy to produce several orders of mag-
nitude of more bubbles using nucleating sites.7 This explains why polysty-
rene produces more bubbles when rubber particles are added, as shown in
Figure 3.5A and Figure 3.5B. These examples show nucleation of bubbles
when the polymer is not sheared. The next section shows the effect of shear
force on the polymer and how it influences the nucleation process. There are
several useful chemical nucleating agents available commercially. They are
as follows:
• Citric acid/sodium bicarbonate (endothermic)
• Azodicarbonate (exothermic)
• p-Toluene sulfonyl hydrazide (exothermic)
• Sodium borohydride (endothermic)
These agents are often added to nucleate cells.

3.7 Shear Effects on Heterogeneous Nucleation

Often, nucleation of cells happens in a shear field. In a typical foam extrusion,
the melt is sheared before exiting the die and shear rate plays an important
role. Lee studied the effect of shear rate on foam nucleation for an LDPE/
CFC-12 system.9 Talc was used to serve as a nucleating site. The key result
of Lee’s experiments is that the nucleation rate increases with an increase in
shear rate. To show the effect of shear rate, the capillary number was
expressed as

© 2004 by CRC Press LLC

www.ketabdownload.com
 r * hg˙
Ca = (3.12)
4s
.
where r* = critical bubble radius, h = polymer melt viscosity, g = average
shear rate, and s = interfacial tension between polymer bubbles.
From Equation 3.12 it is clear that when the shear rate is increased, the
capillary number also increases. It was found that the rate of nucleation of
bubbles grew significantly with an increase in capillary number.9 For exam-
ple, for an LDPE/CFC-12 system at 0.039% talc level concentration, the
nucleation rate rose almost 300% when the capillary number increased from
1 to 2. Possible reasons for increased nucleation can be (a) the breaking up
of agglomerated clumps or (b) the faster release of stable nuclei from cavities.
The basic model describes shear force as a way to reduce the activation
energy barrier so that the bubbles can form more easily through a shorter
traveling path during the nucleation process. A thorough discussion can be
seen in the literature.9

3.8 Microvoid Nucleation

The classical nucleation models in the literature predict a negligible nucle-
ation rate up to a gas saturation pressure of about 34.475 MPa. Kumar has
concluded that the cell nucleation phenomenon in thermoplastics is not well
described by the classical homogeneous nucleation theory.26 In this section,
a new model for nucleation will be presented, based on the heterogeneity
of rubber particles introduced into the polymer where stress-induced micro-
voids on the rubber provide the nucleation sites. The undiluted HPS contains
30% by volume of rubber particles with an average rubber domain size of
2 mm. This polymer was added to polystyrene to a final concentration of
0.3% of rubber gel by volume. At this concentration, the blended polymer
has approximately 109 rubber particles per cm3 of the polymer. The polysty-
rene diluted with HIPS or compounded HIPS/PS blend at 1% HIPS concen-
tration can also be called nucleated PS or 1% HIPS for convenience.

3.8.1 Nucleation Mechanism

Nucleated polystyrene (PS) contains polybutadiene (PB) rubber particles
dispersed in the continuous PS matrix. Composite rubber particles consist
of a collection of occluded stiff PS inclusions in a topologically continuous
minority phase of complaint PB, generally produced by grafted copolymer-
ization as shown in Figure 3.6.
These rubber particles are assumed to have a perfect bond between them-
selves and the surrounding PS matrix transmitting all tractions. The polymer
is in the form of cylindrical pellets. The diameter and the height of the pellets

© 2004 by CRC Press LLC

www.ketabdownload.com
 FIGURE 3.6
Scanning electron micrograph of plasma etched surface of HIPS polymer. (Original magnifica-
tion  3000.) (Reproduced with permission from Ramesh et al., Polym. Eng. Sci., 34, 1685–1706,
1994.)

are 2.5 mm and 3 mm approximately. The pellets are extruded at high

temperature (438 K) and then cooled to room temperature. Since the glass
transition temperatures of PS and PB rubber are 373 K and 181 K, respec-
tively, the PS phase first “freezes” at approximately 373 K. The PB rubber
phase continues to shrink at a faster rate until it reaches room temperature,
298 K. It is well known that cavitation may occur when cooling this polymer,
due to a threefold difference in the coefficients of thermal expansion of the
rubber particle and the matrix resulting in a triaxial tension.
When the polymer is cooled from 373 K to 298 K, PB has a higher coefficient
of thermal expansion than PS. The cooling of a nucleated PS polymer from
the melt, which can be considered stress free, produces stresses in the rubber
particle and in the surrounding matrix. The magnitude of the thermal stresses
can be calculated for an isolated particle from the theory of elasticity as
described in Timoshenko and Goodier.31 Within the particle, thermal shrink-
age produces a triaxial tensile stress given by this formula:32,33

S rr = S qq = S ff = Á
( pb )(
Ê 4 b - b 1 + n G G DT ˆ
ps pb )
pb ps
˜ (3.13)
Ë ( ) ( )
Á 6 1 - 2 n pb Gps + 3 1 + n pb Gpb ˜
¯

© 2004 by CRC Press LLC

www.ketabdownload.com
 And within the matrix:10,15

Ê Rp ˆ
S rr = Á ˜
3

Á
( pb )(
Ê 4 b - b 1 + n G G DT ˆ
ps pb )
pb ps
˜ (3.14)
Ë r ¯ ( ) ( )
Á 6 1 - 2 n pb Gps + 3 1 + n pb Gpb ˜
Ë ¯

where

b = Thermal expansion coefficient

n = Poisson ratio
DT = Temperature drop
G = Modulus
Rp = Particle radius
r = Radial distance from center of inclusion
subscript ps = PS phase
subscript pb = PB phase

The physical properties needed for Equations 3.13 and 3.14 are given in
the literature.2,7 The calculated value of thermal shrinkage stress within the
particle is 2.63  106 Pa. Schmitt suggested that these thermally induced
stresses built up inside the rubber particle might cause the rubber to cavitate
and microvoids to nucleate inside the rubber particle near the PS inclusions.34
This hypothesis was confirmed later by Keskkula et al.35 by studying the
internal morphology of rubber particles using a transmission electron micro-
scope (TEM). Figure 3.7 shows the presence of microvoids in the rubber
membranes inside the rubber particle.10

0.5 µm

FIGURE 3.7
Transmission electron micrograph (TEM) of internal structure of HIPS rubber particles showing
the presence of microvoids in the rubber membranes inside the particle. (Reproduced with
permission from Ramesh et al., Polym. Eng. Sci., 34, 1685–1706, 1994.)

© 2004 by CRC Press LLC

www.ketabdownload.com
 It is important to note that the radial thermal stress within the matrix (Equa-
tion 3.14) depends strongly on the particle radius and thus on the particle size.
This means that large particles should induce very high stresses and that they
tend to nucleate microvoids with large radii. Thus, microvoid distribution is
proposed to be analogous to rubber particle size distribution. Also, the micro-
voids are formed without penetration into the PS phase of the matrix. Similar
results showing microdamage within rubber membranes was also reported
by Donald and Kramer.36 Recently, Adams et al. reported similar results for
the influence of thermal and pressure history on microvoid formation in poly-
carbonate.37 They concluded that samples that cooled quickly under low pres-
sure resulted in more microscopic flaws and voids in homopolymers. Their
analysis indicates that the polymer is self-constrained due to the temperature
effects on modules and the thermal coefficient of expansion. Hence, it is clear
that a population of microvoids exists inside the nucleated PS composite
material and that the preexisting microvoids are present both in the matrix
and inside the rubber particles. To obtain cell density, the population of viable
microvoids with a sufficiently large radius to survive and overcome surface
and elastic energy forces has been modeled. Here the assumption is that the
voids in the rubber will dominate during the process in a heterogeneous
system. The features of the mathematical model based on this hypothesis are
presented in the following section.

3.8.2 Nucleation Model

The schematic of the nucleation model is shown in Figure 3.8. This diagram
shows the presence of a single microvoid surrounded by a finite amount of
polymer. The pressure inside the microvoid tends to grow into a cell while
the opposing surface and elastic forces tend to collapse the microvoid. The
scanning electron microscope (SEM) picture of microcellular foams obtained
by foaming the nucleated PS at 388 K is also shown in Figure 3.9. These
micrographs (Figures 3.6 and 3.9) indicate that both particles and cells vary
in their diameters significantly over the cross section of the micrograph. The
SEM micrographs were digitized and the resulting rubber particle size and
cell size distributions were plotted against the cumulative number of parti-
cles, as shown in Figure 3.10 and Figure 3.11.
Two observations can be made from the data. First, the linearity in the data
indicates that both rubber particles and cells exhibit log-normal distributions.
Second, the nature of the rubber particles’ distribution seems to influence
the cell size distribution because the slopes of the distribution are nearly the
same. This is not surprising because Equation 3.14 shows the role of particle
size on the thermally induced stresses in the matrix, as already discussed.
Since particle size distribution is one of the variables, in addition to the
temperature difference and thermal expansion coefficients, that distribution
controls the thermal stresses in the matrix, and it influences the size distri-
bution of microvoids formed during the cooling of the polymer from its

© 2004 by CRC Press LLC

www.ketabdownload.com
 Surface +
Elastic
Effects

R
P(R)
GAS
S

POLYMER + GAS

FIGURE 3.8
Schematic diagram of a single microvoid surrounded by a finite amount of polymer. (Repro-
duced with permission from Ramesh et al., Polym. Eng. Sci., 34, 1685–1706, 1994.)

10 µm
Saturation Pressure = 13.8 MPa (2000 psig)
Foaming Temperature = 388 K
Gas: Nitrogen
Heating Medium: Ethylene Glycol Bath

FIGURE 3.9
Microcellular foams produced from nucleated PS. (Reproduced with permission from Ramesh
et al., Polym. Eng. Sci., 34, 1685–1706, 1994.)

© 2004 by CRC Press LLC

www.ketabdownload.com
 10

Rubber particles size, µm

0.1
.01 .1 1 5 10 20 30 50 70 80 90 95 99 99.9 99.99
Cumulative number of rubber particles (%)

FIGURE 3.10
Particle size in the nucleated PS polymer as determied by the plasma etching process. (Repro-
duced with permission from Ramesh et al., Polym. Eng. Sci., 34, 1685–1706, 1994.)

molten state to room temperature. Based on the aforementioned reasons, it

was concluded that the log-normal distribution is appropriate to model the
distribution of microvoids and that the microvoids in this heterogeneous
process depend on the rubber particle size distribution. Though each particle
may contain multiple voids, only the largest void among them will survive
owing to its high radius, and all other smaller microvoids both inside the

100
Cell size, µm

10

1
.01 .1 1 5 10 20 30 50 70 80 90 95 99 99.9 99.99
Cumulative number of cells (%)

FIGURE 3.11
The cell size distribution after the microcellular foaming process. (Reproduced with permission
from Ramesh et al., Polym. Eng. Sci., 34, 1685–1706, 1994.)

© 2004 by CRC Press LLC

www.ketabdownload.com
 particles and in the matrix will disappear. Therefore, it is reasonable to
assume that each particle potentially contributes a single microvoid, whose
radius is influenced by its own particle size. Mass transfer during the nucle-
ation process was neglected because the time scale for nucleation is negligible
in comparison with that of the foam growth process (0.0004 s vs. 120 s). In
addition, the following assumptions were made to derive the nucleation
model:

• The nucleation process is assumed to be governed by the mechanics

and thermodynamics of the system.
• A Neo-Hookean model is considered here for modeling the elastic
energy that opposes the growth of a preexisting microvoid. This is
widely used for rubberlike solids since the foaming temperature is
close to the glass transition temperature of the experimental obser-
vations. These models have been shown to be very useful in pre-
dicting the nucleation of bubbles in natural and PB rubber.10
• Spherical symmetry is assumed for simplicity.
• The loss of gas to the surrounding vapor during nucleation is
neglected because it is limited to the skin of the polymer pellet.26
• Solubility of the gas to the polymer is a function of both saturation
pressure and temperature. The solubility of nitrogen in PS and PB
is nearly the same.38

With these assumptions, Ramesh et al.10 derived the following expression

for the nucleation rate:

No log( R / x)
N= erfc (3.15)
2 s 2

where No = total number of potential microvoids (which can be estimated

by the number of particles present in the continuous phase); R = instanta-
neous bubble radius; x = mean of the distribution of the void size; and s =
standard deviation of the distribution. More details are given elsewhere.27

3.8.3 Results and Discussion

Ramesh27 investigated the effect of gas saturation pressure on cell size for
the HIPS/PS blend (1% HIPS), which has approximately 109 rubber particles
per cm3. The average cell size obtained from this experiment for PS and 1%
HIPS (containing 2 mm size particles) has been plotted (Figure 3.12) as a
function of saturation pressure. The average cell size decreases with an
increase in the saturation pressure of PS. On the other hand, it remains nearly
the same for 1% HIPS.
It is interesting to note that particle size distribution plays a role in dictating
the slope of cell size distribution data. It appears that there is a strong

© 2004 by CRC Press LLC

www.ketabdownload.com
 60
Polystyrene (PS)
1% HIPS in PS
50

40

Cell size, µm
30

20

10

0
4 6 8 10 12 14 16
Saturation pressure, MPa

FIGURE 3.12
Cell size as a function of saturated pressure for polystyrene and 1% HIPS foamed at 388 K.
(Reproduced with permission from Ramesh et al., Polym. Eng. Sci., 34, 1685–1706, 1994.)

correlation between particle size and cell size prior to nucleation. This clearly
indicates the ability of the rubber particles to control cell size during nucle-
ation. In general, there is good agreement between theoretical predictions
and the experimental data, as can be seen from Figure 3.13.

1010

109
Cell density, cells/cm3

108

107
model (surface forces only)
model (surface and elastic), y = finite
model (surface and elastic), y = infinite
experimental data
106
0 2 4 6 8 10 12 14 16
Saturation pressure, MPa

FIGURE 3.13
Comparison of theory with experiment for the influence of saturation pressure on cell density
for 1% HIPS. (Reproduced with permission from Ramesh et al., Polym. Eng. Sci., 34, 1685–1706,
1994.)

© 2004 by CRC Press LLC

www.ketabdownload.com
 3.9 Effect of Concentration of Particles on Nucleation
To investigate the effect of particle concentration on cell density, HIPS/PS
blends, 0.1% HIPS, 1% HIPS, and 10% HIPS were saturated with nitrogen
at 13.78 MPa (2000 psi) and allowed to foam in an ethylene glycol bath at
388 K for sufficient time (5 min) to develop a limiting bubble size. The results
were plotted on a log-log scale. The plot revealed that the foamability of the
polymer was directly related to the number of rubber particles present in
the sample, as illustrated in Figure 3.14. The theory (solid line) predicts a
one-to-one increase in the number of cells with increases in concentration
for the same level of supersaturation of the polymer. Comparison with
experimental data shows that the predictions are reasonable, although some
deviations are observed.

3.10 Effect of Rubber Particle Size on Nucleation

To study the effect of rubber particle size on cell density, a series of nucleated
PS samples containing rubber with average particle size of 0.3, 0.5, 2.1, 4.2,

1011
Theory
Experiment
10% HIPS in PS
1010
Cell density, cells/cm3

1% HIPS in PS
109

0.1% HIPS
in PS
108

Saturation Pressure = 11.032 MPa

Foaming Temperature = 388 K
107
107 108 109 1010 1011
Concentration, rubber particles/cm3

FIGURE 3.14
Comparison of theory with experiment for the influence of particle concentration on cell density
for 1% HIPS foamed at 388 K. (Reproduced with permission from Ramesh et al., Polym. Eng.
Sci., 34, 1685–1706, 1994.)

© 2004 by CRC Press LLC

www.ketabdownload.com
 40
5.52 MPa
36 8.27 MPa
11.03 MPa
32 13.80 MPa

28
Cell size, µm
24

20

16

12

4
0 1 2 3 4 5 6 7 8
Rubber particle size, µm

FIGURE 3.15
Influence of rubber particle size on cell size of the nucleated PS cells foamed at 388 K for various
nitrogen saturation pressures. (Reproduced with permission from Ramesh et al., Polym. Eng.
Sci., 34, 1685–1706, 1994.)

and 6.8 mm were selected. Figure 3.15 shows the trend. The cell size appears
to increase with the rubber particle size in a batch foaming process. Its impact
on the extrusion process still needs to be explored.

3.11 Nucleation in Polyurethane Foams

Polyurethane (PU) foam is well known for its wide range of commercial
applications. PU foams are used in packaging, nonpackaging, carpet backing,
sports and leisure applications, seat cushions, and insulation. PU foam is a
thermosetting foam, whereas PS and polyethylene foams are thermoplastic
foams that are easy to recycle. The typical reaction steps are shown below:

2R-NCO + n1 H2O + n2 R-OH + C + S Æ PU Foam + n3 H2O + n4 R-OH

+ n5 CO2

where n1 to n5 are the number of moles of reactant and product species. R-

NCO, R-OH, C, and S are polyisocyanate, polyol, catalyst, and surfactants
respectively, and

© 2004 by CRC Press LLC

www.ketabdownload.com
 1 1
R-NCO + § 2 H2O Æ Polymer + § 2 CO2 + heat

R-NCO + R-OH Æ Polymer + heat

Nucleation of bubbles in PU systems can happen due to the following:

• Evolution of gases like carbon dioxide due to chemical reaction

• Additional gases dissolved in polyol component, including air
• Evaporation of water vapor to participate in the expansion of the
foam

PU is formed because of a reaction between isocyanate and polyol. PU foam

is produced when the PU polymer is expanded by the formation of carbon
dioxide that is produced when reaction isocyanate and water react. The
reaction is exothermic in nature, and the heat generated is used to change
water into steam to help expand the foam.
Both carbon dioxide and steam act as blowing agents. While making very
low density foams, below 2 lb per cubic ft (or 32 kg/m3), the water vapor is
the dominant blowing agent.
Kanner and Decker demonstrated the impossibility of self-nucleating car-
bon dioxide from a mixture of polyol and toluene diisocyanate.39,40 They also
showed that bubble formation was aided by physically mixing air into the
above mixture with an appropriate amount of surfactant for proper stabili-
zation. The addition of air bubbles having diameters in the range of 10–100
mm was found effective in forming cells. Elevated temperature due to pre-
heating raw materials under pressure and the heat of reaction releases dis-
solved air. Thus, the nucleation of cells in PUs can be attributed, in part, to
chemical mixing and agitation with the accompanying dissolution of inert
gases.
Several other gases were considered besides air. These include carbon
dioxide, hydrocarbons, and halogenated blowing agents. While carbon diox-
ide can be dissolved in a polyol side of a PU formulation, it is detrimental
to foam formation as it works to neutralize the amine catalysts, preventing
the reaction from occurring. Due to the detrimental effects of halogenated
blowing agents on the earth’s ozone layer, they are not used in the formation
of PU foams. In some cases, hydrocarbons are used to assist in foam nucle-
ation and growth. The largest commercial occurrence of hydrocarbons is the
use of cyclopentane for appliance insulation foams in Europe.
As foam reaction is initiated, the viscosity of the polymer increases signif-
icantly due to urethane linkages between polyols and polyisocyanates, and
this stabilizes cell formation and growth. The key difference between other
thermoplastic systems and PU foam nucleation is a chemical reaction fol-
lowed by viscosity increase for foam stabilization. The thermosetting nature
of PU foams offer superior thermal stability over polyolefin foams.

© 2004 by CRC Press LLC

www.ketabdownload.com
 3.11.1 Bubble Nucleation
Equation 3.16 applies to bubble nucleation.41 In order for the bubbles to
nucleate, the Gibbs free energy of the system must exceed the mechanical
work needed. Mathematically, this can be written as:

3◊s
DG > (3.16)
r◊r

where s is the surface tension, r is the nucleation bubble radius, and r is the
gas density.
Surface tension is regulated with the use of silicone surfactants. The type
of surfactant used varies with the formulation application. This area of
formulation art is outside the scope of this chapter. However, silicone sur-
factants are almost always used to regulate cell count or cell size. In the cases
of low density foams, below 30 kg/m3, surfactants plays key roles in pro-
ducing open cells to maintain dimensional stability upon cooling to ambient
temperature.

3.11.2 Bubble Growth Model for Polyurethane Foam

Ramesh and Malwitz42 proposed a material and energy balance model based
on the work by Wang, Frisch, and Malwitz.43 It predicts accurately the foam
expansion characteristics. The polyurethane foam process includes isocyan-
ate, polyol, water, catalysts, and surfactants as reactants. The products are
PU foam, unreacted reactants, carbon dioxide, and steam. The foam expands
due to gas evolution and heat of reaction between isocyanate, polyol, and
water. The heat vaporizes water and does the expansion work. The heat of
reaction between isocyanate and polyol can be described as:

312.6
Heat of reaction (Kcal/mol) = - 20.47 (3.17)
T

The heat of reaction between isocyanate and water can be described as:

543.9
Heat of reaction (Kcal/mol) = - 20.65 (3.18)
T

where T is the temperature in Kelvin. The following equation applies to

bubble growth:

2◊s
DP > (3.19)
r

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 3.3
Model Prediction vs. Experiment for Different Formulations
Model Prediction Density Experimental Density
Foam (pcf) (pcf)

PU Formulation 1 2.037 2.0

PU Formulation 2 0.754 0.75
PU Formulation 3 0.43 0.4

where DP is the pressure from blowing agent generated by CO2 and steam
due to exothermic reaction.
Table 3.3 shows the theoretical prediction against experimental data. Based
on the data, it seems possible to predict the PU foam expansion and density
from simply knowing the chemical formulations, using the mathematical
model.42
Foam expansion with respect to percentage of isocyanate reacted for one
formulation is shown in Figure 3.16.
It is important to recognize the difference between thermoplastic and ther-
mosetting foams. Table 3.4 shows the differences in bubble nucleation,
growth, and properties between thermoplastic and thermosetting foams. The
differences in the table show that nucleation and bubble growth processes
vary significantly.

100
Theory
Experimental
Foam density (pcf)

10

0
10 100
% Isocyanate reacted

FIGURE 3.16
Predictions of foam density—theoretical vs. experimental.

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 3.4
Differences between Thermoplastic and Thermosetting Foams
Bubble Nucleation, Thermoplastic Thermosetting
Growth, and Properties Foam Foam
System Non-crosslinked PS Polyurethane foam
and polyolefin foam
Easy to recycle Yes; melting and repelletizing More difficult; regrind,
is often used chemical hydrolysis,
glycolyis, and pyrolysis
options are available
Bubble nucleation Thermodynamic instability of Thermodynamic instability
reason molten polymer-gas system with gas evolution due to
chemical reaction
Bubble growth Inertial and diffusion- Diffusion-controlled
controlled
Cell structure Mostly closed cells are desired Closed and open cells;
open-cell foam is desired
to have a balance of
properties; cell opening
and gelling due to
crosslinking reaction
Gel point No Yes
Cell-opening process No Yes; influenced by kinetics
of bubble growth,
catalysts, surfactants, and
surrounding conditions
Morphology Amorphous and semi- Urea hard domains
crystalline distributed in polyether
soft phase

3.12 Foam Growth in Polymers

After cell nucleation, the bubbles grow due to the diffusion of excess gas in
the polymer. The viscosity of the polymer, the gas concentration, the foaming
temperature, and the amount of nucleating agent and its nature are some of
the variables that control the foam growth process. The expansion dynamic
of polymeric foam is very complicated. It has been increasingly scrutinized
by investigators for the past 90 years. Key milestones in the development of
foam growth models are summarized in the following.

3.12.1 Bubble Growth Models

Several investigations addressing the theoretical and experimental analysis
of bubble growth and collapse in fluids and polymers have been available
in the literature since 1917. The historical development of important models
is listed in Table 3.5. Most of the models can be classified into three groups:

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 3.5
History of Important Models
Bubble
Momentum Heat Rheological Count Ref.
Author Year Transfer Transfer Model Study No.

Rayleigh 1917 Inertial Isothermal — Single 44

Epstein, 1950 — Isothermal — Single 45
Plesset
Scriven 1959 Yes Nonisothermal Newtonian Single 46
Street, 1971 Viscous Conduction Power Law Single 47
Fricke,
Reiss
Han, Yoo 1981 Yes Isothermal Viscoelastic Single 48
Upadhyay 1985 Viscoelastic Nonisothermal Viscoelastic Single 49
Amon, 1984, Viscous Conduction Newtonian Swarm 50, 51
Denson 1986
Arefmanesh, 1991 Viscoelastic Conduction Viscoelastic Swarm 52
Advani (Maxwell)
Ramesh 1991, Viscoelastic Conduction Maxwell, Swarm 27, 53
1992 Power
Law
Lee, 1993 Viscoelastic Conduction Convected Swarm 54
Ramesh, Maxwell
Campbell
Ramesh, 1998 Viscoelastic Nonisothermal Maxwell Swarm 55
Malwitz
Shafi, Joshi, 1997 Viscous — Newtonian Single 56
Flumerfelt
Ohshima 2000 Viscous — Newtonian Swarm 57

(1) the single bubble growth model; (2) the cell model (a swarm of bubbles
growing without interaction); and (3) the combined nucleation and bubble
growth model.

3.12.1.1 Single Bubble Growth Models (1917–1984)

Between 1917 and 1984,44–51 the published models focused on the growth or
collapse of a single bubble surrounded by an infinite sea of fluid with an
infinite amount of gas available for growth. Although these models gave
several insights into bubble growth phenomena, their practical application
in industry was severely limited because the foaming process involves the
growth of numerous bubbles expanding in close proximity to one another
with a limited supply of gas. This led to the development of a new model
called the “cell model,” which will be discussed in the next section.

© 2004 by CRC Press LLC

www.ketabdownload.com
 3.12.1.2 Cell Model (1984–1998)
The cell model has been widely used to describe devolatilization, batch
microcellular, and foam extrusion processes. Gross motion, no bubble
motion, and motion without shear have different implications in the cell
model.
The concept of a cell model was first introduced by Amon and Denson in
1984.50 Their study involved the growth of a group of gas bubbles separated
by a thin film of polymer and dissolved gas during the injection molding
process. The foam was divided into spherical microscopic unit cells of equal
and constant mass, each consisting of a liquid envelope surrounding a single
bubble; the gas available for growth was thus limited. Because it relied on
a more realistic assumption of gas availability, the cell model yielded a final
radius while single bubble growth models showed the growth of bubble
radius was infinite. This fundamental improvement caused more interest in
this area of research, and several studies have emerged since then. However,
the Newtonian viscosity equation was used to describe the rheology of the
polymer due to the complexity of the problem. Later it was modified to
account for the non-Newtonian viscoelastic effects to make it more suitable
for polymeric systems.51
Cell models27,50–55,59 can be broadly classified into two groups: (a) the cell
model for a closed system with no blowing agent and gas loss effects; and
(b) the modified cell model for foam extrusion with a blowing agent and gas
loss effects. These groups are described in the next section. Recently, Shafi
et al. have worked on combining nucleation and foam growth processes.56
They considered bubble size distributions in freely expanded polymer foams.
The final bubble size distribution depended upon nucleation rate and bubble
growth dynamics. However, their work focused only on the initial stages of
bubble growth and the model does not predict an equilibrium radius, which
is important when considering real industrial processes.

3.12.1.2.1 Viscoelastic Cell Model for Injection Molding Process (VCM-IM)

The validity of the cell model was initially tested in experimental work by
Amon and Denson,50 Arefmanesh and Advani,52 Ramesh and Malwitz,58 and
Ramesh et al.54 The first two research groups concentrated on experiments
to verify prediction of the cell model during the injection molding process,
where the bubble growth occurs in a closed system. Amon and Denson50
used low-density polyethylene with a chemical blowing agent whereas Aref-
manesh and Advani52 used polycarbonate with a chemical blowing agent.
The assumption of no gas loss from the mold was adequate because the foam
growth occurred in a closed mold. These studies showed a qualitative agree-
ment between experimental data and theoretical results and have led to a
better understanding of the processing of polymeric foam materials.
Ramesh et al. tested the validity of the cell model by conducting a cell
growth experiment during the microcellular foaming process.27,53 PS was
used with physical blowing agents such as nitrogen and carbon dioxide. In

© 2004 by CRC Press LLC

www.ketabdownload.com
 all the above cases, the blowing agent loss was considered to be negligible,
and therefore the no-gas-loss boundary condition was adequate to predict
the foam growth kinetics reasonably well. However, in the polymeric foam
extrusion process of interest here, the foam sheet or rod is allowed to expand
freely under atmospheric conditions. The assumption of no gas lost to sur-
roundings no longer applies for the actual processing. Therefore, such gas
escape must be accounted for, especially with thin foam shapes. Further-
more, the previous models do not include further complicating effects such
as the influence of the blowing agent on polymer viscosity and the concen-
tration dependence of the blowing agent on diffusion during the bubble
growth process.

3.12.1.2.2 Modified Viscoelastic Cell Model for Foam Extrusion Process

Recently, Lee and Ramesh studied the effects of foam sheet thickness and
nucleation cell density on thermoplastic foam sheet extrusion.59 Low-density
polyethylene was used with HCFC-22 and HCFC-142b to produce foam
sheets of various thickness and nucleation characteristics. It was concluded
that solubility, rheology, and gas loss transport mechanisms play an impor-
tant role in determining the foaming efficiency during foam sheet extrusion.
While predicting the foam density as a function of final cell density (or
number of cells/unit volume) and foam sheet thickness, Lee and Ramesh
used mathematical equations from packing theory to calculate the ratio of
bubbles present closer to the surface to the bubbles present in the core portion
of the foam sheet where there is no gas loss. The value of this ratio, of course,
strongly depends on the nucleation density (which is defined as the number
of bubbles nucleated per unit volume of the polymer) and the thickness of
the foam sheet. With this theoretical background, they attempted to check
the validity of the modified model by comparing with the experimental data.
Although the improved cell model predicted the experimental data very
well, between 919 kg/m3 and 60 kg/m3, it was observed that the agreement
decreases for increased expansion ratios and volatility of blowing agent,
where the nonspherical nature of the bubble, and bubble–bubble interaction
phenomena, become pronounced. Higher nucleation rates and extrusion of
thicker foam sheets seem to enhance foam efficiency. Correlation with cell
geometry, nonspherical bubble dynamics, and consideration of interaction
between bubbles during the growth process would greatly benefit under-
standing of the low-density foam process and assist in product design.

3.13 Foam Extrusion Modeling

A typical foam extrusion process is used here to demonstrate the carefulness

of bubble growth modeling and its ability to predict experimental data. For

© 2004 by CRC Press LLC

www.ketabdownload.com
 LDPE RESIN + Blowing Agent Metering Pump
Nucleating
Agent Blowing
Agent
Gravimatric
Feeder
Blowing Agent Injection

Computer Feed Cooling Capillary

Extruder
Instrumentation Foam
Zone
and Control Die
Video
System
Distance
Expending Measurement
Rod Foam

FIGURE 3.17
Rod extrusion from a die.

example, in a foam rod extrusion from a die shown in Figure 3.17, one-
dimensional heat transfer is assumed.
Heat conduction becomes the dominant heat transfer mechanism. The heat
loss effects on foam growth are less important when the diameter of the
foam rod exceeds 10 mm. The average temperature of the thin foam as it is
exposed to ambient conditions can be found from the standard heat transfer
equation in the literature.60
The schematic diagram of the cell model is shown in Figure 3.18. The
appropriate bubble growth modeling equations are listed there.

∂c = K (C −C ) (Gas Loss)
D ∂r m 1 ∞

∂c = 0 (No Gas Loss)

D ∂r

RF

Pg R

GAS
POLYMER
+ BLOWING
AGENT

c = Kw Pg (Henry’s Law)

FIGURE 3.18
Gas loss boundary condition in the cell model.

© 2004 by CRC Press LLC

www.ketabdownload.com
 3.13.1 Foam Growth Equations
To define gas pressure inside the bubble, the integrated momentum equation
reduces to:
Rf
2s dr
Pg - Ps -
R
+
Ú
R
(t rr - tqq )
r
=0 (3.20)

The rheological equations are

h*o
t+ t = - h*o g (3.21)
G (1)

È E Ê 1 1 ˆ˘
h*o = ho exp Í v Á - ˜ ˙ * f (C) (3.22)
ÍÎ Rg Ë T To ¯ ˙˚

where

f (C) = viscosity reduction factor

Polymer viscosity with blowing agent (3.23)
=
Viscosity of polymer at same temperature

The growth of radius is

d È ∂C ˘
(
dt g
)
r R 3 = 3rDR 2 Í ˙
Î ∂r ˚ r = R
(3.24)

and the concentration-dependent diffusion equation is as follows:

∂C ∂C 1 ∂ Ê 2 ∂C ˆ
+ Vr = Á Dr ˜ (3.25)
∂t ∂r r 2 ∂r Ë ∂r ¯

where diffusivity is a function of blowing agent concentration and it changes

within the polymer envelope during the gas-diffusion process. The diffusiv-
ity also depends strongly on the temperature. Mathematically,

D(C , T ) = [1 + aC]10 -7 e ( )
b- c T
(3.26)

where a, b, and c are constants fitted to the experimental data that depend
on the nature of the blowing agent. The values of a, b, and c need to be
obtained experimentally. For this work, the values used based on our exper-
imental data and the data provided by the resin supplier are listed in Table 3.6.

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 3.6
Experimental Parameters for Bubble Growth Calculations
Process Variables Values
mN-
Surface tension of low-density polyethylene (LDPE) 30 --------
m
Molecular weight of butane (blowing agent) 58 kg/kg mol
Initial blowing agent concentration 6.0% by weight
Foaming temperature, To 383 K
Henry's law constant, Kw 2.235 ¥ 108 Pa
Density of LDPE 920 kg/m3
No. of bubbles per 1 cm3 of foam (from experiment) 136
No. of bubbles across strand diameter 5
Diffusion empirical constant, a 0.5334
Diffusion empirical constant, b 21.93
Diffusion empirical constant, c 7090 K

3.13.2 Boundary Conditions

The initial and boundary conditions are

C(r , o) = Co = K w Pg (3.27)

C( R, t) = K w Pg (3.28)

For bubbles situated in the core of thick sheet expansion:

∂C
=0 (3.29)
∂r

For bubbles on the surface of foam sheet that undergo gas loss:

∂C
D
∂r
(
= k m Csurf - Cs ) (3.30)

where

D
km = 2 (3.31)
pt

© 2004 by CRC Press LLC

www.ketabdownload.com
 and where km is defined as the mass transfer coefficient derived from pen-
etration theory60 and t is the foam growth time. The surface evaporation or
gas loss condition is mathematically expressed in Equation 3.32:

R(t = 0) = Ro (3.32)

The boundary condition, Equation 3.29, represents a closed cell system

where there is no gas loss or escape to its surroundings. This is true in the
injection molding process because the foam expansion occurs inside the
mold.
But, in the practice of producing foams with unrestricted rise, gas loss is
encountered from the bubbles that are adjacent to the top and bottom surfaces
of the foam. The influence of gas loss on foaming efficiency is especially
profound when the extruded shape is thin. Some of the gas close to the
surface, instead of diffusing into the cell, diffuses to the surface and transpires
into the atmosphere. As a result, the final surface cell size is smaller than that
of the core cell. As the final foam density is lower, or the shape is thinner,
more gas escape is anticipated. In other words, high concentrations of blowing
agents or high surface-to-volume foam sheets accelerate gas loss from a sheet
surface, thereby lowering the blowing agent effectiveness. This phenomenon
is compounded at lower atmospheric temperatures that decrease the foam
growth rate. A similar situation is encountered when a small diameter foam
rod is extruded.
The above system of equations was solved numerically by an iterative
numerical scheme. To facilitate the numerical simulation, a definite value of
initial radius Ro is required. Bubble growth originates from thermodynamic
and mechanical instability in the system with perturbed initial radius.50,56,61
Hence, the selection of initial radius is important to achieve accuracy in the
predicted results. Great care is needed to model successfully. This principle
has not been clearly emphasized in the literature so far. The bubble growth
seems to be a strong function of the value of the perturbed initial radius
during the induction period, which is less than 8% of the total expansion
time. Beyond this initial growth time, i.e., greater than 8% of the total growth
time, the foam growth rate is found to be independent of the growth radius,
according to Amon and Denson,50 Upadhyay,49 Ramesh et al.,53 and Aref-
manesh and Advani.52 For all practical reasons, it is better to calculate the
initial radius from fundamental mechanical equilibrium principles, which
involve the calculation of a critical bubble radius. The value of the critical
bubble radius r* can be easily calculated from the force balance at the bubble
wall, which can be written as:

2s
r* = (3.33)
DP

© 2004 by CRC Press LLC

www.ketabdownload.com
 where s is the surface tension and DP is the pressure difference between the
liquid polymer phase and the dissolved gas phase.
The simulation parameters used for all cases in this study are listed in
Table 3.6. The model yields the result in terms of bubble radius as a function
of time. The growing bubble reaches an equilibrium bubble radius once the
gas is depleted. Finally, the bubble diameter is multiplied by the number of
cells that are simultaneously growing across the diameter of the expanding
rod to calculate the rod diameter. The simulated results are plotted in Figure
3.19.

3.13.3 Comparison of Theory with Experiment

Figure 3.19 shows the comparison model predictions with the viscoelastic
model52,62 and experimental data. The solid line denotes the current model,
and the dashed line denotes the previous model. The prediction of the foam
strand diameter by the new model seems to agree well with the experimental
data when the concentration and surface gas loss effects are included. The
previous model predicts longer growth time due to neglecting concentration
effects. Furthermore, the predicted induction time is longer and the shape
of the curve does not seem to match with the characteristics exhibited by
the experimental data. The previous model also predicts a higher strand
diameter than is experimentally observed because it assumes no gas loss
from the bubbles that are close to the surface. A significant difference in

12,000
µ

8,000

4,000 Concentration
dependent (55)
Viscoelastic (52)
Experimental data
Initial rod diameter
0
0.00 1.00 2.00 3.00 4.00 5.00
Rod growth time, seconds

FIGURE 3.19
Comparison of experiment vs. theory for PE foam. (From Ramesh, N.S. and Lee, S.T., Foams,
99, 91, 1999.)

© 2004 by CRC Press LLC

www.ketabdownload.com
 strand diameter leads to serious discrepancies in predicting the final foam
density as it depends on the square of the rod diameter. Therefore, analysis
including blowing agent concentration and surface escape effects into the
thermoplastic foam expansion model is essential to accurately predict foam
growth rate and density.

3.14 Conclusion
This chapter gives an overview of nucleation and bubble growth models
and experiments since 1917. A large volume of literature exists for scientists
to read and understand the complicated foaming process. The more we think
we know about the foaming process, the more we realize how little we know,
due to the complex variables involved in gas dissolution, bubble nucleation,
bubble growth, and foam stabilization. Recent experimental studies by
Ohshima30 have opened up new opportunities to capture a nucleation and
foam growth event. Similarly, the concentration-dependent bubble growth
model55 can be easily applied to describe the foam extrusion process. These
are important steps in understanding foaming polymers to generate new
products to serve automotive, packaging, and nonpackaging markets.
Although 80 years of research has been done, much more work is needed
to describe foaming mechanisms.

Acknowledgment
The author wishes to thank Kevin Lee for his assistance in preparing this
manuscript.

Nomenclature
a Experimental constant as per Equation 3.26
A Interfacial area
b Experimental constant as per Equation 3.26
c Experimental constant as per Equation 3.26
C Concentration of gas molecules
D Diffusion coefficient (cm2/s)
Ev Activation energy for viscosity equation
f(C) Viscosity reduction factor
fo Frequency factor for the rate at which gas molecules join a
critical nucleus
f(q) Heterogeneity factor

© 2004 by CRC Press LLC

www.ketabdownload.com
 G Modulus
DG* Gibbs free energy of forming a critical nucleus
k Boltzman’s constant
km Mass transfer coefficient
Kw Henry’s Law constant
n Number of moles
N Number of nucleation sites per unit volume
No Total number of potential microvoids
P Pressure
DP Pressure drop
r Radial distance from the center
r* Critical bubble radius
R Instantaneous bubble radius
Rf Cell outer radius
Rg Gas constant
S Outer radius of cell as shown in Figure 3.8
T Foam growth time
T Absolute temperature
DT Temperature drop
Vr Radial component of velocity
y Variable that signifies volume of polymer surrounding micro-
void = S3 – R3
Z Zeldovich non-equilibrium factor
b Thermal expansion coefficient
b* Rate at which gas molecules are added to the critical nucleus
g Rate of strain tensor
gg
˙˙ Average shear rate
h Polymer melt viscosity
ho * Modified zero shear viscosity
n Poisson ratio
q Wetting angle
s Surface tension of the liquid or polymer
t Induction period for establishing steady-state nucleation conditions
trr Stress in radial direction
t Stress in circumferential direction
t(1) Convective time derive of stress tensor
x Mean of the distribution of the void size

Superscript
* Critical point

© 2004 by CRC Press LLC

www.ketabdownload.com
 Subscripts
g Gas
hetero Heterogeneous
homo Homogeneous
o Initial
p Particle
pb Polybutadiene
ps Polystyrene
s Final or surounding
sat Saturation
surf Surface

References
1. Cole, R., Boiling nucleation, Advances in Heat Transfer, vol. 10, Academic Press,
New York, 1974.
2. Blander, M. and Katz, J.L., AIChE J., 21, 833–848, 1975.
3. Springer, G.S., Homogeneous nucleation, Advances in Heat Transfer, vol. 14,
Academic Press, New York, 1978.
4. Uhlmann, D.R. and Chalmers, B., Nucleation phenomena, ACS Symposium,
June 21–22, 1–12, 1965.
5. Becker, R., Ann. Physik, 32, 128, 1938.
6. Han, C.D. and Yang, H., J. Appl. Polym. Sci., 29, 4465–4470, 1984.
7. Colton, J. and Suh, N., Polym. Eng. Sci., 27, 485–492, 1987.
8. Cole, R., Bubble nucleation, growth and departure in boiling heat transfer,
Handbook of Heat and Mass Transfer, Gulf Publishing, Houston, TX, 1986.
9. Lee, S.T., Foam Extrusion: Principles and Practice, CRC Press, Boca Raton, FL,
108–109, 2000, chap. 4.
10. Ramesh, N.S., Campbell, G.A., and Rasmussen, D.H., Polym. Eng. Sci., 34,
1685–1706, 1994.
11. Han, C.D. and Han, J.H., J. Polym. Sci., Polym. Phys. Ed., 28, 743, 1990.
12. Boureau, R.J., Mod. Plast., 44, 133, 1967.
13. Bailey, F.E., Polymeric Foams, Hanser, New York, 1991.
14. Youn, J.R. and Byon, S.K., Polym. Eng. Sci., 30, 147, 1990.
15. Einstein, A., Ann. Physik, 38, 1275, 1910.
16. Volmer, M. and Weber, A., Z. Phys. Chem., 119, 277, 1926.
17. Onsager, L., Phys. Rev., 37, 405, 1931.
18. Becker, R. and Doring, W., Ann. Physik, 24, 719, 1935.
19. Turnbull, D. and Fisher, J.C., J. Chem. Phys., 17, 71–73, 1949.
20. Reiss, H., J. Chem. Phys., 18, 840, 1950.
21. Lothe, J. and Pound, G.M., J. Chem. Phys., 36, 2080–2085, 1962.
22. Gibbs, W., The Scientific Papers, vol. 1, Dover, New York, 1961.
23. Zettlemoyer, A.C., Nucleation, Marcel Dekker, New York, 1969.
24. Blander, M., Hengstenberg, D., and Katz, J.L., J. Phys. Chem., 75, 3613, 1971.

© 2004 by CRC Press LLC

www.ketabdownload.com
 25. Martini, J.E., The production and analysis of microcellular foam, S.M. thesis,
MIT, 1981.
26. Kumar, V., Process synthesis for manufacturing microcellular thermoplastic
parts: a case study in axiomatic design, Ph.D. thesis, MIT, 1988.
27. Ramesh, N.S., Investigation of the foaming characteristics of nucleation and
growth of microcellular foams in polystyrene containing low glass transaction
particles, Ph.D. thesis, Clarkson University, 1992.
28. Ruengphrathuengsuka, W., Ph.D. thesis, Texas A&M University, 1992.
29. Park, C.B., Behravesh, A.H., and Venter, R.D., SPE ANTEC, 2, 1958, 1998.
30. Ohshima, M. et al., Visual observations of batch and continuous foaming pro-
cesses, Foams, 141–151, 2002.
31. Timoshenko, S. and Goddier, J.N., Theory of Elasticity, McGraw-Hill, New York,
1970.
32. Beck, R.H. Gratch, S., Neuman, S., and Rusch, K.C., Polymer Melt, 6, 1998.
33. Bucknell, C.B., Toughened Plastics, Applied Science Publishers, London, 1977.
34. Schmitt, J.A., J. Appl. Polym. Sci., 12, 533, 1968.
35. Keskkula, H., Schwara, M., and Paul, D.R., Polymer, 27, 210, 1986.
36. Donald, A.M. and Kramer, E.J., Phil. Mag. A, 43, 873, 1981.
37. Adams, M.G., Cohen, A., and Campbell, G.A., Polym. Eng. Sci., 31, 1036, 1991.
38. Van Krevelen, D.W., Properties of Polymer, Elsevier, Amsterdam, 1976.
39. Kanner, B. and Decker, T.G., J. Cell. Plast., 5, 32, 1969.
40. Klemper, D. and Prisch, K.C., Polymeric Foams, Hanser, New York, 1991.
41. Harrington, R. and Hock, K., Flexible PU Foams, Dow Chemical, 1991.
42. Ramesh, N.S. and Malwitz, N., SPE ANTEC, 1941–1945, 1994.
43. Wang, S.L., Frisch, K.C., and Malwitz, N., 30th Annual Polyurethane Technical
Conference, 363–368, 1986.
44. Lord Rayleigh, Phil. Mag., 6th series, 34, 94, 1917.
45. Epstein, P.S. and Plesset, M.S., J. Chem. Phys., 18, 1505, 1950.
46. Scriven, L.E., Chem. Eng. Sci., 10, 1, 1959.
47. Street, J.R., Fricke, A.L., and Reiss, L.P., Ind. Eng. Chem. Fundam., 10, 54, 1971.
48. Han, C.D. and Yoo, J.J., Polym. Eng. Sci., 21, 518, 1981.
49. Upadhyay, R.K., Adv. Polym. Tech., 5, 1, 55, 1985.
50. Amon, M. and Denson, C.D., Polym. Eng. Sci., 24, 1026, 1984.
51. Amon, M. and Denson, C.D., Polym. Eng. Sci., 26, 255, 1986.
52. Arefmanesh, A. and Advani, S., Rheo. Acta, 30, 274, 1991.
53. Ramesh, N.S., Campbell, G.A., and Rasmussen, D.H., Polym. Eng. Sci., 31, 1657,
1991.
54. Ramesh, N.S., Lee, S.T., and Campbell, G.A., SPE ANTEC, 3033, 1993.
55. Ramesh, N.S. and Malwitz, N., SPE ANTEC, 1907, 1998.
56. Shafi, M.A., Joshi, K., and Flumerfelt, R.W., Chem. Eng. Sci., 52, 635, 1997.
57. Ohshima, M., Foam Extrusion: Principles and Practice, ed. S.T. Lee, CRC Press,
Boca Raton, FL, 2000, chap. 6.
58. Ramesh, N.S. and Malwitz, N., SPE ANTEC, 2171, 1995.
59. Lee, S.T. and Ramesh, N.S., Cellular and microcellular materials, ASME ‘96,
71–80, 1996.
60. Welty, J.R., Wicks, C.E., and Wilson, R.E., Fundamentals of Heat and Mass Transfer,
John Wiley & Sons, New York, 1969, 556–557.
61. Lee, S.T., Study of foam-enhanced devolatilization: experiments and theories,
Ph.D. thesis, Stevens Institute of Technology, 1986.

© 2004 by CRC Press LLC

www.ketabdownload.com
 62. Arefmanesh, A., Numerical and experimental study of bubble growth in highly
viscous fluids, Ph.D. thesis, University of Delaware, 1991.
63. Ramesh, N.S. and Lee, S.T., Foams, 99, 91, 1999.

© 2004 by CRC Press LLC

www.ketabdownload.com
 4
Material Properties Affecting
Extrusion Foaming

Q. Zhang and M. Xanthos

CONTENTS
4.1 Introduction
4.2 Importance of Material Properties in Extrusion Foaming
4.3 Physical Blowing Agents
4.3.1 Types and Properties
4.3.2 Blowing Agent Solubility
4.3.2.1 Main Factors Affecting Blowing Agent Solubility
4.3.2.2 Measurement of Blowing Agent Solubility
4.3.2.3 Gas Dissolution Inside Foaming Extruders
4.4 Resin Properties
4.4.1 Importance of Viscoelasticity
4.4.2 Modification of Resins to Enhance Foamability
4.5 Concluding Remarks
Acknowledgments
References

4.1 Introduction
Extrusion foaming for the production of medium- and low-density thermo-
plastic foams has been carried out successfully for some time with physical
blowing agents (PBAs). PBAs are atmospheric gases, volatile hydrocarbons,
or hydrochlorofluorocarbons (HCFCs) that are metered and dissolved in the
polymer melt during processing. The extrusion foaming process can use a
single extruder or two extruders operating in tandem. The basic steps in the
process involve (a) melting of the solid polymer; (b) injection and dissolution
of the blowing agent into the polymer melt, which is a diffusion controlled

0-8493-1728-2/04/$0.00+$1.50
© 2004 by CRC Press LLC

© 2004 by CRC Press LLC

www.ketabdownload.com
 Hopper Die
Drive
Melting Mixing Cooling
Section Section Section

Check Valve

Injection
Seal
Point
Section

Metering Pump Blowing Agent Supply

Single (Large L/D)

Gas
Hopper Injection
Drive
Port

Crossover

Die
Primary Extruder

Secondary Extruder

Tandem

FIGURE 4.1
Schematics of single barrel and tandem foam extruders.

process; (c) cooling of the blowing agent laden melt; (d) expansion through
the nucleation (often in the presence of nucleating agents) and growth of
bubbles due to super-saturation and diffusion of gas into the nucleated
bubbles (this is, generally, achieved by suddenly releasing the system pres-
sure); and (e) stabilization of the resultant cellular structure, which is accom-
plished by subsequent cooling and solidification of the cell walls.
The schematics of single barrel and tandem foam extruders are given in
Figure 4.1. In the single extruder case, high length to diameter ratio (L/D)
equipment is normally required to provide sufficient elements for polymer
melting, gas dissolution, cooling and pressurization, and foaming through the
die. In such a process, typically a two-stage screw is used. The polymer com-
position fed in the extruder feed section is melted and mixed in the transition
section of the screw and then the blowing agent is injected. The blowing agent
is dispersed and dissolved in the melt; this process depends on several factors
including the type of resin and blowing agent, temperature, pressure, and

© 2004 by CRC Press LLC

www.ketabdownload.com
 conditions such as screw configuration, screw speed, and so on. The gas-laden
melt is then pressurized and extruded through the die. The temperature profile
on the extruder is set in such a way that for the zones prior to the gas injection,
the temperature is high enough to ensure proper melting; for the zones past
the injection section, temperature is gradually lowered substantially so as to
allow for adequate cooling of the gas-laden melt prior to exiting the die, a
necessary condition for cell expansion without coalescence. It is believed that
bubble nucleation is heterogeneous and begins at the shaping die as the gas-
laden melt emerging from the die experiences a sudden pressure drop; this
thermodynamic instability causes phase separation. The escaping gas leads
to expansion within the fluid matrix in such a manner that individual minute
bubbles merge into cells and, through subsequent solidification, stable
expanded structures are produced.
In the tandem process, the polymer and additives are melted and mixed in
a primary extruder, which is typically equipped with a high shear mixing
screw. The blowing agent is injected near the end of this primary extruder.
The melt containing the blowing agent is then transferred to the feed section
of the secondary extruder through the so-called crossover section. The main
purpose of the secondary extruder is to cool the gas-laden melt to an optimum
temperature and pressure for foam formation without cell coalescence. The
tandem extruder setup normally allows for excellent control of process vari-
ables and is used in industry to produce very low-density foams.
The overall basic requirements for polymers, blowing agents, nucleators,
equipment, and processing parameters for making extruded polymeric
foams are shown schematically in Figure 4.2 (adapted from Khemani1). The
processes to be discussed in this chapter involve macrocellular foams (typ-
ically 100 mm or larger in cell size); we will not cover so-called microcellular

Polymer Blowing Agent Cell Nucleators Additives Extruder Die Screw

-Solubility -Cell size -Permeability

-Melt strength -Permeability & Type modifiers -Single/Tandem -Die Design -Screw design
-Viscosity diffusivity Distribution -Processing aids -Throughput Sheet Mixing
-Tm, TC, Tg -Inj. pressure Concentration -Flame retardants -L/D ratio Annular elements
-End groups -Point of injection -Temperature profile Arc Seal ring
-Concentration -Pressure -Temperature -Output
-Dispersion -Pressure

Foam Characteristics

Cell Structure (shape, size, distribution, open vs. closed)

Density and Gas Retention in Cells
Mechanical Properties (impact, tensile, bending)
Thermal Properties
Extractables (for food contact)

FIGURE 4.2
Parameters affecting the thermoplastics foam extrusion process.

© 2004 by CRC Press LLC

www.ketabdownload.com
 foaming where resins, presaturated with blowing agents such as CO2, are
heated at or above Tg and cooled rapidly to stabilize the cellular fine mor-
phology and prevent excessive cell growth. Such morphology is character-
ized by closed cells up to 25 mm in size and a cell density of 108 cell/cm3
and above.2 Continuous microcellular processes involving extrusion equip-
ment as described in various patents3 are under development,4 and some are
already commercially available.

4.2 Importance of Material Properties in Extrusion Foaming

The solubility and diffusivity of the gas in the melt and the viscoelastic
properties of the molten polymer are among the material parameters playing
an important role during the various stages of low-density extrusion foam-
ing. Material properties that influence these processes include melt shear
and elongational viscosity, melt strength, and, if the polymer is crystallizable,
the rate of crystallization. Gas/polymer interfacial properties also play a
significant role in all these stages of operation. To summarize pertinent
information from Throne:5

• PBAs need to be highly soluble (to alleviate the need for high extru-
sion and die pressures) and uniformly distributed in the polymer,
since undissolved microbubbles may promote premature nucleation
before the die. In addition, PBAs must diffuse at a high but controlled
rate to allow rapid cell inflation.
• Melts with high viscosity, particularly extensional viscosity, are
needed in the presence of dissolved gas which may act as a plasti-
cizer, reducing the overall viscosity.

• Melts with high elasticity and high normal stress differences are
needed during the extensional portion of bubble inflation, in addi-
tion to high resistance to stretching for stabilization of the growing
bubbles.

• At the end of bubble growth during the final biaxial stretching, high
melt elasticity to resist blow-out and moderately high elongational
viscosity (the latter increased by cooling) and rapid crystallization
are needed.

The first requirement for successful foaming is to form a uniform gas-laden

melt. This means that all gas molecules are dispersed on a molecular level
throughout the melt. The key here is the solubility of the gas in the polymer
at the melt temperature. If inert gases of low solubility such as nitrogen and

© 2004 by CRC Press LLC

www.ketabdownload.com
 argon are used as blowing agents to produce low-density foams of many
commercial polymers, then very high extrusion and die pressures would be
normally required.
The rheological properties of the gas-laden melt become very important;
for example, the viscosity of the mixture has to be high enough to achieve
the pressure required. The viscosity of the gas-laden melt is a strong function
of the gas concentration. It has been shown that at a given temperature, the
gas-laden polymer melt viscosity decreases exponentially with gas content
throughout the practical range of gas concentration for many polymer–gas
systems.6 The required high pressure is normally achieved by cooling the
polymer–gas mixture substantially before reaching the extrusion die. Tem-
peratures can be set to below the normal polymer melting point and near
the crystallization temperatures, in cases of semi-crystalline polymers, and
near Tg for amorphous polymers. This is possible because the inclusion of
the gas causes depression of the glass transition temperature (Tg) of an
amorphous polymer or the crystallization temperature (Tc) of a semicrystal-
line polymer. Depression of Tg , or Tc (the latter to a lesser extent) is a well-
known phenomenon and experimental results are available for a number of
polymer–gas systems.7–11
A variety of materials can be used during extrusion foaming, mostly to
control nucleation of bubbles and produce uniform cells of the desired den-
sity and dimensions. Fine inorganic minerals such as talc are commonly
used, although polymers immiscible in the foamable matrix may play the
same role. Glycerol stearate esters have also been used as additives in certain
polyolefin foams to control gas diffusion through the cells and to control the
“aging” time of the foams.

4.3 Physical Blowing Agents

4.3.1 Types and Properties

The optimal blowing agent must be environmentally acceptable and eco-
nomically viable, or readily available at low cost. It must also meet safety
regulations as required in storage, usage, and handling of both the blowing
agent and the foams made from it. The ideal blowing agent should be non-
toxic, non-flammable, and chemically and thermally stable. Other require-
ments include appropriate solubility in polymer melt, low vapor pressure at
room temperature, low boiling point, and slow diffusion through the poly-
mer compared to air.
Continuous extrusion foaming processes primarily used chlorofluorocar-
bons (CFCs) as blowing agents prior to their ban in mid-1990s. The release
of CFCs and other ozone-depleting substances (ODSs) had been identified
to be the main cause for rapid erosion of the stratospheric ozone layer, and

© 2004 by CRC Press LLC

www.ketabdownload.com
 the foam industry was one of the main industries that used and released
ozone-depleting chemicals. According to the U.S. Environmental Protection
Agency’s estimation,12 in the U.S., foam plastics accounted for approximately
18% of all U.S. consumption of ozone-depleting chemicals in 1990.
Following the ban on the use of CFCs for foam production in 1996, the
foam industry began to use other groups of chemicals. Currently, HCFCs,
hydrofluorocarbons (HFCs), and volatile organic compounds (VOCs), mostly
hydrocarbons, are the major chemicals being used as blowing agents. HCFCs
are also ozone-depleting chemicals and are used only as transitional alter-
natives to CFCs facing staged phase-out. HFCs are being used as chlorine-
free substitutes of HCFCs, and currently a number of HFC blowing agents
are commercially available. Examples include HFC-134a (CH2FCF3), HFC-
143a (CH3CF3), and HFC-152a (CHF2CH3). Hydrocarbons (such as butane
and pentane) are still very popular but highly flammable and can produce
explosive mixtures with oxygen despite sophisticated equipment and safety
measures.
In response to increasing regulation of halocarbons and increasing safety
and environmental concerns over the use of flammable, volatile hydrocar-
bons, recent research has focused towards replacing the traditional foaming
agents with various inert blowing agents, including atmospheric gases such
as CO2, N2, and Ar, but also water. Currently, CO2 is being commercially
used as a blowing agent in polyurethane foaming, along with CFC alterna-
tives such as HCFC and HFC. Some thermoplastic foaming processes using
CO213 and other atmospheric gases as blowing agents have also been devel-
oped, while most thermoplastic foams are still being manufactured using
HCFCs, HFCs, and hydrocarbons as blowing agents. The use of inert gases
in foam extrusion has been shown to demand a very narrow window of
processing conditions as opposed to the use of halocarbons.14 Inert gases are
much less soluble in the polymer melt and have much higher vapor pressures
than CFCs or hydrocarbons.
A comparison of some of the properties of different blowing agents is given
in Table 4.1. Comparison of solubility data of R-22 CFC and 142-B HCFC
versus CO2 in polystyrene shows higher values at a magnitude of four times
or more.15 The effects of the lower CO2 solubility coupled with its much
higher vapor pressure and permeability versus conventional blowing agents
are reflected in almost all stages of the foaming process, including gas dis-

TABLE 4.1
Comparison of PBA Properties in Polystyrene15
Blowing Agent Vapor Pressure Solubility at 24˚C Permeability
(psia) (pph/atm) (Dow units)
CFC (R-22) 151 1.60 19
HCFC 142-B 49 6.25 0.21
CO2 930 0.40 1430

© 2004 by CRC Press LLC

www.ketabdownload.com
 solution, bubble nucleation and growth, and cell structure stabilization. In
efforts to produce zero ozone-depleting-potential (ODP) polystyrene formu-
lations, the issues in replacing HCFCs with CO2 on an industrial scale have
been addressed through rheological modification of the resin, the use of
processing aids, the use of PBA mixtures, and modification of machine and
process conditions.15

4.3.2 Blowing Agent Solubility

A very important parameter in the production of these multicomponent,

multiphase systems is the solubility of physical blowing agents in the pres-
surized polymer melt, as affected by concentration, pressure, and tempera-
ture. Knowledge of the solubility characteristics of the gas in polymer melts
is highly desirable in order to estimate the pressure requirements to keep it
in the melt throughout the extruder all the way to the die, thus avoiding the
formation of premature bubbles which would result in uneven cell growth.
For molten polymers, the solubility of a blowing agent dictates the condi-
tions under which bubbles will start to nucleate and grow. After the foaming
process, the long-term characteristics of the foam are still influenced by the
solubility and diffusivity of the blowing agents. Diffusion of the blowing
agent out of the cells and the polymer matrix will change the thermal and
mechanical properties of the foam. This is especially the case when the
blowing agent plasticizes the polymer and the material stiffens as the blow-
ing agent diffuses out.16

4.3.2.1 Main Factors Affecting Blowing Agent Solubility

It is generally accepted that solubility is dependent on the free volume
available to the gas molecules in the polymer, and therefore dependent on
temperature and pressure. Free volume increases with temperature; there-
fore, solubility should increase with temperature. For many cases, however,
interactions between gases and polymers cause solubility to decrease with
increasing temperatures. For most gases, it turns out that more easily con-
densed gases have higher solubility.
Solubility is usually expressed through Henry’s law17 as:

X = Kp Pg (4.1)

where X is the concentration of gas in the melt (cm3 (STP)/g), Pg is partial

pressure of dissolved gas (kPa), and KP is the Henry’s law constant (cm3
(STP)/g∑kPa). For a number of gases, the logarithm of the experimental
Henry’s law constant has been shown to correlate linearly with (Tc/T)2,
where Tc is the critical temperature of the gas.18,19
Features in the molecular structure of polymers which disrupt chain
packing increase their free volumes. Therefore, polymers with branched

© 2004 by CRC Press LLC

www.ketabdownload.com
 structures or bulky pendant groups in their repeat unit have increased
solubility, relative to polymers with more linear structures. Specific inter-
actions such as hydrogen bonds between the gas molecules and the poly-
mer repeat structure can also promote solubility.
Experimental data reported by Durrill and Griskey20,21 indicate that
Henry’s law constants rank as N2<CO2<ClF3C for each polymer from the
group of polyethylene (PE), polypropylene (PP), polyisobutylene (PIB), poly-
styrene (PS), and polymethyl methacrylate (PMMA). Thus, for these systems
comprising nonpolar or low polarity polymers, solubility was primarily
dependent on the type of gas rather than the type of polymer, with CFC
showing much higher solubility than atmospheric gases. From the above
data, other researchers22 working on microcellular foaming estimated very
similar solubilities of CO2 at 200∞C and 27.6 MPa in various polymers, such
as PP and PS (12%) and PMMA (15%). The heat of the solution was found
to be positive for nitrogen in PE but negative for CO2 in PE, PP, and PS.18,20
For the PS/CO2 system, recent data indicate good agreement between theory
and experiment with correct prediction of decreasing solubility with increas-
ing temperature.23
For amorphous polymers without strong polar groups, the linear relation
between log-solubility and critical temperature or boiling point of gas18 may be
used as a first approximation. For polybutadiene (PB), for example, solubility
of gases decreases with decreasing critical temperature in the order
CO2>CH4>Ar>N2,19 with a positive temperature dependence for N2 and a neg-
ative dependence for CO2, CH4, and Ar. However, for amorphous polymers with
strong polar groups, significant effects of the polymer polarity on gas solubility
have been observed.24,25 For example, the solubility of CO2 at 25∞C in butadiene/
acrylonitrile rubber increases with an increasing content of acrylonitrile polar
groups in the polymer, whereas that of nitrogen, hydrogen, and oxygen (being
already much lower than CO2 by an order of magnitude) decreases.24 Equilib-
rium uptakes of CO2 at room temperature range from 3% in PE to 27% in
PMMA to over 50% in PVAc. Equilibrium sorption of CO2 in styrene/acry-
lonitrile copolymers increases with AN content and in ethylene/vinyl acetate
with increasing vinyl acetate monomer. For a variety of polar polymers,
solubility of CO2 ranks as EVA>PMMA>PC>PVC.25 The relatively high sorp-
tion values of CO2 in polar polymers may result from enhanced specific
interactions of CO2 with carbonyl or nitrile groups; these could be related to
its quadrupole moment or its H-bonding basicity.25 Recent work26–29 on melt
mixed PS/PMMA blends demonstrated that the solubility of CO2 in the
PMMA phase is, indeed, significantly higher than that in PS. Higher affinity
with the carbonyl group of the PMMA polymer implies also higher plasti-
cization capacity, which is manifested in lower Tg8 and lower viscosity.

4.3.2.2 Measurement of Blowing Agent Solubility

Given the importance of blowing agent solubility, recent experimental efforts

have focused on the measurement of solubility of gases in molten polymers

© 2004 by CRC Press LLC

www.ketabdownload.com
 at high temperatures and pressures. Efforts have also been made to model gas
solubility using, for example, equations of state (EOS). The Simha–Somcynsky
theory has been used to model gas solubility in polymer melts and has found
success in predicting gas solubility characteristics for a number of polymer–gas
systems.30
Experimental data on gas solubility in polymer melts have been reported
by many investigators.19,20,31–34 These investigators used either the pressure
decay method or a high-pressure vessel batch process to determine thermo-
dynamic equilibrium data. More recently, a gravimetric method employing
an electrobalance was also used.35,36
On-line/in-line methods have drawn attention in recent years, since they
allow measurements under dynamic conditions. Dual transmission infrared
sensors have been used for on-line monitoring of foaming processes, the sen-
sors being linked with fiber-optic cables to a Fourier Transform Infrared (FTIR)
spectrophotometer which records spectra of the melt in the near-infrared
region. Infrared probes were mounted on a slit die to monitor the polymer–gas
mixtures during the extrusion of foams.16,37 This technique was also used to
detect degassing in the melt. Ultrasonic sensors were also used in the same
slit die for measuring the phase behavior of molten polymers.38 Tsujimura et
al.39 studied the dynamic behavior of bubble nucleation in a slit die with quartz
glass windows using a long-distance microscope and a high-speed video sys-
tem. The effect of nucleating agents on bubble nucleation dynamics for a PP-
butane system was also studied using this method. An in-line method has
been developed in the laboratory of the authors which is capable of measuring
PBA solubility in polymer melts under extrusion conditions. This method
combines an optical flow cell with a foaming extruder (single or twin-screw)
and generates apparent solubility data by observing the onset of bubble for-
mation and dissolution in the melt, via a microscope-CCD (charged coupling
device) camera-monitor-recorder system. The schematic of the experimental
setup is shown in Figure 4.3. The onset of gas bubble presence or absence in
the cell was monitored and recorded using an optical system consisting of a
microscope, a CCD camera, a monitor (TV), and a recorder (VCR) (Figure 4.4).
The important parameters, such as melt pressure and temperature at the die
and the window, gas injection pressure, and flow rate, were digitized and
recorded in a computer for real-time monitoring and off-line analysis.
At the beginning of each experiment, the window pressure is low, and the
gas laden melt passing through the window is a two-phase system with
many large bubbles dispersed in the continuous melt phase, as shown in
Figure 4.5A. The window pressure is then slowly increased by adjusting
(closing) the restricting valve. The number and size of the bubbles decrease
with increasing melt pressure as indicated in Figure 4.5B. When the pressure
is high enough, all bubbles will disappear and the system will appear as a
single-phase system. Figure 4.5C shows this tendency; this pressure is con-
sidered to be the lowest pressure required to keep the gas in solution under
the specified conditions. This method generates apparent solubility data that
could be used directly to characterize and control a foaming process and

© 2004 by CRC Press LLC

www.ketabdownload.com
 Solid Metering Feeder
Gas
Cylinder F

Manual Pressure P
P
Control Valve

Microscope
Extruder
T P

A/D
Converter
Video Camera

Serial
Interface

TV
Personal
Computer

Video Signal
VCR

FIGURE 4.3
Schematic of the experimental setup to measure the dynamic solubility of blowing agents.

Extrudate
TV
Valve

CCD Flow Light

VCR Microscope
Camera Source
Cell

Traveling Device Extruder

FIGURE 4.4
Schematic of the flow cell—monitoring/recording system used to measure the dynamic solu-
bility of blowing agents.

© 2004 by CRC Press LLC

www.ketabdownload.com
 A.

B.

C.

FIGURE 4.5
Monitoring of CO2 bubbles in PS melt at 215˚C. (Original magnification ¥ 60.) (A) Low pressure.
(B) Pressure close to the targeted pressure. (C) High pressure.

serve as guidance in blowing agent selection and process design and opti-
mization. Unlike traditional methods that generally require long hours for
equilibrium, the apparent solubility measurements are carried out within the
single or tandem extruder residence times (typically 5–20 min); thus, they
are very useful in solubility measurements for temperature-sensitive poly-
mers. They can be used to study effects of processing conditions on gas
solubility characteristics and can also be potentially useful for studying the
effects of polymer modification on solubility if used with a reactive extruder.

© 2004 by CRC Press LLC

www.ketabdownload.com
 16

14

12

S (cm3[STP]/g-melt)
10

0
0 5000
Melt Pressure (kPa)

FIGURE 4.6
Solubility vs. melt pressure of CO2 in PS melt at 215˚C.

TABLE 4.2
Comparison of Experimental Results with Literature Data
Temperature H/Slope Pressure Range
(˚C) cm3 [STP]/g∑kPa (kPa)
In-line data 190 0.0037 Up to 5575
Sorption experiment results (20) 188 0.0022 Up to 2030
Data from low temperature 180 0.0020 Up to 17000
experiments (19)

Solubility data generated for CO2 in polystyrene at 190oC are presented in

Figure 4.6. These results are compared with literature values obtained from
sorption experiments as shown in Table 4.2.
Temperature dependence of gas solubility in PS was also studied. Some
results of CO2 and nitrogen are presented in Figure 4.7 and Figure 4.8. These
results show that for carbon dioxide, the solubility decreases with increasing
temperature. This tendency has been generally observed in many gas–poly-
mer and solute–solvent systems.17,19,20 For nitrogen, however, the solubility
increases with increasing temperature. This tendency called reverse solubil-
ity has been observed for gases of low critical temperatures such as helium,
hydrogen, nitrogen, and oxygen.19
Solubilities of three different atmospheric gases (CO2, nitrogen, and argon)
in polyethylene terephthalate (PET) were also obtained. The results given in
Figure 4.9 indicate that carbon dioxide has the highest solubility in PET,
nitrogen has the lowest, and Ar is somewhere in between. This trend is
consistent with literature results obtained for the same systems under tem-
peratures below the polymer Tg.32

© 2004 by CRC Press LLC

www.ketabdownload.com
 FIGURE 4.7
Solubility vs. melt pressure of CO2 in PS melt at two different temperatures.

FIGURE 4.8
Solubility vs. melt pressure of N2 in PS melt at two different temperatures.

4.3.2.3 Gas Dissolution Inside Foaming Extruders

Dissolution of physical blowing agents is affected by solubility, diffusivity,
pressure, mixing, and residence time. Shear forces and interfacial tension
may also play a significant role in affecting the gas dissolution process.
Considerable research has been conducted on processes that involve the
formation and growth of gas bubbles in polymer solutions or melts during
foam formation or devolatilization of polymer solutions. Studies on gas
dissolution processes in foaming extruders have been largely ignored, likely
due to the fact that when traditional blowing agents (CFCs, HCFCs, and

© 2004 by CRC Press LLC

www.ketabdownload.com
 10
CO2
N2
8
Ar

S {cm3[STP]/g-melt}
6

0
0 2000 4000 6000
Melt Pressure (kPa)

FIGURE 4.9
Solubility vs. melt pressure of inert gases in PET at 290˚C.

hydrocarbons) are used, gas dissolution has never been a major concern
thanks to the high solubility of those agents. However, when inert gases such
as CO2 and N2, which have limited solubility in polymer melts, are used as
blowing agents, solubility and dissolution become major factors in deter-
mining the process efficiency, feasibility, and product quality, since in these
cases the system needs to operate at much higher pressures.
Solubility results obtained from the in-line method for a PS–CO2 system
along with gas injection conditions were used to study the gas dissolution
behavior over the gas injection section of the foaming extruder, in an effort
to elucidate the mechanisms involved in this complex process. In this case,
the extruder used was a 34-mm intermeshing, corotating twin-screw
extruder (Leistritz LSM34) having a L/D of 36.
Generally speaking, gas dissolution in the injection zone is dictated by
solubility and diffusivity of the PBA and polymer/gas flow dynamics. The
main mechanism for gas-melt mass transfer in a partially filled extruder is
diffusion through the surface areas provided by the rotating melt pool and
the barrel melt film40 as shown schematically in Figure 4.10. In extruders the
pool rotates in the partially filled channel under the influence of the cross
channel component of the screw velocity, while being conveyed through the
extruder in drag flow by the down channel component. A stationary melt
film is deposited on the barrel wall due to the clearance between screw and
barrel. The screw rotation causes continuous generation of the free surface
of the rolling pool, significantly enhancing gas diffusion. Other possible
mechanisms have also been speculated for some time. Some researchers41
believe that some degree of foaming or frothing happens inside the injection
zone, which would increase the mass transfer area and, hence, enhance the
rate of gas dissolution.

© 2004 by CRC Press LLC

www.ketabdownload.com
 The actual amount of gas dissolved in the melt as measured downstream
at the window in the twin-screw extruder is compared with the maximum
possible amount that could be dissolved in the gas injection zone under a
particular injection pressure. The latter is calculated using the gas injection
pressure and assuming equilibrium is reached in the short gas injection zone.
Comparative data for CO2 in the PS melt under three different injection
pressures are given in Figure 4.11. The results show that the actual total
amount of gas in the polymer melt is more than that calculated from equi-
librium under the corresponding injection pressure. Given the short resi-
dence time and the relatively low diffusivity values of CO2 , on the order of
10–6 cm2/sec,42 it is much less likely that the diffusion process could reach
equilibrium in the short gas injection section of the screw. Calculations based
on mechanism shown in Figure 4.10 and using typical solubility and diffu-
sivity values have shown that under the applied experimental conditions,
the system can only reach about 40–60% of the equilibrium values.43 This
indicates that the gas is carried downstream, possibly in the form of foam,
and gets dissolved somewhere downstream under increased pressure over
a longer residence time. Todd observed the formation of bubbles in twin-
screw extruders when small amounts of air or water were injected into the
melt.41 Some amount of gas may be trapped into the melt by the circular
flow of the rotating melt pool, and this amount cannot readily dissolve into
the melt due to limitations of mass transfer rates, limited solubility and
diffusivity, and short contacting time. Thus, it can be envisioned that this
trapped gas serves as the source for bubble nucleation and growth in the
gas injection zone inside the extruder, and the extent of foaming continues
to grow as the melt advances downstream. Data from Figure 4.11 show that
the “carryover” or entrainment is quite significant and may play an impor-
tant role during the foaming process.

Barrel

Barrel film
Gas
δ
Vs

Rotating pool
Screw

FIGURE 4.10
Schematic of the mass transfer areas in a particular filled screw channel (d = flight clearance,
VS = melt circular flow velocity).

© 2004 by CRC Press LLC

www.ketabdownload.com
 0.015
Exp. total gas in solution
Calc. from Henry’s law

0.010
Gas weight fraction

0.005

0.000
758 1161 1441 1718
Gas injection pressure (kPa)

FIGURE 4.11
The amount of CO2 dissolved in PS (at 215˚C in twin-screw extruder) vs. that calculated from
equilibrium.

The effects of the extrusion process parameters on solubility characteristics

were also studied using the same in-line method. Detailed results can be
found in Zhang et al.44

4.4 Resin Properties

4.4.1 Importance of Viscoelasticity

During the cell growth phase of extrusion foaming, the polymeric melt
undergoes intense elongational deformations. Rheologically inadequate res-
ins cannot sustain these deformations and their foaming process is charac-
terized by unstable cell growth, nonuniform cell size, and membrane tearing
(blowout) leading to cell collapse and open cells. Many commercially avail-
able low molecular weight (MW) resins including PP, PET, and high-density
polyethylene (HDPE), usually having narrow molecular weight distribution
(MWD), exhibit inadequate rheology for low density foaming and a very
narrow process foaming window. Such resins are characterized rheologically
by low values of melt viscosity, particularly elongational viscosity that is not

© 2004 by CRC Press LLC

www.ketabdownload.com
 dependent on increasing stress or time; poor shear sensitivity in viscosity
curves; and low “melt elasticity” expressed with criteria such as melt strength
and extrudate swell.45
Effects related to melt rheology and melt elasticity of thermoplastics can, in
general, be controlled through additives or through changes in MW and MWD
during reactor or post-reactor processing by chain extension, grafting and
branching, cross-linking, or controlled degradation. “Melt elasticity” changes
can be related to changes in the values of viscoelastic functions such as normal
stress difference, storage modulus, extensional viscosity, or recoverable shear
strain; and also to changes in parameters of practical importance such as
extrudate swell and melt strength.46
For example, foamable polypropylene has higher “melt elasticity,” melt
strength, and extensional viscosity with strain hardening behavior than con-
ventional resins of the same melt flow index (MFI) (Figures 4.12 and 4.13);
foamable PPs may also have higher Mz and high Mz/Mw , and possibly bimodal
MW distribution with the higher MW component highly branched.47 Strain
hardening behavior, defined as the rapid increase of the elongational viscosity
with time or strain under a constant strain rate, is also required for other
processes such as thermoforming or blow molding. Particularly, high melt
strength or elasticity is an important parameter usually associated with
increases in the average relaxation time and broadening of relaxation time
distribution as a result of long chain branching. Most polymers with long chain
branches exhibit marked strain hardening.48,49 Such behavior was found in
bimodal resins or resins of high polydispersity also having higher melt strength
and higher extrudate swell. Improved melt elasticity may also be correlated
with improved foamability of PET resins.50 Long chain branching has also been

3.5 MELT FLOW CONVENTIONAL POLYPROPYLENE

1E5
GROWTH FUNCTION (Pa*SEC)
EXTENSIONAL VISCOSITY

1E4 STRAIN RATE

0.51 1/SEC
0.10
0.51
1.0
2.0
TEMPERATURE = 180°C
1000
0.100 1.000 10.000 100.000
TIME (SEC)

FIGURE 4.12
Extensional viscosity of linear PP.47

© 2004 by CRC Press LLC

www.ketabdownload.com
 1E6
STRAIN RATE
0.05 1/SEC

GROWTH FUNCTION (Pa*SEC)

EXTENSIONAL VISCOSITY
0.1
0.51
1E5 1.0
2.0

1E4

TEMPERATURE = 180°C
1000
0.100 1.000 10.000 100.000

FIGURE 4.13
Extensional viscosity of high melt strength PP. 47

suggested to enhance the strain hardening behavior of the elongational viscos-

ity because of the depression of chain contraction, which has been well recog-
nized by theoretical approaches based on the Doi–Edwards “tube” model.51–53
Yilmazer et al. studied the chain extension and branching of PET and
reported that modified PET was characterized by low MFI, high extrudate
or die swell, high viscosity, high shear sensitivity, high storage modulus, and
pronounced non-Newtonian behavior54 (Figure 4.14). The modified resins
were suggested to have higher mean relaxation times and relaxation strength
values than the unmodified ones.
Yamaguchi and Suzuki55 studied the rheological properties and foam pro-
cessability of blends of linear and crosslinked polyethylenes. They found

100000

10000
modulus (Pa)

1000

100 G′ for unfoamable PET

G′′ for unfoamable PET
10 G′ for foamable PET
G′′ for foamable PET
1
0.1 1 10 100
ω (rad/s)

FIGURE 4.14
Comparison of G¢ and G≤ at 290˚C of unmodified (unfoamable) and modified (foamable) PET
resins.

© 2004 by CRC Press LLC

www.ketabdownload.com
 PP
107 PP/gEHDM (97/3)
107
463 K 0.013
463 K
0.078
106
106

ηE+ (t, ε)/Pa s

ηE+ (t, ε)/Pa s

0.34
. .
0.067 0.015

105 3η+ (t)

0.034 105
0.14

3η+ (t)

104 104
10−1 100 101 102 103 10−1 100 101 102 103
t/s t /s

FIGURE 4.15
Time variation of elongational viscosity for PP and foamable PP/crosslinked elastomer blend.55

that blending a small amount of crosslinked LLDPE enhanced strain hard-

ening of the elongational viscosity (Figure 4.15) and enhanced extrudate/
die swell and elastic modulus, whereas it had little effect on the steady-state
shear viscosity. The foams produced from these blends showed higher
expansion ratio and more homogeneous cell size distribution. According to
the studies by Yamaguchi et al., the conditions under which a small amount
of crosslinked polymer enhances the strain-hardening in elongational vis-
cosity of a polymer are (1) the crosslinked polymer has a network chain with
a low density of crosslink points; and (2) the precursor of the crosslinked
polymer is miscible with the linear polymer.56–58 The origin of the resulting
rheological properties has been suggested to be the result of the chain stretch-
ing of network polymers via trapped entanglements with a linear polymer.
Recent work has suggested that the presence of gel-like structures in the
amount of at least 70% is required for foaming resins, with the resultant
foams having minimal shrinkage.56,58

4.4.2 Modification of Resins to Enhance Foamability

Physical entanglements and entanglements originating from aromatic ring
interactions are believed to be responsible for the foamability of LDPE and
PS, respectively. Contrary to the case for LDPE, the role of strain hardening
has not been clearly defined for PS. Both resins are easy to foam and do
not require any further modification.6 However, for other resins such as PP,
HDPE, and PET, long-chain branching modification in the polymerization
reactor or by post-reactor reactive extrusion has been used to enhance
foamability and broaden the processing window during extrusion foaming;

© 2004 by CRC Press LLC

www.ketabdownload.com
 foamability in such cases is usually related to enhanced melt strength and
strain hardening or extrudate swell. Other methods proposed in the litera-
ture to enhance foamability include:

• Introduction of bimodality as with a high MW component in a

bimodal MWD PS resin
• Nonreactive blending of immiscible polymers, e.g., LLDPE + LDPE
• Reactive processing, e.g., grafting of styrene on PP; compatibilization
of PP/PET blends through functionalized polyolefins and coagents;
incorporation of a dispersed crosslinked polymeric phase in a poly-
meric matrix as, for example, in certain thermoplastic vulcanizates;
PP blends with crosslinked diene based elastomers; LLDPE contain-
ing small amounts of cross-linked LDPE; PP + crosslinked EHDM

Specific methods that have been used to improve the foamability of PP are
• Long chain branching through low temperature free radical modi-
fication with radiation or peroxides in the absence of oxygen
• Controlled crosslinking, grafting with multifunctional monomers,
coagents, vinyl silanes, or styrene (by irradiation or peroxides)
Modified PP resins are characterized by high melt strength and drawability
in a Rheotens apparatus and can be foamed to very low densities (0.05 g/cc).59,60
A significant body of literature exists describing the post-reactor modifi-
cation of linear PET resins to increase MW and broaden the MWD for
foaming and extrusion blow molding applications.61 Modification is carried
out through the use of suitable multifunctional chain extenders with anhy-
dride or glycidyl functionality which are able to react with the polymer end
groups during solid stating or, alternatively, when added during a reactive
extrusion process. Reactions of polyester end groups with anhydride and
epoxy functionalities are shown in Figures 4.16 and 4.17, respectively. The
production of extrusion foamable polyesters modified by solid state poly-
addition with a premixed multifunctional modifier, e.g., pyromellitic dian-
hydride (PMDA) is discussed in a series of patents and publications.62–64
These resins are distinguished from non-foamable resins by increased MW,
as measured by intrinsic viscosity I.V. (>0.9 dl/g), increased melt strength,
high extrudate swell, increased complex viscosity with higher shear sensi-
tivity, and increase in storage modulus. Post-reactor modification by reactive
extrusion of PET carried out using PMDA or other branching additives also
results in significant increase in zero shear viscosity, increase in melt strength
and die swell, increase in molecular weight, and increases in the polydis-
persity ratio Mz/Mn from about 4.5 up to 11. Property changes depend on
the choice of process conditions and additive concentration and its type of
carrier.65,66 Foams with densities of 0.2 g/ml were obtained by injecting
hydrogenated chlorofluorocarbons in the primary extruder of a tandem
line.65 In another patent on PET foaming,66 isopentane was injected into the

© 2004 by CRC Press LLC

www.ketabdownload.com
 n
+

PET

O O
FIGURE 4.16
Esterification of PET with PMDA.64

molten blend of PET and PMDA to produce rod-shaped foams. The expan-
sion of the extrudate was proportional to the amount of PMDA added (up
to 0.4% by weight); the density values for the PMDA modified foams ranged
from 0.35 to 0.13 g/ml vs. 0.7 g/ml when no PMDA was used.
In a recent publication65 it was shown that production of low-density PET
foams by one-step reactive modification/gas injection foaming is feasible at
process conditions not significantly different than those employed in the

Etherification

FIGURE 4.17
PET end group reactions with epoxides.67

© 2004 by CRC Press LLC

www.ketabdownload.com
 simple PET reactive modification with low MW additives. In initial experi-
ments PET was chain extended with a dianhydride, benzophenone tetracar-
boxylic anhydride (BTDA, solid, tetrafunctional), and a triepoxide,
triglycidyl isocyanurate (TGIC, solid, trifunctional), in a 40 : 1 L : D SSE. It
was assumed that only initial OH groups react with dianhydrides and only
initial COOH groups react with epoxides at the process conditions. Table 4.3
shows that changes in the reactive extrusion related parameters (die pressure,
extrudate swell) and product characteristics (melt flow index and melt
strength) as a result of increasing MW and MWD depend on the type of the
modifier. A one-step reactive extrusion/foaming with 2–3 phr isobutane
injected at 19 L : D length was carried out with the PET and the PET/BTDA
system under the same operating conditions (Figure 4.18). The unmodified
PET showed 120% expansion with only a few irregular cells; by contrast the
BTDA modified extrudate showed 540% expansion, low density of 0.13 g/
cc, and uniform cells. Enhanced foamability of PET may be obtained through
the addition of a lower melting temperature polymer such as PP. In this case
it has been shown that proper compatibilization of the immiscible blend
through a functionalized polyolefin and a coagent is important to produce
foams with lower density and smaller cell size (Figure 4.19).67
With respect to reactive blending as a means of enhancing foamability, low
density and uniform cell structure foams were successfully produced from
blends containing LLDPE and peroxide crosslinked LDPE (xLDPE) in a batch
foaming process.68 It was shown that high loadings (50%) of xLDPE were
necessary to bring up properties (including melt strength) to foamable levels
(Figure 4.20). Blending a small amount (10 wt%) of xLDPE also resulted in
significant improvements in LLDPE foam extrusion processability and prod-
uct quality; foams had larger expansion, uniform cell structure, and ~30%
density reduction versus the unmodified LLDPE. Blending recycled xLDPE
foam with LLDPE was also effective in producing foamable resins, opening
new applications for recycling xLDPE foams.

TABLE 4.3
PET Extrusion Characteristics with Different Modifiers (Solid Extrudates)67
Additive Die Die Extrudate Average Melt Melt Melt
Pressure Tempe- Appea- Die Flow Index Flow Strength
(Mpa) rature rance Swell (g/10 min), Index (10–3 N)
(˚C) (Std. Dev.) Initial (g/10 min),
Reading Average
No addi- 0.96 270.6 Smooth, 0.82 >200 >200 Not
tive soupy (0.04) measu-
rable
BTDA 3.38 274.1 Smooth 2.65 5.18 6.69 13.0
(0.19)
TGIC 15.20 279.6 Wavy, 3.07 1.06 1.09 332.0
variable (0.11)
diameter

© 2004 by CRC Press LLC

www.ketabdownload.com
 F

Solid Metering Feeder

(PET + TGIC dry-blend) P Gas
Cylinder

Gas P
Injection
Point
Motor Extruder Barrel
P P

T
PC

FIGURE 4.18
One step reactive extrusion/foaming of PET.

0.3

0.25
Foam Density (g/cm3)

30% Density
4% Density Reduction
0.2 Reduction

0.15

0.1

0.05

0
No Comp. 5% Comp. 5% Comp. +
Coagent
Percent Additive

FIGURE 4.19
Foam density of 80 : 20 (PET : PP) blend as a function of additive content.

4.5 Concluding Remarks

The objective of this chapter was to summarize the importance of certain
properties of physical blowing agents and polymeric resins that are used to
produce low density foams by PBA injection in extruders. Emphasis was
placed on (a) the importance of solubility of the PBA in the polymer melt,
and (b) the viscoelastic characteristics of the polymer melt.

© 2004 by CRC Press LLC

www.ketabdownload.com
 25

20

15
Force (N)

10

0
0 10 20 30 40 50 60
% xLDPE

FIGURE 4.20
LLDPE resin property enhancement—melt strength vs. added xLDPE (170˚C; die L/D = 10;
draw ratio = 100).68

With respect to current efforts in the field of PBAs, activities mostly focus
on the commercialization of foams containing alternative environmentally
friendly PBAs that are nonflammable and nontoxic. The best examples are
atmospheric gases, particularly supercritical gases such as CO2, which has
already been used with polystyrene, and water, which has been used in
certain thermoplastic vulcanizates. Commercialization of such foams may
require modification of extrusion process conditions as well as changes or
adjustments in the polymer formulation to accommodate the lower solubility
and different permeability of the alternatives versus conventional PBAs. The
alternative PBAs are often used in combination with other low boiling point
liquids (e.g., alcohols) to balance solubility and diffusivity requirements. As
a result of such activities, the need has emerged to measure accurately
solubilities of the PBAs in the melt under extrusion conditions and the
concomitant effects on the viscosity of the gas laden polymer. Such data are
necessary to assist the process and die design requirements; this has lead to
the development of a variety of in-line/on-line instrumentation. Adoption
of alternative PBAs has presented more challenges in the case of multiphase,
multicomponent systems (immiscible polymer blends, thermoplastic elas-
tomers, composites) where solubilities and diffusivities of the PBA in the
various phases may differ.
With respect to efforts in the field of enhancing the foamability of a variety
of resins, the recognition of the importance of structural modification to
enhance melt strength has led to the commercial availability of several resins,
such as PP, PET, and miscellaneous polyolefins that are now foamable to low
densities. Modification can be accomplished in the polymerization reactor
by irradiation, by reactive extrusion, or through polymer/polymer blending.
In addition to the production of conventional closed-cell foam structures,

© 2004 by CRC Press LLC

www.ketabdownload.com
 such techniques are currently under evaluation for the production of open
cell structures with high fluid absorption characteristics.

Acknowledgments
The authors would like to express their appreciation to Dr. S.K. Dey of the
Polymer Processing Institute for his contributions to the measurement of the
“dynamic” solubility of the blowing agents.

References
1. Khemani, K.C., Extruded polyester foams, in Polymeric Foams, ed. K.C. Khe-
mani, American Chemical Society, Washington, D.C., 54–80 1997.
2. Handa, Y.P. and Zhang, Z., New pathways to microcellular and ultramicrocel-
lular polymeric foams, Symposium on Porous, Cellular and Microcellular Ma-
terials, ASME Intern. Mech. Eng. Congr. Exp., 82, 1, 1998.
3. Martini-Vvedensky, J., Suh, N.P., and Waldman, F.A., U.S. Patent 4,473,665,
1984.
4. Trexel Inc., Technical literature on MuCell process, Woburn, MA.
5. Throne, J.L., Proceedings of Foam Conference 96, LCM Public Relations, Somerset,
NJ, 2, 1996.
6. Gendron, R. and Daigneault, L.E., Rheology of thermoplastic foam, in Foam
Extrusion: Principles and Practice, ed. S.T. Lee, CRC Press, Boca Raton, FL, 2000.
7. Wang, W.C., Kramer, E.J., and Sachse, W.H., Effects of high-pressure CO2 on
the glass transition temperature and mechanical properties of polystyrene, J.
Polym. Sci., Polym. Phys. Ed., 20, 1371, 1982.
8. Wissinger, R.G. and Paulaitis, M.E., Glass transition in polymer/CO2 mixtures
at elevated pressures, J. Polym. Sci., Part B: Polym. Phys., 29, 631, 1991.
9. Dey, S.K., Jacob, C., and Biesenberger, J.A., Effect of physical blowing agents
on crystallization temperature of polymer melts, Proceedings of 1994 SPE AN-
TEC, San Francisco, CA, 2197, 1994.
10. Handa, Y.P., Kruus, P., and O’Neill, M., High-pressure calorimetric study of
plasticization of poly(methyl methacrylate) by methane, ethylene, and carbon
dioxide, J. Polym. Sci., Part B: Polym. Phys., 34, 2635, 1996.
11. Zhang, Z. and Handa, P.Y., CO2-assisted melting of semicrystalline polymers,
Macromolecules, 30, 8505, 1997.
12. Anon., Acceptable exposure limit lowered, Health Hazard Alert, Office of
Health, U.S. Department of Energy, Washington, D.C., 92–1, September 1992.
13. Gabriele, M.C., New foam extrusion technology is “environmentally friendly,”
Plast. World, 42, 30, 1996.
14. Jacob, C. and Dey, S.K., Inert gases as alternative blowing agents for extruded
low density polystyene foam, J. Cell. Plast., 31, 38, 1995.

© 2004 by CRC Press LLC

www.ketabdownload.com
 15. Paquet, A. and Vo, C.V., Foams 99 Conference, Parsippany, NJ, October 1999.
16. Moulinie, P., Gendron, R., and Daigneault, L.E., Gas solubility as a guide to
physical blowing agent selection, Proceedings of the Foamplas '98 Conference,
123, 1998.
17. Han, C.D., Dispersed flow of gas-charged polymeric systems, in Multiphase
Flow in Polymer Processing, Academic Press, New York, 1981.
18. Stiel, I.L. and Harnish, D.F., Solubility of gases and liquids in molten polysty-
rene, AIChE J., 22, 117, 1976.
19. Sato, Y. et al., Solubilities of carbon dioxide and nitrogen in polystyrene under
high temperature and pressure, Fluid Phase Equilib., 125, 129, 1996.
20. Durrill, P.L. and Griskey, R.G., Diffusion and solution of gases in thermally
softened or molten polymers: Part I. Development of technique and determi-
nation of data, AIChE J., 12, 1147, 1966.
21. Durrill, P.L. and Griskey, R.G., Diffusion and solution of gases in thermally
softened or molten polymers: Part II. Relation of diffusivities and solubilities
with temperature, pressure and structural characteristics, AIChE. J., 15, 106,
1969.
22. Park, C.B. and Cheung, L.K., A study of cell nucleation in the extrusion of
polypropylene foams, Polym. Eng. Sci., 37, 1, 1997.
23. Xie, H. and Simha, R., Theory of solubility of gases in polymers, Polym. Int.,
44, 348, 1997.
24. Van Krevelen, D.W., Properties of Polymers, 2nd ed., Elsevier, Amsterdam, 1976.
25. Berens, A.R. and Huvard, G.S., Interactions of polymers with near critical
carbon dioxide, in Supercritical Fluid Science and Technology, ACS, Washington,
D.C., 1989.
26. Elkovitch, M.D., Lee, L-J, and Tomasko, D.L., Proceedings of 56th SPE ANTEC,
Atlanta, GA, 1855, 1998.
27. Elkovitch, M.D., Lee, L-J, and Tomasko, D.L., Proceedings of 57th SPE ANTEC,
New York, 2811, 1999.
28. Lee, M., Tzoganakis, C., and Park, C.B., Effects of supercritical carbon dioxide
on PE/PS blend viscosity and morphology, Proceedings of 56th SPE ANTEC,
Atlanta, GA, 2570, 1998.
29. Lee, M., Tzoganakis, C., and Park, C.B., Extrusion of PE/PS blends with super-
critical CO2 in a twin-screw extruder and a twin/single tandem system, Pro-
ceedings of 57th SPE ANTEC, New York, 2806, 1999.
30. Simha, R. and Moulinie, P., Statistical thermodynamics of gas solubility in
polymers, in Foam Extrusion: Principles and Practice, ed. S.T. Lee, CRC Press,
Boca Raton, FL, 2000.
31. Lundberg, J.L., Wilk, M.B., and Huyett, M.J., Sorption studies using automation
and computation, I&EC Fundam., 2, 1, 37, 1963.
32. Micheals, A.S., Vieth, W.R., and Barrie, J.A., Solution of gases in polyethylene
terephthalate, J. Appl. Phys., 34, 1, 1, 1963.
33. Rein, D.H. et al., CO2 diffusion and solubility in a polystyrene-polybutadiene
block copolymer with a highly oriented lamellar morphology, Macromolecules,
23, 4456, 1990.
34. Dagli, S.S., Staats-Westover, R.F., and Biesenberger, J.A., Measurement of gas
solubility and relationship to extruded foam properties: a new apparatus and
some data on nitrogen in polyolefins, Proceedings of 47th SPE ANTEC, New
York, 889, 1989.

© 2004 by CRC Press LLC

www.ketabdownload.com
 35. Wong, B. and Handa, Y.P., A high-pressure electrobalance based in-situ tech-
nique for determining the solubility and diffusivity of gases in polymers,
Proceedings of 55th SPE ANTEC, Toronto, Canada, 1979, 1997. Also in J. Polym.
Sci., Part B: Polym. Phys., 36, 2025, 1998.
36. Kamiya, Y., Hirose, T., Mizoguchi, K., and Naito, Y., Gravimetric study of high-
pressure sorption of gases in polymers, J. Polym. Phys., 24, 1525, 1986.
37. Thomas, Y. et al., In-line NIR monitoring of composition and bubble formation
in polystyrene/blowing agent mixtures, J. Cell. Plast., 33, 516, 1997.
38. Sahnoune, A. et al., Application of ultrasonic sensors in the study of physical
foaming agents for foam extrusion, J. Cell. Plast., 37, 429, 2001.
39. Tsujimura, I., Zenki, T., and Ishida, M., Visual analysis of bubble nucleation in
foam extrusion die, Proceedings of PPS-14, Yokohama, Japan, 1998.
40. Biesenberger, J.A. and Sebastian, D.H., Principles of Polymerization Engineering,
John Wiley & Sons, New York, 1983.
41. Todd, D.B., Private communication, Polymer Processing Institute, Newark, NJ,
November 1999.
42. Park, C.B., Continuous production of high-density and low-density micro-
cellular plastics in extrusion, in Foam Extrusion: Principles and Practice, ed. S.T.
Lee, CRC Press, Boca Raton, FL, 2000.
43. Zhang, Q., In-line measurement of solubility of physical blowing agents in
thermoplastic melts as related to extrusion foaming, Ph.D. thesis, New Jersey
Institute of Technology, 2000.
44. Zhang, Q., Xanthos, M., and Dey, S.K., Effects of process conditions on the
dissolution of carbon dioxide in PS in foaming extruders, Intern. Polym. Proc.,
16, 223, 2001.
45. Xanthos, M., Yilmazer, U., Dey, S.K., and Quintans, J., Melt viscoelasticity of
polyethylene terephthalate resins for low density extrusion foaming, Polym.
Eng. Sci., 40, 3, 554, 2000.
46. Xanthos, M., Tan, V., and Ponnusamy, A., Measurement of melt viscoelastic
properties of polyethylene and their blends—a comparison of experimental
techniques, Polym. Eng. Sci., 37, 6, 1102, 1997.
47. Bradley, M.B. and Phillips, E.M., Proceedings of 48th SPE ANTEC, Dallas, TX,
36, 717, 1990.
48. Meissner, J., Raible, T., and Stephenson, S.E., Rotary clamp in uniaxial and
biaxial extensional rheometry of polymer melts, J. Rheol., 25, 1, 1, 1981.
49. Meissner, J. and Hostettler, J., A new elongational rheometer for polymer melts
and other highly viscoelastic liquids, Rheol. Acta, 33, 1, 1, 1994.
50. Japon, S., Leterrier, Y., and Mason, J. A. E., Recycling of poly(ethylene tereph-
thalate) into closed-cell foams, Polym. Eng. Sci., 40, 8,1942, 2000.
51. McLeish, T.C.B., Larson, R.G., Molecular constitutive equations for a class of
branched polymers: the pom-pom polymer, J. Rheol., 42, 1, 81, 1998.
52. Doi, M. and Edwards, S.F., The Theory of Polymer Dynamics, Clarendon, Oxford,
1986.
53. Wagner, M.H., Rubio, P., and Bastian, H., The molecular stress function model
for polydisperse polymer melts with dissipative convective constraint release,
J. Rheol., 45, 6, 1387, 2001.
54. Yilmazer, U., Xanthos, M., Bayram, G., and Tan, V., Viscoelastic characteristics
of chain extended/branched and linear polyethylene terephthalate resins, J.
Appl. Polym. Sci., 75, 3171, 2000.

© 2004 by CRC Press LLC

www.ketabdownload.com
 55. Yamaguchi, M. and Suzuki, K-I., Rheological properties and foam processabil-
ity for blends of linear and crosslinked polymer blends, J. Polym. Sci., Part B:
Polym. Phys., 39, 18, 2159, 2001.
56. Yamaguchi, M. and Miyata, H., Strain hardening behavior in elongational
viscosity for binary blends of linear polymer and crosslinked polymer, Polym.
J. (Tokyo), 32, 2, 164, 2000.
57. Abe, S. and Yamaguchi, M., Study on the foaming of crosslinked polyethylene,
J. Appl. Polym. Sci., 79, 2146, 2001.
58. Folland, R. et al., Low density extruded polypropylene foams, Proceedings of
SPE FOAMS 2002 Conference, Houston, TX, 1, 2002.
59. Mispreuve, H. et al., Novel high melt strength polymers for foam applications,
Proceedings of SPE FOAMS 2002 Conference, Houston, TX, 9, 2002.
60. Xanthos, M. and Dey, S.K., Foam extrusion of polyethylene terephthalate (PET),
in Foam Extrusion: Principles and Practice, ed. S.T. Lee, CRC Press, Boca Raton,
FL, 2000.
61. Al Ghatta, H.A.K., Severini, T., and Astarita, L., Polyester foams and their
manufacture, World Patent WO 9,312,164, 1993.
62. Al Ghatta, H.A.K., Process for production of high molecular weight polyesters
from resins with low intrinsic viscosity, World Patent WO 9,308,226, 1993.
63. Al Ghatta, H.A.K. and Pizzetti, M., Continuous production of high-molecular-
weight polyesters, and fibers therefrom, World Patent WO 9,325,600, 1993.
64. Khemani, K.C., Jaurez-Garcia, C.H., Boone, G.D., Foamable polyester compo-
sitions having a low level of unreacted branching agent, U.S. Patent 5,696,176,
1997.
65. Khemani, K.C., Rheologically improved polyesters for foams, Proceedings of
56th SPE ANTEC, 1934, 1998.
66. Hayashi, M., Amano, N., Taki, T., and Hirai, T., Process for producing polyester
foam resin and polyester resin foam sheet, U.S. Patent 5,000,991, 1991.
67. Zhang, Q., Dhavalikar, R., Wan, C., Dey, S.K., and Xanthos, M., Sequential
reactive blending/foaming of PET in extrusion equipment, Proceedings of SPE
FOAMS 2002 Conference, Houston, TX, 47, 2002.
68. Zhang, Q., Dey, S.K., and Xanthos, M., Batch and continuous foaming of blends
of LLDPE and cross-linked LDPE, Proceedings of SPE FOAMS 2002 Confer-
ence, Houston, TX, 19, 2002.

© 2004 by CRC Press LLC

www.ketabdownload.com
5
Foam Stability in Flexible Polyurethane
Foam Systems

X.D. Zhang, R.A. Neff, and C.W. Macosko

CONTENTS
5.1 Introduction
5.2 Fundamentals of PU Foams
5.3 Cell Opening Mechanisms
5.3.1 Aqueous Foam
5.3.1.1 Film Rupture Mechanisms
5.3.1.1.1 Capillary-wave mechanism
5.3.1.1.2 Nucleation of Pores
5.3.1.1.3 Solute Transport Mechanism
5.3.1.1.4 Solid Particle Defoaming
5.3.1.2 Factors for Short-Term Stability
5.3.1.2.1 Film Drainage
5.3.1.2.2 Prevention of Local Thinning
5.3.2 Critical Review of Proposed Cell Opening Mechanisms
in Flexible Polyurethane Foam
5.3.2.1 Urea Precipitation (Solid Particle Defoaming)
5.3.2.2 Surfactant Phase Separation (Insoluble Liquid
Droplets Defoaming)
5.3.2.3 Foam Matrix Viscosity Rise
5.3.2.4 Surface Phase Transition
5.3.2.5 Spontaneous Film Rupture
5.3.2.6 Extensional Thinning and Rupture
5.3.2.6.1 Bingham Plastic
5.3.2.6.2 Extensional Thinning
5.3.2.6.3 Cell Opening Mechanism
5.4 Controlling Cell Opening in Flexible Polyurethane Foam
5.4.1 Silicone Surfactant: Surface Rheology Effect
5.4.2 Other Formulation Ingredients: Bulk Rheology Effect
5.5 Summary
References
0-8493-1728-2/04/$0.00+$1.50
© 2004 by CRC Press LLC

© 2004 by CRC Press LLCwww.ketabdownload.com

5.1 Introduction

The addition polymerization reaction of diisocyanates with alcohols was

discovered by Dr. Otto Bayer and coworkers in 1937. This discovery set up
the fundamental chemistry for the polyurethane industry, which exceeded
4 billion pounds by 1994 in the U.S. alone.1 Through a variety of starting
materials, polyurethanes are now formulated into a wide variety of end uses
including both nonfoam and foam applications. Nonfoam applications
include coatings, adhesives, sealants, and elastomers. Among the numerous
foam applications are rigid foam for insulation or construction, semirigid
foam for various automotive interior applications, and flexible foam for
cushioning. Flexible polyurethane foam can be either molded or produced
continuously on a conveyor in a slabstock process. Of all the applications
for polyurethanes, flexible slabstock foam accounts for the largest volume
produced. Depending on the formulation, a wide range of densities and
mechanical properties are available for flexible polyurethane foam. Such
versatility is an advantage shared by no other type of cushioning material.
The process that produces polyurethane foam is complex and involves
simultaneous gas generation, polymerization, and phase separation. A
unique cell structure results (see Figure 5.1) comprised of struts (of thickness
on the order of microns) and window films (of thickness on the order of
tenths of microns). This cell structure is in contrast to that of low density
thermoplastic foam, whose cell membranes and struts are of similar and
uniform thickness and are well represented by intersecting planes.2 Low
density rigid foam, whether polyurethane or thermoplastic, is typically
closed celled. In such materials, the blowing agent remains trapped and good
insulating properties are retained. However, without sufficient porosity (air
flow), flexible polyurethane foam would have an undesirable pneumatic feel
upon compression, in addition to exhibiting shrinkage after production.
Shrinkage results as the foam temperature cools from its maximum exotherm
of about 160˚C to room temperature.3 The diffusion of carbon dioxide out of
the foam cells (which occurs faster than the diffusion of air into the cells)
also contributes to shrinkage. Thus, for flexible polyurethane foam, a high
content of open cell windows is required in order to obtain both dimensional

Cell Window
Plateau border
(Strut)

FIGURE 5.1
Illustration of structural features in a foam cell.

© 2004 by CRC Press LLCwww.ketabdownload.com

stability and the desired mechanical properties for cushioning. Molded flex-
ible polyurethane foam normally requires a crushing operation in order to
open the cell windows. In slabstock foam, the porosity results from a sudden
and rapid cell opening event occurring within the process, which is unique
to its special chemistry.
In the slabstock foam process, the foam initially goes through a stage where
the continuous phase is a flowable liquid. Liquid foam is intrinsically unsta-
ble; thus the liquid film separating bubbles can become thin and rupture.
Without proper control of film thinning, gas bubbles will coalesce, possibly
leading to foam collapse. One key to foam stability is preventing localized
thinning within the thin film. This can be accomplished through either the
Marangoni effect, which is driven by interfacial properties, or the bulk rheo-
logical properties of the continuous phase. Most of the previous film rupture
studies were conducted on aqueous foams where interfacial properties are
dominant, while bulk rheology is less important. In polyurethane foams,
both interfacial properties and bulk properties are important in providing
foam stability. Nevertheless, concepts from the considerable volume of avail-
able literature on aqueous foams are still relevant in order to understand the
early stages of the polyurethane foaming process.
In flexible polyurethane foam, a silicone surfactant is added to alter the
interfacial properties. The surfactant controls the rate of window film thin-
ning (through drainage into the struts) and prevents rupture by localized
thinning until the cell opening event occurs. Thus the surfactant helps to
control the open cell content. The timing of the cell opening event is criti-
cal—if it occurs too early (before the struts have developed adequate strength
through bulk rheology), the struts will fail and the foam will collapse. If the
event occurs too late, then the polymer in the cell windows will have
strengthened to the point where few or none of the cell windows will rupture.
The correct timing of cell opening is obtained through proper balance of
polymerization (gelling) and gas generation (blowing). Catalysts are added
to promote both of these reactions, and adjustment of these additive levels
normally provides the proper chemical balance.
The mechanism by which the extraordinary cell opening event occurs in
slabstock foam has been postulated in much recent scientific literature. Crit-
ical reviews of this literature, in addition to relevant literature covering
aqueous foam, are included in this chapter. The cell opening mechanism is
then clarified, in order to explain all of the observed phenomena. Finally, a
discussion of the roles of catalysts and surfactants is provided to illustrate
the functions of the tools available to a foam manufacturer for controlling
cell opening.

5.2 Fundamentals of PU Foams

The formation of flexible polyurethane foam is an intricate process employ-
ing two simultaneous reactions: urethane formation, which produces a

© 2004 by CRC Press LLCwww.ketabdownload.com

covalent polymer network, and CO2 blowing, which secondarily forms
polyurea.3,4 The reaction of the isocyanate group with active hydrogen-
containing compounds to form urethanes is the basis of polyurethane chem-
istry. In water-blown flexible polyurethane foams, two active hydrogen-
containing compounds are involved, water and polyol. The reaction
between isocyanate and polyol (secondary alcohol used in this study) is
called the gelling reaction. Usually, the polyol and isocyanate used are
polyfunctional, and this reaction leads to crosslinked polymers as shown
in Figure 5.2.
Over 90% of the foam volume is the carbon dioxide generated from the
reaction of an isocyanate with water. This reaction is called the gas-producing
reaction or blowing reaction. A flexible foam formulation can contain many
additives to provide the desired properties of the final product. The basic
ingredients include polyol, water, isocyanate, silicone surfactant, and cata-
lyst. The formulation used in this study is listed in Table 5.1. A detailed
discussion of these ingredients can be found in the literature.1,5
The nonionic silicone surfactant is used to achieve stable cells. It can retard
drainage flow in cell windows and prevent cell coalescence until the material
gains sufficient strength to be self-supporting. The stabilization of cell win-
dows has a great effect on the cell opening event and thus affects the porosity
of the final foam product.
Catalysts are used for both the blowing reaction and the gelling reaction.
The tertiary amine catalyst can form a complex in which the tertiary nitrogen

A. Blowing Reaction:

catalyst
H-O-H + 2 OCN CH3 CO2 + CH3 NHCONH CH3

NCO NCO NCO

(Isocyanate) (Urea)

B. Gelling Reaction:

HO HO
catalyst
OH + OCN CH3 OCONH CH3

HO NCO HO NCO

(Polyether (Isocyanate) (Urethane Linkage)

Polyol)

H+
R OH +

O RO O

FIGURE 5.2
Chemical reactions during polyurethane foam formation.
© 2004 by CRC Press LLCwww.ketabdownload.com
TABLE 5.1
Typical Flexible Foam Formulation
Parts by
Component Description Company Weight
Polyol A 3100 Mw polyether polyol with Dow Chemical 100
(Voranol™ 3137) functionality of 3, 87% propylene
oxide and 13% ethylene oxide
Water Deionized water U of Minnesota 2–6
Isocyanate (Voranate™ An 80:20 mixture of the 2,4- and 2,6- Dow Chemical 40–60
T-80) isomers of toluene diisocyanate
Silicone surfactants Silicone-polyether copolymer OSi Specialty 0.2–5
Amine catalyst (Dabco® 33% triethylenediamine in 67% Air Products 0.1–1.0
33LV) dipropylene glycol
Tin catalyst T-9 Stannous octoate catalyst Air Products 0.0–0.5

donates electrons to the carbonyl carbon of the isocyanate group. Thus the
amine catalyst can accelerate both the isocyanate–water (blowing reaction)
and isocyanate–hydroxyl reaction (gelling reaction). Organometallic tin cat-
alysts are used primarily to accelerate the reaction of isocyanate with
hydroxyl groups (gelling reaction). These catalysts are more active than the
amine catalysts and thus are used in smaller amounts. The overall mecha-
nisms of catalysis by the tertiary amine catalyst and the organometallic tin
catalyst are discussed by Herrington and Hock.1 By adjusting the catalyst’s
level and ratio, the desired balance between gelling reaction and blowing
reaction can be achieved.

5.3 Cell Opening Mechanisms

In the past 30 years, many researchers have studied the foaming process,
and various hypotheses have been proposed on the cell opening mechanism.
Some researchers attempted to extend classical aqueous foam defoaming
mechanisms to polyurethane foams. Others attributed the polyurethane
foam cell rupture to a specific factor in the polyurethane system. In this
section, various cell opening mechanisms are discussed and examined.
Aqueous foam mechanisms are reviewed, and mechanisms previously and
currently proposed for polyurethane systems are discussed in the context of
the aqueous foam literature and more recent work.

5.3.1 Aqueous Foam

Liquid foam is a thermodynamically unstable system. Liquid cannot support
stress and will flow under external forces such as gravity and capillary
pressure. Aqueous foam systems have been studied extensively and a rich

© 2004 by CRC Press LLCwww.ketabdownload.com

literature regarding film rupture and film stability can be found in this area.
A brief summary of aqueous film rupture mechanisms and factors that
control short-term stability are reviewed here.

5.3.1.1 Film Rupture Mechanisms

To achieve foam stabilization, it is imperative to understand the mechanism
of film rupture. Several mechanisms of rupture of liquid films have been
proposed.

5.3.1.1.1 Capillary-wave mechanism

This is the most widely used mechanism for film rupture in liquid foam
systems and emulsion systems. This mechanism assumes that in an ideal
liquid film with no heterogeneous structures such as particles and droplets,
the film will rupture by the growth of surface waves on the film surfaces.
The waves originated from thermal motion will grow rapidly when the film
reaches a critical thickness. This model was proposed by Vrij.6 The original
model has been greatly modified. The most recent modification by Sharma
and Ruckenstein considers the effects of film thinning on the growth of the
surface waves.7 The mean film thickness at rupture (hcm) is calculated by the
following equation:

Ê 0.015 A 2 (2x )0.164 ˆ

0.24

hcm =Á 0
˜˜ + 797 R0.25 - 209
Á 24 p 2
msv
Ë t ,c ¯

where vt,c is the film thinning velocity, x0 is the thermal wave amplitude, A
is the Hamaker constant, m is the liquid matrix viscosity, s is the surface
tension, R is the film radius in cm, and the final film thickness calculated
from this equation is in angstroms.

5.3.1.1.2 Nucleation of Pores

This mechanism was propsed by Derjaguin and Gutop.8 It is used primarily
to explain rupture of ultra-thin films between two surfactant monolayers.
The driving force for nucleation of pores is the decrease of the surface energy
that occurs when two interfaces merge. This is opposed by the edge energy
of the pore periphery, which can be described as the line tension. Above a
critical pore size, the deduction of surface energy will overcome the increase
of the edge energy and the pore will grow and cause film rupture. Below
this pore size, the pore will shrink and film will remain stable.

5.3.1.1.3 Solute Transport Mechanism

The solute transport mechanism was investigated experimentally and theo-
retically by Ivanov et al.9 When solute transfers across the film by diffusion

© 2004 by CRC Press LLCwww.ketabdownload.com

process, an instability can develop. For systems with high levels of soluble
gas, transport of gas across the film can cause significant instability and lead
to the growth of capillary waves and cause film rupture at greater thickness.

5.3.1.1.4 Solid Particle Defoaming

The presence of solventphobic solid particles can destabilize a liquid foam
system. For example, silica particles are often used as an important part of
commercial antifoaming agents.10,11 The defoaming mechanism for solid par-
ticles was proposed by Frye and Berg12 (Figure 5.3).
When the thickness of the liquid film is smaller than the particle size, the
particle will bridge the film as shown in Figure 5.3. If the particle is solvent-
phobic, then the contact angle will be larger than 90˚. This will cause liquid
drainage away from the particle and start the dewetting process, which causes
thinning and rupture of the film. The rate of dewetting is directly related to
the contact angle and hence the degree of solventphobicity of the particle.
When the solid particle is solventphilic, the contact angle will be less than
90˚. This small contact angle will prevent dewetting and film rupturing.
Solventphilic particles will actually increase the stability of many foam sys-
tems.14 This stabilization effect is mainly attributed to two factors. First,
particles can increase the viscosity of the liquid matrix and reduce the drain-
age rate. Slower drainage will increase the lifetime of a foam system. The
second stabilization factor is that solid particles can stabilize the foam system
by introducing a structural force.15,16 The oscillatory structural forces occur
when solid particles are confined between two surfaces of a thin film. Each

HYDROPHOBIC PARTICLE HYDROPHILIC PARTICLE

θ > 90° Air

θ θ < 90°
Liquid Film
Liquid Flow θ
Liquid Flow
Air No Film Drainage
No Rupture

Film Drains
Due to Capillary
Suction

Rupture

FIGURE 5.3
Schematic representation of liquid film rupture process by hydrophobic particles.13 (Reprinted
by courtesy of Marcel Dekker, Inc.)

© 2004 by CRC Press LLCwww.ketabdownload.com

surface will induce ordering among the neighboring particles. When the two
film surfaces approach, the two structured zones next to the surfaces will
overlap. This structured particle array between film surfaces will require
more energy for the film to thin further, since thinning of the film will disrupt
packing of the particles. This is the origin of the structural force in a solid
dispersion system. This effect tends to stabilize the film.

5.3.1.2 Factors for Short-Term Stability

Maintenance of proper foam stability is important to avoid catastrophic
coalescence of the bubbles and foam collapse. Much literature exists in the
area of aqueous foam stability, most in the area of long-term stability, where
disjoining pressure within the film is the key for film stability. During poly-
meric foam formation, short-term stability and stability under expansion are
important. Long-term stability in time scale of hours and above is not rele-
vant.
Two factors are important in short-term foam stability: film drainage rate
and ability to prevent local thinning in liquid film. Drainage rate determines
how long the film can remain above the critical rupture thickness and what
percentage of the film is below the critical rupture thickness. Localized
thinning can be detrimental to film stability and needs to be controlled to
provide uniform film thinning and foam stability.

5.3.1.2.1 Film Drainage

In the foam matrix, the liquid lamella will drain to the Plateau borders under
capillary suction and gravity.17 How fast liquid drains out of the film will
determine film thickness and foam density.
Liquid in thin films will drain into the Plateau border (Figure 5.2) under
capillary pressure. Assuming the liquid lamella is a plane-parallel film with
a tangential immobile surface, the axisymmetric drainage rate VRe can be
described by the Reynolds equation:

dh 2 Dph
VRe = - =
dt 3 hR 2

where h is the film thickness, h is the bulk viscosity, R is the radius of the
film, and Dp is the pressure drop that drives film draining. This pressure
drop is a combination of Plateau border suction and disjoining pressure.
However, in actual foaming process, liquid film surfaces are not immobile
since the bubbles are constantly subject to shearing and stretching. Surfactant
diffusion will limit the surfactant surface coverage.
For films with a mobile surface, drainage will cause the movement of the
surface layer and a surfactant surface concentration gradient will be gener-
ated. The surfactant surface concentration gradient will cause a surface ten-

© 2004 by CRC Press LLCwww.ketabdownload.com

sion gradient along the surface that will retard the film drainage. This effect
is called the Gibbs–Marangoni effect and is reviewed by Scriven and Stern-
ling.18 The surface tension gradient can be affected by the diffusion of sur-
factant along the surface as well as the diffusion of surfactant from the bulk.
Thus the surface rheology of the film, such as surface viscosity, becomes
important since it will affect the surface diffusion of the surfactant. The
drainage can be obtained by solving the Navier–Stokes equations under the
lubrication approximation accounting for the mobile surface.19

( )
-1
h 3 Dp È
•
6 h + hs kn2 Mh J 2 (l n ) ù
Vo =
dh
= Í Â( )
ú
dt 24 hR 2 Í n=1 6 h + 6 M + hs kn2 Mh l3n J1 (l n ) ú
Î û

where J0, J1 and J2 are Bessel functions, eigenvalues ln are solutions of equa-
tions J0(ln) = 0 and kn = ln/R. The surface mobility of the film is accounted
for by surface viscosity hs and mobility parameter M. The mobility parameter
M includes the contributions of both bulk and surface diffusion to the
Marangoni effect and is defined as the following:

È Ê ∂G ˆ ù
Í 2Ds Á ˜ ú
3Dh Ë ∂c ¯ 0
M= Í1 + ú
Ê ∂l ˆ
gÁ ˜ Í ú
Dh
Ë ∂c ¯ 0 ÍÎ úû

where D and Ds are the bulk and surface diffusion coefficients of surfactant,
respectively. C is the surfactant concentration in the bulk and G is the surface
excess concentration of the surfactant on the air/liquid interface. Increasing
either D or Ds will cause a faster drainage rate since it will increase the
mobility parameter and reduce the surface tension gradient. Interfacial vis-
cosity will change the Ds value and the ability to maintain a surface tension
gradient. This is one of the reasons high interfacial viscosity can help foam
stabilization. The analysis for drainage rate in a thin liquid film with constant
area is shown in the above section. It will be impractical to obtain all the
parameters shown in these equations. However, the analysis indicates which
factors will contribute to film drainage.

5.3.1.2.2 Prevention of Local Thinning

The other key to foam stability is the prevention of local thinning and rupture
of liquid films under deformation. This can be accomplished by either the
Gibbs–Marangoni effect or bulk rheology effect.
The short-term foam stability can be provided by the Gibbs–Marangoni
effect or the dilational film elasticity of foam lamella. Prevention of local
thinning will reduce the rupture thickness of the liquid film and increase
foam stability. Variables like surface shear viscosity, surface dilational vis-
cosity, and surface transport can affect both the drainage rate of the liquid

© 2004 by CRC Press LLCwww.ketabdownload.com

film and the ability to prevent local thinning in the film. Without the mini-
mum level of surface elasticity or surface viscosity, foam will be unstable
upon any disturbance. Local thinning can be a result of drainage or a sudden
disturbance force. Local thinning of the film will result in an increase of
surface area at the thinning point. The surfactant concentration in this area
will decrease. As a result, a surface tension gradient is established. This
surface tension gradient will tend to pull the two surfaces together. A second
function of the surface tension gradient is that it will promote surface trans-
portation of surfactant into the thinning point. This surfactant transportation
will drag the underlying layer of the liquid into the thinning point as well.
As a result, the thinning point is “healed.” There has been little attempt to
measure the dilational film elasticity for personal wash products.
Bulk rheology can have a significant effect on local thinning as well. For
aqueous foams, the bulk rheology is usually Newtonian and low viscosity.
In this system, the Gibbs–Marangoni effect is the key mechanism for uniform
thinning. When bulk viscosity is high, the surface effect becomes less signif-
icant. Viscosity of the continuous phase and its extensional rheological
behavior become more important. In general, strong extensional thinning
systems will promote localized thinning of films and lead to film rupture
and bubble coalescence. On the other hand, strain hardening of the foam
matrix is shown to have a positive effect on foam stability of thermoplastic
foams.20

5.3.2 Critical Review of Proposed Cell Opening Mechanisms in Flexible

Polyurethane Foam
5.3.2.1 Urea Precipitation (Solid Particle Defoaming)
In a polyurethane foam system, the urea precipitation occurs a few seconds
before the cell opening. These particles are stiffer than the liquid matrix and
may cause rupture of the film. This hypothesis was originally proposed by
Rossmy and coworkers.21 This mechanism is a direct extension of the classical
solid particle defoaming mechanism and is consistent with the fact that the
cell opening always occurs after the urea precipitation occurs regardless of
the formulation used. However, the urea precipitation is linked to the matrix
by urethane linkages and is unlikely to dewet the surrounding liquid. It is
also known that the highly polar urea groups have high surface energy and
will be wetted by the low surface energy liquid matrix. As a result, the urea
precipitate defoaming mechanism should not be valid.
To test this mechanism, two experiments were performed. In the first
experiment, water was reacted with toluene diisocyanate in a flask to gen-
erate solid urea particles. These solid particles were then collected. Solid
particles of 0.2 pphp were added into a surfactant solution in polyol. Foam
column tests were performed to test the stability of liquid foam with and
without the addition of the urea particles. The result of this experiment
shows that the foam system with urea solid particles has a higher stability
and longer foam life than that without the urea particles.

© 2004 by CRC Press LLCwww.ketabdownload.com

 The second test was performed using a similar idea to the first test, except
the particles were made inside an end-capped polyol. The end-capping was
prepared by capping the hydroxyl group with dihydropyran:22

H+
R OH +

O RO O

350 g of Voranol 3137 polyol, 100 index of dihydropyran, 400 mL methylene

chloride, and 0.1 g p-toluenesulfonic acid were placed in a 2000 mL flask.
The mixture was stirred using a 1.5 in. magnetic stir bar for 2 h at room
temperature; 0.5 g sodium bicarbonate and 50 g of 4 molecular sieves were
added to the flask and allowed to stir overnight to neutralize the acid and
remove residual water. After filtering, the mixture was placed in a rotary
evaporator and heated to 70˚C to remove the methylene chloride. The mix-
ture was then placed in a flask and stored under vacuum overnight to remove
traces of volatile components. The completion of end-capping can be mon-
itored by the disappearance of the hydroxyl group (3450 cm–1).
Water (0.25 pphp) and isocyanate (100 index) were added to a surfactant
solution in end-capped polyol and toluene. The end-capping prevented reac-
tion between polyol and isocyanate, and toluene was added to keep the
viscosity of the solution low and the same for solutions with and without
urea particles. The solutions with urea particles showed significantly higher
stability than those without urea particles using the static foam column test.
This again shows that a small amount of in situ formed urea particles
enhances the stability of the liquid film system. The solid dewetting mech-
anism does not apply in a flexible polyurethane foam system.

5.3.2.2 Surfactant Phase Separation (Insoluble Liquid Droplets

Defoaming)
Insoluble liquid droplets can dewet the film and cause foam rupture in much
the same way as the solid particles. The complication in liquid droplet
defoaming is that the liquid droplets are deformable and can spread along
the film surface. This will create some problem for the liquid droplets to
properly bridge the film. There are numerous studies on how insoluble liquid
droplets spread and bridge the film.12,13 Once the insoluble liquid droplets
bridge the film, they will either stabilize or destablize the film. If the three-
phase contact angle is larger than 90˚, then these droplets will destabilize
the film and cause poor foam stability.
For the polyurethane foam, one mechanism proposes that the cell opening
may be due to the silicone surfactant added.23,24 This mechanism assumes a
decrease in solubility of the surfactant in polyol with increasing temperature.
As the temperature rises above a critical value, the surfactant will phase
separate and act like a defoamer.

© 2004 by CRC Press LLCwww.ketabdownload.com

 This mechanism suggests a simple model. The surfactant structure will
affect the critical temperature of surfactant phase separation and cell opening
time. At a different cell opening time, the modulus of the foam material will
be different, thus resulting in a different final foam porosity. The effect of
the silicone surfactant on cell opening will be easy to predict.
In order to test this hypothesis, two experiments were performed. First,
surfactant was dissolved in polyol and the temperature of the solution was
controlled. The cloud point was not observed in the temperature range of 0
to 110˚C. Also, the solubility of the surfactant was observed to increase with
temperature. These observations oppose the proposed theory.
To further examine the surfactant phase separation hypothesis, element
mapping of silicon was obtained for final cured foams. The surfactant dis-
tribution could be obtained because silicon is present only in the surfactant.
A Jeol Electron Probe Microanalyzer 8900 was used for Si mapping. As
electrons struck the surface of the sample in a vacuum chamber, x-rays were
emitted by the sample. The concentration of silicon at each point was deter-
mined by measuring the intensity of x-rays at the characteristic wave length
for silicon. Two samples were used in this study: sample A was a collapsed
foam made by adding silicone oil, a commercial defoamer; sample B was a
normal open cell foam made with a silicone surfactant. Results are shown
in Figure 5.4. The silicon concentrations are indicated by the brightness of
the pixels. In sample A, silicone oil is shown to be phase separated into
micrometer size spherical particles. In sample B, silicone surfactant is shown
to be very homogeneously distributed. No bulk segregation with a large
enough size to cause film rupture can be observed. This experiment shows
that the silicone surfactant is quite homogeneously distributed in the cured
open cell foam. It appears that no surfactant phase separation in the bulk
occurred. Thus the surfactant phase separation mechanism is also not likely
valid.

5.3.2.3 Foam Matrix Viscosity Rise

This hypothesis suggests that as the urea phase separates, a rapid increase
in the foam matrix viscosity will occur. Higher viscosity will limit the move-
ment of the surfactant diffusion to the surface and reduce the surface tension
gradient-driven transport “self-healing” by the Marangoni effect. Once the
stabilizing mechanism is disturbed, the film will experience local thinning
and rupture. Generally, the drainage rate of a thin liquid film will decrease
linearly as the viscosity of the system increases. Higher bulk viscosity will
further increase the surface tension gradient since bulk diffusion will be
lower. Thus, the rise of foam matrix viscosity alone cannot explain the cell
opening event.

5.3.2.4 Surface Phase Transition

In some systems, as the temperature of the foam reaches a critical tempera-
ture, the surfactant layer will experience a surface phase transition. This
surface phase transition will disrupt foam stability. It will lead to faster

© 2004 by CRC Press LLCwww.ketabdownload.com

 A.
Si Counts
15
12
9
6
3
0

Si 5 µm
5µm
Pixel size 0.22 µm
B.
Si Counts
8
7
6
5
4
3
2
1
0

Si 5 µµm
Pixel size 0.11 µm

FIGURE 5.4
(A) Silicon map for Foam A (with silicone oil); (B) Silicon map for Foam B (with silicone sur-
factant).

drainage rate and results in film rupture.25 This mechanism is theoretically

sound and surface phase transition does occur for some systems. In these
systems, the surfactant layer is typically a rigid layer with high intermolec-
ular interactions since the system temperature is below the melting point of
the surfactant. This rigid layer will give an immobile boundary condition
for the drainage flow and will create foams with great stability. When the
system temperature rises above the melting point of the surfactant, the
surfactant surface layer will experience a phase transition and will become
mobile. As a result of the mobile interface, the foam system will have faster
drainage rate and poor foam stability. This is the basis for the surfactant

© 2004 by CRC Press LLCwww.ketabdownload.com

surface phase transition mechanism. In a polyurethane foam system, the
silicone surfactant is very mobile even at room temperature. A rigid surfac-
tant surface layer never exists. When the system temperature is raised, there
will not be a phase transition. Thus, the surface layer phase transition mech-
anism may not be suitable for a polyurethane foam system.

5.3.2.5 Spontaneous Film Rupture

As mentioned previously, spontaneous film rupture, and specifically the
capillary wave mechanism, is most often employed to explain film rupture
in liquid foam and emulsion systems. Therefore, it will be considered here
for the flexible polyurethane foam system. However, one limitation of this
mechanism is that it is normally applied to homogeneous films. This is
especially significant in the case of the flexible polyurethane foam system,
which can have inhomogeneities such as phase-separated urea. Neff5 applied
the relatively recent model of Sharma and Ruckenstein7 to estimate the upper
bound for the critical film thickness at rupture for flexible polyurethane
foam. He calculated this to be 15 nm, which is in the thickness range for
black films.
Using light interference microscopy, the thinnest film found in polyure-
thane is thicker than 150 nm. This is an order of magnitude higher than the
value estimated by Neff. Thus the spontaneous rupture mechanism is not
valid for polyurethane foam systems.

5.3.2.6 Extensional Thinning and Rupture

The first attempt to quantify the internal cell opening time was done by
Neff.5 Using a parallel-plate rheometer, Neff recorded the normal force that
a polyurethane foam exerted on the plate during the reaction. Different
stages during foaming are identified and shown in Figure 5.5.

120
Cell opening
Normal Force (grams force)

100
Urea microphase
80 Initial
separation
bubble
60 growth

40
Sample
injection Foam
20 expansion

−20
0 20 40 60 80 100
Time (s)

FIGURE 5.5
A typical normal force profile during polyurethane foaming.

© 2004 by CRC Press LLCwww.ketabdownload.com

 Fully open Partially open

Closed Pin holes

FIGURE 5.6
Different states of cell windows in polyurethane foams.26

The sudden drop in the normal force is a result of the sudden stop of
foam expansion due to cell opening. The onset of cell opening occurs
several seconds after the onset of urea phase separation and the timing
does not depend on the structure of the surfactant. This implies that cell
opening is triggered by urea precipitation. However, urea particles cannot
cause cell rupture by the dewetting mechanism. In water-blown foam, the
particles are generated within the matrix and are chemically grafted to
the matrix through urethane linkages. Neff proposed that the precipitated
urea will trigger cell opening by affecting the rheological properties of the
foam matrix. This is the Bingham plastic mechanism. Most suspensions
show strong shear thinning and extensional thinning behavior rather than
the Bingham plastic behavior. An extensional thinning mechanism is also
discussed in the following.

5.3.2.6.1 Bingham Plastic

Figure 5.6 shows the different states of cell windows observed in polyure-
thane foams. Some cell windows have completely disappeared without any
debris. This suggests that the material is capable of flow at least in the early
stage of the cell opening period. The fact that the material has solidlike
behavior and is capable of flow suggests that the material is described by a
Bingham plastic model.27,28 The material behaves as a solid until a yield stress
is reached; then the material yields and flows as if it were a liquid. This
mechanism was proposed by Neff in 1995 and is shown schematically in
Figure 5.7.
Window films are stretched during urea microphase separation. Urea
aggregates formed in the film may act as stress concentrators and local

© 2004 by CRC Press LLCwww.ketabdownload.com

 E

Urea microphase
separation
Urea aggregates
t
σ
σy
Cell window

σlocal > σf, σy

ε Over-
Bingham plastic
stabilization
(Chemistry,
Surfactant)

Open window Closed or broken window

FIGURE 5.7
Proposed model for cell opening.5

thinning will occur. The thinned part of the film may form a hole. The hole
will grow in the film, and the material flows back to the Plateau’s borders.
The material keeps stiffening rapidly after the onset of cell opening. The film
will lose its flowability. The holes formed earlier will have enough flowability
and will form fully open cell windows, while films ruptured later cannot
flow very well and result in the torn-off films or partially open films. How-
ever, there is no experimental data performed to show that the phase sepa-
rated polyurethane system behaves as a Bingham plastic.

5.3.2.6.2 Extensional Thinning

True Bingham plastics are rare. The more common behavior for a solid
suspension is shear thinning or extensional thinning behavior. Figure 5.8
describes the development of the foam matrix during foaming. When the
foam is in a liquid state, expansion of the liquid thin film will cause the film
to thin down uniformly until a black film (film thickness less than 60 nm)
is reached. Then the surface waves will cause spontaneous film rupture. To
reach a black film, a much longer time than the foaming process is needed.
When the foam is cured to an elastomer foam, then the expansion will only
stretch the film and the film will relax back to the original state once the
stress is released as the pressure inside the foam bun releases to the atmo-
sphere. The cell windows rupture when the foam matrix is a pastelike mate-
rial due to the precipitation of the urea aggregates.5 The foam matrix becomes
a suspension with urea particles since the crosslinked covalent network is
not yet formed.

© 2004 by CRC Press LLCwww.ketabdownload.com

 It is known that most suspensions display shear thinning behavior.27 To
verify this, a model compound is made to simulate a phase separated mate-
rial. 100 g of Voranol 3137, 2 parts water, and 70 index toluene diisocyanate
(TDI) were placed in a beaker and mixed. The reactants were allowed to
react for 24 h with a magnetic stir bar in the beaker. This material will be
similar to the foam material at cell opening time for higher water part foams
(4–6 parts). Since the limiting reactant TDI will be consumed mainly by
water, polyol will have low conversion and the system will be practically a
urea suspension in polyol. Shear viscosity is tested using parallel plate geom-
etry with serrated surfaces. This compound shows significant shear thinning
behavior (Figure 5.9). This means the foam matrix will have higher viscosity
at low stress and lower viscosity at high stress.
As seen in Figure 5.10, the cell windows usually have a dimple shape.
Cell windows are found to be dimple shape using the light interference
microscopy method and is discussed later in this text. Once the urea phase
separation occurs, the cell window will exhibit shear thinning behavior. The
thinner part of the window (edges) will have much lower viscosity than the
thicker part of the cell window (center). As a result, the film thinning is
localized to the edge of the cell window. The localized thinning will cause
the cell windows to thin rapidly and form pinholes. The cell windows
containing pinholes can further develop into partially open and fully open
cell windows depending on the matrix viscosity and time available for cell
opening. This is why the cell opening always happens a few seconds after
urea phase separation regardless of what surfactant is added into the system.
A cell window between growing bubbles, however, experiences a biaxial
expansion motion rather than a simple shear. It is known that some shear
thinning material actually exhibits an extensional thickening effect.27 There-
fore, it is important to prove that the phase separated polyurethane matrix

105 Elastomer Foam

Onset of Cells Covalent Curing/
Phase Separation Open Hard Segment
104 Annealing

1000
Pastelike Material
G′ (Pa)

Urea Microphase
100 Separation
Liquid Foam

10

1
0 50 100 150 200
Time (s)
FIGURE 5.8
Foam morphology and modulus development during foaming reaction.

© 2004 by CRC Press LLCwww.ketabdownload.com

 6 105

5 105

Viscosity (Pa*s)
4 105

3 105

2 105

1 105

0
100 1000
Stress (Pa)

FIGURE 5.9
Shear viscosity of the model compound.

shows extensional thinning as well. This test was performed on a lubricated

squeeze flow apparatus.
A lubricated squeeze flow apparatus has been used to obtain biaxial vis-
cosity data.28,29 By proper choice of the lubricant viscosity, a sample under
compression will experience biaxial extension without shear on the plate
surfaces.27 By applying different weights on the top plate, the sample will
experience different stresses and deform at different extension rates. Figure
5.11 shows a schematic of the lubricated squeezing geometry.
The model compound is tested using this instrument and a strong exten-
sional thinning behavior is observed (Figure 5.12). Thus, the phase separated
model compound does exhibit extensional thinning behavior and can cause
cell window rupture as illustrated in Figure 5.10. In a transient extensional
viscosity test, no strain hardening is observed.30 Since the shear viscosity is
performed in a strain-controlled rheometer, a very limited stress range is

low stress high stress

high viscosity low viscosity

stretch

FIGURE 5.10
Schematic of the localized film thinning due to the extensional thinning behavior of the foam
matrix.

© 2004 by CRC Press LLCwww.ketabdownload.com

 F

L0 L

2R

FIGURE 5.11
Schematic of the lubricated squeezing geometry.

covered and detailed correlation between shear and extensional viscosity

data cannot be performed.

5.3.2.6.3 Cell Opening Mechanism

Foam formation occurs with a high exotherm. As the temperature of the
foam bun gets higher, heat will be lost from the top surface of the foam bun
to ambient air. As a result of the heat loss, the temperature at the top surface
is lower than that of the center. This lower temperature profile will yield
slower reaction kinetics and delayed urea phase separation. When the inter-
nal cell opening occurs, the top surface of the foam remains liquid. The foam
will continue to expand until the internal pressure is high enough to break
the liquid surface layer. This is the skin blow-off time. Cell opening is a
continuous process between the internal cell opening time and the skin blow-
off time. In other words, as long as foam expansion continues, the cell
opening will continue.

107
Biaxial Extensional Viscosity (Pa)

106

105

106

1000
1000 104
Stress (Pa)

FIGURE 5.12
Biaxial extensional viscosity of the model compound.

© 2004 by CRC Press LLCwww.ketabdownload.com

 Pinhole Open
(smooth edge)

Closed

Partial Open Open

(rough edge)

FIGURE 5.13
Various states of cell windows.

SEM shows that the cell windows are in various states. Windows with
only small pinholes were observed. Fully open windows and partially open
windows were also observed. In some fully open windows, the film debris
is totally missed. Other fully open windows show the film debris wrinkling
at the strut edge. Images of the different states of cell windows are shown
in Figure 5.13.
Partially open windows showed both drained-off (circular) holes and torn-
off holes. These various states of the cell windows imply that the cell win-
dows are in a transient status from liquid to solid during the cell opening
period.
It is known that different surfactant types and concentrations provide
different rates of liquid film drainage.31 At the onset of cell opening, the cell
window will have a thickness distribution. The window thickness distribu-
tion could change the fraction of broken cell windows and air flow. Figure
5.14 shows how cell windows with different morphologies are obtained. The
thinner windows will rupture first when the foam matrix viscosity is low.
At this stage, the foam matrix still has high flowability. As a result, the
ruptured cell window flows back to the struts, and smooth edges will be
obtained. The medium thickness windows will form pinholes later in the
foaming process when the viscosity of the matrix is higher. At this point, the
window has only limited flowability due to the high viscosity. As a result,
pinholes will remain with limited growth or connected by tearing the mate-

© 2004 by CRC Press LLCwww.ketabdownload.com

 Extensional Thinning
and Stress Concentration
Caused by Urea Aggregate

Urea Phase Thin Film

Separation A
Low viscosity
period
Open Window
Overstabilization (smooth edge)
High viscosity
Thick Film by Surfactant
period
Medium
D Film
η
B

B, C Partial Open Window

A
Closed Windows Open Window
Time D (rough edges)
C

FIGURE 5.14
Cell window status development.

rial between them to form partially open cell windows and cell windows
with rough edges. Finally, the very thick windows will not open since the
time for window thinning is less than the time for gel reaction. These win-
dows will remain closed after the foam is cured.

5.4 Controlling Cell Opening in Flexible Polyurethane Foam

Polyurethane foams have a variety of applications depending on their den-
sity and mechanical properties.1 High density polyurethane foams (10–20
lb/ft3) are used in applications such as carpet backing, whereas low density
foams have applications in furniture (e.g., foam in mattress), transportation
(e.g., car seats), and packaging. Many physical properties of flexible poly-
urethane foam, such as mechanical response for cushioning, are highly
affected by its air flow.32 It can be significantly affected by both reaction-
related additives and silicone surfactants.
Air flow is difficult to control in practice because the role of silicone
surfactant in the cell opening process has not been completely elucidated.
The cells in polyurethane formed by carbon dioxide are unstable unless a
silicone surfactant is used. The silicone surfactant is shown to have signifi-
cant effects on the macroscopic cellular structure such as cell size and foam

© 2004 by CRC Press LLCwww.ketabdownload.com

air permeability. A typical surfactant is a copolymer composed of a silicone
backbone and poly(ethylene oxide-co-propylene oxide) grafts. The amount
of surfactant, the backbone length, frequency and composition of the grafts
are varied for different polyurethane formulations.1 The surfactant must
stabilize the carbon dioxide bubbles; yet, in flexible polyurethane foam, the
“windows” between the cells must rupture during polymerization to prevent
shrinkage of the product. Some trends have been observed between the
structure of the surfactant and the air permeability of the final foam product.
Decreasing the molecular weight of the silicone backbone of the surfactant
seems to increase the percentage of open cell windows. However, when the
backbone is too short, the cells will collapse before the material is strong
enough to support itself. Is this controlled by solubility or mobility of the
surfactant in the reacting mixture? Understanding the role of silicone sur-
factant during foaming, especially its effect on the ruptured cell window
percentage and air permeability of the final foam product, can help us design
new surfactants and better control the macroscopic cellular structure.
Additives that change the reaction kinetics can also play important roles
in cell opening. Since the urea separation and extensional thinning of the
bulk liquid is the trigger for cell opening, controlling the balance between
blowing reaction and gelling reaction to achieve proper bulk viscosity and
rheological profile at urea phase separation time is a very effective way to
control the degree of cell opening. A common practice in industry is to tune
polyol structure, catalyst level, water level, and so forth in order to alter the
cell opening process.

5.4.1 Silicone Surfactant: Surface Rheology Effect

Traditionally, it is believed that the silicone surfactant can lower surface
tension, emulsify incompatible formulation ingredients, promote nucleation
of bubbles during mixing, and stabilize cell windows.1,3 Among these func-
tions, cell window stabilization is the most important. During foaming, the
bubbles initially introduced by mechanical mixing will grow. As the volume
fraction of the gas bubbles exceeds 74%, the spherical bubbles will distort
into multisided polyhedra and cell windows are formed. Due to capillary
forces, the liquid pressure inside the Plateau borders will be lower than that
in the cell windows. This pressure difference will cause cell window thinning
by sucking liquid from the windows into the struts. Without adding silicone
surfactant, this drainage rate will be such that film rupture and bubble
coalescence occur rapidly. It is shown in this chapter that the silicone sur-
factant added can retard the cell window drainage rate. Different surfactant
structures will have different abilities to retard drainage, resulting in differ-
ent cell window thickness distributions at the time of cell opening.26 The
porosity or the air permeability of the solid polyurethane foam is signifi-
cantly affected by the structure of the silicone surfactant used in the formu-
lation. Thus, it is essential to find the optimized silicone surfactant that will
result in foam products with desired physical properties. For example, foam
© 2004 by CRC Press LLCwww.ketabdownload.com
 TABLE 5.2
Designed Surfactant Series
Surfactant Polyether Siloxane Backbone
Structure X Y Percentage Length
MDxD¢yMa 119 21 75 140
MDxDyM 85 15 75 100
MDxD¢yM 51 9 75 60
MDxD¢yM 54 6 66 60
MDxD¢yM 57 3 50 60
MDxD¢¢yMb 51 9 88 60
a Branched copolymer of polydimethylsiloxane and polyether where: M = CH3,
D = (CH3)2Si-O, D¢ = (CH3)RSi-O, R = 60% propylene oxide and 40% ethylene
oxide polyether, 1500 Mw, acetoxy capped.
b D¢¢ = (CH )R¢Si-O, R¢ = 60% propylene oxide and 40% ethylene oxide polyether,
3
4000 Mw, acetoxy capped.

used in filters needs high air permeability and durability, while insulation
foam needs low air permeability and low flammability. However, the rela-
tionship between silicone surfactant structure and final foam properties is
not well understood.
A series of surfactants (Table 5.2 and Figure 5.15) were designed to study
the effects of surfactant structures on the cell opening event and the final
open cell percentage.
The surfactants were prepared by reacting the polysiloxane backbone with
polyether branches by Witco Corporation, OSi Specialties Group. Two vari-
ables were changed: the surfactant backbone length and the polyether branch
percentage. The silicone backbone length is defined as x + y. The polyether
branch percentage is defined as (y * Mwpolyether/Mw). Figure 5.15
describes the surfactant structure in terms of the backbone length and poly-
ether branch percentage.
The ruptured window percentage of a foam is directly related to its air
permeability.26 Using the formulation listed in Table 5.1 with different sur-
factants, the air permeability of the resulting foam products are determined
and listed in Table 5.3.

CH3 CH3

(CH3)3 SiO (SiO)D′ Si(CH3)3

(SiO)D

CH3 R

FIGURE 5.15
Typical silicone surfactant structure. [M, Dx, D¢y, M, and D¢ are the units that are connected to
polyether branches.]

© 2004 by CRC Press LLCwww.ketabdownload.com

TABLE 5.3
Air Permeability for Foams Made with Different Surfactants
Surfactant MD119D¢21M MD85D¢15M MD51D¢9M MD54D¢6M MD57D¢3M MD51D¢¢9M
Air Flow 0.3 0.9 1.9 1.0 Collapse Collapse
(CFM)

It is believed that different surfactants have different drainage retardation

abilities, thus resulting in different cell window thickness distributions at
the time of cell opening. This will cause the final foam product to have a
different percentage of ruptured cell windows as well as a different thickness
distribution for the remaining windows. To verify this, the cell window
thickness in the cured solid foam was measured for foams made with two
different surfactants. Surfactants with different backbone lengths
(MD85D¢15M and MD51D¢9M) were used since they result in foams with the
same cell size and different air permeability.
The cell window thickness was determined using the light interference
method. The incident light is normal to the cell window. Light was reflected
from the top surface and the bottom surface of the cell window. The two
reflected beams had interference due to their difference in optical paths. The
optical path difference was directly related to the cell window thickness.
This path difference then leads to the various interference colors observed.
Figure 5.16 shows a sample image as well as the thickness profile.
The thickness profile shows a dimple formation, which is a strong indica-
tion of the existence of a surface tension gradient. The thickness of the thin
part of the window was recorded. Fifty windows were tested for foams made
with these two surfactants. Figure 5.17 shows the thickness distribution of
remaining cell windows.

0.33 micron

1.41 micron

2nd order 9th order

crimson green

FIGURE 5.16
A cell window thickness profile determined by light interference microscopy. (Left: interference
pattern of the cell membrane; Right: cross section of window and thickness profile.)

© 2004 by CRC Press LLCwww.ketabdownload.com

 30

25
MD85D′ 15M (0.27 µm)
Number of Cell Windows

20 MD51D′ 9M (0.22 µm)

15

10

0
0.15 0.25 0.35 0.45 0.55 0.65 0.75
Window Thickness (micron)
FIGURE 5.17
Window thickness distribution for foams made with different surfactants.

Different distributions are observed: foam made with surfactant

MD85D¢15M has a higher film density toward the thicker films, while foam
made with surfactant MD51D¢9M has a higher film density toward the thinner
films. The average thickness for foam made with surfactant MD85D¢15M is
also thicker than that with surfactant MD51D¢9M. The ruptured cell window
percentage is counted for these two foams. Foam made with surfactant
MD85D¢15M has a rupture cell window percentage of 18%. Foam made with
surfactant MD51D¢9M has a higher rupture cell window percentage of 32%.
It is observed that by simply varying the surfactant used in the formula-
tion, foams with different cell window thickness and different percentages
of ruptured cell windows can be obtained. This verifies the importance of
the cell window drainage rate.
Polyurethane foam develops from a shaving-cream-like liquid foam to a
solid foam in about 2 min. It is very important for the surfactant to stabilize
the foam lamella before the gelation of the foaming material. This is accom-
plished by establishing a surface tension gradient along the foam lamella
and reducing the rate of thinning of the cell window. A schematic diagram
is shown in Figure 5.18.
Due to the capillary suction in the Plateau border, the liquid inside the
lamella will drain into the Plateau border. Without the addition of surfactant,
the gas/liquid interface cannot support any stress and this interface will
have a slip boundary condition. The thin liquid film will rapidly thin and
lead to cell window rupture and foam collapse. When a silicone surfactant
is added, a surface tension gradient will develop along the gas/liquid
interface and this interface can have a finite surface stress. As the bubbles
are expanding, the film is expanding. This expansion, in addition to surfac-
tant depletion, will lead to reduction of the surfactant concentration on the

© 2004 by CRC Press LLCwww.ketabdownload.com

 Bubble
expansion
Higher surface tension
Lower
Film drainage− Surfactant
surface
adsorption
tension
Surface tension
gradient
2 *d Γ
film elasticity =
dln(A/A0)

FIGURE 5.18
Schematics of cell window drainage and surface tension gradient formation.

gas/liquid interface. As a result, the surface tension will increase as the

bubble expands. Meanwhile, the Plateau border experiences much less
expansion and retains a low surface tension value. Thus, a surface tension
gradient will exist along the film surface. The surface tension gradient will
establish a surface stress and will reduce the drainage flow rate. This surface
tension gradient is related to a physical parameter, the dynamic film elas-
ticity, which is defined as the rate of surface tension increase as the film
area expands. The effect of surface elasticity on thin liquid film drainage
rate was investigated by Malhotra and Wasan using a numerical simula-
tion.33 A significant effect of the surface elasticity was observed. A large
surface elasticity leads to lower surface mobility and lower film thinning
rate. Measuring the film tension of the surfactant solution can provide
information on the film elasticity and the surface tension gradient. The
surface tension gradient will determine the boundary condition for the
drainage flow in cell windows and the short-term foam stability.
The film rheometer is used to measure the film elasticity of different
surfactants.34 The Gibbs elasticity was obtained by recording the surface
tension after the sudden expansion of a bubble. The Gibbs elasticity for
different surfactant solutions are listed in Table 5.4.
The Gibbs elasticity of the silicone surfactant solution is not affected by
the structure of the surfactants. In a polyurethane foam system, bubbles grow
at a finite rate of about 0.02 mm3/min rather than sudden expansion. The
finite expansion rate allows some surfactant diffusion and adsorption to
reduce surface tension as the bubble expands. As a result, the dynamic film
elasticity obtained under the finite expansion rate is always lower than the
Gibbs elasticity that is obtained by sudden expansion. Figure 5.19 shows

TABLE 5.4
Gibbs Elasticity for 0.2 wt% Surfactant Solution in Voranol 3137
MD119D¢21M MD85D¢15M MD51D¢9M MD54D¢6M
Gibbs Elasticity 22.6 23.1 23.1 26.0
(dyne/cm)

© 2004 by CRC Press LLCwww.ketabdownload.com

 58

57 Sudden
Film Tension (dyne/cm) Expansion
Expansion Rate
56

55

54

53

52

51
0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4
ln (A/A0)
Film Elasticity (dyne/cm)
dA/dt = 0.021 mm2/sec 0.432
dA/dt = 0.053 mm2/sec 2.957
dV/dt = 0.084 mm2/sec 5.195
dV/dt = 0.161 mm2/sec 9.703
FIGURE 5.19
Effect of bubble expansion rate on dynamic film elasticity of 0.2 wt% MD51D¢9M solution in
Voranol 3137.

that as the expansion rate increases, the dynamic film elasticity increases
because of the decreased time for surfactant diffusion and adsorption.
The information above shows that the film rheometer can give accurate,
reproducible results. Next, the dynamic film elasticity of different surfactants
were evaluated with an expansion rate of 0.084 mm2/sec, which is close to
the actual polyurethane foam system. The higher the dynamic film elasticity,
the higher a surface tension gradient can be obtained in the expanding cell
window.
As shown in Figure 5.20 and Figure 5.21, as the silicone backbone length
of the surfactant increases, the surface elasticity increases. The film tension
after expansion is determined by the overall surfactant mobility including
both bulk mobility and surface mobility. The larger surfactant will have a
lower surface mobility, which will lead to a lower overall mobility. The fact
that a larger surfactant has higher film elasticity suggests that the surface
diffusion is significant for a thin liquid film. As a result, a higher surface
tension gradient and lower surface mobility can be obtained for foams made
with these larger silicone surfactants. This will give a slower drainage rate
and smaller open cell window percentage in the cured polyurethane foam.
On the other hand, when the percentage of polyether is varied, the observed
trend is less clear.
The surfactant with very low film elasticity results in a collapsed foam due
to its low drainage retarding ability. The surfactant with reasonable elasticity
will result in a stable foam. When the polyether percentage of the surfactant
is too high (90%), the film elasticity test cannot be performed since the bubble

© 2004 by CRC Press LLCwww.ketabdownload.com

 11 2

10

1.5
Film Elasticity (dyne/cm)

Air Flow (cubic ft/min)

8
1
7

6
0.5

4 0
50 100 150
Backbone Length
FIGURE 5.20
Effect of silicone backbone length of silicone surfactant on dynamic film elasticity. (1 wt%
surfactant solution in Voranol 3137 as well as air permeability of the cured solid polyurethane
foam. Polyether percentage is kept constant at 75%.)

7 6

Foam Collapsed
6
5
Dynamic Film Elasticity (dyne/cm)

stable foam
region
5
Air Flow (cubic ft/min)

3
3

2
2

1 1
40 50 60 70 80 90
Percentage of Polyether
FIGURE 5.21
Effect of polyether percentage of silicone surfactant on surface tension of 0.2 wt% surfactant
solution in Voranol 3137 as well as air permeability of the cured solid polyurethane foam.
Silicone backbone length is kept constant at 60 (SiO) units.

© 2004 by CRC Press LLCwww.ketabdownload.com

 32

28
Drainage Rate

24
(A/min)

20

16

50 100 150
Siloxane Backbone
FIGURE 5.22
Effect of silicone backbone length of the surfactants on vertical film drainage rate 1 wt% surfac-
tant solution in Voranol 3137.

on the tip of the capillary is highly unstable. This surfactant will cause
polyurethane foam collapse since the surfactant is not sufficiently surface
active and no surface tension gradient can be obtained.
As discussed above, a film with a higher elasticity has a higher surface
tension gradient along the film to retard drainage. As a result of the slower
drainage rate, a thicker film will be obtained. The higher the film elasticity,
the thicker the average cell window and the lower the percentage of ruptured
cell windows. As discussed in the previous sections, the remaining cell
window thickness in the cured solid foam is measured for foams made with
different surfactants and different film elasticities. Surfactants with different
backbone lengths (MD85D¢15M and MD51D¢9M) are used since they result in
foams with the same cell size. Figure 5.17 shows the thickness distribution
of the remaining cell windows. MD85D¢15M has a higher film elasticity value
(7.1 dyne/cm) than that of MD51D¢9M (5.2 dyne/cm). Polyurethane foam
made using a surfactant with a high elasticity has thicker films than that of
a foam made using a surfactant with a low elasticity. Also, a different dis-
tribution is observed: foam with a high elasticity has a higher film density
toward the thicker films, while foam with a low elasticity has a higher film
density toward the thinner films.
The short-time foam stability for different surfactant solutions can be eval-
uated by obtaining the actual thinning rate of thin liquid. Since there are
many vertical lamellae in a foam, where gravity plays an important role, a
vertical film drainage test is performed.34 The results of the vertical film
drainage test for different surfactant solutions are shown in Figure 5.22.

© 2004 by CRC Press LLCwww.ketabdownload.com

 TABLE 5.5
Formulations
Unit: pphpa
Formulations A1 A2 A3

Voranol 3137b 100 100 100

Water 4.0 4.0 4.0
Dabco 33LVc 0.30 0.30 0.30
Dabco T-9d 0.32 0.24 0.16
Silicone surfactant 1.0 1.0 1.0
Voranate T-80 index 110 110 110
a Parts per hundred parts polyol (by weight).
bA 1000 equivalent weight triol containing 13% ethylene oxide, 87% propy-
lene oxide and all secondary hydroxyl groups.
c 33% triethylenediamine in dipropylene glycol.
d Stabilized stannous octoate.

As shown, the drainage rate slows as the backbone length of the surfactant
is increased. In vertical films, the drainage flow toward the Plateau borders
creates a stretching motion along the surface. A surface tension gradient
along the film will form to counteract the stretching drainage flow. The
higher the surface tension gradient a surfactant can establish, the slower the
drainage will be. It is shown that the surfactant with the longer silicone
backbone can provide a higher surface tension gradient, and a slower film
drainage rate will be obtained.
Silicone surfactant must provide short-time stability for the reacting poly-
urethane foam system before the matrix material solidifies. Such stabilization
is realized by establishing a surface tension gradient along the film surface
and reducing the rate of drainage. The ability to build a surface tension
gradient is characterized by the dynamic film elasticity measured using the
film rheometer. Both bulk mobility and surface mobility have significant
effects on the film elasticity. The higher the overall surfactant mobility
(mobility parameter M), the lower the film elasticity will be. Drainage rate
and foam stability are observed to increase as the film elasticity increases.
It is shown that a surfactant with a longer backbone will have a lower overall
mobility, higher film elasticity, and slower drainage rate and will result in
thicker films and a lower ruptured window percentage. When the silicone
content is varied, the surfactant performance will vary significantly. Surfac-
tants with very high silicone content give very low film elasticity and low
foam stability. Surfactants with very low silicone content are not surface
active enough and also give low foam stability. As a result, a surfactant with
either a high or a low silicone content results in a collapsed foam system.

5.4.2 Other Formulation Ingredients: Bulk Rheology Effect

Cell opening can be manipulated by additives that change the reaction
kinetics. As previously shown, urea phase separation and extensional thin-
ning of the bulk liquid are key in triggering cell opening. Thus the degree

© 2004 by CRC Press LLCwww.ketabdownload.com

 104
0.16 parts T-9
0.24 parts T-9
0.32 parts T-9
G′ (Pa)

103

Fz Sudden
Decrease
(Cell opening)

102
40 50 60 70 80 90 100 110 120
Time (s)
FIGURE 5.23
Modulus rise for foams made with different tin catalyst levels.

of cell opening is effectively controlled through proper balance of the blow-

ing and gelling reactions, which in turn determine the bulk rheology at the
time of urea phase separation. Accordingly, it is common to adjust formula-
tion variables such as polyol structure, catalyst package, and water level in
order to control cell opening. The effect of tin catalyst on cell opening is
shown in the example here. The tin catalyst concentration was changed
within the formulation, which is given in Table 5.5. This change will alter
the chemistry of the system. The same surfactant type and concentration
were employed at all times. Air flow and cell opening data for all the for-
mulations are given in Table 5.6.
In this series, the time to cell opening onset (t0) changes significantly with
formulation. This shows that the foaming chemistry has a significant influ-
ence on the cell opening process. One important observation is that the cell
opening always happens a few seconds after the modulus sudden increase
(Table 5.6 and Figure 5.23).

TABLE 5.6
Time to Cell Opening, G¢ at Cell Opening and Air Flow
t0 tr G¢ at t0
Formulation # (sec) (sec) (Pa) Air Flow (CFM)
A1 62 53 102 0.07
A2 70 64 82 0.32
A3 80 74 60 4.60
t0: Time to normal force sudden decrease (internal cell opening time).
tr: Time to modulus sudden increase (blow-off time).
Sudden decrease in normal force was not observed because the foam collapsed.
Air flow could not be measured due to foam collapse.

© 2004 by CRC Press LLCwww.ketabdownload.com

 101

0.16 parts T-9

100
Air Flow (CFM)

0.24 parts T-9

0.32 parts T-9

10−1

10−2
50 60 70 80 90 100 110
Modulus at Onset of Cell Opening (Pa)

FIGURE 5.24
Air flow vs. G¢ cell opening time for foams made with different tin catalyst levels.

G¢ variation at the cell opening onset is higher at high tin catalyst levels.
This is due to the different reaction kinetics as well as the different blowing
reaction to gelling reaction ratio. As more tin catalyst is added to the system,
the overall reaction kinetics will be faster as observed from the modulus
development profiles. More importantly, the relative ratio between the gel-
ling and blowing reactions will increase as the tin catalyst level increases.
This implies that the foam matrix will be more crosslinked at the cell opening
time when the tin catalyst level is high. As a result, a higher modulus at cell
opening time is observed for foams made with higher tin catalyst levels.
This leads to lower air permeability since a higher gelling reaction increases
the strength of the foam matrix and less film rupture will occur. Figure 5.24
shows this trend.
It is shown that the rheology of the matrix material plays a very important
role in controlling the degree of cell opening. The higher the bulk viscosity
at cell opening time, the more stable the films are and the lower the open
cell percentage will be.

5.5 Summary
Flexible polyurethane foam is produced in a process with complex chemistry
involving simultaneous gas generation, polymerization, and urea phase sep-
aration. The cell opening event, which occurs suddenly during the process
by which foam for slabstock applications is produced, is especially remark-
able. This process leads to a stable, but porous, material with desirable

© 2004 by CRC Press LLCwww.ketabdownload.com

mechanical properties. We have comprehensively reviewed and tested pre-
viously proposed explanations for the cell opening event, including theories
originating from both the aqueous foam and polyurethane literature. A new
mechanism was proposed that explains both the resulting distributions of
cell window structures and results of in situ measurements such as online
rheometry. In addition, the roles of silicone surfactant and catalysts in the
cell opening event were illustrated.
Spontaneous film rupture mechanisms were eliminated since the observed
window film thickness in polyurethane foams is too large. Furthermore, a
uniform distribution of surfactant within cured foams was demonstrated;
thus cell opening also was not explained by phase separation of surfactant.
Urea phase separation leads to a heterogeneous morphology and sudden
viscosity increase of the foamed liquid. However, the mere presence of solid
urea particles will not cause cell opening since the particles will not dewet
from the liquid. This was illustrated in a liquid foam stability experiment
with added polyurea particles. It was further shown that after phase sepa-
ration, the foamed liquid exhibits strong extensional thinning. As this change
in rheology occurs during foam expansion, it leads to cell windows that are
torn, partially open, or completely missing. Such features were previously
observed by microscopy. The silicone surfactant impacts the process by
retarding film drainage and increasing film elasticity. Using light interference
microscopy, the cell window thickness distribution was measured for foams
employing a series of designed surfactants. Film elasticity, an indicator of
the drainage rate, was measured on surfactant/polyol mixtures through use
of a film rheometer. Structure–performance correlations were thus made with
respect to surfactant structure. Surfactants producing films with higher elas-
ticity resulted in thicker films within foams and more closed cells. Catalysts
were shown to further affect the cell opening event by modifying the bulk
rheology of the foamed liquid at the time of cell opening.

References
1. R.M. Herrington and K. Hock, Flexible Polyurethane Foams, The Dow Chem-
ical Company, Midland, MI, 1991.
2. E.A. Blair, Resinography of Cellular Plastics, ASTM STP, 414, 84, 1967.
3. J.H. Saunders and K.C. Frisch, Polyurethanes: Chemistry and Technology, Applied
Science Publishers, London, 1962.
4. G. Woods, Flexible Polyurethane Foam: Chemistry and Technology, Applied Science
Publishers, London, 1987.
5. R.A. Neff, Reactive Processing of Flexible Polyurethane Foam, Ph.D. thesis,
University of Minnesota, 1995.
6. A.J. Vrij, Disc. Faraday Soc., 42, 44, 1958.
7. A. Sharma and E. Ruckenstein, Langmuir, 3, 760, 1987.
8. B.V. Derjaguin and Y.V. Gutop, Kolloidn. Zh., 24, 431, 1962.
9. I.B. Ivanov, K.D. Danov, and P.A. Kralchevsky, Coll. Surf. A, 152, 161, 1999.

© 2004 by CRC Press LLCwww.ketabdownload.com

10. P.K. Prud’homme and S.A. Khan, Foams, Marcel Dekker, Inc., New York, 1988.
11. S. Ross and G. Nishioka, Coll. Polym. Sci., 255, 560, 1977.
12. G.C. Frye and J.C. Berg, J. Coll. Int. Sci., 130(1), 54, 1989.
13. R.D. Kulkarni, E.D. Goddard, and P. Chandar, in Foams: Theory, Measurements
and Applications, R.K. Prud’homme and S.A. Khan, Eds., Marcel Dekker, Inc.,
New York, 1996.
14. J.B.M. Hudales and H.N. Stein, J. Coll. Int. Sci., 140(2), 307, 1990.
15. A.D. Nikolov, P.A. Kralchevsky, I.B. Ivanov, and D.T. Wasan, J. Coll. Int. Sci.,
133, 13, 1989.
16. A.D. Nikolov, D.T. Wasan, N.D. Denkov, P.A. Kralchevsky, and I.B. Ivanov,
Prog. Colloid Polym. Sci., 82, 87, 1990.
17. I.B. Ivanov, Coll. Polym. Sci., 252, 982, 1974.
18. L.E. Scriven and C.V. Sternling, Nature, 187, 186, 1960.
19. I.B. Ivanov, Thin Liquid Films, Marcel Dekker, Inc., New York, 1988.
20. P. Spatel and C.W. Macosko, Poster Section, The Society of Rheology 74th
Annual Meeting, Minneapolis, October, 2002.
21. G.L.W. Rossmy, H. Schator, M. Wiemann, and H.J. Kollmeier, J. Cell. Plast., 17,
319, 1981.
22. C.B. Reese, in Protective Groups in Organic Chemistry, J.F.W. McOrnie, Ed., Ple-
num Press, New York, 104, 1973.
23. D.L. Schmidt, in Foams, R.K. Prud’homme and S.A. Khan, Eds., Marcel Dekker,
Inc., New York, 1996.
24. W. Grabowski and M.C. Desnier, J. Cell. Plast., 31(3), 39, 1995.
25. R.E. Stevens and S.A. Snow, Personal communication, 1995.
26. K. Yasunaga, R.A. Neff, X.D. Zhang, and C.W. Macosko, J. Cell. Plast., 32, 427,
1996.
27. C.W. Macosko, Rheology: Principles, Measurements and Applications, VCH Pub-
lishers, Inc., New York, 1994.
28. R.W. Whorlow, Rheological Techniques, Ellis Horwood Ltd., London, 1992.
29. S. Chatraei, C.W. Macosko, and H.H. Winter, J. Rheol., 25, 433, 1981.
30. X.D. Zhang, Role of Silicone Surfactant in Polyurethane Foaming Process, Ph.D.
thesis, University of Minnesota, 1998.
31. P.J.M. Baets and H.N. Stein, Langmuir, 8, 3099, 1992.
32. R.E. Jones and P. Fesman, J. Cell. Plast., 1(1), 200, 1965.
33. A.K. Malhotra and D.T. Wasan, Chem. Eng. Commun., 55, 95, 1987.
34. X.D. Zhang, C.W. Macosko, H.T. Davis, A.D. Nikolov, and D.T. Wasan, J. Coll.
Int. Sci., 215(2), 270, 1999.

© 2004 by CRC Press LLCwww.ketabdownload.com

6
Flexible Polyurethane Foams

J. Bicerano, R.D. Daussin, M.J.A. Elwell, H.R. van der Wal, P. Berthevas,
M. Brown, F. Casati, W. Farrissey, J. Fosnaugh, R. de Genova,
R. Herrington, J. Hicks, K. Hinze, K. Hock, D. Hunter, L. Jeng, D. Laycock,
W. Lidy, H. Mispreuve, R. Moore, L. Nafziger, M. Norton, D. Parrish,
R. Priester, K. Skaggs, L. Stahler, F. Sweet, R. Thomas, R. Turner, G. Wiltz,
T. Woods, C.P. Christenson, and A.K. Schrock

CONTENTS
6.1 Historical Background
6.1.1 Application and Markets
6.2 Basics of Chemistry
6.2.1 Overview
6.2.2 Polymerization Reaction
6.2.3 Gas-Producing Reaction
6.3 Foam Components
6.3.1 Polyol
6.3.2 Further Discussion of Copolymer Polyols (CPPs)
6.3.3 Isocyanates: Broad Overview
6.3.4 Isocyanates: In-Depth Review of TDI
Manufacturing Processes
6.3.4.1 Introduction
6.3.4.2 Nitration of Toluene to DNT
6.3.4.3 Hydrogenation of DNT to TDA
6.3.4.4 Phosgenation of TDA to TDI
6.3.5 Prepolymers
6.3.6 Isocyanate Index
6.3.7 Fillers
6.3.8 Water
6.3.9 Surfactants
6.3.10 Catalysts
6.3.11 Additives
6.4 Fundamentals of Foaming

0-8493-1728-2/04/$0.00+$1.50
© 2004 by CRC Press LLC

© 2004 by CRC Press LLC

www.ketabdownload.com
 6.4.1 Introduction
6.4.2 Bubble Nucleation
6.4.3 Bubble Growth
6.4.4 Bubble Packing
6.4.5 Cell Opening
6.5 Classical Physical Picture of Morphology
6.5.1 Heterophasic Nature of Flexible Polyurethane Foams
6.5.2 Analytical Tools for Characterizing Flexible
Polyurethane Foams
6.5.3 Essential Aspects of Classical Physical Picture
of Morphology
6.6 Research Frontiers in Microstructure and Property
Development
6.6.1 Synopsis
6.6.2 Key Aspects of Morphology Development during
Reactive Processing
6.6.3 Reaction Kinetics: Understanding Copolymerization
6.6.4 Phase Separation Kinetics: Understanding the Evolution
of Polymer Morphology
6.6.5 Solidification: Understanding the Arrest
of Phase Separation
6.6.6 Post-Reaction Analyses: Understanding Properties
in End-Use Applications
6.6.7 Summary: Overview of Polymer Microstructure and
Morphology Development
6.7 Foam Preparation
6.7.1 General Considerations
6.7.2 Slabstock Foams
6.7.3 Molded Foams
6.7.4 Carpet Backing Foams
6.8 Evaluation and Testing of Basic Foam Properties
6.9 Evaluation and Testing of Foam Durability
6.10 Control of Noise and Vibrations
6.11 Flammability
6.12 Recycling
6.13 Polyurethane Foams from Renewable Resources
6.14 Summary and Conclusions
References

6.1 Historical Background

Polyurethanes are chemically complex polymers, usually formed by the reac-
tions of liquid isocyanate components with liquid polyol resin components.

© 2004 by CRC Press LLC

www.ketabdownload.com
Although these products are all termed “polyurethanes,” other identifiable
chemical linkages may also be present (such as urea, allophanate, and biuret).
The fundamental addition polymerization reaction of diisocyanates with alco-
hols to produce high polymers was discovered in 1937 by Otto Bayer and
coworkers in the laboratories of the German I. G. Farben Industrie.1 From their
early work pulling “High quality fibers and bristles . . . from the polymer melt”
came the further research and experimentation that led to the first laboratory-
produced flexible polyurethane foam in 1941.2
The first pilot plant production of flexible polyurethane foam was
announced in Germany in 1952.3 Commercial production began in 1954.4,5
Using the German technology, numerous companies worldwide raced to
introduce polyurethanes commercially to fill many industrial and consumer
needs. The first foams were based primarily on the reaction of an aromatic
isocyanate with a polyester polyol. However, these foams proved unable to
withstand many in-use temperature and humidity conditions and often
failed by crumbling away.6,7 Better performance was obtained with foams
based on polyether polyols. These polyols, introduced by BASF Wyandotte
and The Dow Chemical Company in 1957, produced a flexible foam that was
less affected by hydrolysis and was more comfortable and considerably more
durable. By 1958, the combination of polyether polyols, new catalysts, and
new silicone-based surfactants made possible the so-called one-shot foam
technology, which paved the way for commercial quantities of more eco-
nomic foams with significantly improved physical properties.8
In the early 1960s, Dow was first to introduce ethylene-oxide-capped poly-
ols for flexible foam. The resulting faster reactivity (due to the presence of
primary hydroxyl groups) permitted the fast cure needed for hot cured and
high resiliency foams. Dow was also the first to produce amine-capped
polyether polyols. Continuing advances in polyol development and process-
ing by the polyurethanes industry have played an important role in the
expansion of the technological applications and markets for polyurethane
foams.
Concurrent with the advances in the development of polyols, whole fam-
ilies of diisocyanates have come on the global scene. By the early 1950s, the
two basic starting products were toluene diisocyanate (TDI) and diphenyl-
methane diisocyanate (MDI). These two isocyanates remain the most widely
used in the polyurethanes industry. However, many additional isocyanates
(including several aliphatic isocyanates) have also gained acceptance since
then for various applications.

6.1.1 Applications and Markets

The greatest advantage offered by polyurethanes is their versatility, both
in their finished product properties and in their ease of production and
application. Figure 6.1 is a product chart for flexible polyurethane foams.
With proper choice of isocyanate and polyol, products can be made with
© 2004 by CRC Press LLC
www.ketabdownload.com
properties ranging from the downy softness of very low-density flexible
foams to the high strength of elastomeric bubble-free castings. The phe-
nomenal success of polyurethanes in capturing applications previously
held by alternative cushioning materials is recorded in a large number of
references.9
Polyurethane foams can be generated from liquid components on the job
site by pouring, frothing or spraying. However, more commonly, foams are
produced in large quantities in plants designed specifically for the desired
type of foam. Foam formulations can be mixed by hand or by machinery
designed specifically for that purpose. They can be molded to shape or
produced in large bunstock form for later cutting and fabrication.
Their high mechanical strength allows these foams to endure rigorous
fabrication processes. Contemporary flexible polyurethane foams are also
resistant to moisture, odor formation, and normal cleaning procedures, and
are considerably more resistant to degradation than older latex-based foams,
and hence easily used in modern upholstery applications. Their potential
uses and multiplicity of properties are practically endless.
Such versatility has resulted in the rapid expansion of their markets. The
major markets are cushioning materials in furniture, bedding, carpet under-
lay, automotive, and packaging applications. With the very broad range of
available firmness (load-bearing capacity) and resiliency (including also

FLEXIBLE FOAM
POLYETHER PRODUCTION ISOCYANATES
POLYOLS

35% 60% 5%

PACKAGING
TRANSPORTATION FURNISHINGS
+ OTHER

AIRCRAFT FURNITURE
SEATING
HOME
TRIM BEDDING OFFICE
RAILWAY
OUTDOOR

AUTOMOTIVE FULL FOAM CARPET ELECTRONICS

CRASH PADS TILE

HEAD-/ARMRESTS PADDING
TRIM HOME
SEATING INSTITUTIONAL BACKING SPORT MATS
SOUND INSULATION RECREATIONAL
VIBRATION DAMPENING
CARPET BACK HOME TOYS
SUN VISORS INSTITUTIONAL
INSTRUMENT PANELS
ENERGY MANAGEMENT
FILTERS
HEADLINERS

FIGURE 6.1
Applications of flexible polyurethane foams.

© 2004 by CRC Press LLC

www.ketabdownload.com
“specialty” foams such as semiflexible, semirigid, and integral skin foams),
these foams offer degrees of comfort to humans and protection to inanimate
objects not approached by any other single cushioning material.
Most flexible polyurethane foams reach consumers through their use in
products such as upholstered furniture, mattresses, automobiles, and cloth-
ing. Consumers also benefit from their use in all forms of public transpor-
tation and commercial seating. From a modest commercial start in the early
1950s, these foams have grown in sales volume to occupy the sixth position
among all major plastics sold today. The growth of the industry was slow
in the beginning. It gained momentum gradually as the advantages of these
new foams became more evident. The total global sales volume was only
about 400 million lb in 1960. The industry exceeded the 1.5 billion lb mark
by 1970. It continued growing to over 9 billion lb by 1995.10

6.2 Basics of Chemistry

6.2.1 Overview
Polyurethane chemistry is based on the reactions of isocyanates with active
hydrogen-containing compounds. Isocyanates have one or more of the
highly reactive -N== C== O (isocyanate) group. This group reacts readily
with hydrogen atoms that are attached to atoms more electronegative than
carbon. Of the many compounds fitting this description, those that are of
primary interest for polyurethane-forming reactions are listed in Table 6.1.
The normal reactions of isocyanates involve addition to the carbon-
nitrogen double bond. A nucleophilic center from an active hydrogen-
containing compound attacks the electrophilic carbon. The active hydrogen
atom then adds to the nitrogen atom. Electron-withdrawing groups attached
to the isocyanate molecule increase the reactivity of the NCO group toward
nucleophilic groups. Electron-donating groups reduce reactivity. Thus, in
most reactions, aromatic isocyanates are more reactive than aliphatic
isocyanates. Steric hindrance on the isocyanate and/or on the active
hydrogen compound affects the reaction. Extensive reviews of isocyanate
reactions are available.11
Flexible polyurethane foam formation involves many ingredients. The two
competing reactions (polymerization and gas-producing reaction) are dis-
cussed below.

6.2.2 Polymerization Reaction

Polymerization occurs via an addition reaction (Structure 6.1) between an
isocyanate and an alcohol. Its heat of reaction is roughly 24 kcal/mol of
urethane.12 Depending on the choice of starting materials, the R and R¢

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 6.1
Active Hydrogen Compounds Ordered by Decreasing Isocyanate
Reactivity
Typical Reaction Rate
Active Hydrogen Compound Structure Uncatalyzed at 25oC
Primary Aliphatic Amine R-NH2 100,000
Secondary Aliphatic Amine R2-NH 20,000–50,000
Primary Aromatic Amine Ar-NH2 200–300
Primary Hydroxyl R-CH2-OH 100
Water H-O-H 100

O
Carboxylic Acid || 40
R-C-OH

R
Secondary Hydroxyl | 30
R-CH-OH

O
Ureas || 15
R-NH-C-NH-R

R
Tertiary Hydroxyl |
R-C-OH 0.5
|
R

O
Urethane || 0.3
R-NH-C-O-R

O
Amide || 0.1
R-C-NH2

groups may also contain isocyanate or isocyanate-reactive groups respec-

tively. When extended to polyfunctional reactants, this reaction provides a
direct route to crosslinked polymers.

O
−∆
R−N C O + R′−CH2−OH R−N−C−O−CH2−R′
−

Isocyanate Alcohol H Urethane

(Structure 6.1)

© 2004 by CRC Press LLC

www.ketabdownload.com
 The hydrogen on the nitrogen atom of the urethane group is capable of
reacting with additional isocyanate to form an allophanate group (Structure
6.2).

O O
∆
R−N−C−O−CH2−R′ + R−N C O R−N−C−O−CH2−R′

−
−

H 110°C C O

− −
H−N
R
Urethane Isocyanate Allophanate
(Structure 6.2)

The formation of allophanate is a high temperature, reversible reaction. If

actually formed in normal flexible foams, the allophanate linkage can
crosslink the polymer further. The catalysts generally used in foam formu-
lations do not promote this reaction, and temperatures greater than 110˚C
are necessary for significant allophanate formation. Other comments on the
likelihood of allophanate formation can be found elsewhere.13

6.2.3 Gas-Producing Reaction

A polymer must be expanded or “blown” by the introduction of bubbles and
a gas to make a foam. A convenient source of gas is the CO2 produced from
the reaction of an isocyanate group with H2O (Structure 6.3). The total heat
release is roughly 47 kcal/mol of H2O.

O
R−N C O + H−O−H R−N−C−OH
−

H
Isocyanate Water Carbamic Acid

R−NH2 + CO2 + HEAT

Amine Carbon
Dioxide

(Structure 6.3)

The intermediate product of this reaction is a thermally unstable carbamic

acid which decomposes spontaneously to an amine and CO2. Diffusion of CO2
into bubbles previously nucleated in the reacting medium causes expansion
of the medium to make a foam. Further reaction of the amine with additional
isocyanate gives a disubstituted urea, as shown in Structure 6.4.
© 2004 by CRC Press LLC
www.ketabdownload.com
 O
∆ II
R-N=C=O + R' - NH R-N-C-N-R'
2
H' H'
Isocyanate Amine Disubstituted
Urea
(Structure 6.4)

Again, if the isocyanate and amine molecules are polyfunctional, a

crosslinked polymer will result. Another conceptual method of crosslinking
the polymer is by the reaction of a hydrogen from the disubstituted urea
with a free isocyanate group to form a biuret linkage (Structure 6.5).14 How-
ever, since this reaction is also reversible, there is some debate about whether
allophanates and biurets actually exist in the final polyurethane foam.15

O ∆ O
R−N−C−N−R′ + R−N C O R−N−C−N−R′

−
−

−
H H 110°C H C O

− −
H−N
R

Disubstituted Isocyanate Biuret

Urea

(Structure 6.5)

Blowing can also be achieved by physical addition of a low-boiling non-

reactive liquid to a foam formulation. Historically, the most commonly used
physical blowing agents were the chlorofluorocarbons, urethane-grade
methylene chloride, and trichloroethane. The vaporization of these liquids
by heat from the exothermic reactions produces gas molecules which diffuse
into the nucleated bubbles and contribute to foam expansion. Nowadays,
CO2 produced from the reaction of isocyanate with H2O (see above) and/or
CO2 introduced physically as a liquid is the predominant blowing agent.
Small amounts of methylene chloride are still used in South America, but
even there liquid CO2 is the auxiliary (physical) blowing agent of choice.
Liquid CO2 has the advantages of being cheap, readily available, familiar,
and derived from existing industrial processes; and simply supplementing
the CO2 that is already generated by the reaction of water with isocyanate
without introducing yet another type of compound into the formulation.

6.3 Foam Components

Flexible polyurethane foam recipes normally contain many ingredients
selected to aid in achieving the desired grade of foam.16 Table 6.2 lists the
© 2004 by CRC Press LLC
www.ketabdownload.com
 TABLE 6.2
Formulation Basics for Flexible Polyurethane Foams
Component Parts by Weight
Polyol 100
Inorganic Fillers 0–150
Water 1.5–7.5
Silicone Surfactant 0.5–2.5
Amine Catalyst 0.1–1.0
Tin Catalyst 0.0–0.5
Chain-Extender 0–10
Cross-linker 0–5
Additive Variable
Auxiliary Blowing Agent 0–35
Isocyanate 25–85

most common ingredients and typical concentration ranges used in the pro-
duction of flexible polyurethane foam.

6.3.1 Polyol
Polyol is a source of hydroxyl or other isocyanate reactive groups. Processing
and properties of the resultant foam can be markedly influenced by the
choice of starting polyol structure. Polyether-type polyols are used in 90%
of the flexible foams produced today. Most of the remaining 10% are made
from polyester-type polyols.
The polyether polyols may be broadly grouped into the following categories:

• Polyoxypropylene diols.
• Polyoxypropylene triols.
• Polyoxypropylene tetrols and higher analogs.
• Ethylene oxide capped diols, triols, tetrols, and higher analogs.
• Random and block copolymers of the above in which the polyol is
made with both ethylene oxides and propylene oxides. When the
oxides are fed as a mixed feed, the products are termed heteropolyols.
• Graft or “copolymer” polyols (CPPs) which contain stable disper-
sions of a solid particulate polymeric phase in the liquid polyol
phase. Because of the critical importance of CPP in flexible polyure-
thane foam technology, the preparation of CPPs will be discussed in
some detail in the next subsection.
• Crosslinkers, which are typically short-chain polyfunctional mole-
cules added to increase load bearing and/or initial foam stability.

There are so many polyol variations partially because what works well in
a given commercial plant or geographic location may not work well in another
location with seemingly identical conditions. The exact polyol composition

© 2004 by CRC Press LLC

www.ketabdownload.com
will be selected to tailor the foam for a given application and to allow its
processing in commercial plants that differ widely in capabilities.
A polyether polyol is the polymeric reaction product of an organic oxide
and an initiator compound containing two or more active hydrogen atoms.
The active hydrogen compound in the presence of a base catalyst initiates
ring opening and oxide addition. Oxide addition is continued until the
desired molecular weight is obtained. If the initiator has two active
hydrogens, a diol results. If a trifunctional initiator such as glycerine is used,
oxide addition produces chain growth in three directions, and a triol results.
These reactions are exothermic. The addition of propylene oxide, for
example, releases approximately 22 kcal/mol. Many initiator compounds,17
as well as blends of these compounds,18,19 are used for the manufacture of
flexible foam polyols. In practice, side reactions may preclude the production
of a pure polyol from a chosen initiator. Many additional references20
discussing the synthetic chemistry, manufacturing processes, and properties
of polyether polyols are available.
Polyether polyols can be “filled” with other organic polymers to produce
viscous white to off-white fluids useful in making foams of higher hardness
than can be obtained by using the unmodified polyols alone. Filled polyols
also aid foam processing by improving cell openness. The filler polymer is
normally formed by in situ polymerization of monomers in a polyol base,
through free radical or step addition processes.21 The successful preparation
of stable colloidal dispersions requires particles to have some type of steric
stabilization to prevent flocculation.22 This stabilization is typically provided
by a “grafting” process in which either the base polyol or a special added
stabilizer molecule copolymerizes with the monomers to form a copolymer.
There are two important composition classes of graft copolymer polyols.
Chain-growth copolymer polyols are obtained by using steric stabilization to
prepare dispersions in nonaqueous media via free radical polymerization21,23
and include styrene-acrylonitrile copolymers which are the most widely used
graft copolymer polyols. Step-growth copolymer polyols include the Poly-
harnstoff dispersion (PHD) polyols20hh,24 which are dispersions of polyurea
particles in conventional polyols, and the polyisocyanate polyaddition (PIPA)
polyols20hh,25 which are dispersions of polyurethane particles in conventional
polyols. In addition to these two major classes of copolymer polyols, filled
polyols based on dispersions of cured epoxy resins,26,27 polyisocyanurate par-
ticles,28 and melamine-based copolymers29,30 in conventional base polyols
have also been reported.

6.3.2 Further Discussion of Copolymer Polyols (CPPs)

CPPs are used both to increase flexible foam load bearing and as processing
aids to help in cell opening. CPPs are made by free radical polymerization
of ethylenically unsaturated monomers in a polyol. The most commonly
used monomers are styrene (S) and acrylonitrile (AN). An important CPP
property is its load-bearing efficiency which is related to the volume fraction
of rigid polymer in the polyol.31,32
© 2004 by CRC Press LLC
www.ketabdownload.com
 The viscosity of the final CPP is related to the volume fraction of polymer
and the initial viscosity of the carrier polyol.33 The particle size distribution
determines the maximum packing volume fraction of particles and also the
maximum attainable solids level for a liquid CPP.34 CPPs can be made by a
semi-batch or a continuous process.
In the semi-batch process, all ingredients are fed to a batch reactor with a
certain ramp profile. In the continuous process, all ingredients are fed at a
constant rate into a continuously stirred tank reactor. The semi-batch process
produces rather narrow and uniform particle size distributions. In the con-
tinuous process, the particle size distribution is broader, resulting in lower
final CPP viscosities at a given particle volume fraction as a result of having
a higher maximum particle packing volume fraction.
The amount and structure of the SAN-polyol grafts created during the
process also play an important role in the rheology of the final CPP. When
the particles are very small, the protective layer of polyol surrounding the
particles begins to contribute to the final CPP viscosity, and the cell opening
efficiency is also affected adversely. Particles that are too large (with diam-
eters on the length scale of a foam strut) cause foaming instabilities. It is
important to design CPPs with optimum particle size distributions.
The key to the “high solids” CPP process is the design of the stabilizer
system in order to deliver the optimum particle size distribution in combi-
nation with the lowest possible levels of the unwanted serum phase polymer.
The free radical polymerization of S and AN in polyol automatically results
in some “natural” SAN-polyol grafting, through the presence of some unsat-
urated polyol species and by direct radical attack of the polyol chain. Inert
organic solvents and/or chain transfer agents are being used to limit the
formation of unwanted SAN-polyol grafts in the serum phase. The use of
solvents limits the SAN molecular weight through particle swelling and
generally results in smoother particle surfaces. The use of a chain transfer
agent reduces the SAN molecular weight and is especially useful for the
“high solids” CPPs. A chain transfer agent allows control over the molecular
weight of the polymer present in both the serum phase and the particle. The
major part of the reaction happens in the solid SAN particles. Controlling
the molecular weight of the SAN phase allows some control over the fine
balance between solid phase and serum phase polymerization. (A higher
SAN molecular weight will result in a higher reaction rate. The grafting
reaction only happens in the serum phase. Too fast a polymerization will
limit the grafting reaction and thus the production of new material to support
the stabilization of the growing particle surface.)
Generally, the particle stabilizer precursors (macromer) are polyether poly-
ols containing some sort of unsaturation induced by reacting a polyether
polyol with an ethylenically unsaturated monomer. There are many exam-
ples of these materials in the patent literature; for instance, the coupling of
vinyltrialkoxy silane with a polyether polyol35 or creating a vinyl-capped
polyol via unsaturation through the use of dimethyl m-isopropylbenzyliso-
cyanate.36 It has also been shown37 that a stabilizer can be a high molecular
© 2004 by CRC Press LLC
www.ketabdownload.com
weight polyether polyol without induced unsaturation. The natural grafting
by radical chain transfer to the polyether polyol and through the allylalcohol
alkoxylates present in the polyether polyols is adequate for obtaining suffi-
cient SAN polyol grafts to support the dispersion polymerization.
A “seed” is often used to increase the efficiency of the stabilizer. The seed
is a low solids CPP produced by reacting higher levels of macromer with
the S and AN in the polyol carrier. The seed product made in this way
contains very small particles in combination with a high level of SAN-polyol
grafts. These SAN-polyol grafts are especially important to support the dis-
persion polymerization process.
The first CPPs were produced by using azo-type initiators such as 2,2’-
azobis(isobutyronitrile). However, because of the difficulties in handling
these materials and the degradation products that are formed, the industry
is moving towards organic peroxy-type initiators. Peroxide initiators have a
stronger hydrogen abstraction tendency towards the polyether polyol, result-
ing in more SAN grafting and smaller particle sizes but also in somewhat
higher final CPP viscosities. In moving from azo to peroxide initiators, the
macromer and seed structure need to be adapted somewhat to compensate
for the higher natural grafting levels.
Chain transfer agents are used to obtain stable dispersions at higher solids
levels. The polymerization rate inside the particles will be a function of the
conversion. Very high rates are seen at high conversions. The Trommsdorff
effect (reduced radical termination by low chain mobility) changes the del-
icate balance between the grafting reaction in the polyol phase and the solid
SAN phase. Using a chain transfer agent reduces the molecular weight inside
the particles and the final monomer conversion.
In general, higher monomer conversion leads to higher final CPP viscosi-
ties. This effect is thought to be related to the decreasing solubility of the
SAN polyol grafts during the residual monomer removal step (forcing sol-
uble SAN-grafts into the solids phase). Another effect is that the molecular
weight of the SAN-polyol graft moves towards lower SAN molecular
weights (the SAN molecular weight is proportional to the monomer concen-
tration) as the conversion increases, making the SAN-polyol grafts more
soluble and increasing the final CPP viscosity.
The use of polystyrene (PS) made by the direct free radical polymerization
in the polyol results in high final CPP viscosities, poor monomer conversion,
and more difficult particle stabilization, related to the high solubility of PS
in the polyol and the lower reactivity of the S radical compared with the AN
radical. CPPs based entirely on acrylonitrile result in darkly colored prod-
ucts. The low solubility of polyacrylonitrile (PAN) in polyol, combined with
the high reactivity of the AN radical, makes it possible to produce a PAN
polymer polyol without using an additional macromer. Unfortunately, the
final PAN CPP has a rather high viscosity (natural grafting). The vast major-
ity of the CPPs are based on the copolymers of styrene and acrylonitrile,
which provide better CPP characteristices than can be obtained by using
either the PS or the PAN homopolymers.
© 2004 by CRC Press LLC
www.ketabdownload.com
 The pre-mixing of all ingredients is important in both the continuous
process and the semi-batch process.38 The concentration of unreacted mono-
mers at any point in the reactor should be low because the solubility of the
formed particles in the unreacted monomers causes particle destabilization.
The reactors for both type of process are normally equipped with a (pitched)
turbine blade agitator and the reaction mixture is fed into the high shear
zone. The reaction temperatures range from 90 to 140˚C.

6.3.3 Isocyanates: Broad Overview

The isocyanate permits the NCO groups to react with functional groups from
the polyol, water, and crosslinkers in the formulation. All isocyanates used
in industry today contain at least two NCO groups per molecule. The most
commercially viable methods of producing isocyanates involve amine phos-
genation, as shown in Structure 6.6:

H Cl
−

R−NH2 + COCL2 R−N: C O R−N C O + 2HCl

−

H Cl
Amine Phosgene Isocyanate Acid

(Structure 6.6)

The reaction is normally carried out in a chlorinated aromatic solvent

chosen for its utility in removing excess phosgene in later purification
steps.11d The process is complicated by several side reactions.39 Many reviews
of isocyanate manufacturing processes are available.40
More than 90% of flexible polyurethane foams are based on TDI. The global
production capacity for TDI exceeds 1.5 million metric tons. The two most
important isomers of TDI are shown in Figure 6.2. An 80:20 blend ratio of the
2,4-TDI and 2,6-TDI isomers, respectively, is obtained by the standard man-
ufacturing process. Other mixture compositions, as well as the pure isomers,
can be obtained by modifications of this process and/or by purification.
Detailed reviews of typical TDI manufacturing processes are available.41
Because of the central role of TDI in the fabrication of flexible polyurethane
foams, a review of these processes will also be provided below.
2,4-TDI is more reactive than 2,6-TDI. Within a given isomer, steric hin-
drances affect the relative reactivities of the positions.11b,42 Varying the isomer
ratios can have dramatic effects on both polymer and foam properties.43
Higher-load-bearing foams can be prepared by using the 65:35 blend ratio
of the 2,4-TDI and 2,6-TDI isomers, respectively. Catalyst adjustments nec-
essary to overcome the lower reactivity of this blend ratio have been
described.44 In some systems, adding a modified TDI or a polymeric diphe-
nylmethane diisocyanate (PMDI) to the TDI reduces the hardness of the foam
© 2004 by CRC Press LLC
www.ketabdownload.com
 CH3

NCO

NCO
2,4 Toluene Diisocyanate

CH3

OCN NCO

2,6 Toluene Diisocyanate

FIGURE 6.2
Isomers of TDI.

by interfering with the alignment of the polyurea hard segments and the
effective degree of hard and soft phase separation.
Various forms of MDI are used in high resiliency, semiflexible, and micro-
cellular foams. MDI is prepared from aniline, formaldehyde, and phosgene.
The undistilled crude reaction mixture resulting from the phosgenation of the
polyamine mixture contains varying amounts of di, tri, and higher functional
isocyanates. Market demand for the pure MDI isomers dictates that the reac-
tions be optimized for maximum production of two-ring material. Detailed
reviews of typical MDI manufacturing processes are available.45 PMDIs are
differentiated by viscosity, functionality, and reactivity and have lower vapor
pressure than TDI. Blends of PMDI with TDI are also used, as mentioned above.

6.3.4 Isocyanates: In-Depth Review of TDI Manufacturing Processes

6.3.4.1 Introduction
In 1848, Wurtz first synthesized two aliphatic isocyanates (methyl and ethyl
isocyanate) by reacting the alkyl sulfates with potassium cyanate.46 Shortly
thereafter, Hoffman prepared an aromatic isocyanate by pyrolysis of the
symmetrical diphenyloxamide.47 Use of phosgene for conversion of an amine
to an isocyanate was first carried out by Hentschel48 and has become the
process of choice for production of the major commercial aromatic isocyan-
ates including TDI.11a,c,49 Current commercial TDI production involves a
sequence of three process steps which will be reviewed below: (a) nitration
of toluene to dinitrotoluene (DNT), (b) reduction of DNT to toluene diamine
(TDA), and (c) phosgenation of TDA to TDI.

© 2004 by CRC Press LLC

www.ketabdownload.com
6.3.4.2 Nitration of Toluene to DNT
Most commercial toluene nitration processes use a mixture of acids (HNO3/
H2SO4). The nitronium cation is formed by reaction of nitric acid with the
stronger acid (sulfuric acid). Nitration of toluene follows as a bimolecular
reaction involving aromatic electrophilic substitution with the nitronium
cation, NO2+ .50 Under typical mixed acid nitration conditions (65–98% HNO3,
80% H2SO4), toluene is substituted as 57.5% ortho, 4.0% meta, and 38.5% para,
as shown in Structure 6.7.51

NO2
HNO3 HNO3
H2SO4 H2SO4
NO2 NO2

(Structure 6.7)

Subsequent reaction leads to a mixture of DNT isomers. Tillet52 determined

the isomeric mixtures of DNTs produced from each of the three nitrotoluene
compounds. The o-nitrotoluene reacts with nitric acid/sulfuric acid to give
2,4-DNT (66.5%), 2,6-DNT (33.2%), and 2,3-DNT (0.3%). The m-nitrotoluene
upon further nitration leads to 3,4-DNT (55%), 2,3-DNT (25%), 2,5-DNT
(18.6%), and 2,4-DNT (1.4%). The p-nitrotoluene gives rise to 2,4-DNT (99.7%)
and 3,4-DNT (0.3%).
The DNT product distribution thus includes isomers in which the nitro
groups are ortho and para to each other and which upon hydrogenation
produce o-T0DA and p-TDA. In the final TDI product mixture, the meta
isomers are preferred. The o-TDA and p-TDA compounds are considered
byproducts and are normally removed by distillation from the product mix-
ture. Since nitration of m-nitrotoluene is most responsible for the formation
of the undesirable o- and p-DNT derivatives, attenuation of m-nitrotoluene
in the first nitration of toluene is crucial to reduce formation of the undesired
o- and p-DNTs.
Nitrocresols and nitrophenols are other impurities formed in the nitration
of toluene. Several proposed mechanistic routes may account for their for-
mation.53 The cresols are very reactive with respect to aromatic electrophilic
substitution and readily nitrated under conditions of mixed acid nitration
to the di- and trinitrocresols. These phenolic impurities are commonly
removed by aqueous extraction. The wastewater process stream is then
treated by purification processes.54
In addition to the classical mixed acid method for toluene nitration, new
methods such as the use of lanthanide(III)triflates as catalysts are under
development.55

© 2004 by CRC Press LLC

www.ketabdownload.com
6.3.4.3 Hydrogenation of DNT to TDA
The overall product reaction pathway for DNT reduction is shown in Struc-
ture 6.8. In typical commercial processes, the isomeric mixture of DNTs is
comprised of roughly 20% 2,6-DNT and 80% 2,4-DNT. The hydrogenation
product mixture contains a similar isomeric ratio (~20% 2,6-TDA and ~80%
2,4-TDA).

(Structure 6.8)

Several processes have been commercialized for both gas and liquid phase
hydrogenation of nitro compounds. In general, DNT hydrogenation is medi-
ated by either Raney nickel or noble metal (e.g., palladium or platinum) on
carbon catalysts. In reductions of DNT with palladium on carbon, it has been
proposed that iron facilitates reduction by improving dispersion of the metal
phase and aiding in hydrogen transfer.56
It has been reported that the use of tetrahydrofuran in the hydrogenation
of DNT over Raney nickel results in less byproduct formation than the use
of alcoholic solvents.57
Musolino et al. developed methods to quantify major intermediates in
catalytic hydrogenation of 2,4-DNT58 and proposed a general mechanism
based on the analytical results.59 Westerterp et al.60 proposed a mechanism
with two parallel reaction pathways.

6.3.4.4 Phosgenation of TDA to TDI

Amines react with excess phosgene to give an initial reaction mixture com-
prised largely of carbamyl chlorides and amine hydrochlorides. Undesired
secondary reactions are initiated when amines react with the product (iso-
cyanates) to form ureas. Amine phosgenation reaction yields are not only
reduced by the formation of ureas, but the yield losses are compounded by
further reactions of the urea byproducts. The overall reaction is illustrated
in Structure 6.9.
TDI purity, particularly with respect to reduction of chloride compounds
and high molecular weight isocyanate byproducts, was the major goal of
early work on TDI process improvements. A TDA phosgenation process
incorporating the use of low initial reaction temperatures and high phosgene
to TDA mole ratios was thus developed.61

© 2004 by CRC Press LLC

www.ketabdownload.com
 CH3 CH3

NH2 NCO
COCl2
+ 4HCl

NH2 NCO
toluenediamine toluenediisocyanate
(isomeric mixture) (isomeric mixture)
(Structure 6.9)

Many highly efficient mixing devices were designed by the 1960s to

improve the initial reaction of TDA with phosgene, typically using an initial
temperature lower than 60∞C at the point of mixing of the TDA solution
with the phosgene solution to produce a mixture of TDA dihydrochloride
and carbamyl chlorides in a first reaction stage.62 In these early inventions,
completion of the reaction was then achieved by heating the initial reaction
mixture to 100–200∞C in a second reaction stage. Latourette and Johnson63
and Denton et al.64 showed that good yields may also be obtained at signif-
icantly higher initial reaction temperatures by optimizing some of the other
process parameters.
Many inventors have patented impingement, jet, and venturi mixers that
combine the amine and phosgene solutions at high flow velocities to reduce
byproduct formation.65 Others have proposed the use of mechanical mixing
devices to reduce mixing times in the first stage of the reaction.66 In general,
mechanical mixing devices provide high shear at the point of contact of the
amine solution with the phosgene solution. High shear mixing may also be
attained by using a multistage rotary pump.67 High speed mixing devices
may be combined with mixing temperatures below 90∞C to improve overall
reaction yields.68
Preparation of TDA dihydrochloride prior to reaction with phosgene elim-
inates the need for highly efficient mixing and reduced initial reaction tem-
peratures. Relative rates of reaction of TDA dihydrochlorides with phosgene
were investigated by Soos and Petnehazy.69 The TDA dihydrochloride reac-
tion is heterogeneous due to the insolubility of the amine salts. The reaction
of TDA dihydrochloride with phosgene results in improved overall TDI
yields.70 Overall TDI yields as high as 96% have been reported by this so-
called salt phosgenation process.
Partial conversion of the TDA to a mixture of hydrochlorides may also be
carried out to improve yield relative to phosgenation of the free TDA.71
Immediately prior to the initial mixing of the phosgene solution with the
amine solution, the TDA hydrochloride salts may also be formed in situ by
addition of HCl to the phosgene solution.72 The addition of HCl to prevent
byproduct formation has also been utilized in the second stage of phosgena-
tion.73 The formation of TDA dihydrochloride and subsequent phosgenation
to TDI can be carried out in the absence of a reaction solvent.74

© 2004 by CRC Press LLC

www.ketabdownload.com
 Salts other than hydrochloride salts may also be formed prior to reaction
of TDA with phosgene in order to improve reaction yields. For example,
sulfur dioxide salts of TDA have been prepared and phosgenated to provide
TDI with improved yields.75
Dimethylformamide has been reported to catalyze the phosgenation reac-
tion of TDA.76 The use of tetra-N-substituted thioureas has been shown to
decrease the reaction time of the phosgenation reaction.77 Other N,N-disub-
stituted amides such as dimethylacetamide have been used as weak bases
to remove HCl and decompose the carbamyl chlorides to reduce overall
reaction times.78
Staged-temperature phosgenation of TDA in dichlorobenzene with non-
ylphenol polyethylene oxide increased the TDI yield from 73 to 84%.79 The
typical solvents for commercial phosgenation of TDA are chlorohydrocar-
bons, such as chlorobenzene and dichlorobenzene.80 Numerous reaction sol-
vents have been reported to facilitate the reaction of TDA with phosgene
and improve the overall yield of TDI.81
Gas-phase phosgenation of TDA has also been described by several
groups.82 Its advantages include the small amounts of excess phosgene (less
than one equivalent) required and the relatively high overall yields. Follow-
ing phosgenation of the diamine to the diisocyanate, phosgene and solvent
are removed to provide a crude TDI. The crude TDI is then distilled to
provide a colorless TDI product of relatively low viscosity.

6.3.5 Prepolymers
Prepolymers are liquid intermediates between the monomers and a final
polymer. They are formed by reacting an excess of either polyol or isocyanate
so that the product is still a liquid and contains the reactive functional group
of the reagent in excess.
The use of prepolymers offers several advantages. Since the vapor pressure
of a free isocyanate is reduced according to its mole fraction, there is a
reduced vapor hazard because of increased molecular weight of the isocy-
anate component. Increased viscosity leads to better mixing of foam com-
ponents. There may be improved hard segment formation arising from pre-
reaction of some chain extenders that are incompatible with normal foam
components. Pre-reacting some of the ingredients under precisely regulated
conditions may also result in improved control of desired or undesired side
reactions.
Hydroxyl-tipped prepolymers were popular in the early days of
polyurethanes. Most prepolymers in use today are isocyanate tipped. A strict
prepolymer is formed when just enough isocyanate is added to react with all
available hydroxyl sites. If excess or residual free isocyanate monomer is
present, the product is called a quasi-prepolymer. Just like the other types of
isocyanates, the final prepolymer can react to form all of the normally
expected structures. A typical procedure for prepolymer preparation has been
described.83 The key variables involved in producing low-free-isocyanate
prepolymers have been discussed.84 Stable, non-gelling prepolymers can only
© 2004 by CRC Press LLC
www.ketabdownload.com
be prepared when the stoichiometry, moisture, base contamination, and
temperature are all closely controlled. Many additional reviews and foaming
procedures are available.85

6.3.6 Isocyanate Index

The amount of isocyanate required to react with the polyol and any other
reactive components of the formulation is calculated in terms of theoretically
stoichiometric equivalents. This amount may then be adjusted upwards or
downwards, depending on the foam system and the required final proper-
ties. The amount of isocyanate used relative to the theoretical equivalent
amount is known as the isocyanate index.

Actual amount of isocyanate used

Isocyanate Index ∫ 100 ◊
Theoretical amount of isocyanate required

Increasing the isocyanate index over a reasonable range has a pronounced

effect in increasing the hardness of the final foam. This effect is related
directly to increased covalent crosslinking resulting from the more complete
consumption of the isocyanate reactive sites caused by the presence of excess
isocyanate groups. The isocyanate index normally ranges from 105 to 115
(where foam hardness can be controlled readily and safely) in flexible slab-
stock foam formulations. There is, however, a point beyond which foam
hardness does not increase and other physical properties begin to suffer.

6.3.7 Fillers
Finely divided inert inorganic fillers are often added to foam formulations
to increase density, load bearing, or sound attenuation. All other foam phys-
ical properties are generally sacrificed. Typical results have been described
in the literature.86 Depending on the nature of the filler, the overall cost of
the final foam may be reduced.
Typical fillers include the many grades of barium sulfate and calcium
carbonate. Care must be taken to dry the fillers or to know the precise water
content available and factor that information into the formulation calcula-
tions. Normal concentrations used are from 20 to 150 parts per hundred parts
of polyol. The fillers are heavy and will settle out of a polyol mixture unless
constant agitation is used. Due to their abrasive nature, mineral fillers can
also increase wear on machinery components. Calcium carbonate fillers have
a significant effect on the gelling reaction and necessitate a rebalancing of
the blowing and gelling reactions. A general review of fillers for plastics
applications is available.87

6.3.8 Water

Water is a source of active hydrogens. Only demineralized water should be

used for foam production. Isocyanate reacts with water to provide CO2 gas
© 2004 by CRC Press LLC
www.ketabdownload.com
(blowing agent) and polyurea molecules. The CO2 gas diffuses into nucleated
bubbles and aids foam expansion. The polyurea molecules enter into and
contribute to the properties of the final polymer. The compatibility of water
with various polyether polyols has been studied.88
The amount of water used in the formulation is one of the most important
variables at the disposal of a flexible polyurethane foam manufacturer who
wants to optimize the properties of a foam for a given application. The typical
effects of the water level in the formulation on the morphology, density, and
stiffness (modulus) of a flexible polyurethane foam will be discussed and
illustrated later. It suffices to say, at this point, that if all other factors are
kept constant, the use of a larger amount of water in the formulation nor-
mally results in an increased foam modulus because of the increasing volume
fraction of the polyurea-rich hard phase, and a decreased foam density
because of the generation of a larger amount of the CO2 blowing agent.

6.3.9 Surfactants

Almost all flexible polyurethane foams are made with the help of silicone-
based nonionic surfactants. Surfactants lower the surface tension. They emul-
sify incompatible formulation ingredients. They promote the nucleation of
bubbles during mixing. They stabilize the rising foam by reducing stress
concentrations in thinning cell walls. They counteract the defoaming effect
of any solids added (such as fillers) or formed (such as precipitated polyurea
structures) during the foaming reaction.
The stabilization of cell walls is the most important of these functions. By
doing this, the surfactant prevents the coalescence of rapidly growing cells
until those cells have attained sufficient strength through polymerization to
become self-supporting. Cell coalescence would otherwise lead to total foam
collapse. Surfactants also help to control the precise timing and the degree
of cell opening.
The judicious choice of a surfactant or surfactant combination for any given
foam requires literature review, supplier consultation, and small-scale foam-
ing tests.89 Many additional extensive discussions of surfactant production
methods, surfactant selection, influence of surfactant type and amount on
foam structure and properties, and other aspects of surfactant technology
are available in the literature.90

6.3.10 Catalysts

Virtually all commercially manufactured flexible polyurethane foams are

made with the aid of at least one catalyst. Among the many classes of
compounds investigated, the amines and the organometallics have been
found to be the most useful. Various combinations of catalysts are used to
establish an optimum balance between the chain propagation (isocyanate
with hydroxyl) reaction and blowing (isocyanate with water) reactions. The

© 2004 by CRC Press LLC

www.ketabdownload.com
polymer formation rate and the gas formation rate must be balanced so that
gas is entrapped efficiently in the gelling polymer and the cell walls develop
sufficient strength to maintain their structure without collapse or shrinkage.
Catalysts are also important for assuring completeness of reaction or “cure”
in the finished foam.
The selection of a catalyst system for a particular grade of foam is a
complex task requiring literature review, supplier consultation, and small-
scale foaming tests.91,92 Further refinement is usually needed at full commer-
cial production scale.
Tertiary amines are the most commonly used flexible polyurethane foam
catalysts. Generally regarded as blowing catalysts, most amines also offer
some contribution to the polymerization reaction.93 The catalytic activity of
amines is due to the presence of a free electron pair on the nitrogen atom.
Steric hindrance about the nitrogen atom and the electronic effects of sub-
stituent groups are the main factors influencing the relative catalytic activity
of various amines. In some foam systems, combinations of various amines
are used in an attempt to balance the polymerization and blowing reactions
so that the foaming process can be adequately controlled. The type and
concentration of amine catalyst(s) can be selected to satisfy process require-
ments such as cream times, rise profiles, gel times, and even cure of the outer
surface skin. The choice of amine may also affect the foam properties such
as air flow and load bearing through influences on the primary and second-
ary foam reactions.94 In general, the requirements for good catalytic activity
include that the catalyst should be a strong nucleophile capable of attacking
the carbon of the isocyanate group, capable of readily forming an active
hydrogen amine complex, and soluble in water (forming stable hydrogen
bonds with water). Detailed reviews of amine catalysis are available.95
The polymerization reaction between the isocyanate and a polyol is pro-
moted by organometallic catalysts.96 Of the many available metals, tin com-
pounds such as stannous octoate are the most widely used. These compounds
act as Lewis acids and are generally thought to function by interacting with
basic sites in isocyanate and polyol compounds.
In recent years, a growing desire for better in-mold flowability and faster
curing along with a need to balance foam reactivities in the dual-hardness
process has lead to the development of delayed action catalysts.97 These are
compounds that are not very active at room temperature. They become
effective when the initial reaction exotherm warms up the foaming mass.
The most common approach is to use a tertiary amine salt in a suitable
solvent such as a low-molecular-weight glycol or water. Tertiary amine salts
with organic acids such as carbonic, formic, acetic, 2-ethylhexanoic, and lactic
acids have been introduced. Catalyst activity, phase separation, and metal
corrosion problems have been reported when acid-blocked amines are for-
mulated with normal amine catalysts in the presence of water. Other types
of delayed action catalyst include the thermosensitive catalysts whose activ-
ities increase dramatically with increasing temperature.

© 2004 by CRC Press LLC

www.ketabdownload.com
6.3.11 Additives
Various additives may be incorporated into a foaming system to impart
specific desired properties. A general review of additives for plastic products
is available.98 The following are the more common types of additives used
in flexible polyurethane foams: colorants,99 UV stabilizers,100 fire retar-
dants,101 bacteriostats,102 plasticizers, cell openers, antistatic agents,103 com-
patibilizers,104 and auxiliary blowing agents.105

6.4 Fundamentals of Foaming

6.4.1 Introduction
Foams are three-dimensional agglomerations of gas bubbles separated from
each other by thin sections of the host medium. For finished open-celled
flexible polyurethane foams like that shown in Figure 6.3, the void areas are
artifacts of expanded gas bubbles introduced into the reacting medium early
in its existence. In classical discussions of foams, physical chemists identify
cell walls and struts as lamellae and Plateau borders. Polyurethane foam
chemists usually call them windows and struts. Foams can be further char-
acterized as having a very large surface area per unit volume and a widely
variable cellular structure. The properties of a foam result from both the
geometry of the cells and the properties of the polymer. The formation of a
flexible polyurethane foam is an intricate process using unique hardware,
multiple ingredients, and at least the two simultaneous reactions discussed
above. Successful foam preparation requires the formation, growth, and
stabilization of gas bubbles in the reacting medium.

FIGURE 6.3
Open-celled flexible polyurethane foam. (Original magnification
50.)

© 2004 by CRC Press LLC

www.ketabdownload.com
6.4.2 Bubble Nucleation
The development of bubbles within a liquid is called nucleation. For poly-
urethane foams, the gas in the bubbles has several potential sources; namely,
dissolved gases in the liquid reactants, low-boiling liquids added as auxiliary
blowing agents, gases such as CO2 produced by the reactants, and gases
released from thermally decomposable additives. In addition, air or other
gases may be entrained into the system by mechanical mixers or frothing
devices. The number and size of any entrained bubbles are important to
foam performance since gas from the other sources may preferentially diffuse
into these sites.
The development of the fundamental understanding of bubble dynamics
began with Young106 and Laplace107 in the early nineteenth century. Plateau
is also credited with a major share of our knowledge about foams.108 An
excellent review of bubble dynamics has been published by Bikerman.109
Foams can be described conceptually in terms of either a dispersion model
or a condensation model. In a dispersion model, the gas would be present
in a large separate phase such as in the head space of a vessel. With sufficient
added work, aggregate molecules of the gas can be entrained and dispersed
uniformly throughout the continuous liquid phase. A common example is
the meringue topping on a lemon pie. (Meringue is the French word for the
foam formed by whipping air into a mixture of egg whites and sugar.) In a
condensation model, soluble gas molecules are forced to condense to form
a bubble as the solute is dissolved in a continuous liquid phase. Again, with
added work, the soluble gas molecules of a sufficiently supersaturated liquid
may be caused to aggregate and condense into a bubble within the contin-
uous liquid phase. An interesting apparent example of this model is the foam
that forms on a glass of beer.
See Yazdi and Beckman110 for a review of nucleation theory in the context
of bubble nucleation in polymer mixtures. Nucleation requires an input of
energy. For example, the impossibility of self-nucleating CO2 from a super-
saturated solution of polyol and TDI was demonstrated by Kanner and
Decker.111 A bubble of CO2 in a liquid will differ from an equal number of
gas molecules in solution by its possession of enough surface energy to
produce that bubble as a new phase. Foams can, therefore, be characterized
as being thermodynamically unstable until other events (such as polymer
gelation) provide permanent stabilization.
LaMer112 has reviewed how the introduction of various interfaces within
a gas-saturated liquid phase can greatly facilitate nucleation. Major changes
in the number of cells per unit volume and the average cell size can result
from apparently small changes in dissolved gas content. Bessette and Sund-
strom have shown that, for a given volume of foam, a larger number of cells
results in a more viscous and more stable foam.113 The impact of cell count
on foam physical properties has been discussed extensively.114
In practice, both entrainment and supersaturation are done by foam man-
ufacturers. Some bubble air into the polyol day tank to make an entrained

© 2004 by CRC Press LLC

www.ketabdownload.com
gas/polyol mixture of reduced density which is then pumped into the mix-
head. Others simply ensure that the polyol is at equilibrium with dry air of
atmospheric pressure and use the pressure drop after high pressure impinge-
ment mixing to cause bubble nucleation. In low pressure stirred mixing, air
can be injected into the mixhead, and cavitation on the low pressure side of
a multipin mixer will induce nucleation. The polyols must be near saturation
with air to produce foams with high cell count (small cell size).

6.4.3 Bubble Growth

Producing spherical bubbles in a desired liquid is only the first step towards
arriving at a useful foam. The initial mixture of reaction components has a
low viscosity which would result in an unstable foam without input of
additional energy. The possible fates of a foamed system include the floccu-
lation and/or sedimentation of the bubbles (possibly followed by a return
to phase-separated polymer and gas phases) as alternatives to the desired
polymer gelation that produces a stable dry foam. Unless permanent stabi-
lization occurs soon after formation, the foam disintegrates. In polyurethane
foams, the energy for bubble growth comes from the exothermic and gas-
producing reactions. Factors promoting bubble expansion include the pro-
duction of CO2 gas, the volume expansion work provided by an increasing
temperature, and bubble coalescence.
An important consequence of the existence of a surface free energy in gas
bubbles is the presence of a pressure difference across the curved gas-liquid
interface. This pressure difference under a concave curved meniscus gives
rise to the well-known capillary elevation effect on a liquid in a small tube.
If the curved liquid surface fully encloses a volume of gas, a bubble results.
The pressure excess DP of the gas in the bubble is given by the Laplace
equation, where g is the surface tension of the liquid and rc is the critical
radius of the bubble:

2g
DP =
rc

It is evident from this equation that the pressure inside a bubble is inversely
proportional to the radius of the bubble. In real foams, there is always a
distribution of bubble sizes so that the pressures in different bubbles will
not be the same. This will lead to the diffusion of gas molecules from regions
of higher pressure (small bubbles) to regions of lower pressure (large bub-
bles). The rate at which diffusion proceeds will be proportional to the pres-
sure difference, the permeability, and the thickness of the liquid film
separating bubbles of unequal size. Therefore, each bubble will grow or
shrink depending on its diameter and the diameters of the bubbles in its
environment. The rate at which a small bubble shrinks and disappears then
depends mainly on its radius and the permeability of the gas through the
liquid separating adjacent bubbles.
© 2004 by CRC Press LLC
www.ketabdownload.com
 The net observation in a static system is that, with time, the average
bubble size in a polydisperse system will increase while the total number
of bubbles will decrease. A common example of this event can be found in
a bubble bath, as illustrated by Clark and Blackman.115 Similar data were
presented by Smith.116 Frensdorff reported the same phenomena in a model
polyurethane foam system.117 For actual polyurethane-based foams, there
is some argument that diffusion between bubbles is not a major factor in
bubble growth118 because that mechanism is expected to be very slow com-
pared with the rate of gas diffusion from the gas-saturated liquid phase
into the bubbles.
The Laplace equation shows that a low liquid surface tension favors low
pressure differences between bubbles of different sizes. Lowering the system
surface tension is then expected to lead to better bubble stability and smaller
average cell size.
If coalescence were the only growth mechanism, fine-cell foam would not
exist. Coalescence taken to the extreme would result in one bubble. The
mutual approach of two bubbles requires pushing aside the liquid phase
between them. The viscosity of this moving liquid affects the rate at which
the liquid layer thins. Ideally, the viscosity of a thick section of the liquid
would be low enough to allow the bubbles to come together quickly and
then become high enough as a thin film to overcome coalescence. If the thin
film viscosity could be made high enough, coalescence would not occur.

6.4.4 Bubble Packing

The bubbles in a polyurethane foam grow with continued reaction. As their
volume fraction exceeds 74%, they lose their spherical shape and begin
distorting into polyhedra. Matzke reviewed over 300 years of literature
debating the theoretically preferred and actually observed polyhedral
shapes in real foams.119 Early polyurethane researchers described these poly-
hedra as being approximate pentagonal dodecahedra.114a However, pure
structures of that type have 12 five-sided membranes and would not be
expected to account for all available physical space. Real polyurethane
foams also show some four-sided and six-sided cells and may be better
approximated by Lord Kelvin’s 14-faced tetrakaidecahedron space filling
models.120,121 The real and approximate geometries are compared in Figure
6.4. Actual counts of the number of faces per cell closely match 14, as
predicted by Duchartre.122 A distribution of cell sizes is usually found.
Bubble stability during growth is a complex function of surfactant effects,
rate of gas evolution, viscosity, pressure, and presence of cell-disrupting
agents. As growing spherical cells are squeezed into polyhedra, the liquid
phase is initially redistributed between the tetrahedral interstices and the
bubble surfaces. With continued cell volume growth, cell wall thinning takes
place and polymer is drained from the lamellae into the unique structural
features known as Plateau borders (see Figure 6.5). There is strong curvature

© 2004 by CRC Press LLC

www.ketabdownload.com
 PENTAGONAL DODECAHEDRON

FIGURE 6.4
Examples of real foam geometry (left) and approximate cellular shape (right).

in the surface of a lamella at the edges where bubbles meet. The thinning of
the cell windows is promoted by gravity, capillary suction into Plateau
borders, and foam volume expansion. The effects of gravity and capillary
suction forces may be additive in some parts of a lamella and opposite to
each other in other parts.
The means by which foams stabilize themselves is a much debated subject.
The lack of understanding arises in part from the experimental and theoret-
ical difficulties of elucidating the intimate details of the molecular events
occurring in foams. Kitchener and Cooper reviewed the major theories of
foam stabilization.123 The four theories with reasonable applicability to flex-
ible polyurethane foams emphasize the importance of the surface viscosity,124
surface elasticity,125 film elasticity,126 and surface transport.127 These theories
are actually quite similar and complementary. They vary mainly in their
details. Implicit in each theory is the assumption that the rupture of a thin-
ning window is preceded by a pronounced local thinning at some spot on

LAMELLAE PLATEAU BORDERS

FIGURE 6.5
Schematic illustration of the structural features of a foam.

© 2004 by CRC Press LLC

www.ketabdownload.com
the film. This localized thinning results from drainage or from the action of
a sudden force (such as a thermal, pressure, or vibrational shock). Localized
thinning in a film will result in an expansion of the surface in that area,
causing dilution of the locally absorbed surfactant. The expanded film ele-
ment will thus possess a higher non-equilibrium surface tension.
To summarize, Figure 6.6 provides an idealized model for bubble nucle-
ation, growth, and packing in polyurethane foams. The cells are produced
by gas diffusion into bubbles that are nucleated or stirred into the system at
the time of mixing. Their size is related to the total amount of blowing agent
available, the exotherm, and the degree of stabilization (as by a surfactant).
The total number of cells produced is governed by the dissolved gas content
and the amount of gas stirred in. Ultimate stabilization of the polyhedral
framework occurs as a result of the chemical reactions that occur up to the
point of polymer gelation, when all film movement and expansion (except
that induced by outside physical forces and leading to anisotropy) ceases.
At this point, the foam may or may not possess enough open cells to make
it a useful product.

GAS PHASE
POLYOL PHASE
ISO PHASE
WORK, BUBBLE
NEUCLEATION

MIXED PHASE

RX TIME

SOLUTE CO2 DIFFUSION

GAS INTO BUBBLES
MOLECULES

RX TIME

SPHERICAL
BUBBLE GROWTH

RX TIME

TETRAHEDRAL
BUBBLE GROWTH

FIGURE 6.6
Model for bubble nucleation and growth.

© 2004 by CRC Press LLC

www.ketabdownload.com
6.4.5 Cell Opening
It is desirable to have a large fraction of the cell windows rupture through
a natural balance of physical and chemical forces in flexible polyurethane
foam production. A stable and nonshrinking foam may result if this happens
near the chemical gel point of the reacting system. When enough windows
are open, the foam will have less pneumatic character and will be more
suitable for comfort cushioning applications. It has also been postulated that
a more completely open-celled foam will have better fatigue resistance.128 As
Figure 6.7 indicates, cell opening is very complex and formulation depen-
dent. The major theories are that cell opening results from steam formation,129
urea precipitation,130 or balanced catalysis,131 or is due in part to some particle
effects.132 The literature also teaches that surfactants, through their influence
on bubble nucleation and stability, are important contributors to controlled
cell opening.133 The importance of viscosity in controlling the drainage of
polymer from cell windows has also been discussed.118,134 It has also been
reported that foam cells can be opened by the precisely timed application of
a physical shock such as the release of internal pressure from a mold.135
Classical concepts regarding cell opening in flexible foams have been
reviewed (see Saunders and Frisch136). Conceptually, the opening of cells
would occur to a large extent just as the foam reaches its full rise. At that
time, the foamed polymer would have reached high viscosity with very low
elasticity. The high viscosity would not permit the foam structural elements
to flow fast enough to expand and relieve the increasing cell gas pressure.
Low elasticity in the cell window membranes would likewise prohibit revers-
ible stretching of cell windows. Under such conditions, the cell window
membranes burst, leaving an interconnected open-celled network. The poly-
mer in the cell struts must have enough strength to endure this event and
prevent splits or catastrophic foam collapse.
Attaining this ideal natural balance is a complex task due to the large
number of formulation and equipment variables in common use. Shrinkage

POLYMER
PARTICULATES SURFACTANCY
FLOW
CATALYSIS

VISCOSITY
PHASE
SEPARATION

DIFFERENTIAL
NONSHRINK PRESSURES
FOAM
FUNCTIONALITY WINDOW
THINNING
FIGURE 6.7
Factors affecting cell opening in flexible polyurethane foam formation.

© 2004 by CRC Press LLC

www.ketabdownload.com
in a flexible foam results when a large population of cells with intact win-
dows remains at the end of foam manufacture. Just after production, such
cells are filled with hot, pressurized CO2. As the foam cools, the internal gas
pressure diminishes, and CO2 diffuses out of the cells roughly 15 times faster
than air diffuses into the cells.118,137 The net result is a partial vacuum in the
cell. When taken over a population of closed cells, this partial vacuum causes
the foam to shrink and thus lose its physical dimensions.
A unique method for the automatic detection of cell opening has been
described.138 A more accurate understanding of the cell opening mechanism
in conventional slabstock and hot-molded flexible polyurethane foams began
with the work of Rossmy et al.129 They observed that a high polymer molec-
ular weight was not present at the time of cell opening. Shortly before cell
opening, infrared spectroscopy showed negligible formation of urethane
bonds and sudden appearance of strong urea-related absorption.139 Coincid-
ing with that event, a visible phase separation occurred in the bulk liquid.128
The macrophase separation was identified as the precipitation of an agglom-
erated polyurea phase that was noted to increase the viscosity of the bulk
liquid momentarily to a level consistent with the onset of a chemical gel.
Small-angle x-ray studies revealed that the macrophase separation was
accompanied by a microphase separation attributed to the formation of a
bidentate hydrogen-bonded urea hard phase. Furthermore, cell opening was
noted to occur only after a critical amount of hard phase had formed. In
practical terms, the degree of cell openness can be regulated by adjusting the
catalyst package.
The ingredients of higher reactivity used in high resiliency slabstock and
molded foams make these foams more susceptible to having closed cells
and shrinkage. During foam rise, the primary hydroxyl capped polyols and
crosslinkers compete successfully with the water, resulting in higher levels
of urethane and urea bond (and thus molecular weight) formation. The
bidentate hydrogen bonded urea structures form to a lesser degree and
appear only as a microphase. The macrophase separation of urea domain
agglomerates does not occur and thus that potential cell opening mechanism
is not available. In molded foams, cell opening due to continuing volume
expansion is limited due to the restricted volume of the mold cavity. Cell
opening in high resiliency molded foams is thus more dependent on the
chemical methods that can be used to influence it.
In summary, cell opening in a flexible polyurethane foam is a complex
physical and chemical event. It is influenced by the kinetics of chemical
reactions and of cell growth, the type and concentration of surfactant, the
morphology of thinning cell window membranes, phase separation, the
presence of any cell-disrupting agents, and possibly contributions from
physical stresses as in mechanical or vacuum crushing. The proper combi-
nation of these factors at the right moment in a foam’s history leads to the
desired type of foam. Further discussions of cell opening mechanisms are
available.140

© 2004 by CRC Press LLC

www.ketabdownload.com
6.5 Classical Physical Picture of Morphology
6.5.1 Heterophasic Nature of Flexible Polyurethane Foams
The properties of a flexible foam depend both on the macroscopic cell geom-
etry and on the morphology of the polymer in its structural elements. Numer-
ous reviews of the effects of cell geometry on foam properties are available.114
Much of the following discussion of the effects of polymer morphology is
condensed from the literature.32,139,141
Figure 6.8 illustrates how the reactions occurring during the manufacture
of an unfilled conventional slabstock polyurethane foam lead to multiphase
morphology formation. The existence of a heterogeneous (phase-separated)
morphology in polyurethanes was suggested as early as 1966.142 Historically,
it has generally been conceptualized to consist of discrete hard phase rich
regions dispersed within a continuous soft polymer phase. It is also seen
that the amount of water in the formulation can affect the morphology
significantly, as discussed below.
The water-isocyanate reaction proceeds much faster than the polyol-
isocyanate reaction in the early stages of a slabstock foam,130,143 due in part
to the low reactivity of the secondary hydroxyls present in most slabstock-
grade polyols. This disparity in reaction rates leads to the formation of
polyurea oligomers. These oligomers continue to grow until they phase
separate from the continuous liquid phase because of their different
symmetry, thermodynamic incompatibility with liquid phase, and ability to
form strong hydrogen bonds.141m,n A detailed review of hydrogen bonding
in polyurethanes is available.141l In the lower portion of Figure 6.8, hard phase
rich regions are represented by clustered thick lines. These features are tied
into the matrix by covalent attachment to the polymer chains shown fading
off into the continuous phase.
Water is completely soluble in the formulation. It reacts with the isocyanate
to form urea species that are initially found totally in solution at low formu-
lation water levels. At a certain concentration or molecular weight, these
species suddenly separate out as the hard phase rich regions, indicated in
the lower right-hand portion of Figure 6.8. When the water concentration
exceeds the solubility of water in the blend of polyol, isocyanate and addi-
tives, discrete and stable droplets of water are formed in the mixture.141m
With continuing reaction, these regions of locally high water concentration
give rise to locally high urea oligomer concentrations. Once a certain molec-
ular weight is reached, these high urea oligomer concentrations separate into
macrophase urea agglomerates,141n and a third phase consisting of agglom-
erated polyurea hard phase rich regions (polyurea balls shown in the lower
left-hand portion of Figure 6.8) develops.
The continuous soft phase polymer results when the polyol-isocyanate reac-
tion proceeds past the chemical gel point. The covalent links between the soft
phase and the polyurea hard segment rich regions are urethane bonds.
© 2004 by CRC Press LLC
www.ketabdownload.com
 MODEL REACTIONS FOAM MODEL

O
1. R−OH + R′NCO R−O−C−N−R′

−
H
URETHANE

2. H2O + R′NCO R′NH2 + CO2

O
R′NH2 + R′NCO R′−NH−C−NH−R′
UREA

HIGH WATER LOW WATER

LEVEL LEVEL
POLYUREA BALL

CONTINUOUS
SOFT PHASE

HARD
DOMAIN

MORPHOLOGY MODELS
FIGURE 6.8
Conceptual representation of polyphase structure in conventional slabstock polyurethane
foams.

Evidence for the existence of this polyphase structure in slabstock poly-

urethane foams was reported first by Lidy et al.144 and later confirmed by
Turner and Wilkes.141h
Macrophase polyurea agglomerates (urea balls) are not found in typical
high-resiliency foams, where the primary hydroxyl-capped polyol and low
molecular weight crosslinker compete successfully with water in isocyanate-
consuming reactions. Water is also typically totally soluble in such polyol/
isocyanate/additive blends, so that polyurea oligomers are dispersed much
more uniformly in the continuous phase. The polyol reacts earlier with the
© 2004 by CRC Press LLC
www.ketabdownload.com
isocyanate to give prepolymer species. (Note that the solubilities and the
relative reactivities in typical high-resiliency foams are quite different from
those in the typical slabstock foams that were discussed in an earlier para-
graph of this subsection. Water has a higher solubility in the types of polyols
used in typical high-resiliency foams. These polyols also react faster than the
ones used in typical slabstock foam formulations because of the reduced
steric hindrance of their primary hydroxyl caps as well as the increased
compatibility due to the high local concentration of their ethylene oxide
repeat units.) These species react further with water and with the isocyanate
groups in growing hard segments, to couple these segments with the con-
tinuous phase and further prevent agglomeration and macrophase separa-
tion. Crosslinker molecules also react with hard segments to further build
molecular weight, disrupt symmetry, and thus hinder hydrogen bonding
between hard segments.
When the water level is increased, the sizes of both the hard segment rich
regions and the urea precipitates should increase. Both structures will con-
tribute to an increase in the load bearing capacity of the foam. The efficiency
and the permanence of load bearing enhancement may, however, differ
between these two types of structures.141h
An additional reinforcing phase may be present if the foam has been
formulated to contain particulate fillers. Variations in this polyphase mor-
phology can often account for the wide range of bulk physical properties
attainable in flexible polyurethane foams.

6.5.2 Analytical Tools for Characterizing Flexible

Polyurethane Foams
Many tools of analytical science are useful in investigating the morphologies
of flexible polyurethane foams.139,141 This section discusses some tools that
were especially useful in elucidating the “classical” features of the morphol-
ogy summarized above.
Fourier-transform infrared (FTIR) spectroscopy is used to study the for-
mation and structure of polyurethane foams.143–145 FTIR has been particularly
useful in studying the time-related aspects of foam reaction kinetics and
morphology development.146
Dynamic mechanical spectroscopy (DMS) measures the ability of a vis-
coelastic material to store and dissipate mechanical energy. It thus provides
information concerning stiffness and phase separation. Two excellent instruc-
tional reviews on the use of DMS with polyurethanes are available.147,148 The
application of DMS and related techniques to flexible polyurethane foams is
a fairly new endeavor,141e,149 while DMS has been used routinely to charac-
terize both linear and crosslinked polyurethane elastomers. Many other
references150 are also helpful in interpreting foam data. DMS of typical flex-
ible polyurethane foams made from three different conventional slabstock
foam formulations are provided in Figure 6.9. Despite the possibility of
formation of a more complex morphology, polyurethane foam DMS behavior

© 2004 by CRC Press LLC

www.ketabdownload.com
A. B.
1011 100
2 PPH WATER
3 PPH WATER
STORAGE MODULUS,

1010 5 PPH WATER

G′ (DYNES/CM2)

5 PPH WATER

TAN DELTA
109 10−1

3 PPH WATER
108 2 PPH WATER

107 10−2
−180−110−60 −10 40 90 140 190 240 290 −180−110−60 −10 40 90 140 190 240 290
TEMPERATURE, °C TEMPERATURE, °C

FIGURE 6.9
Examples of the dynamic mechanical spectra of flexible polyurethane foams, for foams made
from three different conventional slabstock foam formulations. The theoretical hard segment
weight fraction calculated from the foam composition under the assumption of complete reac-
tion of water, polyol, and isocyanate is 0.21 for H2O at 2 pphp (parts per hundred polyol), 0.26
for H2O at 3 pphp, 0.30 for H2O at 4 pphp, and 0.34 H2O at 5 pphp.

is usually strikingly similar to that of polyurethane elastomers. The shear

storage modulus G¢ is a measure of polymer stiffness. At low temperatures,
molecular movement is essentially frozen, G¢ is high, and the polymer has
nearly perfect elasticity. The loss tangent tan(d) is at a minimum, indicating
that nearly all of the energy stored in deforming the sample is recovered
when the stress is removed. A sharp transition occurs from a glassy state to
a rubbery state at a low temperature (typically, but not always, somewhere
in the range of –50 to –30°C). This transition is assigned to the onset of
molecular motion in the soft phase. In this region, G¢ decreases sharply with
a corresponding sharp increase in tan(d). The maximum in tan(d) is generally
quoted as the soft phase glass transition temperature (Tg). Tg is related to the
chemical structure and the molecular weight of the pure soft segment.151
Upscale shifts of Tg may result from the chain-immobilizing effects of a
second phase.151a,d,152 Then follows a “rubbery plateau,” extending to the
temperature range normally assigned to dissociation of microdomains of
polyurea hard block rich regions.153 This dissociation leads to a gradual
decrease in G¢. Increasing the water level increases G¢ in the plateau region
since a larger volume fraction of the polyurea-rich hard phase is formed
because of the increasing water-isocyanate reactions per unit volume. Finally,
sample degradation takes place via thermal scission of covalent chemical
bonds and G¢ drops precipitously.
Tg normally increases with decreasing molecular weight of the polyol since
soft segments are connected to hard segments, so that lowering the polyol
molecular weight increases the topological constraints on large-scale coop-
erative motions of soft segments. Tg can be modified further by the presence
of urea groups resulting from the water reaction. It can be broadened by
blending polyols. It can be both lowered and broadened by using plasticizers.

© 2004 by CRC Press LLC

www.ketabdownload.com
 Polyurethane “shape memory polymer” foams 151h,i are fascinating
examples of the unusual performance characteristics that can be obtained by
tailoring the soft segment so that its Tg is near or slightly above room
temperature. These foams are sometimes alternatively referred to as
“memory,” “viscoelastic,” “low recovery,” “slow recovery,” “low resiliency,”
or “pressure relief” foams. They manifest reversible switching behavior
(flexible´rigid transitions) with small changes in temperature near room
temperature. Body heat is normally sufficient to cause them to soften and
conform to the shape of the area that is being contacted. This behavior
provides a unique feel and absence of “fight-back,” making them the
preferred materials in many comfort cushioning and health care applications.
Since a polymer will absorb significant amounts of energy at its Tg, they also
have excellent energy absorption properties, resulting in their use in many
applications where it is important to dampen noise and vibrations. In
practice, making a flexible foam with a low resiliency near room temperature
requires the use of a ~700 molecular weight poly(propylene oxide) triol.
Triols with higher molecular weights will result in a lower Tg. Blends are
often used to widen the low resiliency response and delay the hardening of
the foam as the use temperature decreases. Plasticizers can also broaden the
glass transition and extend the low resiliency response to lower use
temperatures. Slow recovery is probably due to increased hydrogen bonding
in these foams, consistent with the observation that the effect is reduced or
totally lost if the foam is placed in a very humid area or is allowed to get wet.
Hard phase dissociation occurs at a higher temperature for a polyurea than
for the corresponding polyurethane. For a given type of hard phase, the
dissociation temperature increases with increasing segment length. The obser-
vation that the foam polymers exhibit fairly temperature-insensitive moduli
over a broad temperature range in DMS offers indirect support for the pres-
ence of a virtual as well as covalent crosslinked network. The typical position,
sharpness, intensity, and insensitivity of the glass transition argue for a well-
segregated phasic structure with little phase mixing. The co-continuity of the
hard phase with the soft phase (see below) is related to the coextension of
both phases throughout the specimen. This is to be distinguished from phase
mixing, which refers to the presence of segments of each type “buried” within
regions rich in the other phase.
Energy dissipation by hysteresis upon repeated deformation cycles is
related to tan(d) as measured by DMS. It increases with increasing phase
mixing and soft segment linearity. It decreases with increasing covalent cross-
linking in the soft phase.
Differential scanning calorimetry (DSC) is useful in confirming the soft
phase Tg and in evaluating the other thermal response characteristics of
foams.154
Optical microscopy (OM),155 scanning electron microscopy (SEM),156 and
transmission electron microscopy (TEM), can all be used to study the cellular
structures of foams. The maximum possible magnification and the amount

© 2004 by CRC Press LLC

www.ketabdownload.com
of effort required to prepare the samples and to run the experiments both
follow the order TEM>SEM>OM.
SEM and TEM are also used routinely in characterizing the polymer mor-
phology in polyurethane foams, just as in in many other types of multiphase
materials. On the other hand, the use of some of the more recently developed
techniques of microscopy to characterize the polymer morphology is grow-
ing gradually but is not yet routine. The most important such relatively new
tool is atomic force microscopy (AFM) which is being used to study the
polymer morphology in polyurethane foams90p,157 and elastomers.158
Wide-angle x-ray scattering (WAXS) is useful in detecting the possible
presence of crystalline ordering, especially in the hard phase.150a,151,159
Small-angle x-ray scattering (SAXS) provides information on the total con-
tent and average size of hard phase rich regions and the thicknesses of
“interphase” regions between hard and soft phases.160,161
Recent advances have made it possible to conduct dynamic (synchrotron)
SAXS, dynamic FTIR and dynamic rheometry simultaneously during poly-
urethane foam formation, allowing the simultaneous study of the kinetics of
polymerization, hard phase formation, phase separation, and stiffness devel-
opment.162,163 These recent advances at the frontiers of research on micro-
structure and property development in polyurethane foams merit a more
detailed discussion, which will be provided in the next section.

6.5.3 Essential Aspects of Classical Physical Picture of Morphology

Figure 6.10 summarizes the general events leading to a flexible polyurethane
foam. A simplified model of the types of morphological features found in

POLYOLS AIR
WATER
ADDITIVES
ISOCYANATES RAW MIXING & CREAM & RISING PHASE SEPARATION CURE
MATERIALS BUBBLE INITIAL FOAM BLOW OFF AND
NUCLEATION RISE GELATION

0 10 60–90 HOURS TO
TIME
SECONDS SECONDS DAYS

ENTRAINED COLOR VISCOSITY CELL OPENING IFD

PHYSICAL
BUBBLES CHANGE INCREASE MODULUS GROWTH SPREAD
EVENTS
COLLISIONAL WATER RX POLYOL RX DIFFUSIONAL
KINETIC
EVENTS CO SAT. MOLECULAR WT. INCREASE
HARD
DOMAIN UREA LAMELLAE
MORPHOLOGY
BUBBLE FORMATION PRECIP. RUPTURE
EVENTS
EXPAND
FIGURE 6.10
Summary of the general types of events that take place during flexible polyurethane foam
formation.

© 2004 by CRC Press LLC

www.ketabdownload.com
 ° (0.015 INCH)
d2 ~ 4,000,000 A

° (0.024 INCH)
d1 ~ 6,000,000 A
A TYPICAL
SINGLE CELL

° (0.001 INCH)
d3 ~ 5000,000 A
°
d4 ~ 5000−30,000 A

COPOLYMER °
d5 ~ 2000−5000 A
PARTICLE

°
UREA BALL d6 ~ 3000 A (Absent in HR foams)

°
d7 ~ 20−50 A °
d8 ~ 65−125 A

HARD DOMAINS

°
d9 ~ 30−50 A
d10 ~ <2 A°
POLYOL CHAIN
(END ON)
FIGURE 6.11
Simplified model for morphological features found in flexible polyurethane foams.

TDI-based flexible polyurethane foams is provided in Figure 6.11. We will

refer to the perspective provided by Figures 6.8 and 6.11 as the “classical”
model. It was developed mainly as the result of postmortem analyses; that is,
analyses where the morphologies of a foam and of the polymer within the
structural elements of this foam are characterized and the properties of the
foam are measured after the fabrication of the foam has been completed. For
the benefit of readers who may be interested in delving into the frontiers of
research, the next section will discuss the important new fundamental
insights that are being obtained by using dynamic (in situ) measurements to
study morphology and property development.

6.6 Research Frontiers in Microstructure and

Property Development
6.6.1 Synopsis
Flexible copoly(ether-urethane urea) foam is a multiphase cellular material.
It contains nonlinear segmented -(HmS)n- multiblock copolymer chains where
© 2004 by CRC Press LLC
www.ketabdownload.com
Hm represents a polyurea hard segment with a degree of polymerization m;
S is the poly(ether-urethane) soft segment; and n is the degree of polymer-
ization of the block copolymer.139,141e,i Polyether and polyurea blocks are
joined via covalent urethane links. The process by which the block copolymer
microstructure and resultant morphology evolve is reaction-induced phase
separation.164 The final morphology of the polymer is strongly dependent
upon the kinetic competition between the rates of polymerization, phase
separation, solidification, and the inherent connectivity between the phases.
Various morphologies can be accessed through control of the rates of poly-
merization and phase separation.
A refined description of the physical events during flexible polyurethane
foam formation, deduced from such in situ real-time experiments,163 is shown
in Figure 6.12. In this example, a typical water-blown copolyurethaneurea
foam is made by using a polyether polyol, TDI, or MDI as the isocyanate,
and water (which reacts with the isocyanate to form polyurea hard segments)
as the chain extender. MST denotes the microstructural phase transition, pNCO
is the fraction of conversion of the NCO groups, F is the volume fraction, N
is the degree of polymerization, c is the Flory-Huggins interaction parameter
quantifying the deviations of a mixture from ideal mixing, and the Bergh-

Increasing pNCO = 0,
isocyanate
T = 23°C
conversion, pnco
TMST
Tg − φ line
1/[χ (NHard + NSoft)]

binodal pNCO = 0.55 + 0.05

T = 95 + 6°C
spinodal

. GLASS

φ Hard ’ Volume Fraction, φHardφ Hard ’’

0 1
Berghmans
Polyurea point
Polyetherurethane rich phase
rich phase
FIGURE 6.12
A refined description of some of the key physical events that take place during flexible poly-
urethane foam formation, as deduced from in situ real-time experiments. A typical water-blown
copolyurethaneurea foam is made by using a polyether polyol, TDI, or MDI as the isocyanate,
and water (which reacts with the isocyanate to form polyurea hard segments) as the “chain
extender.” MST denotes the microstructural phase transition, pNCO is the fraction of conversion
of the NCO groups, F is the volume fraction, N is the degree of polymerization, c is the
Flory–Huggins interaction parameter, and the Berghmans point is the point at which the glass
transition temperature of the hard phase reaches the temperature of the reacting mixture.
© 2004 by CRC Press LLC
www.ketabdownload.com
mans point is the point at which the glass transition temperature of the hard
phase reaches the temperature of the reacting mixture. The onset of vitrifi-
cation takes place at the Berghmans point, so that the microphase separation
is intercepted and arrested by the vitrification of the phase that is richer in
hard segment. The product cN is similar to an effective temperature, so that
the effect of the rapid increase of N and hence decrease of 1/(cN) during
polymerization resembles the effect of applying a temperature quench.
Only the soft phase is assumed to be continuous, and discrete hard phase
domains (resembling those shown in Figures 6.8 and 6.11) are assumed to
be dispersed in this soft matrix, in the classical model of the morphology.
However, Figure 6.12 indicates that the MST occurs preferentially by spin-
odal decomposition, which can lead to hard phase co-continuity. (“Spinodal
decomposition” refers to the process of spontaneous and continuous phase
separation and growth that occurs in the unstable region of the phase dia-
gram beneath the spinodal curve in Figure 6.12. It results from the fact that
the mixture is unstable to infinitesimal concentration fluctuations in this
region, so that phase separation does not require the “nucleation” of finite-
sized regions of concentration inhomogeneity.) This relatively recent com-
prehension that the hard phase often manifests a significant amount of co-
continuity with the soft phase is important in understanding and predicting
the properties of polyurethane flexible foams and elastomers.
The interpretation of the mechanical properties of polyurethane elas-
tomers, which are dense materials, is not complicated by the need to decon-
volute the effects of the cellular structure as in foams. The effects of any hard
phase co-continuity are, therefore, generally easiest to identify unambigu-
ously in the observed elastic properties (moduli) of polyurethane elastomers.
See Ryan et al. for an example.165

6.6.2 Key Aspects of Morphology Development during

Reactive Processing
The reactive processing of a flexible copoly(ether-urethane urea) foam from
liquid monomers and oligomers involves a complex combination of chemical
and physical events.166 From room temperature and in less than 5 min, a
liquid mixture of low molar mass components is transformed into a material
with a supramolecular architecture of a “solid” foam. The copolymerization
proceeds via the simultaneous reactions of a diisocyanate with polyether
polyol and water. The copolymerizing mixture is blown into foam by the
cogeneration of CO2 gas evolved from the water-diisocyanate reaction. As
copolymerization proceeds, the core of the foam becomes insulated by the
surrounding polymerizing mixture, creating, in effect, a quasi-adiabatic envi-
ronment. In the initial 5 min of the copolymerization, the reactive mixture
experiences an exotherm of 110 to 140˚C, depending upon the specifics of
the composition of the formulation, an approximately 30-fold volume expan-
sion for a foam with a density of 32 kg/m3, and a ~106-fold viscosity increase.
These physical changes are illustrated in Figure 6.13.167
© 2004 by CRC Press LLC
www.ketabdownload.com
 A. B.
Time/s
160 1100

Elastic shear modulus°C

104

Foam density kg ar−3

140
825
Temperature°C

120
100
550 103
80
60 275
40 102
20 0
0 50 100 150 200 250 0 40 80 120 160 200
Time/s Time/s

FIGURE 6.13
(A) Adiabatic temperature rise profile and (B) the rapid changes in density and elastic shear
modulus during a typical flexible polyurethane foaming process. (Copyright 1994 American
Chemical Society. Reprinted with permission from Elwell, M.J., Mortimer, S., and Ryan, A.J.,
Macromolecules, 27, 5428, 1994.)

Investigation of the intimate and complex competition between copoly-

merization, macrophase separation and microphase separation, and solidi-
fication processes, requires the use of in situ, time-resolved analytical
methods.146,149,163,166–168 However, when such studies are conducted under
isothermal conditions and thus ignore the importance of the process thermal
history, they lead to erroneous results. The realization that analytical inves-
tigations must be conducted under forced-adiabatic conditions166a,168a,b was
crucial to making significant progress in understanding how the polymer
microstructure and architecture actually evolves.
Structure and morphology development can be divided into four categories:

1. The copolymerization itself, and the subsequent development of the

polymer microstructure and architecture.
2. Macrophase and microphase separation processes, and their associ-
ated kinetics.
3. The solidification processes of vitrification, chemical gelation, and
crystallization, and how they impact the development of the mac-
roscopic properties.
4. The polymer morphology of the foam, post-reaction, four weeks after
it has made. It is at this point that postmortem as opposed to in situ
analyses are conducted. Only by appreciating and understanding
how the prevailing morphology evolved (in situ and postmortem
investigations) is it possible to understand, control, and subsequently
improve the end use performance as defined in terms of certain foam
properties.
© 2004 by CRC Press LLC
www.ketabdownload.com
6.6.3 Reaction Kinetics: Understanding Copolymerization

Fast reactions put certain limitations on the techniques that can be employed
for their analysis.168c Time-resolved FTIR spectroscopy is the technique of
choice. The use of a forced-adiabatic reaction cell in combination with a FTIR
spectrophotometer was reported first by Artavia and Macosko168a and has
since been adapted by other research groups. A wide range of information
can be obtained from such experiments, as shown for example in Figure
6.14.163b Such investigations on slabstock foams have demonstrated that the

1.00

Isocyanate conversion, Pnco

0.80
Absorbance

0.60

0.40

2500
Wa 24002300 40 45 0.20
ve 30 35 6
25 0
nu 2200 20 / 1
mb 21002000 0 5 10 15 Time 0.00
er/
cm −
1
0 50 100 150 200 250 300 350 400

A. Reaction kinetics Time/s

N−H
Normalized absorbance, arbitrary units

N−H N−H
o o o
N−H N−H N−H
H o Solubic Urea Hydrogen-bonded
N (1715 cm−1) Urea (1654 cm−1)
o
Urethane
(1730 cm−1) NPnco=0.55+0.05
−
absorbance

27
24
21 1580 0 50 100 150 200 250 300 350 400
1620
Ti 1815 1660 −1 Time/s
m 12
e/ 9
6 1740
1700
be r/cm
10
num
3 1780
6 0
ve
Wa
B. Hydrogen bonding dynamics
FIGURE 6.14
Time-resolved FTIR data obtained under forced-adiabatic reaction conditions during the for-
mation of flexible polyurethane foam from a polyether triol (number-average molecular weight
of 6000 g/mol), water, and an isomeric mixture of MDI. (A) Copolymerization kinetics in terms
of the fraction of isocyanate conversion, pNCO, versus time, determined from the decay of the
NCO stretch band (2270 cm–1). (B) Hydrogen bonding dynamics, in relation to the microphase
separation transition, derived from the variations with time of the carbonyl absorbances asso-
ciated with soluble polyurea (1715 cm–1) (
) and hydrogen-bonded urea (1654 cm–1) (). (Copy-
right 1997 Wiley-VCH Verlag. Reprinted with permission from Stanford, J.L., Ryan, A.J., and
Elwell, M.J., in Processing of Polymers, ed. H.E.H. Meijer, vol. 18 of Materials Science and Technology:
A Comprehensive Treatment, VCH Publishers, Weinheim, Germany, 1997, chap. 9.)

© 2004 by CRC Press LLC

www.ketabdownload.com
formation of urethane and soluble urea moieties occurs simultaneously in
the initial stages of the reaction, as opposed to sequentially.
In early stages of copolymerization, the reaction mixture comprises poly-
ether, isocyanate-terminated polyether oligomers, water, polyurea
homopolymer, and some copoly(ether-urethane urea) block copolymer that
contains the urethane linking group. Such mixtures are similar to ternary
polymer blends.167,168j Reaction between the monomers and oligomers results
in both a molecular weight increase (N) and a concomitant change in the
values of the interaction parameters (c). The onset of microphase separation
of the polyurea hard segment sequence lengths occurs midpoint in the reac-
tion at an isocyanate conversion fraction of pNCO = 0.55 ± 0.05.166 FTIR spec-
troscopy is capable of providing much information on the reaction chemistry
and the sequence of chemical events, both prior to and after microphase
separation of the urea hard segments. However, it does not provide any
indication of the mechanism(s) of phase separation, the length scales of the
evolving morphology, or the extent of microphase separation.

6.6.4 Phase Separation Kinetics: Understanding the Evolution of Polymer

Morphology

The structure development kinetics must be probed by time-resolved

forced-adiabatic synchrotron SAXS measurements to investigate the real-
time morphological changes during fast bulk copolymerization. The use of
this technique to investigate the structure development kinetics in both
conventional and model MDI-based flexible copoly(ether-urethane urea)
foams was first reported by Elwell, Mortimer, and Ryan.167 This was later
followed by McClusky et al.,168d and more recently by Li et al.168i,j who
further developed the technique and the subsequent analysis and also
extended the interpretation and understanding of the scattering data that
are obtained.
Figure 6.15 illustrates the wealth of polymer microstructure and morpho-
logical information that can be obtained from in situ time-resolved synchro-
tron SAXS measurements.163b There have been many studies on the
morphology and properties of flexible copoly(ether-urethane urea) foams
based upon TDI and MDI. Different research groups have observed
particulate149b,166a,b,d,168d,h,169 and bicontinuous146c,157b,163a,c,166c,e,167,168i,j hard seg-
ment structures in a wide range of systems. The evidence suggests strongly
that phase separation proceeds with kinetics associated with spinodal
decomposition. Vitrification of the polyurea rich phase arrests and preserves
a bicontinuous morphology shortly after the onset of microphase separation.
Li et al.168j also investigated high resiliency foam formulations that incor-
porate chain extenders of low molar mass and crosslinking agents. Their in
situ studies showed that incorporation of such components delays the onset
of microphase separation. The extent of disruption to urea group assembly
and development of hydrogen bond interactions is a function of the chemical
structure of the chain extender or crosslinking agent.

© 2004 by CRC Press LLC

www.ketabdownload.com
 REAL TIME EXAMPLE LOCATION OF MICROSCOPE
DENSITY CHANGES SEPARATION TRANSITION
12 0.08

RELATIVE INVARIANT, q
COUNTER, 104 counts/sec

Cells are opening 0.07

REAR IONIZATION

10 0.06
8 0.05

I(q,t)/arbitrary units
6 1.0 0.04
4 0.8 0.03
0.6 0.02
2 0.4 300
0.2 250 0.01
0 0.0 200
150 0
0 50 100 150 200 0.02 0 20 40 60 80 100 120
0.04 0.06 100 Time,sec
ELAPSED TIME, SEC 0.080.10 50 Time, sec
0.12 0.14
q/Å-1 0
0.16

INTERDOMAIN EVALUATION OF MECHANISM OF MICROPHASE

SPACING DIFFUSION COEFFICIENTS SEPARATION
30 1

25
RELATIVE INTENSITY,

20
arbitrary units

[R(q)/q2]

15 I (q)
10
5
0 0.1

−5
−10 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0 8.5
0.01 0.03 0.05 0.07 0.09 0.011 0 0.002 0.004 0.006 0.008 0.010 Time, sec
SCATTERING
VECTOR, q/Å-1
FIGURE 6.15
Application of time-resolved SAXS to the study of structure development in reactive polymer
systems. Three-dimensional SAXS data from a flexible polyurethane foam-forming system,
comprising 38% by weight of polyurea hard segments, provide information on macroscopic
density changes, microphase separation, and domain size scale, with simultaneous cell forma-
tion and copolymerization. The I(q) versus time plot shows data obtained at scattering vector
values of q = 0.036 (+), 0.050 (
), and 0.076 (). (Copyright 1997 Wiley-VCH Verlag. Reprinted
with permission from Stanford, J.L., Ryan, A.J., and Elwell, M.J., Processing of Polymers, ed.
H.E.H. Meijer, vol. 18 of Materials Science and Technology: A Comprehensive Treatment, VCH
Publishers, Weinheim, Germany, 1997, chap. 9.)

6.6.5 Solidification: Understanding the Arrest of Phase Separation

Dynamic rheometry has been employed to investigate the development of
polymer modulus, the nature of the connectivity between the hard and soft
segments, and the macroscopic stability of the foam beyond the onset of cell
opening.166 The development and use of an adiabatic rheometer was reported
first by Mora et al.166b,168b As illustrated in Figure 6.16,149b there are four
regions of differing rheological behavior: (1) bubble nucleation; (2) liquid
foam and microphase separation; (3) stiffening of the polymer, hard segment
vitrification, and cell opening; and (4) chemical gelation.
In slabstock foams, the storage modulus begins to increase at the onset of
the microphase separation of the polyurea hard segments. The increase in
stiffness is derived from the growth in the interconnecting physical network
of hydrogen-bonded, polyurea hard segments. Cookson et al.170 showed that
the onset of microphase separation coincided with a sharp increase in the
normal force, and cell opening coincided with the apex of the peak in the
normal force measured during foaming, as illustrated in Figure 6.17.

© 2004 by CRC Press LLC

www.ketabdownload.com
 1 2 3 4
105

104
G′ (Pa)
1000

100
G′ (Pa)
10 G′′ (Pa)
φ
1
0 50 100 150 200
Time (s)
FIGURE 6.16
Modulus build in a flexible slabstock foam. Four regions of behavior are indicated: (1) initial
bubble growth, (2) bubble network, (3) polymer stiffening, and (4) final curing. Cell opening
occurs a few seconds after the onset of Region 3. (Copyright 1995 Society of the Plastics Industry
[now held by the Alliance for the Polyurethanes Industry]. Reprinted with permission from
Neff, R. and Macosko, C.W., Proceedings of Polyurethanes 1995, Society of the Plastics Industry,
Chicago, 344–352, September 1995.)

The sudden stiffening of the polymer in the absence of the covalent net-
work results in a brittle material. This material has insufficient tensile and
shear strength to accommodate the additional strain resulting from the con-
tinued expansion of the foam. Consequently, the cell windows undergo
rupture and blow-off takes place. Catastrophic collapse occurs if the foam is
too weak to support its own weight after cell rupture, while under other

0.25
cell opening
0.20 initial bubble growth

0.15
Normal Force/N

MST
0.10

0.05

0.00 foam expansion

−0.05
sample injection, t − 15s
−0.10
0 50 100 150 200
Time/s

FIGURE 6.17
Example of normal force versus time during polyurethane foam formation, illustrating identi-
fication of microphase separation transition (MST) and cell opening.

© 2004 by CRC Press LLC

www.ketabdownload.com
circumstances there may even be no blow-off or cell opening. Li et al.168j
showed that the incorporation of chain extenders and crosslinking reagents
delayed the onset of microphase separation and that more block copolymer
could be formed. The compatibility between the hard segments and soft
segments increased. By increasing the degree of connectivity between the
hard and soft segments, the stress on the polyurea interconnecting physical
network can be spread to the soft segment phase, increasing the overall
extensibility and strength of the foamed polymer. The delay in the onset of
microphase separation leads to an increase in the molecular weight of the
polyether at the time of cell opening. The increase in molecular weight and
concomitant increase in viscosity results in a higher degree of closed cells in
the foam.

6.6.6 Post-Reaction Analyses: Understanding Properties in End-Use

Applications

Postmortem analyses of both MDI-based and TDI-based flexible copoly(ether-

urethane urea) foams by scanning transmission x-ray spectromicroscopy and
tapping mode AFM have confirmed the presence of a bicontinuous, inter-
connecting network of polyurea hard segments. Figure 6.18 shows AFM
phase images from an MDI-based flexible copoly(ether-urethane urea) foam
of Li et al.168i and a TDI-based flexible copoly(ether-urethane urea) foam of
Rightor et al.157b The lighter phases are rich in polyurea. The hard segment
rich regions appear to be interconnected. These images agree qualitatively
with those reported by Hamley et al.163c They are also consistent with the
hypothesis that arresting phase separation by vitrification of the hard seg-
ment rich phase prevents the evolution of a morphology comprising dis-
persed isolated domains such as those exemplified by the equilibrium
morphologies (e.g., spheres, cylinders, and lamellae) of diblock and triblock
copolymers.

6.6.7 Summary: Overview of Polymer Microstructure and Morphology

Development

In summary, the reactive processing of a flexible polyurethane foam from

liquid monomers and oligomers involves a complex combination of chemical
and physical events. In early stages of copolymerization, the reaction mixture
comprises polyether, isocyanate terminated polyether oligomers, water,
polyurea homopolymer, and some copoly(ether-urethane urea) block
copolymer that contains the urethane linking group. The copolymerizing
mixture is blown into foam by the cogeneration of CO2 gas evolved from
the water-diisocyanate reaction. As copolymerization proceeds, the core of
the foam becomes insulated by the surrounding polymerizing mixture,
creating, in effect, a quasi-adiabatic environment. Reaction between the
monomers and oligomers results in both a molecular weight increase and
a change in the interaction parameters, resulting in a reaction-induced
© 2004 by CRC Press LLC
www.ketabdownload.com
 750

500

250

0
250 500 750 500 nm
nm

150

100

50

0
0 50 100 150 nm

A. B.
FIGURE 6.18
(A) AFM tapping mode phase images of an MDI foam at low (top) and high (bottom) magni-
fications. (Copyright 2002 American Chemical Society. Reprinted with permission from Li, W.,
Ryan, A.J., and Meier, I.K., Macromolecules, 35, 5034, 2002.) (B) High magnification (5
5 mm2)
AFM phase image of a TDI foam. The arrows indicate some of the larger polyol-rich regions
present within urea aggregates. (Copyright 2002 American Chemical Society. Reprinted with
permission from Rightor, E.G. et al., Macromolecules, 35, 5873, 2002.)

phase separation process. The onset of microphase separation of the poly-

urea hard segment sequence lengths occurs midpoint in the reaction, at
isocyanate conversion fraction pNCO = 0.55 ± 0.05. The evidence suggests
strongly that phase separation proceeds with kinetics associated with a
spinodal decomposition.
At the onset of microphase separation, the system becomes both spatially
and chemically heterogeneous, separating the functional groups. The
microphase-separated polyurea hard segments continue to grow. Association
of these urea moieties takes place via the formation of hydrogen bonds.
Hydrogen bonding is thus a parasitic process, not the driving process of
microphase separation. As copolymerization proceeds further, the concen-
tration of copoly(ether-urethane urea) block copolymer continually increases
until the microphase separation is arrested by vitrification. The copoly(ether-
urethane urea) block copolymer evolves via the formation of a urethane
© 2004 by CRC Press LLC
www.ketabdownload.com
group between the respective polyether and the polyurea blocks. It effec-
tively acts as a surfactant for the immiscible hard polyurea and soft polyether
segments. The strength of the interactions between the multi-block copoly-
mer and the homopolymer phases controls the overall degree of mixing and
the driving force for interfacial segregation.
The storage modulus begins to increase at the onset of microphase sepa-
ration of polyurea hard segments. The increase in stiffness is derived from
the growth in the interconnecting physical network of hydrogen-bonded,
polyurea hard segments. Cell opening occurs within 5 to 10 sec after the
onset of modulus increase. The sudden stiffening of the polymer in the
absence of a covalent network results in a brittle material that has insufficient
tensile and shear strength to accommodate the additional strain resulting
from the continued expansion of the foam. As a direct consequence, the cell
windows undergo rupture and blow-off occurs.
Vitrification of the polyurea-rich phase at a phase composition whose T g
is equal to the temperature of the surrounding medium arrests and pre-
serves the bicontinuous morphology shortly after the onset of microphase
separation. This results in a foam whose internal polymer morphology
comprises a bicontinuous interconnecting physical network of hydrogen-
bonded polyurea hard-segment sequence lengths within a crosslinked
poly(ether urethane).

6.7 Foam Preparation

6.7.1 General Considerations
Polyurethanes are usually processed quite differently from other major poly-
mers. Most thermoplastics are polymerized in large factories by chemical
manufacturers and then sold to processors as pellets or powders. The pellets
or powders are then converted to useful end products by fabrication methods
involving heating, shaping under pressure, cooling, and packaging. The
properties of the plastic thus produced are controlled by the properties of
the base polymer. By contrast, the polymerization of polyurethanes is more
often carried out by independent outside processors who often possess
enough skill to achieve wide variations in the properties of products made
from just a few base resins. In addition to providing the raw materials, the
chemical manufacturer is often required to provide advice, technical infor-
mation, and even manpower assistance to the processor.
There is a vast number of variables in the formation of polyurethane foams.
Many of these variables can have dramatic effects on the foam properties.
For example, the typical effects of varying the water level in the formulation
on the density of the resulting foam are illustrated in Figure 6.19. As was
discussed earlier, the reaction of H2O with isocyanate produces both polyurea
compounds which remain in the foam and CO2 which acts as a blowing
© 2004 by CRC Press LLC
www.ketabdownload.com
A. B.
CORE DENSITY, PCF*
3.5
50 3.0

MDI-BASED FOAMS
3
40 2.5
Density kg/m3

2.5

Density Ib/ft3
2.0
30
2
1.5
20 TDI FOAMS
1.5
1.0
for minimum fill peds
10 1
0 2 4 6 4 5
2 3
Formulation Water Level (phpp)
Formulation Water Level (phpp)
FIGURE 6.19
Typical effects of varying the water level in the formulation on the density of the resulting
foam. The reaction of H2O with isocyanate produces both polyurea compounds which remain
in the foam and CO2 which acts as a blowing agent. Increasing the amount of H2O in the
formulation increases the production of the CO2 blowing agent and hence decreases the density
of the foam. Consequently, in the absence of auxiliary blowing agents, the amount of H2O in
the formulation determines the foam density. (A) Results for typical slabstock foam formula-
tions. (B) Results for typical TDI-based and MDI-based molded foam formulations. The units
(phpp) denote parts per hundred parts polyol.

agent. Increasing the amount of H2O in the formulation increases the pro-
duction of the CO2 blowing agent and hence decreases the density of the
foam. Consequently, in the absence of auxiliary blowing agents, the amount
of H2O in the formulation determines the foam density.
The complexity of foam formulations makes the development or selection
of recipes that will give good results at a minimum cost difficult. This task
requires not only knowledge of polyurethane chemistry but also skill in the
operation of dispensing machines of various types, and in the technology
and art of evaluating final products. Over the roughly 40-year commercial
history of polyurethanes, many successful methods for foam production
have been developed and recorded in the literature. Many of the best foam
recipes are commercially available from suppliers as fully formulated sys-
tems, which are often attractive to manufacturers who (because of volume
requirements, in-house technical limitations, equipment, or other restraints)
limit their involvement to the final mixing and distribution of the polyure-
thane materials.
The chemical reaction between a polyol and an isocyanate begins imme-
diately when the two ingredients are mixed in the presence of suitable
catalysts. Once started, the polymerization is exothermic up to its comple-
tion. Whichever method of mixing and dispensing is used, there are certain
basic requirements for satisfactory foam production:

© 2004 by CRC Press LLC

www.ketabdownload.com
 • Temperature conditioning. The viscosity, density, and chemical reactiv-
ity of polyols and isocyanates vary significantly with temperature.
Controlling the temperature is, therefore, essential to the reproduc-
ible processing of polyurethane foams.
• Accurate metering or weighing. To make a foam reproducibly, the
ingredients must be mixed together consistently in the proportions
required by the formulation.
• Freedom from contaminants. Contamination is defined as the presence
of any of the vast number of possible unintended foreign materials
in the foam preparation system. Defects in polyurethane foams are
often eventually linked to contaminants.
• Effective mixing. Incomplete mixing can produce a foam that has both
cosmetic and physical defects.
• Curing environment. In general, flexible polyurethane foams do not
develop their ultimate properties until some hours or days after
initial mixing. Many foams cure satisfactorily and develop properties
simply through storage at room temperature. Other systems may
benefit from a post-cure heating cycle. The optimum processing and
curing conditions must be investigated for each type of flexible foam.

Foaming at different scales often produces foams with different cell sizes
and hence different resultant properties from the same starting formula-
tion.114a,171 Commercial flexible polyurethane foam development typically
proceeds through three stages:

• Cup foaming.172 Foaming on this small “laboratory” scale typically

requires 1 worker, approximately 1 gal of materials, and 1 day of
work.
• Box foaming.173 Foaming on this intermediate “laboratory” scale typ-
ically requires 1 worker, approximately 1 to 5 gal of materials, and
1 day of work.
• Machine foaming.174 Foaming on this “production” scale typically
requires a number of workers that varies as a function of the size
and the complexity of the machine, approximately 5 to 55 gal of
materials, and 1 to 5 days of work.

Flexible foams are made commercially by various fabrication processes, as

will be discussed briefly in the next three subsections. Typical block flow
diagrams for slabstock and molded foam processes are available.175

6.7.2 Slabstock Foams

More than half of all flexible polyurethane foam is produced as slabstock
buns. Slabstock foam is in high demand because of its wide application and
© 2004 by CRC Press LLC
www.ketabdownload.com
large volume usage. Simple and adaptable manufacturing processes, coupled
with the wide array of products that can be produced with high quality, have
made these foams economical and versatile choices for many markets. Pro-
cess and formulation changes that slabstock foam manufacturers made in
the 1990s will allow the industry to continue to produce these products while
also addressing the environmental concerns of customers and the public.
Approximately 1.5 billion lb (680 million metric tons) of flexible slabstock
foam were produced in North America during 1994.176 The U.S. market
breaks down into furniture (40%), bedding (15%), transportation cushioning
(12%), carpet underlayment (25%), and packaging (5%). The balance is used
in textiles, household goods, and miscellaneous other applications.
Most slabstock foams are made on continuous foam machines, but discon-
tinuous box foaming is viable in less-developed countries or for low-volume
specialty foams.
Slabstock foam buns can be cut into a variety of desired shapes after curing.
The foam pieces cut from the bun can then be used “as is,” or cut, sewn,
glued, or stapled to each other or to other objects or materials. The relative
ease of fabrication helps ensure the complete cushioning of delicate objects,
allows manufacturers to produce complex sizes and shapes, and makes
flexible slabstock foam one of the most adaptable among all cushioning and
comfort products. Cushioning and comfort properties can be tailored to meet
the requirements of various applications by simply modifying the formula-
tion used to produce the foam. Formulation changes (whose effects are well
understood) can be made easily, to increase or decrease the density and the
load-bearing characteristics of the foam (resulting in a softer or harder foam).
Additives can be used to modify the polymer chemistry and cell structure
and allow manufacturers to fine-tune the properties. Other additives can
modify foam flammability, dissipation of static electricity, and many other
important properties. Mechanical processes such as partially crushing the
foam to open cells or to cause foam densification can modify foam charac-
teristics after a foam has been produced, to meet the needs of specialized
applications.
There are four main types of flexible polyurethane slabstock foam: conven-
tional, high resiliency, filled,177 and high load bearing.178 The properties can be
tailored to the requirements of various applications by properly selecting the
level and type of polyol and the other formulation components. The most
important properties are the density, hardness, and porosity. The hardness
grades that are typically specified for conventional flexible polyurethane foams
in North America are listed in Table 6.3. The formulation variables that most
affect the key properties are the isocyanate index, amount of water, amount
of auxiliary blowing agent, and catalyst level and type. Recent formulation
changes, dictated by environmental considerations, have resulted in formula-
tions that use a higher water level and a lower isocyanate index. There is
growing belief that these changes may be detrimental to long term fatigue
(loss of load bearing) resistance. Research is in progress to further elucidate
the causes of fatigue. An improved understanding of the fundamental nature
© 2004 by CRC Press LLC
www.ketabdownload.com
 TABLE 6.3
Hardness Grades for Conventional Flexible Slabstock
Foams in North America
Density Range Indentation Force
Foam Grade (lb/ft3) Deflection (lb/50 in.2)
Supersoft 1.0–1.8 10–20
Soft 1.0–1.5 20–30
Intermediate 1.2–1.4 30–40
Hard 1.2–3.0 >40

of fatigue in foams will enable the design of foam structures and polymer
morphologies that result in improved fatigue resistance.31,179

6.7.3 Molded Foams

In contrast to the large buns produced in conventional slabstock foams,
molded foams are made one at a time in production plants designed specif-
ically for the repetitive (batch) production of discrete foam shapes. These
plants often operate two or three working shifts per day and may produce
more than one style or part. Molding involves pouring the reacting mixture
into a suitable mold, closing the mold, and allowing the foam to fill the mold.
In addition to replicating the internal dimensions of the mold, the final
product can also permanently record its part or style number and incorporate
physical inserts of various types. A vast variety of intricate shapes can be
molded successfully. With the appropriate selection of the starting isocyanate
and polyol and with the selective use of additives, it is possible to form
flexible molded foam products meeting very different performance specifi-
cations. It is also possible to “foam in place” various reinforcements or
otherwise produce foams with multiple zones of hardness.
Molded foams find application in all forms of transportation seating and
trim parts, as well as in some upholstered furniture,180 bedding, packaging,
and novelty items. While a small percentage of the seating cushions used in
automobiles today is cut from slabstock foam buns,181 most are made from
molded foams. Typical methods of constructing a final automobile seat
assembly have been reviewed.182
Molded foam production is considerably more complicated than slabstock
foam production, because of factors such as the use of formulations of higher
reactivity, pour-pattern limitations, the discrete size of each individual shot,
the need to open and close a mold lid mechanically, the choice of release
agent, mold temperature latitudes, and the available curing cycle. Foam
formulations used for molded and slabstock foams differ mainly with respect
to their catalyst packages and the compositions of their polyols. More reac-
tive polyols of higher molecular weight are generally used to enable the
production of molded foams with good cure behavior and rapid property
development.
© 2004 by CRC Press LLC
www.ketabdownload.com
 The two major molding techniques used in the polyurethane industry are
the hot cure and the cold cure (or high resiliency) processes. The hot cure
molding process was established commercially more than 30 years ago. High
temperature is applied to the mold during the cure cycle, driving the foaming
reaction to a sufficient degree of surface cure to permit the foam to be
demolded in a timely manner and then subjected to additional downstream
fabrication steps. The newer high resiliency foam molding process utilizes
ingredients (especially polyols) of higher reactivity so that less oven curing
(and hence less externally applied energy) is needed. The essential differ-
ences between hot and high resiliency molding are summarized in Table 6.4.
Detailed discussions of other processing and physical property differences
are available.183
Most flexible polyurethane foam molding is done in cast aluminum molds,
but there has also been some use of reinforced epoxy molds. Molds are
usually made in two sections, with mechanical opening and closing of the
lid. A mold release agent (most often, a commercial blend of various natural
and synthetic waxes) is used to help remove a polyurethane foam article
from a mold. Mold release agents can be characterized as water based or
solvent based. Detailed review articles are available.184 Recent developments
in water-based systems have been reported.185
Specialty molded foams include semiflexible, semirigid, and integral-skin
types of foams. In general, semiflexible molded foams are used in instrument
panels, arm rests, console covers, door panels, and other foam parts found
in automobile interiors. If the mold is lined with a preformed plastic skin or
textile fabric, dispersed foam will adhere to it and simplify the assembly
requirements of many composite products. Technology for directly foaming
onto a cover material preformed into a seat cushion mold has existed since
1955.186 Recent developments in vacuum-assisted foam-in-place technology

TABLE 6.4
Essential Differences between Hot Cure and High Resiliency Flexible
Polyurethane Foam Molding Technologies
Hot Cure High Resiliency
Isocyanate 80/20 TDI 80/20 TDI
Blends with MDI
Polymeric MDI
MDI prepolymers
Polyol
Molecular Weight (g/mole) 2800–3500 4500–6500
Ethylene Oxide Capping Yes Yes
Copolymer Polyol Optional Optional
Cure Oven 180–300˚C 75–200˚C
Temperature (356–572˚F) (165–392˚F)
Postcuring No Optional
Mold Temperature 25–45˚C 50–70˚C
At Pour (77–113˚F) (122–158˚F)

© 2004 by CRC Press LLC

www.ketabdownload.com
for auto seating have been discussed.187 Alternative schemes for bonding the
cover fabric to the foam pad have been developed.188

6.7.4 Carpet Backing Foams

An adhesive backing is required to anchor the tufts in manufacturing tufted

carpets. The most popular carpet backing systems are based on styrene-
butadiene latexes which give good tuft-lock at low cost. Another backing
approach is the application of a resilient foam backing over the anchor coat.
Foamed materials used for this backing are poly(vinyl chloride) (PVC) plas-
tisols, polyurethanes, EVA, and latex.
The concept of forming a flexible polyurethane foam directly on the back
of a carpet was described as early as 1959.189 Advantages in improved tuft-
lock, energy absorption, heat and sound absorbance, non-slip character, and
appearance retention time of the carpet face were recognized rapidly. Com-
mercial success for polyurethanes came at the expense of traditional backing
materials (foamed rubber latex and foamed PVC).190
Polyurethane raw materials are more expensive than latex. Nonetheless,
advantages in processing, physical properties, and durability have resulted
in polyurethanes becoming the foams of choice.191 A processing advantage
is lower energy consumption due to the lower curing oven operating tem-
peratures. Unlike polyurethane foams, large amounts of water must be evap-
orated in latex foams. In addition to their physical property and plant capital
advantages, frothed polyurethanes are also more energy efficient than alter-
native backings.192 Detailed reviews of the processes and applications for
mechanically frothed polyurethane foams are available.191,193

6.8 Evaluation and Testing of Basic Foam Properties

Properties are tested to characterize a foam and thus determine its suitability
for applications of interest, as well as for quality control. The historical
development of basic test methods for flexible polyurethane foams can be
found elsewhere.194
ASTM D 3574-95 presents a standard method for conditioning foam sam-
ples and testing the basic physical properties:195 density, indentation force
deflection, compression force deflection, constant deflection compression set,
tensile strength, tear resistance, airflow, and resilience. Some of these tests
have been automated successfully.196 Most of these basic foam properties
have also been the subject of math modeling exercises.114a,197
ASTM D 3574-95 calls for conditioning a sample by waiting for 1 week or
longer after fabrication and then storing the undistorted foam at a specified
temperature (generally 23°C ± 2°C) and humidity (generally 50% ± 5% rela-
tive humidity) for 12 h prior to testing. In some tests, mechanical condition-
© 2004 by CRC Press LLC
www.ketabdownload.com
ing is also used in an attempt to obtain data under conditions resembling
the actual performance of a product in service.
The density measurement provides an apparent or bulk density (and not
the true polymer density). A representative sample is measured carefully
and weighed accurately. The bulk density (normally reported in pounds per
cubic foot or kilograms per cubic meter) is simply the weight of the sample
divided by its volume. The relative density (or the solid volume fraction) of
the foam is defined as its bulk density divided by the true density of the
solid (dense) polymer from which it is made. The void fraction is defined as
[1 – (relative density)]. The relative density is relevant both to cost and to
load bearing. Higher relative densities generally result in improved load
bearing at higher cost. A typical example of how foam properties vary with
density198 and simple models for such dependences31 have been presented.
The indentation force deflection (IFD) test measures the load-bearing prop-
erties. The force required to depress a circular plate into the foam (indenta-
tion resulting from a specified force, force required to reach a specified
indentation) is measured. The results can depend on the temperature and
relative humidity conditions under which the foam was made, cured, and
tested. Detailed discussions of such effects are available.199
The compression force deflection (CFD) test is similar to the IFD test. The
force required to compress the foam to a specified level is measured and
reported for a plate large enough to apply force across the entire top surface
of a 50
50
20 mm foam sample.
Variations on these basic load-bearing tests include changes in sample dimen-
sions and indentor foot area. Such changes affect the results of the tests.200
In the constant deflection compression set test, three samples are placed
between metal plates and compressed to 50%, 75%, or 90% of their original
thickness. They are held at 70˚C for 22 h, removed, allowed to recover for
30 min, and remeasured. The results are normally reported as a percentage
of the original thickness.
The tensile test measures the stiffness and strength under tension. Samples
are generally die-cut and are “dogbone” or “dumbbell” shaped. The sample
is pulled at a constant rate until it breaks. The stiffness (modulus of elasticity)
is defined as the initial slope of the stress/strain curve in the linear region
of ideal elasticity where Hooke’s law holds. The force recorded at the break-
ing point is the tensile strength. The maximum extension of the sample as a
percentage of its original length is the elongation at break.
Another measure of foam strength is tear resistance. A sample is slit at one
end. The force required to continue this tear at a constant rate is measured.
The results are highly rate dependent. Tear strength is not necessarily related
to tensile strength in foams. The results have also been found to vary with
the sharpness of the sample cutting die.
The air flow test114a measures cell openness (porosity). Air is pulled through
a sample of standard thickness and its flow rate is observed. Air flow is very
sensitive to the orientation of the sample within the foam. Air flow is gen-
erally reported both parallel to and perpendicular to the foam rise direction.
© 2004 by CRC Press LLC
www.ketabdownload.com
In molded foams, air flow through the surface skin is also of interest and
can be measured.201
Resilience is measured in a ball rebound test where a steel ball is dropped
on a foam sample and the height of rebound is measured visually.202 The
results of this test have been correlated to overall cushion comfort.201a,203
Fogging refers to formation of a light-scattering film on interior glass
surfaces. It is important for foams used in the automotive industry. Volatile
ingredients in interior trim parts may contribute to the film. A sample is
typically placed in an apparatus and heated at a desired temperature to
promote the release of volatile components. These components are then
condensed on a cooled glass plate and subjected to quantification by gravim-
etry or light reflectance loss measurement. The details of fogging test
methods204 and the key variables influencing the test results204a,205 have been
documented extensively.

6.9 Evaluation and Testing of Foam Durability

The techniques for measuring fatigue at the laboratory scale206 fall in two
classes: static fatigue and dynamic fatigue. Within each class, individual tests
vary in the degree of loading or deflection, use of indentation or full size
deflection, use or absence of shear, sample thickness, number of cycles, rate
of application of flex, and recovery or rest period after flexing. The sample
type also varies. Measurements from the tests (including height loss, load
loss, physical appearance, or loss in other properties) vary considerably.
Some standard test methods for commonly used laboratory tests are sum-
marized below.
Two types of static fatigue tests are commonly used. (1) The constant
deflection compression set test is used most often. It is described in ASTM
D 3574-95, Test I1.207 It is similar to the simple constant deflection compres-
sion set test. The sample is removed and allowed to recover when the com-
pression time is complete. Its height is measured. The results are provided
as a percentage of the original height. (2) The constant load creep test mea-
sures the reversible change in the shape of a sample with time under a
sustained load. A foam block is placed under a weight platform. The platform
is loaded to the desired mass and lowered onto the foam. The percentage of
vertical deflection (creep) is measured over 10 min and plotted versus time.
A creep rate (slope of curve showing the data) is calculated. It varies with
the percent penetration into the foam. A high creep rate corresponds to
cushion comfort loss.11c,208
Dynamic fatigue tests simulate the alternating loads experienced by foam
seating cushions under actual service conditions. Two general types of tests
have been used; namely, those in which the sample is indented to a fixed
percentage of its thickness by an oscillating indentor, and those in which the
sample is loaded and unloaded repeatedly with an indentor exerting a fixed
© 2004 by CRC Press LLC
www.ketabdownload.com
load. Tests such as the constant load pounder, in which the extent of
indentation increases as the foam softens, have reportedly shown a high
correlation with height loss (and to a lesser extent, also with hardness loss)
in the end use performance of foams. Two laboratory-scale dynamic testing
protocols are commonly used by furniture and automotive foam suppliers:
(1) constant-load pounding and (2) roller-shear. A standardized constant-
load-pounding fatigue test, ISO 3385-1975(E), is a common choice for many
furniture and automotive foam application areas.209 Another frequently used
test protocol is the ASTM D 3574-95, test I3. The Jounce test is used to assess
fatigue performance for seating cushions.210 The standard roller-shear test
and a suggested method of equipment construction are described in ASTM
D 3574-95.211 A modified roller-shear test, the Caster fatigue test, which is
used in carpet backing applications, gives the best correlation with actual
use performance.212 A Society of the Plastics Industry study reports excellent
correlation between the laboratory roller-shear procedure and field
evaluations of fatigue213 but the absolute accuracy was unsatisfactory.
There is an extensive body of literature describing the results of experi-
ments probing into the effects of various chemical and physical factors affect-
ing foam durability as measured by the types of fatigue tests that were
summarized above.114a,214 Most importantly, recent research has begun to
deconvolute fatigue effects due to the cellular nature of the foam from those
that are due to the polymer used in the foam. The main cellular factor
influencing fatigue is found to be the density. If all other factors are kept
constant, loadbearing retention under fatigue decreases with decreasing den-
sity. Increased foam hardness (a larger foam storage modulus) attained by
increasing the hard segment volume fraction (for example, by increasing the
water level in the formulation) or by incorporating SAN CPP particles in the
polyol is also usually detrimental to fatigue performance. Consequently, in
general, if high durability is needed simultaneously with a higher modulus
in a certain grade of foam, then the foam density should also be increased.

6.10 Control of Noise and Vibrations

The ability to control noise and vibrations is crucial to the use of foams in
vehicle seats214g,h as well as in many of the trim parts of vehicles. The noise
and vibrations present inside an automobile contribute to occupant discom-
fort by creating a harsh environment. They have become a major consumer
concern as cars have become smaller and lighter. Interior noise levels of up
to 70 decibels have been reported.215
Grades of specialty molded foams that are helpful in reducing this dis-
comfort are called noise and vibration harshness (NVH) foams. Many new
car and truck models will insulate noisy mechanical parts with shields pad-
ded on the noise-receiving side with molded NVH foam. Commercial exam-
ples of such applications of molded NVH foams have been presented.216 On
© 2004 by CRC Press LLC
www.ketabdownload.com
the other hand, recent advances in cutting and contouring technology have
won new NVH applications for conventional slabstock foams.217
Air flow, cell size, and density are the main factors influencing the effec-
tiveness of sound absorption by simple cut sheets of foam.218 Increasing the
cell count improves sound absorption at all frequencies. Optimum air flow
ranges between 0.1 and 1.0 ft3/min at 0.5 in. of water pressure. Increasing
the relative density generally improves sound absorbance, but does so less
than the previous two factors. The effects of foam thickness, cutting profile,
and surface treatment have also been studied.219 Fundamental mechanisms
and experimental methods of measuring sound absorption by flexible foams
have been reviewed.220 A review of the patent literature prior to 1986 is
available.221 The physiological effects of noise and vibrations on humans have
been described.222

6.11 Flammability
The flammability of flexible polyurethane foams is important because their
growing use in upholstered furniture and bedding presents the possibility
that they could contribute to a life and safety hazard. Since these foams
consist of carbon, nitrogen, hydrogen, and oxygen atoms, an entirely non-
combustible foam cannot be made. Furthermore, furniture and cushion
foams are normally very low in solid volume fraction. The high void space
fraction results in a large surface area which allows a free oxygen pathway
into the foam. It also creates thread-like struts that can be heated easily by
adjacent heat sources. Even small ignition sources such as a match or a
cigarette may hence warm small polymer fibrils in certain types of uncovered
polyurethane foams to their ignition point.223 However, since the foam is
always used as a filler in an upholstered cushion, there is no way for smol-
dering cigarettes or flaming ignition sources to reach the foam without first
penetrating the upholstered cover. Consequently, foam involvement in fires
is related to the lack of resistance of the cover to ignition, and flammability
assessment must be done on the specific construction used to make the article
rather than on any single component of the upholstered furniture.224 There
are many standards and regulations in different parts of the world regarding
the ignition resistance of upholstered furniture and cushion items.
Although a life and safety hazard results from ignition of the entire uphol-
stered cushion, significant efforts have been made by the flexible polyure-
thane foam industry to improve the fire resistance of uncovered foams.
Several methods are prominent in improving the fire resistance of foams. (1)
Addition of fire retardants101 (especially phosphorus-containing and/or halo-
gen-containing ones) during foam preparation. Halogens interrupt the free
radical reactions involved in combustion. Phosphorus helps form a barrier
that restricts the organic material in the foam from feeding the flame. (2)
Addition of fillers along with the fire retardants. The most effective fillers
© 2004 by CRC Press LLC
www.ketabdownload.com
are melamine225 and hydrated alumina.224b (3) Protection of the foam by fire-
resistant upholstery fabric or by liners between the upholstery fabric and
the foam cushion.224c
Many tests exist to evaluate flammability. The basic objective of these tests
is to screen the performance of new materials and formulations. They range
from bench or laboratory-scale tests to full-scale tests on entire rooms or
buildings. Laboratory-scale tests are generally used to compare the ignitability
and the burn rate of small specimens. Their results can vary widely, depending
on the orientation of the specimen, the size of the flame source, and the amount
of radiant energy. Furthermore, laboratory-scale tests cannot be used to predict
performance under actual fire conditions. The performance in a real fire situ-
ation can vary markedly from that observed in small-scale tests. In fires involv-
ing building materials, there are many parameters that affect the course of a
real fire, and many additional hazards (such as smoke) can arise. Most deaths
in fires result from asphyxiation due to lack of oxygen or from inhalation of
smoke or toxic vapors.226 For these reasons, full-scale tests are generally
needed to measure both flame spread and smoke density. In recent years,
many national standard and insurance organizations have been involved in
developing a unified set of tests.227 New tests continue to evolve.

6.12 Recycling
The typical life cycle of polyurethane products is shown in Figure 6.20.
Petroleum-derived chemicals are converted into rigid and flexible foams,
reaction injection molded (RIM) and cast elastomers, and other products.228
Flexible polyurethane foams are thermoset polymers. It is worth emphasiz-
ing that no foamed thermoplastics of any kind have yet found use as ade-
quate replacements for flexible polyurethane foams in practice. While
thermosets have historically not generally been associated with recycling, a
lot of recycling of flexible polyurethane foams actually occurs, to conserve
and recover as much as possible of the invested energy and other resources.
The strategies include material recycling (recovery by foam rebond229 and
regrind230 processes), chemical recycling (recovery of chemicals by polymer
breakdown through hydrolysis,231 glycolysis,232 or pyrolysis233), and energy
recovery by combustion technologies such as the rotary kiln/fluidized bed234
and the high temperature gasification processes.
Virtually all process scrap from flexible slabstock foam cutting operations
and molded foam seating production is recycled by a rebonding process into
high-quality carpet underlay. Other uses for rebond foams include seating
and cushioning applications such as gym mats, seat inserts, headrests, and
school bus and golf cart seats. Additional rebond uses are being developed,
including sound absorption firewall, door, and automotive carpet pads. Post-
consumer foam scrap is less widely recycled because of the diversity of
location and the lack of a collection infrastructure. Several companies in the
© 2004 by CRC Press LLC
www.ketabdownload.com
 POLYURETHANES
LIFE CYCLE

PETROLEUM
URETHANE
CHEMICALS

HYDROLYSIS
RIGID FLEXIBLE RIM
GLYCOLYSIS FOAMS
FOAMS ELASTOMERS
PYROLYSIS

REGRIND REGRIND

COMPRESSION MOLDING COMPRESSION MOLDING

REBOND
ENERGY NEW NEW
POOL APPLICATIONS PARTS

CARPET
SOLID UNDERLAY
WASTE

ENERGY
RECOVERY POST-CONSUMER
ASH
WASTE

LANDFILL

FIGURE 6.20
Typical life cycle of polyurethane foam and elastomer products.

U.S. have found it profitable to collect scrap foam from furniture, bedding,
and carpet underlayment and process it into rebond foam. The volume of
scrap collected has increased steadily as the value of the scrap is being
recognized.
The formation of the Polyurethane Recycle and Recovery Council by the
Polyurethane Division of the Society of the Plastics Industry (now the Alli-
ance for the Polyurethanes Industry, a business unit of the American Plastics
Council) underscores the importance of the recycle issue to the industry.235
This industry group sponsors recycling programs in flexible and rigid foams.
Its membership includes all of the major suppliers to the polyurethane indus-
try. The flexible foam programs are focused on post-consumer scrap, includ-
ing the separation and recycling of flexible foam from automotive shredder
operations, the evaluation of energy recovery options, and the possibilities
for adding scrap regrind to flexible slabstock and molded foam production.

6.13 Polyurethane Foams from Renewable Resources

Efforts to reduce the dependence of flexible polyurethane foam production
on petroleum-based precursor materials by using renewable feedstocks con-
© 2004 by CRC Press LLC
www.ketabdownload.com
stitute a new research frontier. For example, John et al.236 described the
reaction of polyols based on inexpensive soybean-based vegetable oils (as
replacements for conventional polyols) with TDI and MDI to produce flexible
slabstock polyurethane foams. Crude soybean oil does not contain hydroxyl
groups. It did not generate any foam. However, it contains 95 to 97% trig-
lycerides and a fatty acid composition with a high level of unsaturation. It
was functionalized by introducing hydroxyl groups at some of these sites of
unsaturation, to obtain polyols which were reacted with isocyanes to make
foams. Just like foams obtained from conventional polyols, the properties of
these foams could be changed by controlling variables such as water content,
isocyanate index, and catalysts.

6.14 Summary and Conclusions

Polyurethanes are chemically complex polymers. They were first synthesized
early in the twentieth century. They are usually formed by reactions of liquid
isocyanate components with liquid polyol resin components. Their greatest
advantage is versatility, both in the finished product properties and in their
ease of production and application. By proper choice of isocyanate and
polyol, products can be made with properties ranging from the downy soft-
ness of very low-density flexible foams to the high strength of elastomeric
bubble-free castings.
The phenomenal success of flexible polyurethane foams in capturing appli-
cations previously held by alternative cushioning materials is recorded in
an extensive body of literature. Such foams are now ubiquitous and hard to
avoid in the civilized world. Commercial production involves many formu-
lation ingredients (such as polyols, isocyanates, water, auxiliary blowing
agents, catalysts, surfactants, additives, and fillers) and at least the two main
competing reactions (polymerization and gas-producing reaction). Many
variables are, hence, at the disposal of a manufacturer who wants to optimize
the properties of a foam for any one of a vast number of potential applica-
tions. The potential applications are often only limited by our imagination.
Considerable understanding has been developed on the fundamentals of
foaming and on analytical characterization, morphologies, properties, and
practical performance characteristics of flexible polyurethane foams. The
most important progress in the fundamentals in recent years has involved
the development of a new understanding (differing from the “classical” one
at the level of its basic qualitative physical picture) of the evolution and final
form of the polymer morphology in the solid portion of the foam. This
progress has been paralleled by improvements in methods for the fabrication
of slabstock, molded, and carpet backing foams.
Future progress in two areas should augment the environmental friendli-
ness of flexible polyurethane foams and thus help them to continue to flour-
ish in a world that places increasing emphasis on sustainable development.
© 2004 by CRC Press LLC
www.ketabdownload.com
While thermosets have historically not generally been associated with recy-
cling, a lot of recycling of flexible polyurethane foams actually occurs, to
conserve and recover as much as possible of the invested energy and other
resources. Efforts to reduce the dependence of flexible polyurethane foam
production on petroleum-based precursor materials by using renewable
feedstocks constitute a new research frontier.

References
1. Bayer, O., Mod. Plast., 24, 149–152, 250–262, 1947.
2. Bayer, O., Ang. Chem., 59, 257–272, 1947.
3. Hochtlen, A., Kunststoffe, 42/10, 303–310, 1952.
4. Brochhagen, F.K., Kunststoffe, 44/12, 555–558, 1954.
5. Tenhoor, R.E., Chem. Eng. News, 41, 94–104, 1963.
6. Winkler, J., SPE J., 12/11, 23–25, 1956.
7. Ferrari, R.J. et al., Ind. Eng. Chem., 50/70, 1041–1044, 1958.
8. Frisch, K.C., J. Macromol. Sci. Chem., A15(6), 1089–1112, 1981.
9. (a) Mod. Plast., 32/3, 106–108, 214–216, 1954. (b) Editorial, Chem. Week, 75/10,
82–84, 1954. (c) Mod. Plast., 32/12, 102–104, 212–215, 1955. (d) Beutel, A.C. et
al., Harvard Graduate School Report, Polyurethane Associates, Cambridge, MA,
1956. (e) Editorial, Br. Plast., 29/1, 5–9, 39, 1956. (f ) Ind. Eng. Chem., 48/9,
1383–1388, 1956. (g) Mod. Plast., 35/1, 115–121, 238–247, 1957. (h) Buist, J.M.
and Lowe, A., Plast. Int. Trans., 27, 13–27, 1959. (i) Editorial, Br. Plast., 34/8,
408–417, 1961. (j) Erlich, V.L., Mod. Text. Mag., 44/2, 37–40, 1963. (k) Saunders,
J.H. and Frisch, K.C., Polyurethanes Chemistry and Technology, Part II: Technology,
Interscience, New York, 1964. (l) Bender, R.J., Handbook of Foamed Plastics, Lake
Publishing, Libertyville, IL, 1965. (m) Frisch, K.C. and Saunders, J.H., Plastic
Foams, Part I, Marcel Dekker, New York, 1972. (n) Oertel, G., Kunststoffe, 71,
2–7, 1981. (o) Allport, D.C., Plast. Rub. Proc. Appl., 4/2, 173–180, 1984. (p) Oertel,
G., Polyurethane Handbook, Hanser, New York, 1985. (q) Woods, G., The ICI
Polyurethanes Book, John Wiley & Sons, Chichester, England, 1987. (r) Ulrich,
H., An Encyclopedia of Chemical Technology, 3rd ed., John Wiley & Sons, New
York, 1982, vol. 23, 576–608.
10. (a) Nortman, P.B., Plast. Tech., 42–49, December, 1960. (b) Wouters, R.P., Confer-
ence Papers, UTECH 88, Crain Communications, London, 1988. (c) Manno, P. J.,
Conference Papers, UTECH 90, Crain Communications, London, 1990, 5–8. (d)
Anon., Facts and Figures of the U.S. Plastics Industry, Society of the Plastics
Industry, Washington, D.C., 1990, 51–54. (e) Manno, P.J., 1991 U.S. Foamed
Plastics Markets and Directory, Technomic, Lancaster, PA, 1991, 5–9.
11. (a) Pinner, S.H., Plastics (London), 11, 206–211, 215, 1947. (b) Cooper, W., Pear-
son, R.W., and Darke, S., Ind. Chem., 3, 121–126, 1960. (c) Saunders, J.H. and
Frisch, K.C., Polyurethanes: Chemistry and Technology, Robert E. Krieger Publish-
ing, Huntington, NY, 1978. (d) Sayigh, A.A.R., Ulrich, H., and Farrissey, W.J.,
Condensation Monomers, Wiley-Interscience, New York, 1972, 369–476. (e) Satch-
ell, D.P.N. and Satchell, R.S., Chem. Soc. Rev., 4/2, 231–250, 1975. (f ) Richter, R.
and Ulrich, H., The Chemistry of Cyanates and Their Thio Derivatives, Part 2, ed.
S. Patai, John Wiley & Sons, New York, 1977.
© 2004 by CRC Press LLC
www.ketabdownload.com
 12. Lovering, E.G., Laidler, K.J., Can. J. Chem. 1962, 40, 26–30.
13. (a) Smith, H.A., J. Appl. Polym. Sci., 7, 85–95, 1963. (b) Smith, H.A., J. Polym.
Sci.: Part A-1, 6, 1299–1306, 1968. (c) Hartley, F.D., Cross, M.M., and Lord, F.
W., Advances in Polyurethane Technology, ed. J.M. Buist and H. Gudgeon, Ma-
claren and Sons, London, 1968, 127–140. (d) Cole, K.C. et al., Proceedings of the
SPI-31st Annual Technical/Marketing Conference, Technomic, Lancaster, PA, 1988,
2–10. (e) Spirkova, M., Kubin, M., and Dusek, K., J. Macromol, Sci.-Chem., A27(4),
509–522, 1990. (f ) Dusek, K. and Spirkova, M., Cell. Polym., 9/2, 69–83, 1990.
14. Kurzer F., Chem. Rev., 56, 95–197, 1955.
15. Saunders, J.H. and Hansen, R.H., Plastic Foams, Part I, Marcel Dekker, New
York, 1972, 23–108.
16. Abbate, F.W., Plast. Cmpd., 9/4, 20, 23–24, 26–27, 1986.
17. Hill, B.G., Chemtech, 613–616, October, 1973.
18. Sparrow, D.J. and Thorpe, D., Telechelic Polymers, Vol. 2: Synthesis and Application,
ed. E.J. Goethals, CRC Press, Boca Raton, FL, 1989, 203.
19. Friedli, H.R., Reaction Polymers, ed. W.F. Gum, W. Riese, and H. Ulrich, Hanser,
Munich, 1992, 68.
20. (a) Dege, G.J., Harris, R. L., and MacKenzie, J. S., J. Am. Chem. Soc., 81, 3374–3379,
1959. (b) Simons, D.M. and Verbanc, J.J., J. Polym. Sci., 44, 303–311, 1960. (c) Steiner,
E.C., Pelletier, R. R., and Trucks, R.O., J. Am. Chem. Soc., 86, 4678–4686, 1964. (d)
Fogiel, A.W., Macromolecules, 2/6, 581–587, 1969. (e) Livigni, R.A. et al., Polyethers,
ACS Symposium Series No. 6, ed. E.J. Vandenberg, American Chemical Society,
Washington, D.C., 1975, 29–31. (f ) Stanford, J.L. and Stepto, R.F.T., Br. Polym. J.,
6, 124–132, 1977. (g) Miller, D.R., Valles, E.M., and Macosko, C.W., Polym. Eng.
Sci., 19/4, 272–283, 1979. (h) Barksby, N. and Allen, G.L., Proceedings of the Poly-
urethanes World Congress 1993, Technomic, Lancaster, PA, 1993, 445–450. (i) Carr,
R.H., Hernalsteen, J., and Devos, J., J. Appl. Polym. Sci., 52/8, 1015–1022, 1994. (j)
David, D.J. and Staley, H.B., Analytical Chemistry of the Polyurethanes, Vol. 16, Part
III, Wiley-Interscience, New York, 1969, 291–292. (k) Earing, M.H. to Wyandotte
Chemical Company, U.S. Patent 3,271,462, 1966. (l) General Tire Corporation,
U.S. Patent 3,278,457, 1966. (m) Cuscurida, M. to Jefferson Chemical Corporation,
U.S. Patent 3,393,243, 1968. (n) McKellar, R.L. and Musolf, M.C. to Dow Corning
Corporation, U.S. Patent 3,485,861, 1969. (o) Herold, R.J. and Linigim, R. A.,
Polymerization Kinetics and Technology, Advances in Chemistry Series No. 128,
American Chemical Society , Washington, D.C., 1973. (p) Bleijenberg, K.C., Con-
ference Papers, UTECH 88, Crain Communications, London, 1988, 185–187. (q)
Schuchardt, J.L. and Harper, S.D., Proceedings of the SPI-32nd Annual Technical/
Marketing Conference, Technomic, Lancaster, PA, 1989, 360–364. (r) Heuvelsland,
A.J. to Dow Chemical Company, U.S. Patent 5,010,187, 1991. (s) Hinz, W., Pro-
ceedings of the SPI/ISOPA Polyurethanes World Congress 1991, Technomic, Lancaster,
PA, 1991, 519–523. (t) Penati, A., Maffezzoni, C., and Moretti, E., J. Appl. Polym.
Sci., 26, 1059–1071, 1981. (u) Baize, T. H., Ind. Eng. Chem., 53/11, 903–906, 1961.
(v) Powell, D.G., Puig, J.E., and van Leuwen, B.G., J. Cell Plast., 8/2, 90–99, 1972.
(w) Gaylord, N.G., ed., Polyethers, Vol. 1: Polyalkylene Oxides and Other Polyethers,
Interscience, New York, 1963. (x) Furukawa, J. and Saegusa, T., Polymerization of
Aldehydes and Oxides, Interscience, New York, 1963, 125–208. (y) Newton, R.A.,
Kirk-Othmer Encyclopedia of Chemical Technology, 3rd ed., John Wiley & Sons, New
York, 1982, vol. 18, 633–645. (z) Schauerte, K. et al., Polyurethane Handbook, ed. G.
Oertel, Hanser, New York, 1985, 42–50. (aa) Hampton, H.A., Hurd, R., and Shear-
ing, H.J., J. Oil Col. Chem. Ass., 43, 96–123, 1960. (bb) Yost, C. W., Plank, C.A., and

© 2004 by CRC Press LLC

www.ketabdownload.com
 Gerhand, E.R., Ind. Eng. Chem. Prod. Res. Dev., 6/2, 113–115, 1967. (cc) Yen, Y.C.,
Process Economics Program Report No. 45, Stanford Research Institute, Menlo Park,
CA, 1968. (dd) Wright, P. and Cumming, A.P.C., Solid Polyurethane Elastomers,
Gordon and Breach, New York, 1969, 66–68. (ee) Guibert, R.M., Plank, C.A., and
Gerhard, E.R., Ind. Eng. Chem. Prod. Res. Dev., 10/4, 497–500, 1971. (ff ) Christian-
son, L.R., Dheming, M., and Ochoa, E.L., J. Cell Plast., 13/2, 111–117, 1977. (gg)
Hampton, H.A., Hurd, R., and Shearing, H.J., J. Oil Col. Chem. Ass., 43, 96–123,
1960. (hh) Sparrow, D.J. and Thorpe, D., Telechelic Polymers, Vol. 2: Synthesis and
Application, ed. E.J. Goethals, CRC Press, Boca Raton, FL, 1989, 181–228.
21. Barrett, K.E.T., Dispersion Polymerization in Organic Media, John Wiley & Sons,
London, 1975.
22. Napper, D.H., Polymeric Stabilization of Colloidal Particles, Academic Press, Lon-
don, 1983.
23. (a) Kuryla, W.C. et al., J. Cell. Plast., 2/2, 84–96, 1966. (b) Bamford, C.H. and
Tipper, C.F.H., Free Radical Polymerization, Elsevier, New York, 1976. (c) Kahrs,
K.H. and Zimmermann, J.W., Makromol. Chem., 58, 75–103, 1962. (d) Critchfield,
F.E., Koleske, J.V., and Priest, D.C., Rubber Chem. Tech., 45, 1467–1484, 1972. (e)
Cloetens, R. et al., Proceedings of the SPI/FSK Polyurethanes World Congress,
Technomic, Lancaster, PA, 1987, 480–489. (f ) Ramlow, G.G., Heyman, D.A., and
Grace, O.M., J. Cell Plast., 19/4, 237–242, 1983. (g) Shears, J.H. et al., Proceedings
of the 1996 SPI Polyurethanes Conference, Technomic, Lancaster, PA, 1996,
150–155. (h) Yen, Y.C. and Tsao, T.S., Process Economics Program Report No. 45A,
Stanford Research Institute, Menlo Park, CA, 1982, 119–152.
24. (a) Spitler, K.G. and Lindsey, J.J., J. Cell Plast., 17/1, 43–50, 1981. (b) Vehlewald,
P. and Volland, R., Kunstoffe, 73/8, 439–443, 1983. (c) Koshute, M.A. and Freitag,
H.A., Proceedings of the SPI/FSK Polyurethanes World Congress, Technomic, Lan-
caster, PA, 1987, 502–507.
25. (a) Reischl, A., Muller, H., and Wagner, K. to Bayer, U.S. Patent 4,260,530, 1981.
(b) Rowlands, J.P. to Interchem International S.A., U.S. Patent 4,374,209, 1983.
(c) Pickin, K., Urethanes Tech., June 23–24, 1984. (d) Carroll, W.G. and Farley, P.
to Imperial Chemical Industries, U.S. Patent 4,452,923, 1984. (e) Carroll, W.G.
to Imperial Chemical Industries, U.S. Patent 4,554,306, 1985. (f ) Pal, J.M., Cos-
man, J.P., and Tan, K. to Dow Chemical Company, U.S. Patent 5,068,280, 1991.
(g) Wujcik, S.E., Christman, D.L., and Grace, O.M. to BASF Corporation, U.S.
Patent 5,179,131, 1993. (h) Van Veen, K.J. and Blair, G.R. to Woodbridge Foam
Corporation, U.S. Patent 5,292,778, 1994. (i) Yeakey, E.L. and Cuscurida, M. to
Arco Chemical Company, U.S. Patent 4,785,026, 1988.
26. Knight, J.E., Proceedings of the SPI-27th Annual Technical/Marketing Conference,
Technomic, Lancaster, PA, 1982, 225–227.
27. van der Wal, H.R., Proceedings of the SPI/FSK Polyurethanes World Congress,
Technomic, Lancaster, PA, 1987, 493–497.
28. Narayan, T. and Patton, J.T. to BASF Wyandotte Corporation, U.S. Patent
4,326,043, 1982.
29. Sam, F.O., Stefani, D., and Gallo, B., Proceedings of the Polyurethanes World
Congress 1993, Technomic, Lancaster, PA, 1995, 62–69.
30. Kageoka, M., Tairaka, Y., and Kodama, K., Proceedings of the 1995 SPI Polyure-
thanes Conference, Technomic, Lancaster, PA, 1995, 62–69.
31. Gibson, L.J. and Ashby, M.F., Cellular Solids, 2nd ed., Cambridge University
Press, Cambridge, England, 1997.

© 2004 by CRC Press LLC

www.ketabdownload.com
 32. Gier, D.R. et al., Polyurethanes Expo ’98, Technomic, Lancaster, PA, 1998, 227–237.
33. Macosko, C.W., Rheology: Principles, Measurements, and Applications, John Wiley
& Sons, New York, 1994.
34. Lee, D.I., J. Paint Tech., 42, 579, 1970.
35. Cloetens, R.C., Lidy, W.A., and Thanh, H.P. to Dow Chemical Company, U.S.
Patent 4,883,832, 1989.
36. Moore, D.R., Plowman, K.R., and Neill, P.L. to Dow Chemical Company, U.S.
Patent 5,494,957, 1996.
37. Lidy, W.A. and Thanh, H.P. to Dow Chemical Company, U.S. Patent 4,831,076,
1989.
38. Langdale-Smith, R.A. to Union Carbide Corporation, U.S. Patent 4,208,314, 1980.
39. Wright, P. and Cumming, A.P.C., Solid Polyurethane Elastomers, Gordon and
Breach, New York, 1969, 68–70.
40. (a) Sittig, M., Amines, Nitriles and Isocyanates, Processes and Products, 1969, Noyes
Development Corporation, Park Ridge, NJ, 1969. (b) Ozaki, S., Chem. Rev., 72/
5, 457–496, 1972. (c) Ranney, M.W., Isocyanates Manufacture, Noyes Data Cor-
poration, Park Ridge, NJ, 1972. (d) Ulrich, H., Introduction to Industrial Polymers,
Hanser, Munich, 1982, 83–84.
41. (a) Eldib, I.A., Hydrocarbon Proc. Petrol. Ref., 42/12, 121–127, 1963. (b) Chadwick,
D.H. and Cleveland, T.H., Kirk-Othmer Encyclopedia Of Chemical Technology, 3rd
ed., John Wiley & Sons, New York, 1982, vol. 13, 789–818. (c) Lunde, K.E. and
Bratt, L.C., Process Economics Program Report No. 1, Stanford Research Institute,
Menlo Park, CA, 1963.
42. (a) Bailey, M.E., Kirss, V., and Spaunburgh, R.G., Ind. Eng. Chem., 48/4, 794–797,
1956. (b) Lenz, R.W., Organic Chemistry of Synthetic High Polymers, Interscience,
New York, 1967, 180–185. (c) Bender, R. J., Handbook of Foamed Plastics, Lake
Publishing Corporation, Libertyville, IL, 1965, 130. (d) McGinn, C.E. and
Spaunburgh, R.C., Symposium on Isocyanate Polymers, Paper No. 38, Amer-
ican Chemical Society Meeting, Atlantic City, NJ, September 1956, 16/2.
43. (a) Gmitter, G.T. and Gruber, E.E., SPE J., 13/1, 27–31, 1957. (b) Berger, S.E. et
al., SPE J., 17/2, 170–173, 1961. (c) Smith, C.H. and Petersen, C.A., SPE J., 18/
4, 455–459, 1962. (d) Mispreuve, H. et al., J. Cell. Plast., 24/4, 348–358, 1988. (e)
Nierzwichi W. and Walczynski, B., J. Appl. Polym. Sci., 41, 907–915, 1990. (f )
Aneja, A., Wilkes, G.L., and Rightor, E.G., J. Polym. Sci. Polym. Phys. Ed., 41,
258–268, 2003.
44. Vargas, J.B., Loefgren, B.C., and Kato, H., Proceedings of the SPI-18th Annual
Urethane Division Technical Conference, Technomic, Lancaster, PA, 1972, 17–35.
45. (a) Yen, Y.C., Process Economics Program Report No. 1-A, Stanford Research In-
stitute, Menlo Park, CA, 1968–1973. (b) Twitchett, H.J., Chem. Soc. Revs.,
3, 209–230, 1974. (c) Chono, M., Fukuoka, S., and Kohno, M., Proceedings of the
SPI-6th International Technical/Marketing Conference, Technomic, Lancaster, PA,
1983, 394–398.
46. A. Wurtz, Ann., 71, 326, 1849.
47. A. W. Hofmann, Annales, 74, 9, 1850.
48. W. Hentschel, Berichte, 17, 1284, 1884.
49. (a) Saunders, J.H. and Slocombe, R.J., Chem. Rev., 43, 203, 1948. (b) Arnold, R.G.,
Nelson, J.A., and Verbanc, J.J., J. Chem. Ed., 34, 158, 1957.
50. Melander, L., Nature, 163, 599, 1949.
51. Roberts, R.A. et al., J. Am. Chem. Soc., 80, 4285, 1958.

© 2004 by CRC Press LLC

www.ketabdownload.com
 52. Tillet, J.G., J. Am. Chem. Soc., 5142, 1962.
53. (a) Bennett, G.M. and Youle, P.V., J. Chem. Soc., 1816, 1938. (b) Titov, A.I., Uspekhy
Khimii, 27, 845, 1958. (c) Hanson, C., Kaghazchi, T., and Praat, M.W.T., ACS
Symp. Ser., 22, 132, 1976.
54. (a) Carr, R.V.C. et al. to Air Products, U.S. Patent 4,604,214, 1986. (b) Baur, K.G.
to BASF, European Patent 503,387, 1992.
55. Waller, F.J. et al., Chem. Ind., 75, 289, 1998.
56. Pinna, F. et al., J. Catal., 150, 356, 1994.
57. Beckhaus, H., Waldau, E., and Witt, H. to Bayer, German Patent 3,611,677, 1987.
58. (a) Neri, G. et al., J. Mol. Catal. A: Chem., 111, 257, 1996. (b) Musolino, M.G.
et al., J. Chromatogr. A, 818, 123, 1998.
59. (a) Musolino, M.G. et al., Stud. Surf. Sci. Catal., 108, 239, 1997. (b) Neri, G. et al.,
Ind. Eng. Chem. Res., 34, 2226, , 1995. (c) Neri, G. et al., J. Mol. Catal. A: Chem.,
95, 235, 1995.
60. (a) Westerterp, K.R. et al., Chem. Eng. Sci., 43, 2229, , 1988. (b) Janssen, H.J. et al.,
Ind. Eng. Chem. Res., 29, 1822, , 1990. (c) Westerterp, K.R., Janssen, H.J., and van
der Kwast, H.J., Chem. Eng. Sci., 47, 4179, , 1992. (d) Molga, E.J. and Westerterp,
K.R., Chem. Eng. Sci., 47, 1733, , 1992.
61. Allied Chemical, British Patent 1,007,785, 1965.
62. (a) Gemassmer, A.M. et al. to Bayer, British Patent 901,377, 1962. (b) Bayer,
British Patent 761,590, 1954. (c) Bayer, British Patent 761,592, 1954. (d) Vaganay,
J. and Wevert, S., French Patent 1,071,629, 1954. (e) Nicholas, A.J. and Twitchett,
H.J. to ICI, British Patent 1,034,285, 1966. (f ) Kantyka, T.A. and Toyne, C. to
ICI, British Patent 806,903, 1959. (g) Cleveland, T.H. to Mobay, British Patent
979,259, 1965. (h) Slocombe, R.J., Flores, H., and Cleveland, T.H. to Monsanto,
U.S. Patent 2,680,127, 1954. (i) Bloom, A., Freyermuth, H.B., and Normington,
J.B. to GAF, U.S. Patent 2,847,440, 1958. (j) Boehme, O., Mott, F., and Pfirschke,
J. to Bayer, U.S. Patent 2,875,225, 1959. (k) Maeda, S., Yamori, K., and Kaminaka,
H. to Sumitomo, U.S. Patent 3,184,494, 1965.
63. Latourette, H.K. and Johnson, O.H., U.S. Patent 2,908,703, 1959.
64. Denton, W.I., Hammond, P.D., and Wood, J.A. to Olin, U.S. Patent 3,287,387, 1966.
65. (a) Upjohn, British Patent 1,238,669, 1968. (b) Gemassmer, A.M. et al. to Bayer,
British Patent 901,377, 1962. (c) Beck, T.R. to Du Pont, U.S. Patent 2,822,373,
1958. (d) Van Horn, I.B. to Mobay, U.S. Patent 3,507,626, 1970. (e) Horn, P.,
Koehler, W., and Bittler, R. to BASF, U.S. Patent 4,128,569, 1978.
66. (a) Ewald, R.M. to Mobay, U.S. Patent 3,321,283, 1967. (b) Irwin, C.F. to Du
Pont, U.S. Patent 337,167, 1975.
67. Mitrowsky, A. et al. to Bayer, U.S. Patent 3,947,484, 1976.
68. (a) Gemassmer, A.M. to Bayer, U.S. Patent 3,188,337, 1965. (b) Nicholas, A.J.
and Twitchett, H.J. to ICI, U.S. Patent 3,465,021, 1969. (c) Yamamoto, R. et al.
to Mitsui, U.S. Patent 4,422,976, 1983.
69. Soos, R. and Petnehazy, I, Periodica Polytechnica, 15, 119, 1971.
70. (a) Delfs, D. and Muenz, F. to Bayer, German Patent 942,145, 1956. (b) Bircher, B.J.
and Wild, J.H. to ICI, British Patent 1,050,555, 1966. (c) Irwin, C.F. to Du Pont, U.S.
Patent 2,644,007, 1953.
71. Suzuki, S. et al. to Mitsubishi, U.S. Patent 3,484,472, 1969.
72. Lomax, R. and Twitchett, H.J. to ICI, British Patent 1,120,770, 1968.
73. (a) Cooper, J.R. to Du Pont, French Patent 1,368,031, 1964. (b) Cooper, J.R. to
Du Pont, U.S. Patent 3,234,253, 1966.
© 2004 by CRC Press LLC
www.ketabdownload.com
 74. (a) Fuchs, W., Platz, R., and Vogt, V. to BASF, U.S. Patent 3,960,916, 1976. (b)
Fuchs, W., Platz, R., and Vogt, V. to BASF, U.S. Patent 3,978,105, 1976.
75. Hardy, E.E., U.S. Patent 2,839,559, 1958.
76. Pagano, A.S. to Rohm and Haas, U.S. Patent 3,746,746, 1973.
77. Bircher, B.J., Dodman, D., and Wild, J.H. to ICI, British Patent 1,085,824, 1967.
78. (a) Stauffer, British Patent 1,114,085, 1963. (b) Stamm, W. to Stauffer, U.S. Patent
3,440,268, 1969.
79. Arndt, O., Semler, G., and Schaeffer, G. to Hoechst, U.S. Patent 3,835,172, 1974.
80. (a) Flores, H. to Monsanto, U.S. Patent 2,680,129, 1954. (b) Irwin, C.F. and
Swamer, F.W. to Du Pont, U.S. Patent 2,757,183, 1956. (c) Hettich, B.V. and
Latourette, H.K. to FMC, U.S. Patent 3,226,410, 1965.
81. (a) Vaganay, J. et al. to Tolochimie, German Patent 1,175,666, 1964. (b) Michelet,
J. and Alheritiere, L. to Tolochimie, French Patent 1,469,105, 1967. (c) Alheritiere,
L. and Repper, M. to Tolochimie, French Patent 1,476,755, 1967. (d) Burkhardt,
T. and Findeisen, K. to Bayer, French Patent 2,329,051, 1975. (e) Clifford, B.H.
and Hargreaves, J.R. to ICI, British Patent 1,037,933, 1966. (f ) Thomas, R.M.
and Blaich, M.C. to Olin, U.S. Patent 3,119,856, 1964. (g) Clifford, B.H. and
Hargreaves, J.R. to ICI, British Patent 1,054,328, 1967. (h) Bloom, A., Freyer-
muth, H.B., and Normington, J.B. to GAF, U.S. Patent 2,911,429, 1959. (i) Bloom,
A., Freyermuth, H.B., and Normington, J.B. to GAF, U.S. Patent 2,875,226, 1959.
(j) Skiles, B.F. to Du Pont, U.S. Patent 2,908,704, 1959. (k) Bloom, A., Freyermuth,
H.B., and Normington, J.B. to GAF, U.S. Patent 2,999,873, 1961.
82. (a) Biskup, K., Koenig, C., and Waldau, E. to Bayer, European Patent 570,799,
1993. (b) Joulak, F., Revelant, D., and Vacus, P. to Rhone-Poulenc, European
Patent 593,334, 1994. (c) Armand, J. et al. to Rhone-Poulenc, European Patent
699,657, 1996.
83. Anon., Preparation of Prepolymers, Laboratory to Full-Scale Production, The Dow
Chemical Company, Form No. 109-00895-290-SAI.
84. (a) Heiss, H.L. et al., Ind. Eng. Chem., 51/8, 929–934, 1959. (b) Wall, J.R., Chem.
Eng. Prog., 57/10, 48–51, 1961. (c) Saunders, J.H. and Frisch, K.C. Polyurethanes,
Chemistry and Technology, Part II. Technology, Interscience, New York, 1964, 8–43.
(d) Martin, R.A., Hoy, K.L., and Peterson, R.H., Ind. Eng. Chem., Prod. Res. Dev.,
6/4, 218–222, 1967. (e) Peebles, L.H., Macromolecules, 7/6, 872–881, 1974. (f )
Wang, T. and Lyman, D.J., Polym. Bul., 27, 549–555, 1992.
85. (a) Saunders, J.H. and Frisch, K.C. Polyurethanes, Chemistry and Technology, Part
II, Technology, Interscience, New York, 1964, 8–49. (b) Gemeinhardt, P.G., Hand-
book of Foamed Plastics, ed. R.J. Bender, Lake Publishing Corporation, Liber-
tyville, IL, 1965, 173–208. (c) Zwolinski, L.M. and Frink, J.W., J. Cell. Plast., 8/
1, 20–29, 1972. (d) Brydson, J.A., Plastics Materials, Butterworth Scientific, Lon-
don, 1982, 709–712. (e) Hartley, F.D., Cross, M.M., and Lord, F.W., Advances in
Polyurethane Technology, ed. J.M. Buist and H. Gudgeon, H., Maclaren and Sons,
London, 1968, 135–136. (f ) Klempner, D. and Frisch, K.C., Handbook of Polymeric
Foams and Foam Technology, Hanser, Munich, 1991, 53–54. (g) Elwell, R.J. et al.
to Dow Chemical Company, U.S. Patent 5,114,989, 1992.
86. (a) Ferrigno, T.H., SPE J., June, 638–640, 1960. (b) Byall, R.D., paper presented
at American Chemical Society Meeting, Rubber Chemistry Division, Oct. 20,
1965. (c) Reilly, A.F., Chem. Eng. Prog., 63/5, 104–108, 1967. (d) Currier, V., Plast.
Tech., 12/8, 35–37, 1966. (e) Doyle, E.N., The Development and Use of Polyurethane
Products, McGraw-Hill, New York, 1971, 77–82.
© 2004 by CRC Press LLC
www.ketabdownload.com
 87. Katz, H.S. and Milewski, J.V., Hand Book of Fillers for Plastics, Van Nostrand
Reinhold, New York, 1987.
88. Tabor, R.L. et al., Proceedings of the SPI-34th Annual Technical/Marketing Confer-
ence, Technomic, Lancaster, PA, 1992, 514–528.
89. Hersch, P., Plast. Tech., 13/12, 49–53, 1967.
90. (a) Bryant, R.M. and Stewart, H.F., J. Cell Plast., 9/2, 99–102, 1973. (b) Plumb,
J.B., Atherton, J.H., Block Copolymers, ed. D.C. Allport and W.H. Janes, John
Wiley & Sons, New York, 1973, 305–353. (c) Brydson, J.A., Plastics Materials,
Butterworth Scientific, London, 1982, 730–766. (d) Boudreau, R.J., Mod. Plast.,
133–147, January 1967. (e) Frey, J.H., Proceedings of the 1995 SPI Polyurethanes
Conference, Technomic, Lancaster, PA, 1995, 224–229. (f ) Snow, S.A., Fenton,
W.N., and Owen, M.J., J. Cell. Plast., 26/2, 172–182, 1990. (g) Burkhart, G.,
Schlons, H.-H., and Zellmer, V., Proceedings of the SPI-34th Annual Technical/
Marketing Conference, Technomic, Lancaster, PA, 1992, 359–363. (h) Burkhart,
G., Klietsch, J., and Zellmer, V., Proceedings of the SPI/ISOPA Polyurethanes World
Congress 1991, Technomic, Lancaster, PA, 1991, 127–131. (i) Grabowski, W. and
Desnier, M.C., Proceedings of the Polyurethanes World Congress 1993, Technomic,
Lancaster, PA, 1993, 81–88. (j) Brune-Fischer, A., Burkhart, G., and Zellmer, V.,
Proceedings of the 1994 SPI Polyurethanes Conference, Technomic, Lancaster, PA,
1994, 267–273. (k) Brune, A., Klietsch, J., and Zellmer, V., Conference Papers,
UTECH 96, Crain Communications, London, 1996, Paper 12. (l) Dahm, M.,
Cellular Plastics, National Academy of Sciences–National Research Council Pub-
lication 1462, 52–63, 1967. (m) Rossmy, G. R. et al., J. Cell. Plast., 13/1, 26–35,
1977. (n) Woods, G.W., Flexible Polyurethane Foams, Chemistry and Technology,
Applied Science, New Jersey, 1982, 68–72. (o) Stevens, R.E. et al., Proceedings of
the Polyurethanes World Congress 1997, Technomic, Lancaster, PA, 1997, 118–129.
(p) Kaushiva, B.D. et al., Polymer, 41, 285–310, 2000.
91. Brecker, L.R., Plast. Eng., March, 39–42, 1977.
92. Rusch, T.E. and Raden, D.S., Plast. Compd., 61–74, July/August 1980.
93. Listemann, M.L. et al., Proceedings of the Polyurethanes World Congress 1993,
Technomic, Lancaster, PA, 1993, 595–608.
94. (a) Winkler, J., SPE J., 12/11, 23–26, 1956. (b) Alzner, B.G. and Frisch, K.C., Ind.
Eng. Chem., 51/5, 715–716, 1959. (c) Gmitter, G.T., Gruber, E.E., and Joseph,
R.D., SPE J., 15/11, 957–960, 1959. (d) Listermann, M.L. et al. to Air Products
and Chemicals Inc., U.S. Patent 5,508,314, 1996.
95. (a) Burkhart, G., Kollmeier, H.J., and Schloens, H.H., J. Cell. Plast., 20/1, 37–41,
1984. (b) Malwitz, N. et al., Proceedings of the SPI-30th Annual Polyurethane
Technical/Marketing Conference, Technomic, Lancaster, PA, 1986, 338–353. (c)
Malwitz, N. and Kresta, J.E., Proceedings of the SPI/FSK Polyurethanes World
Congress, Technomic, Lancaster, PA, 1987, 826–836.
96. (a) Anon., Plast. Tech., 8/4, 26–32, 1962. (b) Sandridge, R.L., Gemeinhardt, P.
G., and Saunders, J.H., SPE Trans., 3/2, 117–122, 1963. (c) Mack, G.P., Mod.
Plast., 42/4, 148–160, 194, 1964.
97. (a) Letts, J.B., Gerkin, R.M., and Martineau, P., Proceedings of the SPI/ISOPA
Polyurethanes World Congress 1991, Technomic, Lancaster, PA, 1991, 174–181. (b)
Diblitz, K. and Diblitz, C., Proceedings of the Polyurethanes World Congress 1993,
Technomic, Lancaster, PA, 1993, 619–624. (c) Derderian, E.J. et al., Proceedings
of the 1994 SPI Polyurethanes Conference, Technomic, Lancaster, PA, 1994,
274–279. (d) Casati, F.M.H., Arbir, F.W., and Raden, D.S., Proceedings of the SPI-
27th Annual Technical/Marketing Conference, Technomic, Lancaster, PA, 1982,

© 2004 by CRC Press LLC

www.ketabdownload.com
 35–48. (e) Casati, F.M.H., Arbir, F.W., and Tylenda, E.J., Proceedings of the SPI-
6th International Technical/Marketing Conference, Technomic, Lancaster, PA, 1983,
161–171.
98. Stepek, J. and Daoust, H., Additives for Plastics, Springer-Verlag, New York, 1983.
99. (a) Miley, J.W. and Moore, P.D., Proceedings of the SPI-26th Annual Technical
Conference, Technomic, Lancaster, PA, 1981, 83–86. (b) Moore, P.D., Miley, J. W.,
and Bates, S.H., Proceedings of the SPI-27th Annual Technical/Marketing Conference,
Technomic, Lancaster, PA, 1982, 255–261. (c) Bates, S.H. and Miley, J.W., Pro-
ceedings of the SPI-30th Annual Technical/Marketing Conference, Technomic, Lan-
caster, PA, 1986, 160–165. (d) Bates, S.H., Moore, P.D., and Miley, J. W.,
Proceedings of the SPI/FSK Polyurethanes World Congress, Technomic, Lancaster,
PA, 1987, 916–921. (e) Miley, J., Pure & Appl. Chem., 68/7, 1423–1428, 1996. (f )
Vielee, R.C. and Haney, T.V., Coloring of Plastics, ed. T.G. Webber, Wiley-Inter-
science, New York, 1979, 191–204.
100. (a) Davis, A. and Sims, D. Weathering of Polymers, Applied Science, London,
1983, 222–237. (b) Beachell, H.C. and Ngoc Son, C.P., J. Appl. Polym. Sci., 7,
2217–2237, 1963. (c) Nevskii, L.V. and Tarakanov, O.G., Soviet Plast., 9, 47–48,
1967 (translation of Plast. Massy., 9, 1966). (d) Bieneman, R.A., Polyurethane
Technology, Interscience Publishers, New York, 1969, 230–231. (e) Bellettini, A.G.
and Harrison, J.R., Cell. Polym., 3, 241–247, 1984. (f ) Capocci, G., Proceedings of
the SPI-30th Annual Polyurethane Technical/Marketing Conference, Technomic,
Lancaster, PA, 1986, 220–227. (g) Rabek, J.F., Photostabilization of Polymers, Elsevi-
er, London, 1990. (h) Valentine, C., Craig, T.A., and Hager, S.L., Proceedings of
the SPI-34th Annual Technical/Marketing Conference, Technomic, Lancaster, PA,
1992, 149–155. (i) Sutton, R.G. and Steever, J.A., J. Cell. Plast., 19/6, 372–374,
1983.
101. (a) Lyons, J.W., The Chemistry and Uses of Fire Retardants, Robert Krieger
Publishing, Malabar, FL, 1987. (b) Babiec, J.S. et al., Plast. Tech., June, 47–50, 1975.
(c) Treadwell, K. to M&T Chemicals Inc., U.S. Patent 3,635,821, 1972. (d)
Cobbledick, D.S. and Norman, A.J. to Gencorp Inc., U.S. Patent 3,810,851, 1974.
(e) Cobbledick, D.S. and Norman, A.J. to General Tire and Rubber Company,
U.S. Patent 3,876,571, 1975. (f ) Molbert, R.A. to General Tire and Rubber
Company, U.S. Patent 3,884,849, 1975. (g) Fabris, H.J. and Maxey, E.M. to General
Tire and Rubber Company, U.S. Patent 3,925,266, 1975. (h) Cobbledick, D.S. to
General Tire and Rubber Company, U.S. Patent 3,931,062, 1976. (i) Stone, H.,
Proceedings of the 1978 Fire Retardant Chemicals Association Meeting, Houston, TX,
March 12–15, Fire Retardant Chemicals Association, Westport, CT, 1978, 257–273.
(j) Woods, G.W., Flexible Polyurethane Foams, Chemistry and Technology, Applied
Science, NJ, 1982, 88. (k) Grace, O.M., Mericle, R.E., and Taylor, J.D., J. Cell. Plast.,
21/5, 311–317, 1985. (l) Natoli, F.S. et al., Proceedings of the SPI/ISOPA
Polyurethanes World Congress 1991, Technomic, Lancaster, PA, 1991, 762–765. (m)
Wujcik, S.E. et al., Proceedings of the SPI-34th Annual Technical/Marketing
Conference, Technomic, Lancaster, PA, 1992, 672–677. (n) Bhatnagar, V. M., Fire
Retardant Formulations Handbook, vol. 1, Progress in Fire Retardancy Series,
Technomic, Lancaster, PA, 1972. (o) Handbook of Flame Retardant Chemicals and
Fire Testing Services, Technomic, Lancaster, PA, 1992. (p) Kuryla, W.C. and Papa,
A.J., Flame Retardancy of Polymeric Materials, vol. 3, Marcel Dekker, New York,
1975, 1–133.
102. (a) Darby, R.T. and Kaplan, A.M., Appl. Microbiol., 16/6, 900–905, 1968. (b)
Kapan, A.M. et al., Dev. Ind. Microbiol., 9, 201–217, 1968. (c) Rei, N.M., J. Coated

© 2004 by CRC Press LLC

www.ketabdownload.com
 Fab., 8, 21–32, 1978. (d) Tamborini, S.M., Mahoney, C.J., and McEntee, T.C.,
Proceedings of the SPI-27th Annual Technical/Marketing Conference, Technomic,
Lancaster, PA, 1982, 308–313. (e) Baseman, A.L., Plast. Tech., 9, 33–37, 1966. (f )
Patarcity, R., Stern, E., and Murthy, U., Proceedings of the SPI-6th International
Technical/ Marketing Conference, Technomic, Lancaster, PA, 1983, 142–144. (g)
Brophy, J.F., Proceedings of the SPI-29th Annual Technical/ Marketing Conference,
Technomic, Lancaster, PA, 1985, 224–229. (h) Battice, D.R., Hales, M.G., Proceed-
ings of the SPI-29th Annual Technical/Marketing Conference, Technomic, Lancaster,
PA, 1985, 108–112.
103. (a) Knobel, T.M., Kennedy, E.E., and Walker, M.A. to Dow Chemical Company,
U.S. Patent 4,618,630, 1986. (b) Sweet, F.H. et al., J. Cell. Plast., 22/2, 139–146,
1986.
104. (a) Kopsheva, L.M., Shamov, I.V., and Tarakanov, O.G., J. Appl. Chem. USSR.
Translated from Zh. Prikl. Khim. (Leningrad), 57/6, 1358–1361, 1984. (b) White,
K.B. et al., J. Cell. Plast., 22/3, 331–339, 1986. (c) Hager, S.L. et al., Proceedings
of the SPI-34th Annual Technical/Marketing Conference, Technomic, Lancaster, PA,
1992, 529–534. (d) Koehler, C.E., McClellan, T.R., and Murray, P.L. to The Up-
john Company, U.S. Patent 4,246,364, 1981. (e) Grube, L.L. and Horner, C.J. to
GAF, U.S. Patent 4,444,916, 1984. (f ) Wood, R.J. to Stephan Company, U.S. Patent
4,529,744, 1985. (g) Goel, A.B. and Tufts, T.A. to Ashland Oil Incorporated, U.S.
Patent 4,579,877, 1986. (h) Graefe, P.U., Gajewski, V.J., and Jimenez, I.A. to Inter-
Polymer Research Corporation, U.S. Patent 3,929,730, 1975. (i) Nelson, D.L.,
Matijega, R.J., and Miller, D.P. to Dow Chemical Company, U.S. Patent
4,585,803, 1986. (j) Majewski, W., Khalil, H., and Wypych, J. to Tremco Ltd.,
U.S. Patent 5,126,421, 1992. (k) Glynn, K.T. and Doyle, E.N. to Crowley Chem-
ical Company, U.S. Patent 5,104,904, 1992. (l) Wang, D.S.T., J. Cell. Plast., 15/3,
144–151, 162, 1979. (m) Barron, B.G. to Dow Chemical Company, U.S. Patent
3,993,576, 1976. (n) Olstowski, F. and Peffley, R.D. to Dow Chemical Company,
U.S. Patent 4,485,031, 1984. (o) Olstowski, F. and Peffley, R.D. to Dow Chemical
Company, U.S. Patent 4,485,032, 1984. (p) Carswell, R. to Dow Chemical Com-
pany, U.S. Patent 4,981,877, 1991.
105. (a) Klesper, E., Rubber Age, Oct., 84–87, 1958. (b) Burt, J.G. and Brizzolara, D.F.,
Proceedings of the SPI-18th Annual Technical Conference, Technomic, Lancaster,
PA, 1975, 35–39. (c) Sayad, R.S. and Williams, K.W., Proceedings of the SPI-24th
Annual Urethane Division Technical Conference, Technomic, Lancaster, PA, 1978,
1–8. (d) Sayad, R.S. and Williams, K.W., Plast. Tech., June, 71–74, 1979. (e)
Handbook for Reducing and Eliminating Chlorofluorocarbons in Flexible Polyurethane
Foams, United States Environmental Protection Agency, 21A-4002, April 1991.
(f ) Walmsley, G.D. to Hickory Springs Manufacturing Company, U.S. Patent
5,120,771, 1992. (g) Graff, G., Mod. Plast., 70/4, 32–36, 1993. (h) Stone, H. et al.
to PMC Incorporated, U.S. Patent 4,906,672, 1990. (i) Farkas, P.V. and Duley,
J.A. to Woodbridge Foam Company, U.S. Patent 5,081,162, 1992. (j) Doyle, E.N.
and Carson, S., U.S. Patent 5,120,770, 1992. (k) Fiorentini, C. et al., Proceedings
of the 1994 SPI Polyurethanes Conference, Technomic, Lancaster, PA, 1994, 12–16.
(l) Eiben, R.G. et al., Proceedings of the 1995 SPI Polyurethanes Conference, Tech-
nomic, Lancaster, PA, 1995, 70–73. (m) Fiorentini, C. et al., Proceedings of the
1995 SPI Polyurethanes Conference, Technomic, Lancaster, PA, 1995, 491–496. (n)
Eiben, R.G., Conference Papers, UTECH 96, Crain Communications, London,
1996, Paper 31. (o) Noakes, C.W. and Casagrande, G., Conference Papers, UTECH
96, Crain Communications London, 1996, Paper 11. (p) Taverna, M. and Cor-

© 2004 by CRC Press LLC

www.ketabdownload.com
 radi, P., Conference Papers, UTECH 96, Crain Communications, London, 1996,
Paper 10. (q) Taverna, M., Meloth, H., and Griffiths, T., Proceedings of the 1996
SPI Polyurethanes Conference, Technomic, Lancaster, PA, 1996, 164–168. (r) Fran-
klin, R. et al. to Azko N.V., U.S. Patent 5,030,664, 1991. (s) Bradford, L., Franklin,
R., and Williams, B., Proceedings of the SPI-34th Annual Technical/Marketing
Conference, Technomic, Lancaster, PA, 1992, 668–671. (t) Lichtenberg, F.W., ed.,
CFCs and the Polyurethane Industry: A Compilation of Technical Publications, 1986-
1988, Technomic, Lancaster, PA, 1988. (u) Anon., Eliminating CFCS in Foam
Plastics, The Society of the Plastics Industry, 1989. (v) Lichtenberg, F.W., ed.,
CFCs and the Polyurethane Industry, Volume 2: A Compilation of Technical Publi-
cations, 1988-1989, Technomic, Lancaster, PA, 1990. (w) Lichtenberg, F.W., ed.,
CFCs and the Polyurethane Industry, Volume 3: A Compilation of Technical Publi-
cations, 1990, Technomic, Lancaster, PA, 1990. (x) Methven, J. M., RAPRA Review
Reports, 3/1, 1, 5–33, 1990. (y) Zwolinski, L.M., Conference Papers, UTECH 90,
Crain Communications, London, 1990, 23–24. (z) Dwyer, F.J., Knopeck, G.M.,
and Zwolinski, L.M., Proceedings of the SPI-33rd Annual Polyurethane Technical/
Marketing Conference, Technomic, Lancaster, PA, 1990, 400–406. (aa) Decaire,
B.R. et al., Proceedings of the SPI-34th Annual Technical/Marketing Conference,
Technomic, Lancaster, PA, 1992, 2–11. (bb) White, L., Urethanes Tech., 13/4, 42,
1996.
106. Young, T., Phil. Trans. R. Soc. (London), 95, 65–87, 1805.
107. Laplace, P.S., Traité de Mécanique Céleste, vol. 4, Courcier, Paris, 1806.
108. Plateau, J., Mem. Acad. R. Sci. Belgique, 23, Series 2, 1849.
109. Bikerman, J.J., Foams: Theory and Industrial Applications, Reinhold, New York,
1953.
110. Yazdi, A.V. and Beckman, E.J., Plast. Eng. (NY), 33 (Polymer Devolatilization),
67–102, 1996.
111. Kanner, B. and Decker, T.G., J. Cell. Plast., 5/1, 32–39, 1969.
112. LaMer, V.K., Ind. Eng. Chem., 44, 1270–1277, 1952.
113. Bessette, M.D. and Sundstrom, D.W., Polym. Proc. Eng., 3(1 and 2), 25–35, 1985.
114. (a) Jones, R.E. and Fesman, G., J. Cell. Plast., 1/1, 200–216, 1965. (b) Harding,
R.H., Resinography of Cellular Plastics, ASTM STP 414, 1967. (c) Blair, E.A.,
Resinography of Cellular Plastics, ASTM STP 414, 1967. (d) Buist, J.M., Ball, G.
W., and Woods, G., Advances in Polyurethane Technology, ed. J.M. Buist and H.
Gudgeon, Maclaren and Sons, London, 1968, 152–154. (e) Meinecke, E.A. and
Clark, R.C., Mechanical Properties of Polymeric Foams, Technomic, Lancaster, PA,
1972. (f ) Baumhakel, R., J. Cell. Plast., 8/6, 304–310, 1972. (g) van Leuwen, B.G.
et al., Advances in Urethane Science and Technology, vol. 2, Technomic, Lancaster,
PA, 1973, 173–202. (h) Berlin, A.A., Shutov, F.A., and Zhitinkina, A.K., Foam
Based on Reactive Oligomers, Technomic, Lancaster, PA, 1982, 85–87.
115. Clark, N.O. and Blackman, M., Trans. Faraday Soc., 44, 1–7, 1948.
116. Smith, C.S., Met. Rev., 9/33, 1–51, 1964.
117. Frensdorff, H.K., Rubber Age, 812–818, August, 1958.
118. Owen, M.J. and Kendrick, T.C., J. Colloid Interface Sci., 24, 141–150, 1967.
119. Matzke, E.B., Am. J. Bot., 33/1, 58–80, 1946.
120. Lord Kelvin, Phil. Mag., 5s24, 503–514, 1887.
121. Lord Kelvin, Proc. R. Soc. London, 55, 1–16, 1894.
122. Duchartre, P., Elements de botanique, Paris, 1867.
123. Kitchener, J.A. and Cooper, C.F., Q. Rev. (London), 13/1, 71–97, 1959.
124. Plateau, J., Mem. Acad. R. Sci. Belgique, 37, 49, 1869.
© 2004 by CRC Press LLC
www.ketabdownload.com
 125. Marangoni, C., Nuovo Cimento, (2), 5–6, 239, 1871.
126. Gibbs, J.W., Collected Works, vol. 1, Longmans Green, New York, 1928.
127. Ewers, W.E. and Sutherland, K.L., Aust. J. Sci. Res., A, 5, 697–710, 1952.
128. Blair, E.A., Resinography of Cellular Plastics, ASTM STP 414, 1967, 84–95.
129. Rossmy, G.R. et al., J. Cell. Plast., 13/1, 26–35, 1977.
130. Kollmeier, H.J. et al., Plast. Cmpd., July/August, 64–74, 1985.
131. Petrella, R.G. and Kushner, S.A., Proceedings of the SPI-33rd Annual Technical/
Marketing Conference, Technomic, Lancaster, PA, 1990, 186–190.
132. Patten, W., Seefried, C.G., Jr., and Whitman, R.D., J. Cell. Plast., 10/6, 276–282,
1974.
133. (a) Boudreau, R.J., Mod. Plast., 133–138, 143–147, 234–240, January 1967. (b)
Dahm, M., Cellular Plastics, National Academy of Sciences–National Research
Council Publication 1462, 52–63, 1967. (c) Schwarz, E.G., Applied Polymer Sym-
posia, No. 14: Silicone Technology, John Wiley & Sons, New York, 1970, 71–93. (d)
Battice, D.R. and Lopes, W.J., J. Cell. Plast., 23/2, 158–167, 1987.
134. Turner, R.B., Nichols, J.B., and Kuklies, R.A., Proceedings of the SPI-31 Annual
Technical/Marketing Conference, Technomic, Lancaster, PA, 1988, 18–21.
135. (a) Cavender, K.D. to Union Carbide Corporation, U.S. Patent 4,579,700, 1986.
(b) Cavender, K.D., J. Cell. Plast., 22/3, 222–234, 1986.
136. Saunders, J.H., and Frisch, K.C., Polyurethanes, Chemistry and Technology, Part I:
Chemistry, Robert Krieger, New York, 1978, 250–255.
137. (a) Ostrogorsky, A.G. and Glicksman, L.R., Trans. ASME J. Heat Trans., 110,
500–506, 1988. (b) Foreman, P.J., Conference Papers, UTECH 90, Crain Commu-
nications, London, 1990, 268–273.
138. Miller, T.E. and Schmidt, D.L., J. Cell. Plast., 19/5, 326–332, 1983.
139. Armistead, J.P., Wilkes, G.L., and Turner, R.B., J. Appl. Polym. Sci., 35, 601–629,
1988.
140. (a) Artavia, L.D. and Macosko, C.W., Low Density Cellular Plastics, ed. N.C.
Hilyard and A. Cunningham, Chapman & Hall, London, 1994, 47–55. (b) Ya-
sunaga, K., Neff, R., and Macosko, C.W., Proceedings of the 1995 SPI Polyurethanes
Conference, Technomic, Lancaster, PA, 1995, 362–370. (c) Zhang, X.D., Davis,
H.T., and Macosko, C.W., J. Cell. Plast., 35/5, 458–476, 1999.
141. (a) Seefried, C.G., Jr., Whitman, R.D., and Pollart, D.F., J. Cell. Plast., 10/4, 171–180,
1974. (b) Noshay, A. and McGrath, J.E., Block Copolymers, Overview and Critical
Survey, Academic Press, New York 1977, 365–389. (c) Gibson, P.E., Vallance, M. A.,
and Cooper, S.L., Developments in Block Copolymers-1, ed. I. Goodman, Applied
Science Publishers, London, 1982, 217–259. (d) Wilkes, G.L., Abouzahr, S., and
Radovich, D., J. Cell. Plast., 19/4, 248–254, 1983. (e) Turner, R.B., Spell, H.L., and
Wilkes, G.L., Proceedings of the SPI-28th Annual Technical/Marketing Conference,
Technomic, Lancaster, PA, 1984, 244–247. (f) Armistead, J.P., M.S. thesis, Virginia
Polytechnic Institute and State University, 1985. (g) Abouzahr, A. and Wilkes, G.L.,
Processing Structure and Properties of Block Copolymers, ed. M.J. Folkes, Elsevier
Applied Science Publishers, London, 1985, 165–207. (h) Turner, R.B. and Wilkes,
G.L., Proceedings of the SPI/FSK Polyurethanes World Congress, Technomic, Lancaster,
PA, 1987, 935–940. (i) Creswick, M.W. et al., Proceedings of the SPI-31st Annual
Technical/Marketing Conference, Technomic, Lancaster, PA, 1988, 11–17. (j) Artavia,
L.D. et al., Proceedings of the SPI/ISOPA Polyurethanes World Congress 1991,
Technomic, Lancaster, PA, 1991, 509–518. (k) Artavia, L.D. and Macosko, C.W.,
Low density cellular plastics, Hilyard, N.C., Cunningham, A., Eds., Chapman &
Hall, London, 1994, 22–55. (l) Priester, R.D., Turner, R.B., Low Density Cellular

© 2004 by CRC Press LLC

www.ketabdownload.com
 Plastics, ed. N.C. Hilyard and A. Cunningham, Chapman & Hall, London, 1994,
chap. 4. (m) Lidy, W.A., UTECH ‘96-Processing Workshop 3, Crain Communications,
London, 1996. (n) Lidy, W.A. et al., Proceedings of the 1996 SPI Polyurethanes
Conference, Technomic, Lancaster, PA, 1996, 119–135. (o) Lidy, W.A. et al.,
Proceedings of the Polyurethanes World Congress 1997, Technomic, Lancaster, PA,
1997, 95–117. (p) Mertes, J. et al., J. Cell. Plast., 34, 526–543, 1998.
142. Cooper, S.L. and Tobolsky, A.V., J. Appl. Polym. Sci., 10, 1837–1844, 1966.
143. (a) Hauptmann, G. et al., Proceedings of International Conference on Cellular and
Non-Cellular Polyurethanes, Strasbourg, France, June 1980, Carl Hanser Verlag,
Munich, Germany, 1980, 617–634. (b) Bailey, F.E. and Critchfield, F.E., J. Cell.
Plast., 17/6, 333–339, 1981.
144. Lidy, W.A., Phan Thanh, H., and Cadole, D., an unpublished presentation at
the 1984 Gordon Conference on foams.
145. (a) Merten, R., Lauerer, D., and Dahm, M., J. Cell. Plast., 4/7, 262–275, 1968. (b)
Rossmy, G.R. et al., J. Cell. Plast., 13/1, 26–35, 1977. (c) Menges, G., Schwesig,
H., and Hahn, H., Org. Coat. Plast. Chem., 44, 229–233, 1981. (d) Cole, K.C. and
Van Gheluwe, P., J. Appl. Polym. Sci., 34, 396–407, 1987. (e) Van Gheluwe, P. et
al., J. Cell. Plast., 23/1, 73–85, 1987. (f ) Artavia, L.D. and Macosko, C.W., Pro-
ceedings of the SPI-33rd Annual Technical/Marketing Conference, Technomic, Lan-
caster, PA, 1990, 554–561.
146. (a) Cole, K.C. et al., Proceedings of the SPI-31st Annual Technical/Marketing Con-
ference, Technomic, Lancaster, PA, 1988, 2–10. (b) Priester, R.D. et al., J. Cell.
Plast., 26/4, 346–367, 1990. (c) Elwell, M.J. et al., Macromolecules, 29/8, 2960–2968,
1996.
147. Fuest, R.W. and Palinkas, R.L., Conference Papers, UTECH 88, Crain Communi-
cations, London, 1988, 32–35.
148. Clift, S.M., Proceedings of the SPI-33rd Annual Technical/Marketing Conference,
Technomic, Lancaster, PA, 1990, 547–553.
149. (a) Neff, R. and Macosko, C.W., Proceedings of the 1994 SPI Polyurethanes Con-
ference, Technomic, Lancaster, PA, 1994, 386–392. (b) Neff, R. and Macosko,
C.W., Proceedings of the 1995 SPI Polyurethanes Conference, Technomic, Lancaster,
PA, 1995, 344–352.
150. (a) Paik Sung, C.S., Hu, C.B., and Wu, C.S., Macromolecules, 13/1, 111–116, 1980.
(b) Paik Sung, C.S., Smith, T.W., and Sung, N.H., Macromolecules, 13/1, 117–121,
1980. (c) Wang, C.B., Cooper, S.L., Macromolecules, 16/ 5, 775–786, 1983. (d)
Tyagi, D., Structure-property relationships in segmented copolymers, Ph.D.
thesis, Virginia Polytechnic Institute and State University, 1985.
151. (a) Whitman, R.D., Faucher, J.A., and Reding, F.P., Proceedings of the SPI-7th
Annual Technical Conference, The Society of the Plastics Industry, New York,
1963, Section 2-C, 1–8. (b) Darr, W.C., Gemeinhardt, P.G., and Saunders, J.H.,
Ind. Eng. Chem. Prod.Res. Dev., 2, 194–199, 1963. (c) Seefried, C.G., Koleske, J.
V., and Critchfield, F.E., J. Appl. Polym. Sci., 19, 2493–2502, 1975. (d) Seefried,
C.G., Koleske, J.V., and Critchfield, F.E., J. Appl. Polym. Sci., 19, 2503–2513, 1975.
(e) Redman, R.P., Developments in Polyurethane-1, ed. J.M. Buist, Applied Science
Publishers, London, 1978, 47. (f ) Tazewell, J.H. et al. to Firestone Tire and
Rubber et al; U.S. Patent 4,843,138, 1989. (g) Gilch, H. and Rath, W., Conference
Papers, UTECH 90, Crain Communications, London, 1990, 179–184. (h) Tobushi,
H. et al., J. Intell. Mater. Struct., 12, 283–287, 2001. (i) Kintrup, S., Treboux, J.P.,

© 2004 by CRC Press LLC

www.ketabdownload.com
 and Mispreuve, H., Proceedings of the 2000 API Polyurethanes Conference, Tech-
nomic, Lancaster, PA, 2000, 91–95.
152. (a) Illinger, J.L. and Schneider, N.S., Polym. Eng. Sci., 12/1, 25–29, 1972. (b)
Work, J.L., Macromolecules, 9/5, 759–763, 1976. (c) Christenson, C.P. et al., J.
Polym. Sci. Part B: Polym. Phys. Ed., 24, 1401–1439, 1986.
153. (a) Harrell, L.L., Macromolecules, 2/6, 607–612, 1969. (b) Lee, H.S. and Hsu, S.
L., Macromolecules, 22/3, 1100–1105, 1989.
154. Schneider, N.S. et al., Macromolecules, 8/1, 62–67, 1975.
155. (a) Rhodes, M.B., Cellular and Noncellular Polyurethanes, International Conference,
Strasbourg, France, June, 1980, Carl Hansen Verlag, Munich, Germany, 1980,
803–819. (b) Rhodes, M.B. and Khaykin, B., Langmuir, 2/5, 643–649, 1986. (c)
Rhodes, M.B., Ureth. Tech., 5/3, 29–32, V. (d) Rhodes, M.B., Rub. Plast. News,
18/3, 27–28, 1988. (e) Rhodes, M.B. and Khaykin, B., Proceedings of the SPI-32nd
Annual Polyurethane Technical/Marketing Conference, Technomic, Lancaster, PA,
1989, 178–182. (f ) Chaffanjon, P. and Verhelst, G., Proceedings of the SPI/ISOPA
Polyurethanes World Congress 1991, Technomic, Lancaster, PA, 1991, 545–552. (g)
Rhodes, M.B., Proceedings of the SPI-34th Annual Technical/Marketing Conference,
Technomic, Lancaster, PA, 1992, 548–551. (h) Reimann, K.A. et al., Proceedings
of the 1996 SPI Polyurethanes Conference, Technomic, Lancaster, PA, 1996, 112–118.
156. (a) Brumfield, H.L. and Estill, W.B., J. Cell. Plast., 5/4, 212–220, 1969. (b) Kanak-
kanatt, S.V., J. Cell. Plast., 9/1, 50–53, 1973. (c) Hilyard, N.C., Mechanics of Cellular
Plastics, Macmillan, New York, 1982, 107.
157. (a) Aneja, A. and Wilkes, G.L., Polyurethanes Expo 2001, Technomic, Lancaster,
PA, 2001, 195–199. (b) Rightor, E.G. et al., Macromolecules, 35, 5873–5882, 2002.
158. Garrett, J.T., Siedlecki, C.A., and Runt, J., Macromolecules, 34, 7066–7070, 2001.
159. (a) Chang, Y.P. and Wilkes, G.L., J. Polym. Sci. Phys. Ed., 13, 455–476, 1975. (b)
Bonart, R., Morbitzer, L., and Hentze, G., J. Macromol. Sci. Phys., B3(2), 337–356,
1969. (c) Bonart, R., Morbitzer, L., and Muller, E.H., J. Macromol. Sci. Phys., B9(3),
447–461, 1974. (d) Blackwell, J., Gardner, K.H., Polymer, 20, 13–17, 1979.
160. Clough, S.B., Schneider, N.S., and King, A.O., J. Macromol. Sci. Phys., B2(4),
641–648, 1968.
161. Koberstein, J.T., Morra, B., and Stein, R.S., J. Appl. Cryst., 13, 34–45, 1980.
162. McClusky, J.V. et al., Proceedings of the Polyurethanes World Congress 1993, Tech-
nomic, Lancaster, PA, 507–516.
163. (a) Elwell, M.J. et al., Plast. Rub. Comp. Proc. Appl., 23/4, 265–276, 1995. (b)
Stanford, J.L., Ryan, A.J., and Elwell, M.J., Processing of Polymers, ed. H.E.H.
Meijer, VCH Publishers, Weinheim, Germany, 1997, chap. 9. (c) Hamley, I.W.
et al., Polymer, 41, 2569–2576, 2000.
164. (a) Macosko, C.W., RIM: Fundamentals of Reaction Injection Moulding, Hanser,
Munich, 1989. (b) Ryan, A.J., Polymer, 31, 707, 1990. (c) Yang, W.L.P. and Ma-
cosko, C.W., Makromol. Chem., Makromol. Symp. 1989, 25, 23. (d) Girard-Reydet,
E. et al., Polymer, 39, 2269, 1998. (e) Yamanka, K., Takagi, Y., and Inoue, T.,
Polymer, 30, 1839, 1989. (f ) Kim, B.S., Chiba, T., and Inoue, T., Polymer, 36, 43,
1995. (g) Pearson, R.A., Advances in Chemistry Series, 233, ed. C.K. Riew and
A.J. Kinloch, American Chemical Society, Washington D.C., 1993, 405–425.
165. Ryan, A.J., Stanford, J.L., and Still, R.H., Plast. Rub. Proc. Appl., 13, 99–110, 1990.
166. (a) Artavia, L.D., Ph.D. thesis, University of Minnesota, 1991. (b) Mora, E. M.Sc.
thesis, University of Minnesota, 1991. (c) Elwell, M.J.A., Ph.D. thesis, Victoria
© 2004 by CRC Press LLC
www.ketabdownload.com
 University of Manchester, United Kingdom, 1993. (d) Neff, R., Ph.D. thesis, Uni-
versity of Minnesota, 1995. (e) Li, W., Ph.D. thesis, University of Sheffield, 2002.
167. Elwell, M.J., Mortimer, S., and Ryan, A.J., Macromolecules, 27, 5428, 1994.
168. (a) Artavia, L.D. and Macosko, C.W., J. Cell. Plast., 26, 490, 1990. (b) Mora, E.,
Artavia, L.D., and Macosko, C.W., J. Rheol., 35, 921, 1991. (c) Elwell, M.J.,
Thermochim. Acta, 269/270, 145, 1995. (d) McClusky, J.V. et al., J. Cell. Plast., 30,
338, 1994. (e) Bras, W. et al., Science, 267, 996, 1995. (f ) Elwell, M.J.A. et al.,
Polymer, 37, 1353, 1996. (g) Wilkinson, A.W. et al., Polym. Comm., 37, 2021, 1996.
(h) Neff, R. and Macosko, C.W., Rheologica Acta, 35, 656, 1996. (i) Li, W., Ryan,
A.J., and Meier, I.K., Macromolecules, 35, 5034, 2002. (j) Li, W., Ryan, A.J., and
Meier, I.K., Macromolecules, 35, 6306, 2002.
169. (a) Neff, R. et al., J. Polym. Sci., Part B: Polym. Phys., 36, 573, 1998. (b) Kaushiva,
B.D. and Wilkes, G.L., J. Appl. Polym. Sci., 77, 202, 2000. (c) Kaushiva, B.D. and
Wilkes, G.L., Polymer, 41, 6981, 2000. (d) Aneja, A. and Wilkes, G.L., J. Appl.
Polym. Sci., 85, 2956, 2002.
170. Cookson, P., Stanford, J.L., and Ryan, A.J., Real-time measurements of structure
development during the formation of flexible polyurethane foams, Macro-
Group Young Researchers’ Meeting, The Royal Society of Chemistry, Annual
Conference 2001, University of Strathclyde, Glasgow, April 18–20, 2001.
171. (a) Harding, R.H., Resinography of Cellular Plastics, ASTM STP 414, 1967, 3–18.
(b) Blair, E.A., Resinography of Cellular Plastics, ASTM STP 414, 1967, 84–95. (c)
Buist, J.M., Ball, G.W., and Woods, G., Advances in Polyurethane Technology, ed.
J.M. Buist and H. Gudgeon, Maclaren and Sons, London, 1968, 152–154. (d)
Meinecke, E.A. and Clark, R.C., Mechanical Properties of Polymeric Foams, Tech-
nomic, Lancaster, PA, 1972. (e) Baumhakel, R., J. Cell. Plast., 8/6, 304–310, 1972.
(f ) van Leuwen, B.G. et al., Advances in Urethane Science and Technology, vol. 2,
Technomic, Lancaster, PA, 1973, 153–182. (g) Berlin, A.A., Shutov, F.A., and
Zhitinkina, A.K., Foam Based on Reactive Oligomers, Technomic, Lancaster, PA,
1982, 85–87.
172. (a) Rusch, T.E. and Raden, D.S., Plast. Compd., July/August, 61–74, 1980. (b)
Wooler, A.M., Cell. Polym., 8, 296–320, 1989. (c) Burks, S., Autenrieth, R., and
Gutowsky, K., Proceedings of the SPI33rd Annual Polyurethane Technical/Marketing
Conference, Technomic, Lancaster, PA, 1990, 297–305.
173. (a) Rosemund, W.R. and Sanders, M.R., J. Cell. Plast., 13/3, 182–193, 1977. (b)
Motte, P. and Regnault, J., Proceedings of the SPI32nd Annual Polyurethane Tech-
nical/Marketing Conference, Technomic, Lancaster, PA, 1989, 691–692.
174. (a) Davis, B., Urethanes Tech., March, 28–38, 1987. (b) Neary, W.J., Plast. Tech.,
11/3, 50–53, 1965. (c) Lerner, A. and Coen, R., Plast. Tech., July, 97–99, 1966. (d)
Bowyer, R.L., International Progress in Urethanes, vol. 1, ed. K.C. Frisch and A.
Hernandez, Technomic, Lancaster, PA, 1977, 272–285. (e) Anon., How to Buy a
RIM Machine, Gusmer-Admiral, Inc., Lakewood, NJ, 1996. (f ) Jennings, R., Plast.
Tech., 15/3, 43–46, 1969. (g) Schlueter, K., J. Cell. Plast., 17/1, 51–58, 1981.
175. Dworkin, D., SPE J., 1269–1275, December 1961.
176. Anon., End use market survey of the polyurethane industry in the United
States, Allegheny Marketing, September 1995.
177. (a) Currier, V., Plast. Tech., 12/8, 35–37, 1996. (b) Doyle, E.N., The Development
and Use of Polyurethane Products, McGraw-Hill, New York, 1971, 238. (c) Ferri-
gno, T.H., SPE J., 638–640, June 1960.
178. (a) Ramlow, G.G., Hwyman, D.A., and Grace, O.M., J. Cell Plast., 19/4, 237–242,
1983. (b) Spitler, K.G. and Lindsey, J.J., J. Cell. Plast., 17/1, 43–50, 1981. (c)
© 2004 by CRC Press LLC
www.ketabdownload.com
 Kuryla, W.C. et al., J. Cell. Plast., 2/2, 84–96, 1966. (d) Heyman, D.A. and Grace,
O.M., Proceedings of the SPI-28th Annual Technical/Marketing Conference, Tech-
nomic, Lancaster, PA, 1984, 257–260.
179. Wolfe, H.W., Mechanics of Cellular Plastics, ed. N.C. Hilyard, Macmillan, New
York, 1979, 99–145.
180. (a) Stewart, D.A., Westfall, P.M., and Quarles, P.D., Proceedings of the 1995 SPI
Polyurethanes Conference, Technomic, Lancaster, PA, 1995, 90–93. (b) Chadwick,
F. and McCune, L., Conference Papers, UTECH 96, Crain Communications, Lon-
don, 1996, Paper 33.
181. (a) Potzsch, R., Proceedings of the SPI/FSK Polyurethanes World Congress, Aachen,
Technomic, Lancaster, PA, 1987, 887–890. (b) Ubben, R., Conference Papers,
UTECH 90, Crain Communications, London, 1990, 145–148. (c) Schiffler, C.,
Conference Papers, UTECH 96, Crain Communications, London, 1996, Paper 20.
182. Brown, M.J., Proceedings of the SPI-34th Annual Technical/Marketing Conference,
Technomic, Lancaster, PA, 1992, 463–467.
183. (a) Knibbe, D.E., J. Cell. Plast., 21/4, 264–268, 1985. (b) Lambach, J.L. and Gill,
W.A., Proceedings of the SPI33rd Annual Polyurethane Technical/Marketing Confer-
ence, Technomic, Lancaster, PA, 1990, 165–171.
184. (a) Green, H.L., J. Cell. Plast., 5/3, 173–175, 1969. (b) Axel, F., International
Progress in Urethanes, ed. K.C. Frisch and A. Hernandez, Technomic, Lancaster,
PA, 1975, 1, 286–291. (c) Kirkland, C., Plast. Tech., 65–70, Sept., 1980. (d) Collins,
S.H., Plast. Cmpd., 38–44, March/April, 1983. (e) Cox, D.B., Urethane Chemistry
and Applications, ed. K.M. Edwards, ACS Symposium Series 172, 1981, 565–573.
(f ) Lammerting, H., Uret. Tech., 3/1, 38–42, 1986. (g) Sherman, L.M., Plast. Tech.,
42–45, August 1995.
185. (a) Andrews, G.D. and Wasilczyk, G.J., Proceedings of the SPI 31st Annual Tech-
nical/Marketing Conference, Technomic, Lancaster, PA, 1988, 447–454. (b) Schrod-
er, W., J. Elast. Plast., 20, 200–207, 1988. (c) Wochnowski, H., Kunststoffe, 78/8,
680–683, 1989. (d) Block, H.H., Kunststoffe, 79/3, 240–242, 1989. (e) Fountas, G.,
Elastomerics, 122/12, 30–31, 1990. (f ) Moskowitz, M., Plast. World, 49/3, 103,
1991. (g) Andrew, G.D. et al., Proceedings of the Polyurethanes World Congress 1991,
Technomic, Lancaster, PA, 1991, 877–884. (h) Editorial, Elastomerics, 124/2, 29,
1992. (i) Thies, W., Conference Papers, UTECH 94, Crain Communications, Lon-
don, 1994, Paper 15. (j) Michaeli, W. et al., Kunststoffe, 86/6, 838–842, 1996. (k)
Andrew, G.D., Boyer, T.C., and Olari, J.R., Proceedings of the 1996 SPI Polyure-
thanes Conference, Technomic, Lancaster, PA, 1996, 480–486.
186. (a) Abernathy, H.H., Rub. World, 131/6, 765–769, 1955. (b) Wall, J.R., a paper
given before the Cellular Plastics Division, Society of the Plastics Industry, Inc.,
Detroit, June 1960. (c) Healy, T.T., Polyurethane Foams, ed. T.T. Healy, lliffee
Books, London, 1963, 75–92. (d) Bishop, A.C., Plast. Rub., 1/9, 381, 1970.
187. (a) Zick, J., Proceedings of the SPI/FSK Polyurethanes World Congress, Technomic,
Lancaster, PA, 1987, 570–573. (b) Bayer, H.G., Proceedings of the SPI/FSK Poly-
urethanes World Congress, Technomic, Lancaster, PA, 1987, 574–577. (c) Ostertag,
B., Kunststoffe, 79/11, 1144–1145, 1987. (d) Zick, J., Proceedings of the SPI-32nd
Annual Technical/Marketing Conference, Technomic, Lancaster, PA, 1989, 211–215.
(e) Ostertag, B., Proceedings of the SPI-32nd Annual Technical/Marketing Confer-
ence, Technomic, Lancaster, PA, 1989, 209–210. (f ) Heiberger, R., Dipietrantonio,
B.F., and Katz, J.J., Proceedings of the SPI-32nd Annual Technical/Marketing Con-
ference, Technomic, Lancaster, PA, 1989, 44–49. (g) Pierkes, L., Conference Papers,

© 2004 by CRC Press LLC

www.ketabdownload.com
 UTECH 90, Crain Communications, London, 1990, 126–130. (h) Deno, L., Con-
ference Papers, UTECH 90, Crain Communications, London, 1990, 142–144. (i)
Schneider, F.W. and Grasse, H., Conference Papers, UTECH 90, Crain Commu-
nications, London, 1990, 149–154. (j) Schneider, F.W., Grasse, H., and Gillham,
M., Proceedings of the SPI-33rd Annual Technical/Marketing Conference, Technomic,
Lancaster, PA, 1990, 32–39. (k) Gill, W.A., Good, R.M., and Hawker, L.E., Pro-
ceedings of the SPI-33rd Annual Technical/Marketing Conference, Technomic, Lan-
caster, PA, 1990, 40–45. (l) Murray, D., Deno, L., and Katz, J.J., Proceedings of the
SPI-33rd Annual Technical/Marketing Conference, Technomic, Lancaster, PA, 1990,
172–176.
188. (a) Kozlowski, E.F. and Kruger, E.R. to Lear Siegler, Inc., U.S. Patent 4,692,199,
1987. (b) Anon., Seating Technology, Lear Seating Corporation, 1994. (c) Gabriele,
M. and Monks, R., Plast. Tech., 38/4, 15–23, 1992. (d) Anon., Trim Retention
Bonding, The Woodbridge Group, 1993. (e) Andrew, G.D., Boyer, T.C., and Olari,
J.R., Proceedings of the 1996 SPI Polyurethanes Conference, Technomic, Lancaster,
PA, 1996, 480–486.
189. Buist, J.M., Plast. Inst. Trans., 27, 13–27, 1959.
190. Editorial, Mod. Plast. Int., June 14–17, 1973.
191. Editorial, Mod. Plast. Int., 12–13, August 1978.
192. Cravens, T.E., Proceedings of the Carpet Rug Institute Annual Meeting, Hilton
Head, SC, 1975.
193. (a) Editorial, Plast. Rub. Wkly., March 7, 25–26, 1975. (b) Marlin, L., Durante, A.J.,
and Schwarz, E.G., J. Cell. Plast., 11/6, 317–322, 1975. (c) Kallert, W., J. Coated Fab.,
4, 272–278, 1975. (d) Bobe, J., Hurd, R., and Woods, G., Proceedings of the SPI-4th
International Cellular Plastics Conference, Society of the Plastics Industry, New York,
1976, 290–295. (e) Hurd, R., Developments in Polyurethane-1, ed. J.M. Buist, Applied
Science, London, 1978, 211–222. (f ) Bez, W. and Quack, G., Cell. Polym., 2, 31–53,
1983. (g) Berthevas, P., Cell. Polym., 4, 179–193, 1985.
194. (a) Anon., Rub. Age, 804–810, August 1956. (b) Editorial, Br. Plast., 34/2, 73–74,
1961. (c) Jones, R.E., Patten, G, and Steingisen, S., Handbook of Foamed Plastics,
ed. R. Bender, Lake Publishing Corporation, Libertyville, IL, 1965, 52–68. (d)
Rogers, T.H., J. Cell. Plast., 4/10, 392–394, 1968. (e) Touhey, W.J., J. Cell. Plast.,
4/10, 395–397, 1968.
195. Annual Book of American Society for Testing and Materials Standards, Volume 09.02,
D 3574–86, ASTM, Philadelphia, 1990, 320–335.
196. Graf, W.C. and Schondlowski, H., Proceedings of the SPI-3rd International Cellular
Plastics Conference, The Society of the Plastics Industry, New York, 1972,
225–236.
197. (a) Lambert, A., Eur. J. Cell. Plast., 25–35, January 1980. (b) Wang, S.L., Frisch,
K.C., and Malwitz, N., Proceedings of the SPI-30th Annual Technical/Marketing
Conference, Technomic, Lancaster, PA, 1986, 363–368. (c) Lawler, L.F. and Schif-
fauer, R., Proceedings of the Polyurethanes World Congress 1993, Technomic, Lan-
caster, PA, 1993, 288–296. (d) Ramesh, N.S. and Malwitz, N., Proceedings Annual
Technical Conference—Soc. Plast. Eng., 52/2, 1941–1945, 1994. (e) Fabishak, D.M.,
J. Cell. Plast., 8/2, 82–84, 1972. (f ) Schiffauer, R., Proceedings of the 1994 SPI
Polyurethanes Conference, Technomic, Lancaster, PA, 1994, 225–238. (g) Cavender,
K.D. and Kinkelaar, M.R., Proceedings of the 1994 SPI Polyurethanes Conference,
Technomic, Lancaster, PA, 1994, 248–251. (h) Lederman, J.M., J. Appl. Polym.
Sci., 15, 693–703, 1971. (i) Hilyard, N.C. and Collier, P., Cell. Polym., 6/6, 9–26,

© 2004 by CRC Press LLC

www.ketabdownload.com
 1987. (j) Gent, A.N. and Rusch, K.C., J. Cell. Plast., 2/1, 46–51, 1966. (k) Hilyard,
N.C., Low Density Cellular Plastics, Physical Basis of Behaviour, ed. N.C. Hilyard
and A. Cunningham, Chapman & Hall, London, 1994, 235–269.
198. Kreter, P.E., J. Cell. Plast., 21/5, 306–310, 1985.
199. (a) Hogan, J.M. et al., J. Cell. Plast., 9/5, 219–225, 1973. (b) Saotome, K., Mat-
surbara, K., and Yatomi, T., J. Cell. Plast., 13/3, 203–209, 1977. (c) Woods, G.,
Flexible Polyurethane Foams, Chemistry and Technology, Applied Science Pub-
lishers, London, 1982, 119–120. (d) Moreland, J.C., Wilkes, G.L., and Turner,
R.B., Proceedings of the SPI/ISOPA Polyurethanes World Congress 1991, Technomic,
Lancaster, PA, 1991, 500–508. (e) Dounis, D., Wilkes, G.L., and Turner, R.B.,
Polymer Preprints, vol. 35, no. 2, 1994, American Chemical Society, Washington,
D.C., 1994, 781–782. (f ) Brasington, R., Kinkelaar, M.R., and Cavender, K.D.,
Conference Papers, UTECH 96, Crain Communications, London, 1996, Paper 9.
200. (a) Anon., Plast. Tech., 8/4, 26–32, 1962. (b) Lambert, A., J. Cell. Plast., 10/1,
35–42, 1974. (c) Ashe, W.A. and Grace, O.M., J. Cell. Plast., 25/4, 371–390, 1989.
201. (a) Hartings, J.W. and Hagan, J.H., J. Cell. Plast., 14/2, 81–86, 1978. (b) Rose-
mund, W.R. and Sander, M.R., J. Cell. Plast., 13/3, 182–193, 1977. (c) Blair, G.R.
and Wilson, A.R., Proceedings of the 1995 SPI Polyurethanes Conference, Technom-
ic, Lancaster, PA, 1995, 413–417.
202. (a) Jones, R.E. et al., Plast. Tech., 5/9, 55–59, 1959. (b) Polis, S., Proceedings of the
SPI-24th Annual Urethane Division Technical Conference, Technomic, Lancaster,
PA, 1978, 59–69. (c) Time Tech TT502 Ball Rebound Foam Resilience Tester,
Time Tech, Inc. Wilmington, DE, 1996.
203. Wolfe, H.W., Mechanics of Cellular Plastics, ed. N.C. Hilyard, Macmillan, New
York, 1979, 105.
204. (a) McCallum, J.B., Text. Chem. Color., 21/12, 13–15, 1989. (b) American Society
for Testing and Materials, D 5393-93, ASTM, Philadelphia, 1993. (c) General
Motors Engineering Standard GM9305P, January, 1992. (d) Ford Laboratory
Test Method BO-116-03, 1990. (e) Deutsches Institut für Normung (DIN) Test
Method DIN 75201R and DIN 75201G. (f ) SAE Test Method J1756, 1994.
205. (a) Coupland, F.E., Proceedings of the SAE International Congress and Exposition,
Detroit, MI, 1988, SAE, Warrendale, PA, Paper 880504. (b) Baatz, G. and Franyutti,
S., Conference Papers, UTECH 94, Crain Communications, London, 1994, Paper 9.
(c) Bradford, L., Pinzoni, E., and Wuestenenk, J., Proceedings of the 1996 SPI
Polyurethanes Conference, Technomic, Lancaster, PA, 1996, 358–361. (d) Loock, F.,
Lampe, Th., and Bahadir, A.M., Kunstoffe, 83/3, 201–205, 1993.
206. (a) Rothermel, H.M., Polyurethane Handbook, ed. G. Oertel, Hanser, New York,
1985, 455–507. (b) Wolfe, H.W., Mechanics of Cellular Plastics, ed. N.C. Hilyard,
Macmillan, New York, 1979, 99–145. (c) Woods, G., Flexible Polyurethane Foams,
Chemistry and Technology, Applied Science Publishers, London, 1982, 126–128, 296.
207. Annual Book of American Society for Testing and Materials Standards, Volume 09.02,
D 3574-95, ASTM, Philadelphia, 1990, 320–335.
208. Campbell, G.A., Report GMR-2662, General Motors Research Laboratories,
April 1978.
209. (a) Ashe, W.A., Grace, O.M., and Otten, J.G., Proceedings of the SPI-25th Annual
Urethane Division Technical Conference, Technomic, Lancaster, PA, 1979, 107–112.
(b) Flexible Cellular Materials—Test for Dynamic Fatigue by Constant Load
Pounding, International Standard 3385.
210. Hartings, J.W. and Hagan, J.H., J. Cell. Plast., 14/2, 81–86, 1978.

© 2004 by CRC Press LLC

www.ketabdownload.com
211. American Society for Testing and Materials, D3574-86, ASTM, Philadelphia,
1990.
212. Ashe, W.A., Proceedings of the SPI-30th Annual Technical/Marketing Conference,
Technomic, Lancaster, PA, 1986, 320–325.
213. Knight, J.E., Proceedings of the SPI-30th Annual Technology/Marketing Conference,
Technomic, Lancaster, PA, 1986, 48–57.
214. (a) Dwyer, F.J., J. Cell Plast., 12/2, 104–113, 1976. (b) Wolfe, H.W., Brizzolara,
D.F., and Bryam, J.D., J. Cell. Plast., 13/1, 48–56, 1977. (c) Beals, B., Dwyer, F.J.,
and Kaplan, M., J. Cell. Plast., 1/1, 32–41, 1965. (d) Ball, G.W. and Doherty, D.J.,
J. Cell. Plast., 3/5, 223–232, 1967. (e) Stone, H., Proceedings of the SPI-27th Annual
Technical/Marketing Conference, Technomic, Lancaster, PA, 1982, 124–137. (f )
Saunders, J.H. and Frisch, K.C. Polyurethanes, Chemistry and Technology, Part 1:
Chemistry, Robert Krieger, New York, 1962, 327–345. (g) Casati, F.M. et al.,
Polyurethanes World Congress 1997, Technomic, Lancaster, PA, 1997, 402–420. (h)
Casati, F.M. et al., Polyurethanes Expo ’98, Technomic, Lancaster, PA, 1998,
417–431. (i) Broos, R., Casati, F.M., and Herrington, R.M., Polyurethanes Expo
’99, Technomic, Lancaster, PA, 1999, 347–364. (j) Moore, D.R. et al., Proceedings
of the SPI-34th Annual Polyurethanes Technical/Marketing Conference, Technomic,
Lancaster, PA, 1992, 502–511.
215. (a) Bryan, M.E., Tempest, W., and Williams, D., Appl. Ergonom., 9/3, 151–154,
1978. (b) Winklhofer, E. and Thien, G.E., Proceedings of the SAE International
Congress and Exposition, Detroit, MI, 1985, SAE, Warrendale, PA, Paper 850966.
(c) Kryter, K.D., The Effects of Noise on Man, Academic Press, New York, 1970.
(d) Miller, J.D., J. Acou. Soc. Am., 56/3, 729–763, 1974. (e) Howell, T.M. and
Schumacher, R.F., Proceedings of the SAE International Congress and Exposition,
Detroit, MI, 1985, SAE, Warrendale, PA, Paper 850969. (f ) Bagga, K.S. and
Repick, E.P., Proceedings of the SAE International Congress and Exposition, Detroit,
MI, 1985, SAE, Warrendale, PA, Paper 850994. (g) Measurement of Interior
Sound Levels of Light Vehicles, SAE Recommended Practice SAE J1477, SAE,
Warrendale, PA, Jan., 1986. (h) Stacy, J.M., Proceedings of the SAE International
Congress and Exposition, Detroit, MI, 1987, SAE, Warrendale, PA, Paper 871260.
216. (a) Costa, C.E.T. and Pauperio, A., Proceedings of the SPI-32nd Annual Polyure-
thane Technical/Marketing Conference, Technomic, Lancaster, PA, 1989, 579–586.
(b) MacFarland, D.R., Proceedings of the SAE International Congress and Exposition,
Detroit, MI, 1994, SAE, Warrendale, PA, Paper 940700.
217. Grande, J.A., Mod. Plast., 30–31, October 1996.
218. (a) Ball, G.L., Schwartz, M., and Long, J.S., Off. Dig. Fed. Soc. Paint Technol., 32/
425, 817–831, 1960. (b) Buist, J.M. and Gudgeon, H., Advances in Polyurethane
Technology, Maclaren and Sons, London, 1968, 159–162.
219. Imai, Y., and Asano, T., J. Appl. Polym. Sci., 27, 183–195, 1982.
220. (a) Moore, R.S., Encyclopedia of Polymer Science and Technology, 12, 700–724, 1970.
(b) Ludman, W.R., Proceedings of the Plastics and Rubber Institute Urethanes Group
Conference, London, 1976, Gl–G4. (c) Kingsbury, H.B., Cho, K., and Powers, W.R.,
J. Cell. Plast., 14/2, 113–117, 1978. (d) Eeckhaut, G., Pham, T., and Lockwood,
R.J., Proceedings of the SPI/FSK Polyurethanes World Congress, Technomic, Lancast-
er, PA, 1987, 708–711. (e) Gahlau, H.K., Proceedings of the SPI/FSK Polyurethanes
World Congress, Technomic, Lancaster, PA, 1987, 712–726. (f ) Lee, J.R., Conference
Papers, UTECH 88, Crain Communications, London, 1988, 81–82. (g) Corsaro,
R.D. and Sperling, L.H., eds., Sound and Vibration Damping with Polymers, ACS

© 2004 by CRC Press LLC

www.ketabdownload.com
 Symposium Series 424, American Chemical Society, Washington, D.C., 1990. (h)
Lutter, H.D., Mertes, J., and Zschiesche, R., Proceedings of the Polyurethanes World
Congress 1991, Technomic, Lancaster, PA, 1991, 252–256. (i) Lauriks, W., Low
Density Cellular Plastics, Physical Basis of Behaviour, ed. N.C. Hilyard and A.
Cunningham, Chapman and Hall, London, 1994, 319–361.
221. Polymers in Vibration Damping and Soundproofing, PB86-858966, National
Technical Information Service, Springfield, VA, 1986.
222. (a) Goldman, D.E. and von Gierke, H.E., Shock and Vibration Handbook, ed. C.M.
Harris, and C.E. Crede, McGraw-Hill, New York, 1961, 44.1–44.51. (b) McCor-
mick, E.J., Human Factors Engineering, McGraw-Hill, New York, 1964, 474–483.
(c) Kryter, K.D., The Effects of Noise on Man, Academic Press, New York, 1970.
(d) Miller, J.D., J. Acou. Soc. Am., 56/3, 729–764, 1974. (e) Griffin, M.J., Handbook
of Human Vibration, Academic Press, London, 1990.
223. (a) Wooley, W.D., Proceedings of Fire and Cellular Polymers, Elsevier, London,
1984, 37–39. (b) Drysdale, D.D. and Thomson, H.E., Proceedings of Flame Retar-
dants 90, Elsevier, London, 1990, 191–205.
224. (a) Rua, L. et al., Evaluations of the fire performance of flame retarded flexible
polyurethane foam in an upholstered furniture application, J. Consum. Prod.
Flamm., 7, 99–126, 1980. (b) Szabat, J., J. Consum. Prod. Flamm., 8, 189, 1981. (c)
Schuhmann, J. and Hartzell, G., J. Fire Sci., 7, 386–402, 1989. (d) Powell, D., Fire
Prev., 27–30, 1989.
225. Van der Plaats, G., Soons, H., and Snellings, R., The thermal behavior of
melamine, DSM Corporate Research and Patents Department, Letter report
WdB/GBL/860, February 20, 1990.
226. Balkon, J., Proceedings of the SPI-31st Annual Technical/Marketing Conference,
Technomic, Lancaster, PA, 1988, 218–239.
227. (a) Babrauska, V. and Krasny, J.F., NBS Monograph 173, National Bureau of
Standards, Washington, D.C., 1985. (b) Strength, R.S. and Sherman, M., Pro-
ceedings of the SPI-31st Annual Technical/Marketing Conference, Technomic, Lan-
caster, PA, 1988, 268–272.
228. Gum, W.F., Proceedings of the SPI-33rd Annual Polyurethane Technical/Marketing
Conference, Technomic, Lancaster, PA, 1990.
229. (a) Hull & Co., End Use Market Survey of the Polyurethane Industry in the United
States and Canada, prepared for the Society of the Plastics Industry, Inc., Poly-
urethane Division, May 29, 1990. (b) Saunders, J.H. and Frisch, K.C., High
Polymers, Vol. XVI, Part II, Interscience, New York, 1964, 849. (c) Roegler, M.,
Polyurethane Handbook, ed. G. Oertel, Hanser, Munich, 1985, 176–177. (d) Woods,
G. The ICI Polyurethanes Book, John Wiley & Sons, New York, 1987, 203–204. (e)
Miyama, S., Conserv. Recyc., 10/4, 265–272, 1987. (f ) Kettemann, B.U. et al.,
Kunstoffe, 85/11, 33–35, 1995. (g) Editorial, Charlotte Observer, March 26, 1990.
230. Baumann, B.D. et al., Proceedings of the SPI-6th International Technical/Marketing
Conference, Technomic, Lancaster, PA, 1983, 139–141.
231. (a) Gerlock, J., Braslaw, J., and Zimbo, M., Ind. Eng. Chem. Proc. Des. Dev., 23,
545–552, 1984. (b) Campbell, G.A. and Meluch, W.C., Environ. Sci. Tech., 10/2,
182–185, 1976. (c) Meluch, W.C. and Campbell, G.A., U.S. Patent 3,978,128, 1976.
(d) Campbell, G.A. and Meluch, W.C., J. Appl. Polym. Sci., 21, 581–584, 1977.
(e) Salloum, R.J., Proceedings of the SPE ANTEC, Society Of Plastics Engineers,
Brookfield Center, Ct., 1981, 491–494. (f ) Grigat, E. and Hetzel, H., U.S. Patent
4,051,212, 1977. (g) Niederdellmann, G. et al., U.S. Patent 4,399,236, 1983. (h)

© 2004 by CRC Press LLC

www.ketabdownload.com
 Mahoney, L.R., Weiner, S.A., and Ferris, F.C., Environ. Sci. Tech., 8/2, 135–139,
1974. (i) Braslaw, J. and Gerlock, J.L., Ind. Eng. Chem. Pro. Des. Dev., 23, 552–557,
1984.
232. (a) Ulrich, H. et al., Polym. Eng. Sci., 18/11, 844–848, 1978. (b) Frulla, F.F., Odinak,
A., and Sayigh, A.A.R., U.S. Patent 3,709,440, 1973, and U.S. Patent 3,738,946,
1973. (c) Tucker, B. and Ulrich, H., U.S. Patent 3,983,087, 1976. (d) Ulrich, H. et
al., Elast. Plast., 11, 208, 1979. (e) Simioni, F., Bisello, S., and Cambini, M., Macplas
8, (47), 52, 1983. (f ) Simioni, F., Bisello, S., and Tavan, M., Cell. Polym., 2, 281–283,
1983. (g) Simioni, F., Modesti, M., and Navazzio, G., Macplas 12, (88),127–129,
1987. (h) Simioni, F., Modesti, M., Brambilla, C.A., Cell. Polym., 8, 387–400, 1989.
(i) Sheratte, M.B., Proceedings of the SPI-20th Annual Technical Conference, Tech-
nomic, Lancaster, PA, 1977, 59–60. (j) Kinstle, J.F. et al., Polym. Preprints, 24/2,
446–447, 1983. (k) Ulrich, H., Advances in Urethane Science and Technology, vol.
5, ed. K.C. Frisch and S.L. Reegen, Technomic, Lancaster, PA, 1978, 49–57. (l)
Chapman, T.M., J. Polym. Sci. A. Polym. Chem., 27, 1993–2005, 1989. (m) Grigat,
E., Kunststoffe, 68/5, 281–284, 1978. (n) Bauer, G., RECYCLE 1990, Davos, Swit-
zerland, May 29–31, 1990. (o) Bauer, G., German Patent 2,738,572. (p) Bauer,
G., German Patent 19,512,778.
233. (a) Madorsky, S.L. and Straus, S., J. Poly. Sci., 36, 183–194, 1959. (b) Rapp, N.S.
and Ingham, J.D., J. Polym. Sci., A/2, 689–704, 1964. (c) Ingham, J.D. and Rapp,
N.S., J. Polym. Sci., A/2, 4941–4964, 1964. (d) Tilley, J.N. et al., J. Cell. Plast., 4/
2, 56–66, 1968. (e) Levin, B.C., National Bureau of Standards Report, NBSIR
853267, 1986. (f ) Maya, P. and Levin, B.C., Fire Mater., 11, 129, 1987. (g) Wooley,
W.D. and Fardell, P.J., Fire Res., 1, 11–21, 1977. (h) Wooley, W.D., Br. Polym. J.,
4, 27–43, 1972. (i) Wooley, W.D., Fardell, P.J., and Buckland, I.G., Fire Research
Note No. 880, July, 1971. (j) Chambers, J., Jiricny, J., and Reese, C.B., Fire Mater.,
5, 133–141, 1981. (k) Braslaw, J., Gealer, R.L., and Wingfield, R.C., Polym. Pre-
prints, 24, 434, 1983.
234. Myers, J.I. and Farrissey, W.J., Proceedings of the SAE International Congress and
Exposition, Detroit, MI, 1991, SAE, Warrendale, PA, Paper 910583.
235. Fosnaugh, J., Automot. Transp. Int., 76, May 1994.
236. John, J., Bhattacharya, M., and Turner, R.B., J. Appl. Polym. Sci., 86, 3097–3107,
2002.

© 2004 by CRC Press LLC

www.ketabdownload.com
7
Rigid Polyurethane Foams

H.J.M. Grünbauer, J. Bicerano, P. Clavel, R.D. Daussin, H.A. de Vos,

M.J. Elwell, H. Kawabata, H. Kramer, D.D. Latham, C.A. Martin, S.E.
Moore, B.C. Obi, V. Parenti, A.K. Schrock, and R. van den Bosch

CONTENTS
7.1 Introduction
7.2 Raw Materials
7.2.1 Isocyanates
7.2.1.1 Introduction
7.2.1.2 Production
7.2.1.3 Properties
7.2.1.4 Products
7.2.2 Polyols
7.2.2.1 Introduction
7.2.2.2 Production
7.2.2.3 Properties
7.2.2.4 Products
7.2.3 Blowing Agents
7.2.4 Additives
7.2.4.1 Catalysts
7.2.4.2 Surfactants
7.3 Applications and Markets
7.3.1 Appliance
7.3.1.1 Introduction
7.3.1.2 Processing
7.3.1.3 New Blowing Agents
7.3.1.4 Hydrochlorofluorocarbons
7.3.1.5 Hydrocarbons
7.3.1.6 Vacuum Insulation Panels (VIPs)
7.3.1.7 Markets
7.3.2 Construction
7.3.2.1 Introduction
7.3.2.2 Continuous Production

0-8493-1728-2/04/$0.00+$1.50
© 2004 by CRC Press LLC

© 2004 by CRC Press LLC

www.ketabdownload.com
 7.3.2.3 Discontinuous Production
.4 Products and Applications
7.3.2.5 Markets
7.4 Foaming Process
7.4.1 Mixing
7.4.2 Nucleation
7.4.3 Expansion
7.4.4 Rheology
7.4.5 Polymer Morphology
7.5 Physical Properties
7.5.1 Thermal Conductivity
7.5.1.1 K-factor Contributions
7.5.1.2 Condensation
7.5.2 Aging
7.5.2.1 Gas Diffusion
7.5.2.2 Dimensional Stability
7.6 Conclusions
Acknowledgments
References

7.1 Introduction
Urethane foams represent a highly versatile class of expanded thermoset
materials that find application in a wide range of industrial products such
as carpet backings, matresses, insulation panels, and refrigerators. Within
this spectrum, closed-cell rigid foams are an important sub-class owing to
their outstanding thermal insulation properties when expanded with suit-
able blowing agents of low thermal conductivity.1–4 To illustrate, Figure 7.1
compares the thermal insulation performance of rigid polyurethane foam
with that of other materials, such as expanded polystyrene, mineral wool,
cork, softwood, fireboard, concrete blocks, and bricks. Figure 7.1 clearly
shows that to match the insulation performance of 50-mm rigid polyurethane
foam other materials need to be applied at much higher thicknesses, ranging
from 80 mm (~3 in.) for expanded polystyrene to 90 mm for mineral wool
or 1720 mm for bricks. As will be explained in Section 7.5.1, this performance
is primarily due to a combination of blowing agent properties, cell size, and
cell-morphology. The latter is characterized by fully drained, dodecahedron-
shaped cells5,6 with void volumes of 95–97% and 100–500 mm cell size. An
example is given in Figure 7.2.
Notwithstanding its prime importance, excellent thermal insulation per-
formance is not the only property that makes rigid polyurethane foams an
attractive choice in various industrial applications. There are several added
advantages which, alone or in combination, are often decisive to the selection

© 2004 by CRC Press LLC

www.ketabdownload.com
 50 mm Polyurethan-Hartschaum
80 mm Polystyrol
90 mm Mineralwolle
100 mm Kork
130 mm Holzfaserplatte
200 mm Weichholz

760 mm Leichtbetonstein

1720 mm Mauerziegesteine

FIGURE 7.1
Example of the relative thicknesses of alternative building materials needed to match the
thermal conductivity properties of rigid polyurethane foam at 50 mm. Note that the other
materials need to be applied at much higher thicknesses (ranging from 80 mm for expanded
polystyrene to 90 mm for mineral wool and 1720 mm for bricks) to match the insulation per-
formance of 50-mm rigid polyurethane foam. (Reprinted with permission from Koschade, R.,
Die Sandwichbauweise (How to Build a Steel Sandwich), Ernst & Sohn, Berlin, 2000.)

of polyurethane foams in a particular application or process environment.7

The most important advantages are their high mechanical strength, their
strong adhesive propensity, and, above all, their easy processing. The com-
bination of all three is particularly apparent in, e.g., refrigerators and sand-
wich panels where rigid polyurethane foam is virtually unmatched.1–3,8–10
Easy processability allows “pour in place” manufacturing of foams, even in
molds of highly complex shapes and forms. It is most convincingly demon-
strated by a lab scale experiment, as visualized in Figure 7.3. Starting from
two low viscosity liquids, polyol and isocyanate, the first step of foam prep-
aration consists of mixing and pouring the mixture. After 5–20 sec, the react-
ing mass takes a creamy appearance which reflects the fact that nucleation
of gas bubbles has occurred. Subsequently, the foam rises to achieve its final
shape after 30–120 sec. Finally, a curing step (not shown) of several minutes
to hours should be taken into account. The appealing simplicity of this
pictorial should not conceal that the design of such a system that foams “by
itself” often takes years of development and significant investments before
all ingredients and sub-processes are fine tuned.

© 2004 by CRC Press LLC

www.ketabdownload.com
FIGURE 7.2
Cellular structure of rigid polyurethane foam. (Reprinted with permission from Koschade, R.,
Die Sandwichbauweise (How to Build a Steel Sandwich), Ernst & Sohn, Berlin, 2000.)

From a physical point of view, rigid polyurethane foaming represents an

intricate combination of simultaneous processes that starts by mixing two
partially miscible liquids, followed by a multiple co-polymerization process
involving at least the steps of solidification, vitrification, and chemical net-
work formation. Meanwhile, sufficient gaseous compounds are evolving to
expand the reacting mass 30- to 50-fold to achieve a final low density of
20–50 kg/m3 without deteriorating its fine-cell structure. In view of this
complexity, it is not surprising that the poyurethane industry has been
unable to respond with “drop-in” technology to the challenges imposed by
global ozone depletion and global warming issues.11–13 On the contrary, these
issues have raised a wave of research and development throughout academia
and industry which is still going on today. It is the aim of this chapter to
provide a concise account of this work by discussing both raw materials,
major applications, and new fundamental insights.

7.2 Raw Materials

7.2.1 Isocyanates
7.2.1.1 Introduction
Isocyanate types can be subdivided into aromatic, aliphatic, and cyclo-ali-
phatic isocyanates. Due to their higher reactivity and favorable economics,

© 2004 by CRC Press LLC

www.ketabdownload.com
FIGURE 7.3
Lab scale preparation of rigid polyurethane foam. (Reprinted with permission from Koschade,
R., Die Sandwichbauweise (How to Build a Steel Sandwich), Ernst & Sohn, Berlin, 2000.)

aromatic isocyanates account for more than 95% of isocyanate consuption for
rigid polyurethane foams. Early developments were primarily based on tol-
uene diisocyanate (TDI), but this changed during the 1960s when diphenyl-
methane diisocyanates (MDI) and their polymeric byproducts (pMDI) became
available in commercial quantities. Since then, MDI and pMDI production
has seen a very significant capacity growth. In 1984, the global use of MDI/
pMDI overtook that of TDI and has outperformed the latter ever since.

7.2.1.2 Production

The production of MDI/pMDI is based on aniline and formaldehyde as the

main raw materials. Aniline is produced from benzene. The preferred and
lowest cost commercial production method is via nitration of benzene to
nitrobenzene, using a mixed acid process (HNO3/H2SO4)22,23 whereby aniline
is produced from nitrobenzene via a reduction step.24,25 Methylenedianiline
(MDA) and a mixture of its oligomers (pMDA) subsequently results from
an acid-catalyzed condensation of aniline with formaldehyde:26,27

(C6H5 - NH2)2 + H2C = O æ(æ

HCI )
æÆ NH2 - C6H4 - CH2 - C6H4 - NH2

Three MDA isomers are predominantly produced in this reaction process:

4,4’methylenedianiline (4,4'-MDA), 2,4’ methylenedianiline (2,4'-MDA), and
2,2’methylene-dianiline (2,2'-MDA). The relative amounts of these products
depend on the acid catalyst concentration, the aniline/formaldehyde ratio,
and the process temperature (Structure 7.1):

© 2004 by CRC Press LLC

www.ketabdownload.com
 H 2N C NH2
H2

NHCH2 NH2 4,4'-MDA

N-(p-aminobenzyl)aniline H2N

H2N C
H2
H2N
2,4'-MDA
NHCH2
NH2
N-(o-aminobenzyl)aniline

C
H2

H2N
2,2'-MDA

(Structure 7.1)

In addition to the MDA-isomers, the oligomeric mixture of pMDA is

largely comprised of compounds with three (n = 1) to approximately six
(n = 4) phenylamino functional groups (Structure 7.2). They are formed
via further reactions of the MDA-isomers:

NH2 NH2 NH2

C C
H2 H2
n

(Structure 7.2)

The MDA/pMDA mixture is subsequently phosgenated into the corre-

sponding isocyanate mixture of MDI/pMDI in the presence of an inert
solvent (Structure 7.3).

NH 2 NH 2 NH 2 NC O NC O NC O

CO Cl 2
C C C C
H2 H2 H2 H2
n n

MD A/ PM DA MD I/ PM DI

(Structure 7.3)

High efficiency initial mixing of the amine solution and the phosgene
solution is important for obtaining a high quality, consistent MDI/pMDI
product with high yields and low byproduct formation. Important parame-
© 2004 by CRC Press LLC
www.ketabdownload.com
ters are, apart from the relative concentrations of the reactants, reactor
design, temperature, and pressure.28–31 In the initial reaction stage carbamoyl
chloride and amino hydrochloride are formed as an intermediate, which are
subsequently thoroughly phosgenated in a second process step at high tem-
peratures, yielding the isocyanate product (Structure 7.4):

O
RNH2 + COCl2 + HCl
RNH Cl
carbamyl chloride

RNH2 + HCl RNH3Cl

RNCO + HCl
RNH Cl
(Structure 7.4)

Following the reaction of the polymeric amine mixture with phosgene,

excess phosgene and acid are removed from the mixture and the solvent is
distilled off. Undesired secondary reactions are initiated when amines react
with isocyanates to form ureas, which can be dehydrated by phosgene to
form carbodiimides (CDIs) (Structure 7.5):

O
RNH2 + RNCO
RNH NHR
urea

COCl2

RN C NR
carbodiimide

(Structure 7.5)

In the pMDI-process other byproducts can be formed from CDI, such as

uretonimine and iminohexahydro-s-triazine compounds.32 The final MDI/
pMDI mixture contains varying amounts of different functionality isocyan-
ates. Monomeric MDI is typically present between 40 and 70% (w/w) in this
mixture, with a split between 80 and 98% 4,4'-MDI and 2–20% 2,4'-MDI,

© 2004 by CRC Press LLC

www.ketabdownload.com
 35
30
Weight-percentage
25
20
15
10
5
0
N2 N3 N4 N5 Nrest
MDI-functionality
FIGURE 7.4
Typical distribution of MDI and its oligomers in polymeric MDI (pMDI).

sometimes also referred to as para-para and ortho-para MDI, respectively.

Oligomeric MDI with more than two NCO groups per molecule is typically
present at 30–60% of the total mixture.
In production, pure MDI is separated from the MDI/pMDI product mix-
ture by means of distillation. The bottom product from this distillation is
what is known as polymeric MDI (pMDI), which is the isocyanate of choice
for the production of rigid polyurethane foam. It usually contains a mixture
of pure MDI (N2) and higher functional isocyanates (N3, N4, N5, Nrest), as
shown by Figure 7.4. Dependent on market demand and plant capacity,
pMDI is sometimes produced as a straight run product by changing the plant
operating conditions.

7.2.1.3 Properties
Polymeric MDI products are usually characterized by their NCO content,
acidity, and viscosity. The NCO content is defined as the NCO weight as a
percentage of total weight and is indicated by %NCO. The molecular weight
of a reactive NCO group being 42, the isocyanate equivalent weight (I.E.) is
readily calculated from %NCO values via:

I .E. = 42 x100 %NCO

The number of NCO equivalents present in a weighed amount of pMDI

follows in the usual way (by dividing the total weight through I.E.). Subse-
quently, the so-called isocyanate index for a given reaction with polyol(s)
can be calculated via the equation

Equiv.Isocyanate
Index = Equiv.Polyol

where the number of equivalents of polyol component is defined by the

polyol equivalent weight. The isocyanate index is very important in poly-

© 2004 by CRC Press LLC

www.ketabdownload.com
urethane chemistry as it quantifies the number of NCO groups available for
reaction with polyol(s). Rigid polyurethane foaming is usually carried out
at an index of 1.05–1.20 which implies an NCO excess of 5–20% on an
equivalent basis. A notable exception is polyisocyanurate foaming which is
carried out at index values from 2 to 5 in order to stimulate the trimerization
reaction between NCO groups, as discussed in Section 7.4.3.

7.2.1.4 Products
The acidity of pMDI is important for its effect on the reactivity, and hence
processing performance, of rigid foam systems which are base catalyzed,
and its processing performance. The viscosity of pMDI is linked to its average
functionality and oligomeric distribution. As a rule, viscosity increases with
decreasing MDI monomer (N2) content; that is, the higher the viscosity, the
higher the average number of NCO groups per molecule. The average num-
ber of isocyanate groups per molecule is called the “functionality” of a given
pMDI. It ranges from 2.7 to 3.2 in current pMDI products, which corresponds
to a viscosity range of 200 to 2000 cps at 25˚C. Low viscosity pMDIs (200 cps
at 25˚C) are used as a standard throughout rigid polyurethane foam appli-
cations and in those cases where the flowability of the foam is a key param-
eter, such as, for instance, in the appliance industry. In the building industry,
the preferred pMDI type has a viscosity of 550 to 700 cps at 25˚C. High
viscosity pMDIs are favored when mechanical strength or curing rate is an
important processing parameter. This is typically the case for metal-faced
sandwich panels at higher thicknesses, as discussed in Section 7.3.2.2.
A special class of isocyanate products for rigid foams is formed by the
modified isocyanates or so-called prepolymers. These are isocyanates which
are partially prereacted with polyols. Although there are two types of these
prepolymers, only the NCO-terminated polyols, created by the reaction of a
polyol with an excess of isocyanate, are of some technical and commercial
importance for rigid foams, most commonly in rigid block production to
control the exotherm during the foam reaction. It also produces less friable
foams and provides better adhesion to substrates. The prepolymer reaction
is normally carried out in the absence of catalysts at relatively low temper-
atures (40–90˚C) to prevent side reactions like dimerization, trimerization,
or carbodiimide formation. These reactions may create additional heat as
well as unwanted gelation and an uncontrolled pressure buildup. Prepoly-
mer reactions are carried out under neutral or mildly acidic conditions to
favor chain extension via the urethane reaction and to avoid unwanted
gelation through branching.

7.2.2 Polyols
7.2.2.1 Introduction
Isocyanates react with any molecular entity possessing active hydrogens, such
as hydroxylic, carboxylic, and amine groups. For rigid foams, primary and

© 2004 by CRC Press LLC

www.ketabdownload.com
 O O
HO R O C R′ C O R O H
n

Typical polyester polyol

CH3

CH2 O CH2 CH O H
m
CH3

CH O CH2 CH O H
m

CH3 n−2

CH2 O CH2 CH O H
m
Typical polyether polyol

FIGURE 7.5
Structural formula of typical polyether and polyester polyols. (Reprinted with permission from
Grünbauer, H.J.M., Rigid polyurethane foams, in Polymeric Materials Encyclopedia, vol. 10, ed.
J.C. Salamone, CRC Press, Boca Raton, FL, 1996, 7504–7512.)

secondary hydroxyl group terminated polyether polyols are most important,

followed by polyester polyols, which have the oldest application.1–4 Both are
represented by Figure 7.5, where polyether polyols are seen to provide by far
the largest variety of possible molecular structures as far as hydroxyl func-
tionality and molecular weight are concerned. This is expressed in Figure 7.5
by integers m and n, which represent the number of alkoxydes per branch
and the number of branches of the polyol, respectively. The latter, usually
termed the “functionality” of the polyol, can range from 2 to 8, depending on
the original initiator molecule used for alkoxylation. In contrast to polyols for
flexible foams, rigid foam polyols are relatively small in comparison to the
original initiator molecule. This is due to the requirement of high stiffness of
the foamed, rigid end product, which is achieved by means of crosslinking
and urea precipitation, as detailed in Section 7.4.5. As a result, the molecular
weight of the initiator is sometimes as high as 70% of the total polyol molecular
weight. The character of the initiator is thus reflected in the performance of
the polyol during rigid foam formation. This gives the polyol formulator the
ability to tune polyol blends for desired product performance via selection of
suitable polyols and their blends.
Polyester polyols for rigid foam applications are almost always of aromatic
origin and can be characterized in terms of raw material origin. They are
produced by condensation of phthalic anhydride, dimethylterephthalate
(DMT) or polyethyleneterephthalate (PET) recycle streams with ethylene
glycol, diethylene glycol, or triols and higher functional products like penta-
erythritol. These moieties are represented in Figure 7.5 by R¢ and R, respec-
tively. Polyester polyols enjoy a growing interest for rigid polyurethane foam
producers due to their relatively low cost as well as the presence of aromatic

© 2004 by CRC Press LLC

www.ketabdownload.com
groups in their backbones, which gives enhanced flammability performance
of the foam.

7.2.2.2 Production
Polyether polyols are manufactured by the addition polymerization of alky-
lene oxides. The main reaction for rigid polyols is the homopolymerization
of propylene oxide and, to a lesser extent, the homopolymerization of eth-
ylene oxide and the co- or block-polymerizations of both. Rigid polyols are
anionic polymerized. The overall polymerization can be characterized with
four typical steps: initiation, propagation, chain transfer, and termination.
Catalyst concentration, temperature, and type of alkoxide used determine
the reaction rate.
The initiation step is shown below:

Initiation: I NIT - H + K + OH - æ
æÆ I - NIT- K + + H 2 O ≠

An initiator (e.g., glycerine or sugar) is loaded in the reactor with KOH.

Vacuum is applied to force the equilibrium to alkanolate formation. A nucleo-
phile can also be formed from direct reaction of the base with a monomer
to form an alkoxide ion:

O
K + OH - + n H æÆ HO[-CH2 -CHR-O-]n - K +
æ
R

Diols are initiated with this reaction; a KOH solution is heated and the
alkoxides are fed to the reactor. Propagation and chain transfer immediately
follow initiation. After the water has been (sometimes partly) removed the
monomer is added and propagation and chain transfer will take place at a
temperature of 100–120˚C.

Propagation:
O
H æÆ Int[-CH 2 - CH 2 -CHR -O -]n - K +
æ
+ -
Init OH + n
R

Chain transfer:
Init[-CH2 -CHR-O-]n - K + R x OH ¨æææ
æææÆ Init[-CH2 - CHR-O-]n H + R x - K +

Only a few hydroxyl groups are alkanolated and therefore active to grow.
A narrow molecular weight distribution can be obtained for two reasons:
the chain transfer reaction and the dependence on the molecular weight of
the reaction rate of the growing alkoxide chains. In other words, lower
equivalent weight chains grow faster, and hence, narrower distributions are
obtained. Because ethylene oxide reacts more quickly than propylene oxide,
the copolymer chains will be predominantly terminated by secondary

© 2004 by CRC Press LLC

www.ketabdownload.com
hydroxyls. If termination by primary hydroxyl is desired, ethylene oxide is
loaded to the reactor after complete reaction of propylene oxide.
The reaction is finished or terminated by adding water:

Termination:

Init[-CH 2- CHR - O -]n - K + + H 2 O æ

æÆ Init[- CH 2 - CHR - O -]n H + KOH

After the termination reaction the polyol must be neutralized. Rigid poly-
ols are mostly neutralized with acids; the residual salts are filtered out or
left in the polyol. These residual salts (like K-acetate) will catalyze the poly-
urethane reaction if they are left in the polyol; see Section 7.2.4.1.

7.2.2.3 Properties
Polyol structure is usually quantified in terms of equivalent weight, func-
tionality, chain structure, and primary or secondary hydroxyl end groups.
The hydroxyl equivalent weight HEW of a polyol is given by:

HEW = Mpolyol functionality

where functionality represents the number of reactive groups per molecule

available for alkoxylation of the original initiator. Assuming all available
groups have reacted, the functionality corresponds to the nominal number
of branches or arms of the final polyol.
In those cases where a mixture of initiators is used, the average function-
ality is calculated from the molar fractions of each initiator:

 ÈÍÎ(moles moles) ◊ functionality ùúû

i

functionality = i
i
n =1

The alkoxyde chain structure of the polyol may be varied in a number of

different ways, by selecting different alkylene oxides, mostly ethylene oxide
(EO) and propylene oxide (PO), or by using homopolymers or block or
random copolymers. The structure of the end group is of special importance
for the intrinsic reactivity of the polyol. Polyols terminated with primary
hydroxyl groups containing ethylene oxide are more reactive with isocyanate
than their propylene oxide counterparts which predominantly contain sec-
ondary hydroxyl groups. To quantify the number of reactive groups, the
hydroxyl number OHno is frequently used. It is defined as the concentration
of isocyanate reactive hydroxyl groups per unit weight of polyol, expressed
in mg KOH/g.

OHno = Ê ˆ
56.1 * functionality
Ë Mpolyol¯ * 1000

© 2004 by CRC Press LLC

www.ketabdownload.com
It follows that:

HEW = Ê 56.1OH ˆ * 1000

Ë no ¯

which can be readily used to calculate the total number of equivalents avail-
able for reaction with pMDI in a given amount of polyol. The calculation of
the isocyanate index from this number and the number of pMDI equivalents
is explained in Section 7.2.1.3.

7.2.2.4 Products
Except for high index isocyanurate foams, high functionality sucrose or
sorbitol initiated polyols are at the basis of every rigid polyurethane foam.
A typical polyol formulation comprises 40–60% of this polyol type, while
the remaining additives and polyols are selected to balance the processing
needs for the specific application at hand. In the last 10 years, there has been
a notable shift toward sorbitol initiated polyol, especially in Europe, where
these polyols are widely used. Perceived advantages of sorbitol intiated
polyols include improved compatibility with hydrocarbon blowing agents,
enhanced flammability performance, higher foam strength, and faster
strength build up during the first 10 min after reaction (green strength). As
a result, demold times are reduced and thicker panels, up to 20 cm, can be
produced on a double band laminator.
A wide variety of polyol components are used to tune processing and
performance parameters, such as flowability, demold, adhesion, reaction
profiles, and so on. An incomplete listing includes 50 to 100 equivalent
weight EO diols as chain extenders and metal adhesion promotors, 500 to
750 equivalent weight PO diols and triols as hydrocarbon compatibilizers,
plastic adhesion promotors, viscosity cutters and flow enhancers, and so on.
In practice, the improvements always need to be balanced against deterio-
rated foam strengths, increased demold times, and flammability perfor-
mance. A special class of polyols in this context is quatrols or 4-functional
polyols based on amine initiators such as ethylene diamine and toluene
diamine. The amine group gives these polyols a self-catalytic character for
reaction with isocyanate. Apart from the advantage of reduced catalyst lev-
els, these polyols are further beneficial in that they yield foams with shorter
demold times and increased foam strength versus triol-based counterparts.
Polyester polyols are an important raw material for the production of
polyisocyanurate (PIR) insulation panels, especially in North America. These
foams derive their excellent flammability performance from the high content
of aromatic groups which results from isocyanurate formation in combina-
tion with high aromaticity polyester polyols. In other regions, polyester
polyols are seeing increased application for the same reason as well as for
their relatively low cost. The latter is ultimately connected to the origin of
aromatic polyester polyols which are produced from recycled PET bottles,
fibers, and x-ray film sheets, or DMT process waste.

© 2004 by CRC Press LLC

www.ketabdownload.com
7.2.3 Blowing Agents
For almost 30 years, chlorofluorocarbon CFC-11 was the blowing agent of
choice for the rigid polyurethane foam industry, due to its high molecular
weight, in combination with low toxicity, optimum boiling point (23.8˚C),
and very low gas K-factor of 7.8 mW/mK. As a result, CFC-11 can be con-
sidered the best blowing agent ever used for thermal insulation foam appli-
cations: rigid polyurethane foams containing high amounts of CFC-11 could
reach K-factors as low as 15-17 mW/mK.1–4 This performance level is simply
unattainable with current alternatives that have been developed in response
to the global issues of ozone depletion and global warming11–13 raised during
the last two decades of the 20th century. The phase-out of CFCs was com-
pleted in the developed countries by the first of January 1996, 13 although
several countries voluntarily completed the phase-out much earlier.
New blowing agent technology required identification or development of
compounds with low gaseous thermal conductivity. They had ideally to be
cheap and also had to show many other characteristics in order to fit within
rigid polyurethane foam applications. Characteristics vary widely and
include qualities such as being non-toxic, being environmentally acceptable,
having zero ozone depletion potential (ODP), having low global warming
potential (GWP), being chemically inert, preferably having good solubility in
the polyol and isocyanate components, having a usable boiling point (ideally
in the range of –10 to 50˚C), having a low diffusion coefficient in the PU foam
matrix, and of course being readily available in commercial quantities.
Not surprisingly, none of the current alternatives can fully satisfy these
requirements. Nevertheless, hydrocarbons, HCFC, and HFC replacements
are currently employed by industry. Major European manufacturers have
almost completely converted to hydrocarbons while adapting conventional
foaming plants to guarantee safety and compliance with environmental reg-
ulations. North American manufacturers have chosen HFCs as a replacement
for short-term alternative HCFC-141b, which was phased out on January 1,
2003. Both HFC-245fa and HFC-134a blowing agents are considered due to
their relatively good insulation performance and low flammability. Both
blowing agents are properly handled by utilizing a premix station to assure
a constant mixing ratio between polyol and the blowing agent. Other regions
follow the Montreal Protocol implementation plan.13 Based on local regula-
tions, all blowing agents are still in use. Cyclopentane has been chosen as
the main blowing agent by the major original equipment manufacturers
(OEMs) in Japan and Latin America, while HCFC-141b is still widely used
in India as well as in the other Pacific countries. The aid of international
funds such as the United Nations Industrial Organization (UNIDO) has
driven the major African and Middle Eastern OEMs to adopt cyclopentane
as the preferred blowing agent.
An overview of the different blowing agents and their critical properties
and environmentally relevant properties is listed in Table 7.1.

© 2004 by CRC Press LLC

www.ketabdownload.com
TABLE 7.1
Overview of New Blowing Agents and Their Physical Properties
Vapor
Molecular
gas Boiling Pressure Flammability GWP Atmosph. Phase-
Molecular Weight @25∞C Point 20∞C Limits Indicative 100 yr POCP lifetime out
Compound Formula (g/mol) mW/mK (∞C) (bars) (% vol. in air) Price $/lb ODP ITH HGWP or SFP (yrs) target
CFC-11 C-Cl3F 137.5 7.8 24 0.88 none n.a. 1.0 4000 1.0 <0.1 50–60 1995
HCFC-141b CH3C-Cl2F 116.9 9.8 32 0.69 5.6–17.6 ~0.9–1.0 0.11 630 0.12 10 8–10 2003
HCFC-22 CHClF2 86.5 11.2 –41 8.92 none ~1.0 0.05 1700 0.35 6 13–15 2010
HCFC-142b CH3C-ClF2 100.5 9.8 –10 2.50 6.7–14.9 ~1.0 0.06 2000 0.36 5 19–20 2010
HCFC-124 CHClFCF3 136.5 12.3 –11 3.27 none ~2.0 0.02 480 0.10 12 5–6 2020
HFC-125 CHF2CF3 120 14.0 –48 12.41 none ~2.5–3.0 0 2800 0.58 1 32–35 none
HFC-134 CHF2CHF2 102 14.1 –23 4.53 none ~1.8–2.0 0 1000 0.20 8 10–12 none
HFC-134a CH2FCF3 102 14.3 –26 5.62 none ~1.5–1.8 0 1300 0.26 6 14–16 none
HFC-152a CH3CHF2 66 13.8 –25 5.06 3.9–16.9 ~2.0–2.5 0 140 0.03 14 1.5–2 none
HFC-245 fa CHF2CH2CF3 134 12.2 15 1.24 8.9–11.2 ~3.0–4.0 0 820 0.18 11 7–10 none
HFC-365 mfc CH3CF2CH2CF3 148 10.6 40 0.47 3.5–9.0 ~3.0–4.0 0 840 0.18 7 9–12 none
n-Pentane C5H12 72 14.6 36 0.65 1.4–8.3 ~0.25 0 11 0.002 500 0.03 none
Iso-Pentane C5H12 72 13.8 28 0.80 1.4–7.6 ~0.25 0 11 0.002 500 0.03 none
Cyclo-Pentane C5H10 70 12.6 50 0.34 1.4–7.8 ~0.75 0 11 0.002 400 0.05 none
Iso-Butane C4H10 58 15.9 –12 2.99 1.8–8.4 ~0.3 0 5 0.001 650 0.02 none
n-Butane C4H10 58 15.3 –1 2.08 1.6–8.4 ~0.3 0 5 0.001 650 0.02 none
Carbon dioxide CO2 44 16.3 –78 56.55 none n.a. 0 1.0 1.10-4 0 120–200 none
Air N2/O2 28.8 26.5 –193 624.03 none n.a. 0 0 0 0 none
ODP = Ozone Depletion Potential
GWP = Global Warming Potential
ITH = Infinite Time Horizon
HGWP = Halocarbon Global Warming Potential (R-11 = 1.0)
POCP = Photochemical Ozone Creation Potential (methane = 1)
SFP = Smog Formation Potential (methane = 1)

© 2004 by CRC Press LLC

www.ketabdownload.com
7.2.4 Additives
7.2.4.1 Catalysts

Polyurethane foaming is a highly complex event which requires delicate

balancing of several simultaneously occuring processes, such as foam expan-
sion, network kinetics, and phase segregation. Several types of catalysts are
known to influence these processes.1–3,33 To clarify the action of various
catalysts and to facilitate catalyst selection for practical purposes, the reac-
tions described in Section 7.4.3 are usually distinguished:1–3

• The blowing reaction of pMDI with H2O to form CO2 and polyurea
• The gelling reaction of pMDI with polyol to form polyurethane
• The trimerization of pMDI (as well as related reactions, such as
carbodiimide and dimer formation)

In preparing formulations for specific processing and application needs,

catalysts are thought to balance these reactions, albeit synergistic effects of
certain catalyst combinations are known as well. The distinction is never-
theless of some practical use as it provides guidance to the selection of
possible combinations for trial and error experiments. This system is there-
fore used to summarize today's most important catalysts in Table 7.2.

TABLE 7.2
Overview of Frequently Used Catalysts in Rigid Polyurethane Foams
Catalyst Type Code Perceived Action

Tertiary amines blowing

Niax A1 PMDETA blowing
pentamethyl diethylene triamine
triethylene diamine TEDA gelling
dimethyl cyclohexyl amine DMCHA blowing/gelling
triethyl amine TEA curing
Quaternary ammonium salts
2-hydroxy propyl trimethyl ammonium TMR gelling/curing/trimer
(2-ethylhexoate) formation
2-hydroxy propyl trimethyl ammonium TMR-2 delayed action/trimer
(formiate) formation
Alkali metal carboxylates
Potassium acetate K Ac gelling/trimer formation
Potassium octoate K Oct gelling/trimer formation
Sodium N-2-hydroxy, 5-nonylphenol Curithane 52 gelling/trimer formation
methyl N methyl glicinate
Tin complexes
Stannous octoate Sn Oct gelling
Dibutyl tin dilaurate DBTDL gelling

© 2004 by CRC Press LLC

www.ketabdownload.com
 Catalysts for rigid polyurethane foam include a range of chemical struc-
tures, such as tertiary amines, aromatic amines, quaternary ammonium salts,
alkali metal carboxylates, and organo-tin compounds. For the greater part,
their catalytic activity is dependent on their basicity, with steric hindrance
on the active site playing a secondary role. The activity of tertiary amines
is, for example, decreased in the presence of residual acidic compounds in
the polyol formulation. This phenomenon is succesfully exploited with so-
called “delayed action” catalysts which are active only after acid-blocked
amine catalyst decomposes during foaming to give active amine. Applied
as formic acid salts, these catalysts are active only after decomposition of
the formate at some higher temperature. Quaternary ammonium salts are
widely used because of their strong gelling and trimerization action, while
maintaining a relatively smooth foam rise profile. Alkali metal carboxylates
are very active curing and trimerization catalysts. Higher molecular weight
carboxylates are giving smoother rise rate profiles but have to be dosed at
higher amounts for a given catalytic effect.
Tin complexes are the most powerful gelling catalysts known today. They
are not often used in rigid foams, albeit that small quantities are sometimes
used to boost gelation. Stannous octoate is hydrolytically unstable, prevent-
ing its use in aqueous formulations.

7.2.4.2 Surfactants
The presence of surfactants serves a twofold purpose in polyurethane foam-
ing. First of all, they stabilize the foam immediately after mixing polyol and
isocyanate by lowering the surface tension of the emerging gas-liquid
interface34 and presumably also by emulsifying the polyol-isocyanate inter-
face. Mechanistically, this effect arises from a preferred accumulation of
surfactant molecules at interfaces. A second, equally important role of sur-
factants is to stabilize the polymerizing liquid-gas interface during the
roughly 30- to 50-fold volume increase of a rising foam.35 Here, the mecha-
nism is rather dynamic: the expanding foam continuously creates new sur-
face area of high tension that needs to be stabilized by fast migration of
surfactant towards the interface (the Marangoni effect).
The surfactants most commonly used in the polyurethane industry are
polydimethyl siloxane-polyether copolymers.36 Since the late 1950s, these so-
called “silicone surfactants” almost completely replaced other organic, non-
ionic surfactants which were used before.37,38 This is largely due to their
ability to lower surface tensions to approximately 20 dyne/cm, compared to
only 30 dyne/cm for common organic surfactants.39 Today, the largest com-
mercial application of polysiloxane surfactants is their use as additives for
the production of polyurethane foam, with a worldwide volume of about
40,000 MT/Y versus a global pMDI capacity of over 3,000,000 MT/Y.
The surface-active character of siloxane surfactants results from high-molec-
ular weight methyl-rich segments,40–44 attached to a flexible –O–Si–O–Si– back-
bone, in combination with polyether chains that are grafted to the same

© 2004 by CRC Press LLC

www.ketabdownload.com
backbone, either via Si—O—C or Si—C bonds. The latter type is hydrolytically
stable45 and predominant in polyurethane foam applications. They can be
represented as Me3SiO(Me2SiO)x(RMeSiO)ySiMe3 where Me represents a
methyl group (Structure 7.6) :

Me Me Me Me
Me Si O Si O Si O Si Me
Me
Me Me x

CH2

CH2
n
O
CH2

HC Me m

OR y

(Structure 7.6)

Silicone surfactants are typically added in amounts of 0.4–2.0% w/w of

the polyol formulation. To meet specific processing needs of different foam
systems, the molecular structure may be tuned by varying the length and
the composition of the polydimethylsiloxane backbone or the number,
length, and composition of the pendant polyether chains.46 Surfactants for
rigid foam typically contain 10 to 50 Si-units per molecule while the average
molecular weight of a polyether chain ranges from 400 to 1500 g/mol. The
ethylene oxide content typically varies from about 50 to 100% and the poly-
siloxane/polyether ratio lies in the range of 3/1 to10/1. The total molecular
weight of the surfactant amounts to 1500 to 15000 g/mol. In general, surface
activity in rigid polyol formulations increases with increasing polysiloxane
molecular weight. Increasing the polyether molecular weight has a reducing
effect on surface activity but is improving solubility. As a result, emulsifica-
tion and thereby cell stabilization can be improved. High propylene oxide
content tends to increase solubility in isocyanates, whereas high ethylene
oxide content tends to increase solubility in polyols, but simultaneously
reduce surface activity and flow properties.

7.3 Applications and Markets

Rigid polyurethane foams are employed in a rich variety of industrial prod-
ucts and processes. Among these are various buoyancy applications, such
as surfboards and bouys for navigation, and energy absorbing applications,

© 2004 by CRC Press LLC

www.ketabdownload.com
for example, in schock and explosive blast wave dissipation and as industrial
filters and carrier materials for biomedical research purposes. From an indus-
trial point of view, however, none of these applications is nearly as important
as their main use in thermal insulation materials. This is the reason why the
present section is limited to the use of rigid polyurethane foam in the global
appliance and construction market.

7.3.1 Appliance
7.3.1.1 Introduction
The construction of modern domestic refrigerators and freezers would not be
possible without the use of polyurethane foam technology.14,15 While its high
insulation efficiency ensures minimal wall thickness, polyurethane appliance
foams allow the use of very little plastic and steel, the structural strength being
instead guaranteed by the strength of the foam and its rigidity. The foam also
acts as the structural adhesive to bond the inner and outer liners of the
refrigerator during the foaming process, and allows ease of application and
manufacture. In physical terms, the foam itself is typically characterized by a
closed cell content of 95% or higher, low density (30 to 40 kg/m3), and high
compressive strength (ca. 150–200 kPa). It is manufactured by pouring, just
after mixing, a mixture of polyol and isocyanate into a mold, in most cases
between a plastic inner liner material, generally made of high-impact
polystyrene (HIPS) or acrylonitrile butadiene stryrene copolymer (ABS) and
a metal sheet, which will be the external part of the appliance. The location
of the foam in the final product is shown in Figure 7.6. A typical mold is
shown in Figure 7.7.

FIGURE 7.6
Appliance foam example, showing a rigid polyurethane foam component inside a refrigerator.

© 2004 by CRC Press LLC

www.ketabdownload.com
FIGURE 7.7
Appliance foam example, showing a refrigerator mold with a refrigerator in the “breakers up”
position. (Reprinted with permission from Cannon.)

The formulated polyol often consists of a blend of high functionality sucrose

or sorbitol initiated polyols and lower functionality polyols such as glycerine
and/or amine initiated polyols. The high functionality polyols in the system
are chosen to give the foam its high strength and rigidity, and the low func-
tionality polyols to reduce the viscosity, improve the compatibility of the
polyol formulation, and enhance flow characteristics. Other possible types of
polyols, such as polyester polyols, or polyaromatic polyamine (PAPA) or
toluene diamine (TDA) initiated polyols, may have a positive influence on
the thermal conductivity of the foam. Polymeric MDI (pMDI) is commonly
used on the isocyanate side. Blowing agent packages have recently shifted
from CFC-11 to hydrocarbons, HCFCs, and HFCs. This ongoing development
will be described in detail due to its impact on modern appliance technology.

7.3.1.2 Processing

Today’s industry standard for appliance manufacturing is a high pressure

foaming machine, typically equipped with an L-shaped impingement mix
head as described in Section 7.4.1 and depicted by Figure 7.11. The “L” in the
mix head design denotes the 90˚ angle that the foam transverses after mixing,
as it exits the mix head. This “L” shape is used to decelerate the liquid mixture
prior to its departure from the mix head, decreasing the amount of cell
destruction or coagulation prior to the gelling reaction taking place. As
discussed in Section 7.5.1.1, small cells are a key ingredient in creating a foam
which performs well from a thermal conductivity standpoint.16 Processing

© 2004 by CRC Press LLC

www.ketabdownload.com
parameters vary by application but normally run between 70 and 170 bar,
15–25˚C and 20–90 kg/min. These parameters are varied based on the
blowing agent used, the type or size of appliance being filled, and the optimal
flow pattern based on the internal dimensions of the appliance.
Foaming operations for refrigerator cabinets and doors consist of both
stationary and carousel fixture lines. These fixtures must support the exter-
nal and internal liners against the pressure generated by the foam flowing
into the cavity to fill. The methodology used for cabinet injection depends
on the position of the refrigerator during foaming. The “breakers up” posi-
tion shown by Figure 7.7, with the door opening face on top, is the most
commonly used position. This method is considered the most advantageous
because the reacting foam can distribute itself more uniformly in the back
part of the cabinet, then rise and flow to fill the lateral walls. For this flow
position, the flow paths are relatively short. In addition, the foam reaches
the farthest areas when its viscosity is rapidly rising, thereby limiting leak-
age. The “breakers up” foaming technology is typically the foaming position
of choice for the industry when building a new plant. The “breakers down”
position, with the door opening face on bottom, is sometimes chosen when
the mold construction is simpler. In this case, it is extremely important that
a proper cabinet seal is achieved to avoid foam leakage.
In the cabinet manufacturing process, the cabinet is quickly pre-assem-
bled, coupling the external metal sheets with a plastic inner liner made of
thermoformed HIPS or ABS. The inner and outer materials are then placed
into a fixture (or jig) coupled with a mold plug. Once properly set, the
cavity between the metal and plastic liners is injected with the proper
amount of polyurethane foam. During foaming, the position and number
of the injection holes depends on the foaming technology (breakers up or
down) as well as the foam reactivity.
Demolding times depend on many factors, such as type of blowing agent
employed, foam over-pack, wall thickness, and mold temperature. Demold
times as fast as 3–5 min can be obtained with a standard refrigerator’s wall
thickness (27–50 mm), while longer demold times are needed when the wall
thickness increases. After demolding, the cabinet will continue along the
manufacturing assembly line. During this assembly process, the door (pre-
viously foamed with a similar process), the compressor unit, and the other
elements (condenser, handles, sometimes the top cover) of the refrigerator
will be added to the cabinet.
Two types of refrigerator foaming line layouts are currently employed for
both breakers up and breakers down technology: carousel lines and station-
ary jigs lines. Carousel lines comprise a large number of molds in a contin-
uously moving line. The line is composed of a conveyor chain to transport
the molds and several stations where the cabinet is first inserted, heated up
before foaming, foamed, and finally demolded. Line productivity is depen-
dent on the number of molds. This number is determined by the demolding
time and the number of different models produced on the line. Insertion and
removal of molds are therefore quite important limiting factors.

© 2004 by CRC Press LLC

www.ketabdownload.com
FIGURE 7.8
Typical production line for appliances with stationary jigs. (Reprinted with permission from
Cannon.)

Stationary jigs lines allow superior productivity if a fast demold foam

system is employed. The line is composed of (1) a supply line from the
assembling area, (2) a preheating station where the cabinet is heated up to
facilitate foam flow, (3) stationary jigs where the cabinets are foamed and
kept until curing time is completed, and (4) an exit transport system to bring
the cabinet to the final assembling and control stations. In this case, each
mold can be equipped with its own mixing head or a single mix head system
can move, injecting multiple stationary fixtures. An example is given in
Figure 7.8. Compared to the carousel line layout, one of the advantages is
that the lighter cabinets are transported instead of the heavier molds, and
the line is easily automated, providing the maximum production flexibility.
A judicious combination of the mold and the cabinet design can eliminate
the bottlenecking of demolding time: each jig demold time can be appropri-
ately set so the longest necessary demold time for thick wall cabinets would
not affect the full line speed. Moreover, the availability of two to four plugs
already mounted on the fixture system would practically eliminate the time
wasted for model change and practically reduce the need for part storage,
providing obvious logistical advantages.
Doors can be foamed as an open pour or closed pour process. Both multiple
presses and a rotational mold system are currently used in the industry
routinely. In either case, the foaming principles are not different from those
discussed for the cabinet, though they are often an easier application. Open
lid foaming does guarantee an optimum foam distribution due to the ability
to adjust the pour pattern of the mix head over the cross section of the door.
However, closed pour doors often allow a more complete fill of the dikes
and ridges in a door.

© 2004 by CRC Press LLC

www.ketabdownload.com
7.3.1.3 New Blowing Agents
In response to global environmental issues11–13 raised during the last two
decades of the 20th century, the appliance industry has gone through a
considerable effort to replace CFC-11, the traditional blowing agent of choice,
by non-ozone-depleting (ODP) blowing agents that, preferably, exhibit low
global warming potentials (GWP) as well. At this moment, several replace-
ments are used in commercial plants which deserve special attention due to
their current and future (potential) impact on foaming practice in appliances.

7.3.1.4 Hydrochlorofluorocarbons
HCFC-141b is no longer used in Western Europe and North America but
remains an interim alternative in Latin America, in Eastern Europe, and in
the Pacific regions for those manufacturers who have decided not to convert
to hydrocarbon blown foams. HCFC-141b has a residual ozone depleting
potential (ODP = 0.11; see Table 7.1), but its low gas K-factor combined with
the fact that only minor machine/line modifications were required for its
implementation made it a preferred candidate for a certain number of OEMs.
Nevertheless, additional cost had to be considered due to the solvent effect
of HCFC-141b on thermoplastic liners (ABS or HIPS) which necessitates the
use of a protective layer on the inner liner or a modified liner polymer.
Hydrofluorocarbons HFC-245fa, now commercially available, and HFC-
134a are the chosen solutions for the medium and longer term in the U.S.
and possibly in the Asia-Pacific region, although many signs are observed
that hydrocarbons may end up being preferred in this region as well. HFC-
245fa has a boiling point of 15˚C and new systems blown with this material
have insulation properties comparable to HCFC-141b. Work to date has
shown about a 2% energy reduction for the switch from HCFC-141b to HFC-
245fa. Another alternative in use by some appliance manufacturers is HFC-
134a.17 The primary concern with the HFC-134a is that it provides a higher
K-factor at ambient than the other candidates, albeit the difference with
HCFC-141b and HFC-245fa diminishes at actual service temperatures
because of the absence of condensation effects, as discussed in Section 7.5.1.2.
For the same reason, dimensional stability is improved under these condi-
tions; see Section 7.5.1.2. On the other hand, processing and handling of
polyol formulation is challenged due to the low boiling point of –26˚C.

7.3.1.5 Hydrocarbons
The first hydrocarbon systems, launched in 1993, were blown with cyclo-
pentane. Compared to CFC-11 blown foams, their K-factor and density were
significantly higher. Since that time, continuous improvement of the hydro-
carbon formulations has led to the development of systems with thermal
insulation performance similar to 50% reduced CFC-11 systems. More recent
developments aimed to rectify the low vapor pressure of cyclopentane,
which may cause unwanted condensation at lower temperatures with neg-
ative impact on the dimensional stability of cyclopentane blown foams, are
discussed in Section 7.5.1.2.

© 2004 by CRC Press LLC

www.ketabdownload.com
 Long-term successful experiences in the appliance industry with low boil-
ing blowing agents (LBBA) such as HCFC22/142b and HFC-134a provided
the direction for a solution of this problem. This class of blowing agents is
loosely defined as having a normal boiling point below ambient (and fre-
quently even below 0˚C). Based on their boiling characteristics and process-
ing conditions, the low boiling hydrocarbons appeared to be ideal candidates
for use in combination with cyclopentane. The higher vapor pressure was
expected to produce foams with superior cell gas pressure, and hence, better
foam mechanical strength and dimensional stability. For isomeric butanes,
the inferior gas K-factor performance given by Table 7.1, as well as their
more critical solubility characteristic,18 are however limiting their maximum
concentration. A judicious combination of these two hydrocarbons does
nevertheless yield optimum properties for rigid appliance foam. An addi-
tional benefit of cyclopentane/butane blown systems is their excellent
flowability. This feature allows the use of 8–10% lower densities versus the
latest generation of cyclopentane, while maintaining compressive strength
values above the accepted limit to guarantee foam dimensional stability.
The application of cyclopentane/isopentane blends forms an intermediate
solution between cyclopentane and cyclopentane/butane technologies. It
consists of exploiting the low vapor pressure of iso-pentane to increase the
cell pressure and therefore improve the dimensional stability, subsequently
allowing a reduction of foam density and weight. This technique can be
applied most advantageously when the existing production unit does not
allow the use of low boiling hydrocarbon. Machine modifications are not
needed and existing polyol/hydrocarbon premixing units can still be uti-
lized. The amount of isopentane used with cyclopentane cannot exceed a
certain level to avoid degradation of the insulation properties; see Table 7.1.
To be attractive, isopentane levels should, however, be high enough to obtain
a sufficient increase of vapor pressure in the cells to allow reduction of the
foam weight. In practice, a good compromise has been found with a blend
consisting of 8–9 php of cyclopentane and 3–4 php of isopentane.

7.3.1.6 Vacuum Insulation Panels (VIPs)

The technological concept of VIPs is the use of the excellent insulation prop-
erties of high vacuum. Panels 15–25 mm thick made of a covering plastic foil
containing a filling material (silica, polyurethane, etc.) are evacuated in order
to achieve pressures of ca. 0.1–1 mbar. These panels typically have a K-factor
of 6–10 mW/mK. They are glued inside the cabinets and doors (on the plastic
or on the metal liner). Polyurethane foam is then poured the same way as
for conventional foaming. With VIPs the overall K-factor can be reduced
down to the level of 100% CFC-11 foams. VIPs are used by several OEMs.
Their cost is very high, i.e., 30–35 Œ/m2 or ca. 50–75 Œ for a refrigerator with
60% VIP coverage. This implies a doubling of the cost contribution for
insulation. On the other hand, VIPs allow for energy consumption reductions
as high as 20%, provided their reliability over the lifetime of a refrigerator
is guaranteed.

© 2004 by CRC Press LLC

www.ketabdownload.com
7.3.1.7 Markets
The production of cold domestic appliances reached ca. 80 million units in
2002, requiring more than 500,000 MT of polyurethane foam as filling and
insulation material. In appliances, the cost to produce a cabinet varies
between 100 and 250 Πas a function of the size and degree of sophistication
(this number is for the most common models; top end models can be two
to three times more costly). The cost of polyurethane foam raw materials for
the filling of a cabinet is ca. 3–7% of total cost. However, the role of the foam
is essential not only as insulation material but also to insure the structural
integrity of the refrigerator.

7.3.2 Construction
7.3.2.1 Introduction
Rigid polyurethane foam is widely applied for thermal insulation purposes
in the building and construction industry where it is available as foam board,
sandwich panels, sprays, aerosol cans, and shaped and molded foams.1,3,9,10
For their manufacture, polyisocyanurate (PIR) technology is frequently
applied, in addition to polyurethane types of technology, similar to those
described in Section 7.3.1 for appliance foams. PIR technology is particularly
significant in North America where it dominates the construction foams
market. It is characterized by the use of an excess polymeric MDI (pMDI),
with an isocyanate index ranging from 2 to 5, usually in combination with a
polyester polyol as described in Section 7.2.2. Together with suitable catalysts,
high isocyanate indices facilitate the formation of isocyanate trimers, as
described in Section 7.4.3, from which PIR foams derive their excellent flam-
mability characteristics.
Both PIR and polyurethane construction foams are characterized by their
excellent thermal conductivity with K-factors between 20 and 30 mW/mK,
low densities ranging from 30 to 100 kg/m3 at high mechanical strengths of
100–1000 kPa. They are used over a temperature range from –80 to +100˚C
and are especially efficient when one or more of these strengths are combined
with their adhesive bonding, easy mold-filling, and flexible processing capa-
bilities. The very low thermal conductivity enables application of thinner
panels when space filling is an issue. Adhesive bonding in combination with
high strength is exploited successfully in steel-faced sandwich panels. These
panels derive their excellent mechanical strength at low weight from a
unique, foam-based composite structure which explains why the steel sand-
wich market is dominated by the use of polyurethane foam as insulation
material.19 Mold filling latitude is utilized when foaming articles with a
complex shape, such as liquid natural gas (LNG) tank housings, by means
of “pour in place” techniques or spray foams. On-site insulation of cavities
around doors, windows, and beams using commercially available one-com-
ponent spray cans is another well-known example of the processing flexibil-
ity of rigid polyurethane foams. Spray foams combine thermal insulation,
filling, mechanical strength, and adhesion in a very cost-effective way.

© 2004 by CRC Press LLC

www.ketabdownload.com
 Construction foams are manufactured by high efficiency mixing of pMDI
with formulated polyol blends containing catalysts, surfactants, and, option-
ally, flame retardants, as defined by processing and application requirements.
Polyols usually contain relatively high amounts of reactive functional groups
per molecule (functionality) and relative short molecular chains (chain-
length) compared to polyols in flexible foams in order to achieve the desired
high crosslink density, which is key to end-product performance. Examples
of frequently used polyols are described in Section 7.2.2 and comprise sor-
bitol, glycerine, and sucrose/glycerine co-initiated polyols, blended with
polyester polyols and polyether polyols based on initiators containing aro-
matic groups. Reactivity is tuned for a particular manufacturing process by
varying the catalyst package which usually contains amines to balance foam
rise and gelation reactions. Salt complexes are important toward the end of
the foaming process as they have the ability to accelerate curing, particularly
via trimer formation; see Section 7.2.4.1. Together with CO2 from the H2O-
isocyanate reaction, pentane isomers and HCFC-141b are blowing agents of
choice for construction foams, albeit the latter is being phased out.13 Occa-
sionally, HFCs as described in Section 7.2.3 are used as well, especially in
North America, when very low thermal conductivity and severe flammabil-
ity requirements must be satisfied.
Flame retardants are important additives for building and construction
applications.20 Their role is to reduce heat and smoke emission by slowing
down polymer combustion and degradation with the aim to increase escape
time in case of fire. Flame retardants can be divided in reactive and non-
reactive components. Commonly used non-reactive flame retardants include,
in decreasing order of importance, organophosphorous compounds, with or
without halogen substitution; melamine; and alumina trihydrate. Bromine
and phosphorous containing polyols are the main representatives of reactive
flame retardants. Flame retardancy is further enhanced by the use of poly-
ester and aromatic polyether polyols. Together with polyisocyanurate trimer
formation at high pMDI indices, these polyols contribute to good flamma-
bility test performance by virtue of their higher thermal stability.

7.3.2.2 Continuous Production

Foam panels of up to 20 cm thickness are commonly produced by means of
a continuous production unit which is known as a double band laminator
(DBL). Depending on the type of facer on the panel, a further distinction can
be made into flexible-faced (DBL-FF) and rigid-faced (DBL-RF) foam panels.
A typical layout of a DBL-FF unit is given by Figure 7.9. This process consists
of a multi-component dosing unit; a high-pressure impingement mixing
head; a laydown section where, immediately after mixing, the foaming mass
is evenly distributed over the full width of the band; and a heated conveyer
to transport and cure the foaming mass followed by a panel cutting unit
where the panels are cut at the desired length. Before cooling down, insula-
tion boards are typically stacked to finalize the curing (bundle cure). This
technique is often applied in North America where insulation boards are

© 2004 by CRC Press LLC

www.ketabdownload.com
 DOUBLE BAND LAMINATOR

Polyol Polymeric Isocyanate

1 2 1 2

Mixer

High-pressure pump High-pressure pump

1
Catalyst/additive streams
Paper top facing
2
cutter
Blowing agent
2 1 3
3
Static mixer
1. Lay-down section
Paper bottom
2. High-pressure mix-head
facing
Panel stacking
3. Conveyor/oven
FIGURE 7.9
Layout of double band laminator for the continuous production of flexible faced insulation
panels.

frequently produced via PIR technology where sufficient heat is stored in

the panels to boost curing and adhesion to the facer. After cooling down, the
panels are ready for transport. Examples of flexible facings include paper,
aluminum foil, bonded fibers, bitumen, or combinations thereof. DBL-FF
lines may achieve running speeds up to 60 m/min when producing poly-
isocyanurate boards.
For DBL-RF with steel sandwich panels, the rolls with flexible facings are
replaced by steel coils and the unit is extended with steel profiling and pre-
heating sections to pre-shape and prepare the steel for the foaming operation,
which is otherwise identical to that of DBL-FF production. State of the art
machinery including additional equipment can be as long as 150 m for steel
sandwich panel production. The maximum length of steel-faced panels can
be up to 22 m. DBL-RF line speed varies from 4 to 15 m/min, depending on
panel thickness.
Rigid bunstock or block foam is sometimes produced in a continuous
fashion using foam lines that require less investment than typical DBL lines.
Mixing is carried out by a low pressure mix head, which enables continuous
dosing of up to 10 components, including solids and pastes. The foaming
mass is poured on a moving belt conveyor which is open on the top side,
except, in many cases, at the very end of the belt where a moving lid is used
to impose a square shape on the otherwise freely expanding foam in order
to minimize waste. Blocks produced on a continuous line are typically cut
in sizes of 1–2 m width, 2–4 m length, and up to 1 m height, depending on

© 2004 by CRC Press LLC

www.ketabdownload.com
the foam density. After cutting, the blocks need to be cured for 1–2 weeks
before being sold as such or being cut in the desired final shape, typically
panels, shaped shells, or cubes.

7.3.2.3 Discontinuous Production

Discontinuous production of panels is carried out using molds with well-
defined shapes and sizes, ranging from 3 to 12 m in length, 1 to 2 m in width,
and 5 to 20 cm in thickness. In this technology, a reacting polyol-pMDI
mixture is injected through a hole in the side of the panel and the injection
hole is closed immediately afterwards. Alternative technologies include mul-
tiple shot injection, filling tube process, and mix head withdrawal process.
In the latter, the entire mix head is inserted into the mold and withdrawn
during pouring. After injection, the foaming mass fills the mold, whereby
the air in the mold is released through small venting holes. The foam is left
in the mold for 10 to 60 min depending on the thickness of the panel.
Demolding takes place when the stiffness of the foam is sufficient for the
panel to keep its desired shape. Discontinuous panel manufacturing is char-
acterized by its great flexibility to produce almost any desired panel shape
with any kind of facer, which leads to a higher cost per square meter. Con-
tinuous panel manufacturing is less flexible but has a much lower cost per
square meter.

7.3.2.4 Products and Applications

In residential building operations, polyurethane foams are being applied
throughout the entire building, in roofing, walls, and flooring. Roof insula-
tion is either applied onto a roof deck followed by a moisture barrier, between
the roof cover and the tiles, or between the beams on the inside of the roof.
Insulation thickness for roofs varies from 2 to 20 cm depending on the needs
and local building regulations. Wall insulation can be divided into three
areas: cavity filling, and insulation on the outside or the inside of the wall.
External insulation foam, which can be PIR board or sprayed foam, is later
covered, for example, with plaster to provide a waterproof layer or a siding
material such as brick or wood. Insulation panels at the inside wall of a
building are usually applied to upgrade its energy efficiency. Polyurethane
or polyisocyanurate foam is ideally suited to this purpose as it requires the
smallest thicknesses of all available insulation materials to achieve a desired
thermal insulation performance.
Steel sandwich panels are by far the largest application of polyurethane
insulation foams in the commercial building market in Europe, while DBL-
FF is the largest application in North America. In this application, the specific
strengths of polyurethane foam are combined as in no other product to
provide thermal insulation, mechanical strength and building efficiency.
Steel sandwich panels are directly mounted to the steel structure of the
building and doors and windows can be cut out of panels on the building
site. An example is given in Figure 7.10. Sandwich building technology

© 2004 by CRC Press LLC

www.ketabdownload.com
FIGURE 7.10
Construction of industrial building using steel faced panels. (Reprinted with permission from
Koschade, R., Die Sandwichbauweise (How to Build a Steel Sandwich), Ernst & Sohn, Berlin, 2000.)

provides full freedom to architects and builders to design the entire building
using different profiles and colors. Despite their continuous production, there
is a huge variety of colors and profiles available at thicknesses ranging from
2 to 20 cm. The steel facings are giving an extra positive contribution to the
fire rating of steel sandwich elements. This was demonstrated by a series of
tests conducted according to a new European publication on fire testing of
building products.10,21
Garage and entry doors are another example of a smaller but fast-growing
market where aluminum or steel sandwich panels provide a combination of
high energy efficiency, comfort, and appearance. Both polyurethane and PIR
technologies are used in these applications. The thickness of these doors is
on average 35 to 55 mm.
Another typical polyurethane application in the commercial building mar-
ket is heating and ventilation. Specially shaped panels and shells are cut out
of rigid polyurethane foam blocks, produced either continuously on bun-
stock lines or on discontinuous molding facilities.
Spray foam is one of the oldest applications within the construction market.
It provides a simple, direct, and efficient way of applying thermal insulation.
In contrast to the above-discussed applications, the polyurethane chemicals
for spray foam production are brought to the building site and subsequently
foamed on-site using mobile equipment. Spray foam is particularly useful
for the insulation of cavities at the outside of storage tanks, under roofs,

© 2004 by CRC Press LLC

www.ketabdownload.com
underneath floors in cellars, etc. In North America, the predominant spray
foam application is roofing, where it is applied externally to the roof deck.
Industrial applications can be split into pipe and tank on one side and
district heating on the other side. Tanks are usually insulated via spray foam,
as discussed above under commercial applications, but can also be poured
in place; cavities are filled by discontinuous foaming, which is then applied
in ships for cryogenic storage of gases such as LNG. Pipes are mainly
insulated via pre-shaped pipe shells.
District heating serves the thermal insulation of underground pipes for
transportation of warm water and steam from power plants to residential
areas and transportation of liquids over long distances. The diameters of
these pipes may vary from 0.15 to 1.5 m. Pipes for district heating are pre-
assembled in factories whereby the cavity between a steel inner pipe and a
high density polyethylene (HDPE) outer pipe is filled with polyurethane
foam via discontinuous or continuous operations. Applied foam densities
are 70 to 80 kg/m3. The foams provide very good thermal insulation and
excellent mechanical strength, which guarantee the 30-year lifetimes required
by governmental certification procedures.

7.3.2.5 Markets

The total consumption of rigid polyurethane foam in the construction indus-

try on a global basis was ca. 3500 kT in 2000 and is expected to grow to ca.
4400 kT by 2005. The overview per geography and application, given by
Table 7.3, shows Europe to be the biggest consumer, followed by the U.S.,

TABLE 7.3
Market Data Rigid Construction Foam Per Application in kT, Year 2000
Global DBL-FF DBL-RF + DCP Block Spray PIP
Polyol 215 168 78 96 75
pMDI 356 168 78 96 49
Total 571 440 201 256 124
North America
Polyol 118 20 25 33 22
pMDI 199 33 42 57 34
Total 317 53 67 90 56
Europe
Polyol 64 110 25 32 18
pMDI 106 180 40 53 27
Total 170 290 65 85 45
Latin America
Polyol 7.5 1.2 3.8 2.4 1.1
pMDI 11.5 1.8 5.2 3.7 1.8
Total 19.0 3.0 9.0 6.1 2.9
Pacific
Polyol 26 37 24 29 8
pMDI 39 57 24 29 8
Total 65 94 60 75 20

© 2004 by CRC Press LLC

www.ketabdownload.com
the Pacific region, and Latin America, in that order. It is furthermore clear
that the largest segment by far is formed by DBL, followed by block and
spray foam. The U.S. and Europe are the largest consumers. The DBL-FF
market is larger than DBL-RF in the U.S., whereas in Europe this trend is
reversed. This is undoubtedly a reflection of differences in building practice
between the two regions.

7.4 Foaming Process

7.4.1 Mixing

By far the most common route to the manufacture of rigid polyurethane

foams comprises the mixing of pMDI with a formulated polyol containing
blowing agent(s) and additives. In practice, high-velocity jet impingement
mixing at pressures ranging from 50 to 200 bars is the preferred technique
used to create sufficient fluid motion for both components to form a fine
dispersion that allows molecular diffusion of reactants to be faster than their
rate of reaction.47 The initial mixture resulting from the turbulence of jet
impingement has been described as composed of “striations” of about 10–100
mm thickness.48,49 It is generally believed that this degree of mixing intimacy
is sufficient for thermodynamically and kinetically driven interfacial mixing
phenomena to be rapid enough to establish kinetic control of reactivity rather
than diffusion control.
The impingement mixing process itself has hitherto been studied in some
detail for RIM formulations47–54 using, for example, visual inspection of flow
patterns,47–50 adiabatic temperature rise,47,48,50,51,55 emulsion testing,50 physical
properties of molded parts48,49,51–55 or tracers48 as indicators of mixing quality.
Characterizing the intensity of mixing by Renzl, the nozzle Reynolds number
of the component of highest viscosity (usually the polyol), these studies
conclude that the extent of mixing increases sharply with increasing Renzl
until a critical value Recr is reached. Beyond Re mixing efficiency remains
flat for so-called I-heads, whereas for “self-cleaning” L-heads it still increases
further, albeit at a much lower rate.50,55 This is attributed to turbulence
enhancement due to the 90˚ bend that has to be passed by the mixture before
exiting the L-head,55 as shown by Figure 7.11. Some disagreement exists with
regard to the magnitude of Recr which ranges from 20047,53 to 38055 whereas
physical property enhancement flattens only beyond Re = 50055 when the
incremental effect beyond Recr of the L-head design is taken into account.
The relevance of these results for rigid polyurethane systems is not imme-
diately obvious due to the presence of significant amounts of blowing agents
and, eventually, flame retardants that are known to reinforce or counteract
interfacial mixing phenomena observed for linear55 and crosslinked55–58 RIM
systems. For example, chlorofluorocarbon based blowing agents such as

© 2004 by CRC Press LLC

www.ketabdownload.com
 ‘‘I” Shape Short ‘‘L” Shape Long ‘‘L” Shape
D1 D1 D1

A B A B A B
L1
L1
L1
L2

L1 L1 L2
= 1.0 = 1.3
D1 D1 L1
= 2.5
L1 D1
= 0.7
L2 L1
= 2.0
L2

FIGURE 7.11
Schematic diagram of mix heads for high-pressure impingement mixing. (Reprinted with per-
mission from Molnar, J.A., Jr. and Lee, L.J., J. Appl. Polym. Sci., 37, 2295, 1989.)

CFC-11, HCFC-141b, HFC-134a, and HFC-245fa are considered to be com-

patible with rigid polyol formulations,59–63 whereas pentane isomers are
reported to form emulsions with certain aromatic polyester polyols.64 Apart
from reducing polyol viscosity, it is very likely that the solvency of these
additives exerts a significant influence on polyol-pMDI compatibility and
thereby on interfacial mixing as well. Similar arguments hold for surfac-
tants.65 Undoubtedly, there are effects from water which is usually present
at rather high concentrations in the formulated polyol since the polyols in
these blends are necessarily more polar than typical flexible polyols due to
their high OH content for crosslinking.66 For example, a typical pentane
blown polyol may contain around 2% (w/w) water which translates into 46
mol% for a 750 molecular weight polyol. It is very likely that these highly
mobile molecules react early during foam rise,67 thereby enhancing polyol-
pMDI compatibility at the interface due to the concomitant formation of
urea-based oligomers; see Section 7.4.5.

7.4.2 Nucleation
Nucleation is the next noticeable event to occur immediately after mixing
polyol and pMDI. This phenomenon is of great practical interest due to its
impact on the ultimate cell size—and thereby on the insulation perfor-
mance—of a fully cured foam; see Section 7.5.1.1. From a macroscopic point
of view, foam nucleation is evident from a change in color of the reacting
mixture which adopts a creamy appearance due to the evolution of gas
bubbles, becoming big enough in size to scatter light. This change occurs at
a characteristic time, usually about 5–20 sec after the mixing started, which
is known as the cream time. At that stage, nucleation has occurred already
and the entrapped gas is in metastable equilibrium with its surroundings
which implies that, to a good approximation, its free energy is changing

© 2004 by CRC Press LLC

www.ketabdownload.com
linearly with time. By using this assumption, it has been shown that rc, the
critical radius of a gas-bubble or liquid droplet dispersed in a polyol, can be
expressed as:68–70

2sv
rc = (7.1)
v( p b - p a ) + RT (ln x * - ln x S )

where pb – pa represents the pressure difference between the interior of a

dispersed bubble (or droplet) and its ambient surroundings, and s represents
the interfacial tension of the polyol-gas/liquid blend. Molar volume, loading
level, and ambient pressure solubility of the blowing agent in the polyol are
indicated by v, x*, and xs, respectively. The stationary state described by
Equation 7.1 should not be confused with the preceding nucleation phase
which is essentially an unstable, transient stage. Qualitative information
about the actual nucleation phenomena may nevertheless be derived when
it is assumed (1) that each bubble or droplet is produced by one nucleus and
(2) that polyol and pMDI are not yet mixed on a molecular scale when nuclei
are formed. Under these premises, nucleation can be considered to take place
either in the polyol phase or in the pMDI phase of the reacting polyol/pMDI
mixture, and Equation 7.1 can be assumed to represent critical radii before
any chemical reaction has actually changed their phase behavior. Equation
7.1 thus provides indirect information about the preceding nucleation stage,
such that, for a given system, smaller values of rc hint towards more efficient
nucleation and vice versa.70 For simplicity, and for the practical reason that
blowing agents are usually dissolved only in the polyol, the discussion below
will focus on the polyol phase.
Literature reports two classical, limiting forms of Equation 7.1: one in
which the solubility term vanishes because xs = x* and one in which the
pressure term vanishes because pb = pa. These limiting forms are known as
the Laplace and Kelvin equations,71 respectively. The Laplace equation is
valid for one-component systems and for two-component gas-liquid systems
where the gas is essentially insoluble in the liquid phase, such as N2 in most
polyols. In these systems, xs ª x*, and rc is accordingly determined by surface
tension and by the intrinsic kinetic energy of the dispersed gas. For gases
above their critical point, such as N2, pb can take much higher values than
pa and hence rc will be small which explains the well-known suitability of
N2 as a nucleating gas in rigid foam manufacture.
The Kelvin equation is valid for incompressible liquid-liquid and solid-
liquid systems where pa = pb and rc is consequently determined by the ratio
of x* over xs alone. An illustrative example for rigid foams is the nucleating
effect of essentially insoluble perfluorinated liquids72 which are therefore
characterized by very high x*/xs ratios,70 while pa = pb in this case. Together
with their very small surface tension, this explains why highly significant
cell size reductions, presumably caused by very small rc values, have been
found upon addition of even small amounts of perfluorocarbons.72 The oppo-
site is also true: highly soluble blowing agents with vapor pressures close to

© 2004 by CRC Press LLC

www.ketabdownload.com
 0.04

WB 0.03

0.02

Methylformate Pentane

0.01

0.00
30 35 40 45 50 55 60 65 70
T (°C)

FIGURE 7.12
Construction of industrial building using steel faced panels. (Reprinted with permission from
Koschade, R., Die Sandwichbauweise (How to Build a Steel Sandwich), Ernst & Sohn, Berlin, 2000.)

ambient are poor nucleators. In this case, pa ª pb and x*/xs £ 1 due to very
strong attractive intermolecular interactions between the blowing agent and
polyol. The vapor pressure of the blowing agent is accordingly reduced in
the presence of polyol which is reflected by a significant boiling point ele-
vation in polyol-blowing agent mixtures, especially at lower loadings.59–61,73
An example is given by Figure 7.12, where elevations up to 30˚C are observed
for pentane and methyl formate, blowing agents that are known to be rela-
tively poorly soluble. Similar data for highly soluble blowing agents such as
CFC-11 show even stronger effects.59,62
As a result, nucleation with either liquid or gaseous CFC-11 is not likely
to occur until, in real life foam formation, temperature rises sufficiently to
force CFC-11 into the vapor phase. This is nicely illustrated by a study74 of
CFC-11 nucleation in the absence of air, which revealed a 1000-fold decrease
in number density of bubbles and a corresponding 10-fold increase in final
cell size versus air nucleated foams.
The situation is more complex when low-boiling blowing agents (LBBA)
such as HCFC-22, HFC-134a, propane, and butane are considered. In these
cases, both terms of Equation 7.1 are operational and nucleation efficiency
will depend on the steepness of the slope of the LBBA solubility versus
pressure profile.70,75 For example, the solubility of butane is expected to
increase fourfold upon increasing its partial pressure from 1 to 2 bars.70 A
polyol-butane blend loaded with up to several bars of butane in a typical
storage tank will thus be heavily supersaturated with butane once the mix-
ture has been laid down under ambient pressure after leaving the mix head
nozzle. Rigid foams (partially) expanded with LBBAs such as butane and
HFC-134a are therefore expected70 and actually observed69,76,77 to have
reduced cell sizes.

© 2004 by CRC Press LLC

www.ketabdownload.com
 The success of Equation 7.1 in describing nucleation phenomena in a
qualitative sense should not conceal that current understanding of nucleation
in quantitative terms is rather poor. This is due to the complexity of real life
foaming where phase mixing, emulsification, and fast temperature rise due
to chemical reactions all occur simultaneously. It is, for example, unclear
whether nucleation in rigid foams should be characterized as homogeneous,
i.e., by forming nuclei as described above, or as heterogeneous by formation
of new gaseous phases at the interface of insoluble gases, liquids, or solids
such as, respectively, N2 or liquid or solid perfluorinated particles.78 The fact
that very fine, open-celled foam can be obtained via nucleation with poly-
tetrafluoroethylene (PTFE) particles79–81 at least suggests that heterogeneous
nucleation is overriding in this particular case. Similar uncertainties exist
with respect to the importance of coalescence of freshly formed nuclei and
with respect to nucleation by CO2 formed (possibly at the polyol-pMDI
interface) from the H2O-pMDI reaction, which is also known to occur during
early stages of the foam rise process;67 see Section 7.4.5. In all, the conclusion
seems warranted that much more experimental and theoretical research
work needs to be carried out in order to create the next level of understanding
in this field.

7.4.3 Expansion
Rigid polyurethane foams are applied at densities ranging from about 20 to
50 kg/m3. Starting from initial raw material densities around 1 kg/l or
1000 kg/m3, this implies that a 20- to 50-fold volumetric expansion is taking
place during the foaming process which is complete within 1 min, and fre-
quently even faster. During that relatively short time frame, excessive bubble
growth as well as bubble collapse and coalescence of bubbles should be min-
imized in order to produce foams with uniform small cells that are desired for
various commercial applications. In practice, these requirements are fulfilled
by adding 1–2% (w/w) of a branched polydimethylsiloxane-(co)polyether sur-
factant to the formulated polyol. These compounds are thought to provide the
necessary lowering of surface tension and sufficient surface elasticity (or
“dynamic” surface tension) to counteract the thinning tendency, due to Gibbs-
Marangoni drainage in expanding liquid films; see also Section 7.2.4.2.
From a mechanistic point of view, foam expansion is driven by the evapo-
ration of the blowing agent and the liberation of CO2 gas, which is usually
generated via the H2O-pMDI reaction, albeit other reactions have been
exploited for that purpose as well.1–3 In chemical terms, this reaction is an
example of the reaction of -NCO groups with “active” hydrogens attached to
atoms of higher electronegativity than carbon. First discovered by Wurtz83
and Hoffman, this chemistry has later been extended by Bayer,84 who discov-
ered the poly-addition polymerization reaction of di-isocyanates with glycols
which forms the basis of todays polymeric urethane products. The reaction
of pMDI with H2O proceeds via the formation of carbamic acid, an unstable

© 2004 by CRC Press LLC

www.ketabdownload.com
intermediate which dissociates into CO2 and amine. The latter reacts relatively
quickly with pMDI to form urea. The overall reaction sequence is given by:

O O
R N C O
R N C O + H 2O R N C OH R N C N R + CO 2
H
H H

(7.2a)

where R represents the pMDI backbone. The overall reaction is exothermic

with a total heat release of about 45 kcal/mol of H2O.85–87 This mechanism
is commonly referred to by using the somewhat misleading term “water
blowing.”
The polymer backbone further comprises urethane linkages resulting from
the reaction of polyol with pMDI:88–91

R1 N C O + R2 OH R1 N C O R2 (7.2b)
H

This reaction is also exothermic with a reaction heat of 24 kcal/mol

urethane. The reaction rate is strongly influenced by the structure of the
polyol and isocyanate compound. The fact that current formulations
frequently contain H2O up to a CO2 mole fraction of 0.5 thus immediately
leads to the conclusion that most rigid foam polymers are more appropriately
termed copoly(urethane-urea) polymers rather than polyurethanes. This is
most readily detected by on-line FTIR techniques. Figure 7.13 gives an
example showing the kinetic development of urethane and urea linkages in
a typical “water-blown” foam as compared to a CFC-11 blown counterpart.4,67
The broad absorbance centered at 1730 cm–1 is assigned to urethane groups93
while the absorbance of urea groups tends to shift to lower wavenumbers,
depending on their solvation state.92 In the very early stages of co-
polymerization, the reaction mixture comprises polyurea, copoly(urethane),
isocyanate terminated polyether, monomers/oligomers, and non-reacted
water (if used as a co-blowing agent). Reaction between the monomers and
oligomers results in both a molecular weight increase and a concomitant
change in the values of the interaction parameters. It is apparent that urethane
absorbance (1730 cm –1 ), urea absorbance (1715 cm –1 ), and dissolved
hydrogen-bonded urea (1661 cm–1) are developing as the copolymerization
proceeds. It is also worth noting that solidification via chemical gelation and/
or vitrification occurs very quickly (time < 60 sec, at isocyanate conversion,
pNCO < 0.60) in these non-linear, highly branched systems.67
A wide variety of reactions with isocyanates have been described88–91 in
addition to urea and urethane formation, but their importance in the context

© 2004 by CRC Press LLC

www.ketabdownload.com
 A.

1750 1700 1650 1600 1550

B.

1750 1700 1650 1600 1550

FIGURE 7.13
Time-resolved, exploded view of the carbonyl region of the mid-infrared spectrum for a H2O
blown foam (A) and a CFC-11 blown foam (B) using identical polyol compositions and isocy-
anate index. (Reprinted with permission from Grünbauer, H.J.M., Rigid polyurethane foams,
in Polymeric Materials Encyclopedia, vol. 10, ed. J.C. Salamone, CRC Press, Boca Raton, FL, 1996,
7504–7512.)

of rigid foam formation is rather limited. An exception to this rule is the

trimerization of isocyanate (Structure 7.7) which is, especially in the U.S., of
technical and commercial significance, as described in Section 7.3.2:

O
R C R
N N
(
3 R N C O ) C C
O N O
R

(Structure 7.7)

This reaction is again exothermic. The heat generated by urea, urethane, and
trimer formation is responsible for the evaporation of the blowing agent, the
second driving force of the expansion process next to the formation of CO2.
Heat balance models correlating measured exotherms to foam expansion via
© 2004 by CRC Press LLC
www.ketabdownload.com
 simplified kinetic equations and heats of evaporation are relatively successful
in predicting actual foam rise profiles,59–62,73 provided that the boiling point
elevation effects shown by Figure 7.12 are taken into account.
Measurement of foam expansion has been attempted using many tech-
niques. These include floats riding the foam surface93 and recording foam
height using simple rulers94 and optical95,96 or ultrasonic97 sensors. By far the
most elegant method is based on the buoyant force exerted by air on the
expanding foam.98 At 200 g of formulated material this force leads to an
apparent weight loss of ª5 g for foams of about 30 kg/m3 free rise density,
which is accurately measurable by pouring the foaming mass in a
20 ¥ 20 ¥ 20 cm3 cardboard box positioned on a standard laboratory balance.
Fitted with a computer interface, this type of experimental setup is capable
of fully characterizing dynamic foaming profiles by simultaneous recording
of multiple signals. An example is given by Figure 7.14, where typical foam
volume expansion and exotherm development are plotted against time for a
formulation blown with CO2 via Reaction 7.2a. The rise rate profile in this
figure is calculated by differentiating the volume expansion profile and
smoothing. This profile, in turn, follows directly from the apparent weight
loss and the density of air.98 The on/off signal of the stirrer indicates both the
stirring time and the “gel” or “string” time, here 57 sec after start of stirring.
This characteristic time is commonly defined as the time at which strings of
solid material can be pulled from a rising foam using a wooden spatula. It is
interesting to note that, at this time, volume expansion is nearly complete
whereas the temperature in the foam core has only reached about 90˚C, almost
50˚C below the final value of more than 160˚C. This implies that resin forma-
tion is far from complete when the foam is fully risen and it also suggests
that polymer strength is being built up afterwards, despite the early start of
reaction kinetics immediately after mixing polyol and pMDI.

200 6.0
180
5.0
160
T (°C) or dV/dt (ml/s)

140
4.0
Volume (l)

120
100 3.0
80
2.0
60
40 1.0
20
0 0.0
0 20 40 60 80 100 120
Time (s)

String T(°C) Rise rate (ml/s) Volume (l)

FIGURE 7.14
Rise rate, volume expansion, and reaction exotherm of a water-blown rigid foam.

© 2004 by CRC Press LLC

www.ketabdownload.com
7.4.4 Rheology

The balance between foam expansion and polymer resin formation is of great
practical importance, as commercial foamers always strive for smooth expan-
sion profiles, or high “flowability,” in combination with fast curing to reduce
cycle time in molding applications or to increase line speed in laminated
foam manufacture. It is therefore not surprising to find several literature
references reporting measurement of important rheological parameters dur-
ing foam rise. This is experimentally rather difficult due to the combination
of high volumetric expansion rates with fast exothermic chemical reactions.99
Experimental approaches that have met with some success include foam
modulus measurement by means of a pulse rheometer setup100,101 or a flooded
parallel plate fixture in a rheometer,99,102–104 as well as the use of on-line
recording instruments, such as measurement of drag forces on immersed
spheres,105,106 vibrating needles,107,108 spheres,109,110 rods, or collets,111,112 and a
vane rheometer.113
Figure 7.15 shows a typical rigid foam profile obtained by means of a
Nametre vibrating collet.111 This instrument operates by emitting a flat shear
wave of 1–20 mm amplitude into the surrounding medium, which leaves the
growing foam cells intact. Provided proper precautions are taken to ensure
constant immersion depth by using an overflow device, the rheology profile
of a rising foam can thus be recorded starting from a few seconds after mixing
onward. It has been shown111,114 that, mathematically, the system is equivalent
to an electrical AC circuit in resonance, whereby elastic and viscous compo-
nents of the viscous drag forces caused by the surrounding medium act as
“mechanical” capacitance Xm and resistance Rm forces, respectively. The resis-
tance against the shear force exerted by the vibrating rod, Rm , is given by
Equation 7.3:

Rm = pfM hnr (7.3)

where fM represents the resonance frequency of the rod in the medium and
hn and r its nominal viscosity111 and density, respectively. By simultaneous
on-line recording of foam expansion and viscosity rise, the density term can
be eliminated to yield the actual polymer viscosity build-up during foam
rise. The expansion, hn, and Rm profiles of Figure 7.15 are exemplary. Taken
together, these profiles nicely illustrate that Rm , the actual mechanical resis-
tance against shear forces, remains at an almost constant low level during
the first 60 sec of the foaming process, despite the continuously increasing
polymer viscosity. The reason is the concomitant volume expansion which
compensates for the increasing viscosity term in Equation 7.3 via its connec-
tion with the density term. In other words, the flowability of a rising foam
is primarily dependent on its volumetric expansion rate, albeit resin forma-
tion needs to be in balance, such that premature gelation due to too high
polymerization rates has to be avoided.

© 2004 by CRC Press LLC

www.ketabdownload.com
 10 10

8 8

kPa.s/m (Rm)
Pa.s (ηn ) or

6 6

Liter
4 4

2 2

0 0
10 20 30 40 50 60
Time (s)
ηn Rm Volume

FIGURE 7.15
Volume expansion, nominal viscosity rise, and mechanical resistance profile of a typical rigid
foam.

Throughout the initial stages, the resonance frequency fM remains virtually

constant, until appreciable elastic forces start to develop, which are notice-
able as a shift in fM. By using proper calibration constants, it is possible111 to
extract the mechanical capacity term Xm and to convert the profile into more
common rheological parameters, such as dynamic viscosity and modulus h¢
and G¢, respectively. The resulting profile of Figure 7.16 clearly shows the
expected features of a steep increase of h¢ upon gelation, followed by a much
slower build-up of G¢ during the subsequent curing phase.102

6 1000

800
4
600
G′(MPa)
η' (mPa.s)

400
2
200

0 0
0 50 100 150 200 250 300
Time (s)

G′(MPa) η'(mPa.s)

FIGURE 7.16
Buildup of dynamic viscosity h¢ and dynamic modulus of G¢ of a rigid foam polymer during
foam rise.

© 2004 by CRC Press LLC

www.ketabdownload.com
7.4.5 Polymer Morphology

The development of the polymer microstructure and morphology during

rigid foam formation can be divided into four processes: first, the urea-
urethane co-polymerization itself as well as the subsequent development of
the polymer architecture and microstructure;67,115–119 second, the competition
between the co-polymerization and the polyurea macrophase separa-
tion;4,67,119–128 third, the solidification processes of vitrification and chemical
gelation and how these impact the development of macroscopic proper-
ties;4,120–121 and, finally, the build-up of the prevailing polymer morphology
of the foam, several hours to days after the reaction is complete, as the result
of a slow curing process.
The role of water—about 2 to 4% w/w of polyol in, respectively, 50 and
100% CO2 blown foams—is important not only for the foam expansion
process, but for the polyurea that is formed as a result, and has a significant
impact on the evolution of the morphology and on the ultimate mechanical
properties of the polymer. This is nicely illustrated by Figure 7.17, which
compares the loss tangent, tan d, obtained by dynamic mechanical spectros-
copy (DMS) as a function of temperature for CFC-11 and H2O blown foams
over a wide range of isocyanate indices. The once commercial CFC-11 blown
polyol system comprised a fine-tuned optimized blend, while the water-
blown polyol system incorporated a high functionality base polyol and a
high equivalent weight, viscosity cutting polyol.67 The contrast in behavior
is immediately apparent from inspection of Figure 7.17. A smooth tan d shift
with increasing index is observed for the CFC-11 foams, which is typical for
a gradual increase in crosslink density.135 In contrast, the water-blown foam
shows a sharp transition at –40˚C, which gradually disappears at higher
isocyanate indices.
Meanwhile, a broad transition emerges at higher temperatures. At an index
of approximately 1, there is a maximum in the loss peak at approximately
200˚C, accompanied by a broad shoulder between 100 and 200˚C. As the
index is further increased, the transition narrows, which was ascribed to
isocyanurate formation. Combining the DMS profiles from Figure 7.17 with
additional information from DSC, TEM, and solvent extraction studies, the
conclusion seems warranted67,118–121 that, in the water-blown case, polymer
strength build-up is dominated by the high functionality, base polyol
component, and concomitant precipitation of urea. The latter is particularly
evident from TEM images revealing the presence of irregularly spaced dark
spots with a length scale of 50–300 nm. The number of these “spots”
(tentatively ascribed to urea-urethane precipitates) was observed to increase
with increasing isocyanate index, while above an isocyanate index of 0.7,
the “spots” could no longer be discerned.120 Figure 7.18 presents an example
recorded at an isocyanate index of 0.5.

© 2004 by CRC Press LLC

www.ketabdownload.com
 A.
0.8
0.6

0.6
0.7
tan δ
0.4 0.9 1.3
1.1

0.2

0.0
−100 0 100 200 300
Temperature (°C)

B.
0.4
0.3

0.3

0.9 2.0
tan δ

0.2 1.2
0.5

0.1

0.0
−100 0 100 200 300
Temperature (°C)
FIGURE 7.17
DMS spectra of CFC-11 blown foam (A) and water-blown foam (B) at varying iocyanate indices,
as indicated by the curve labels. (Reprinted with permission from Grünbauer, H.J.M., Rigid
polyurethane foams, in Polymeric Materials Encyclopedia, vol. 10, ed. J.C. Salamone, CRC Press,
Boca Raton, FL, 1996, 7504–7512.)

In this view, the low temperature transition at –40˚C is due to non-reacted,

high equivalent weight, low viscosity polyol component. This confirms the
practical notion of its “viscosity cutting” action in polyol formulations: it acts
as a reactive diluent for reacting/precipitating species during the early stages
of the foaming process, thereby replacing the solvency effect of liquid blow-
ing agents that are absent in this case.67
Probing the polymer morphology of (partially) water-blown foams on the
nanometer scale by means of SAXS119–121 revealed a rather complex architec-
ture which is best interpreted by using fractal concepts.130–133 To illustrate,
Figure 7.19 shows Lorentz corrected scattered intensity, I(q).q2 versus q,
plotted for five different water-blown model foams made with single-polyol
formulations of increasing propylene oxide per OH (PO/OH) chain length.121
The shoulder in the scattering curve of the foam derived from the 30 PO/
OH polyol is readily discernible. In this particular foam, the polyether chain
is of a sufficient length that when reaction between the hydroxyl group of

© 2004 by CRC Press LLC

www.ketabdownload.com
this chain and the isocyanate group from a polyurea sequence length takes
place, a block copolymer is evolved. At shorter chain lengths, this tendency
is apparently hampered due to a lack of mobility of the polyer chain segments
involved. The result in SAXS is a steadily decaying scattering intensity I(q)
that becomes linear when plotted as ln I(q) vs. ln q. This behavior has been
interpreted in terms of a fractal concept120,131 such that the fractal dimension
calculated from the slope of the ln I(q) vs. ln q curve indicates whether an
open, lung type structure with a characteristic high surface to volume ratio
or a more densely connected structure is involved.120,121
Figure 7.20 presents a plot of the mass fractal dimension, Fm , as a function
of the equivalent weight of the polyol for the above mentioned series of
model copoly(urethane-urea) foams.121 In general, as the equivalent weight
of the polyol increased (a greater number of propylene-oxide segments and
thus chain length), a concomitant decrease in the mass fractal dimension,
Fm , was observed. In other words, the shorter the chain length, the more
dense and rigid the structure. Furthermore, the higher the functionality of
the isocyanate that was employed, the higher the observed mass fractal
dimension, Fm , and therefore the greater the propensity for branching, and
the more dense and connected was the structure that was evolved. These
results are of great practical interest due to their immediate connection with
effective diffusion coefficients of CO2 and blowing agents as defined in
Section 7.5.2.1. It has been observed121 that foams exhibiting a rigid

FIGURE 7.18
TEM image of the fully water-blown foam of Figure 7.17 at an isocyanate index of 0.5. (Reprinted
with permission from Grünbauer, H.J.M. and Folmer, J.C.W., J. Appl. Polym. Sci., 54, 935, 1994.)

© 2004 by CRC Press LLC

www.ketabdownload.com
 30-MHA-100
q* = 0.04
I(q)q2/a.u.

14-MHA-100

6-MHA-100

3-MHA-100

2-MHA-100
0.00 0.05 0.10 0.15 0.20
Scattering vector, q/λ
FIGURE 7.19
Lorentz corrected scattering profiles of water-blown mode foams from single polyol formula-
tions at nominal PO/OH chain lengths of 2, 3, 6, 14, and 30. (Reprinted with permission from
Megat-Yusoff, P.S., Ph.D. thesis, Victoria University of Manchester, 1996.)

interconnected structure with a characteristic high value of the mass fractal

dimension, Fm , also displayed a low value of effective diffusion coefficient,
Deff.121 Moreover, when the equivalent weight of the polyol component in
the formulation was increased resulting in a concomitant decrease in the
crosslink density and mass fractal dimension, the effective diffusion
coefficient was observed to increase.121 This was explained in terms of the
“looseness” or “openness” of the polymer network. The more open and loose
the network structure, the higher was the propensity for diffusion of the
gaseous blowing agent through the foam structure.
The overall picture of polymer microstructure and morphology develop-
ment emerging from these studies is that of a reactive co-polymerization
process, more specifically a reaction-induced phase separation process. For
the case in which water is employed as a co-blowing agent, upon mixing and
molecular contact between the components the di- or polyisocyanate-water
reaction proceeds, generating polyurea and carbon dioxide. The polyurea
formed undergoes macrophase separation almost instantaneously, and will
subsequently aggregate. The formation of urethane from the di- or polyiso-
cyanate-hydroxyl reaction (derived from the high functionality base polyols),
reacts after and/or in competition with the water. Thus, there exists a kinetic
competition between the polyurea polymerization, the macrophase separa-
tion, and the subsequent association of the polyurea so formed, and the
incipient crosslinking reaction to form the covalent polymer network.
The kinetic competition between multiple processes results in a
“frustration” on the system with respect to the structure development process.
The extremely high crosslink density that results from the short chains (low
equivalent weight) on the high functionality base polyol, generates a very
© 2004 by CRC Press LLC
www.ketabdownload.com
 3.2

monomeric MDI
3.0 polymeric MDI
Mass fractal dimension, Fm
2.8

2.6

2.4

2.2

2.0

1.8
0 500 1000 1500 2000
Equivalent weight of polyol/g mol−1

FIGURE 7.20
Mass fractal dimension Fm as a function of equivalent weight for water-blown-made foam from
single polyol formulations at nominal PO/OH chain lengths of 2, 3, 6, 14, and 30. (Reprinted
with permission from Megat-Yusoff, P.S., Ph.D. thesis, Victoria University of Manchester, 1996.)

rigid and tight polymer network structure (similar to a fine-mesh fishing net).
This in turn prevents the formation of molecular associations (hydrogen
bonds) between segregated polyurea aggregates, beyond a certain size and
length scale; hence, the lack of appreciable concentrations of hydrogen-
bonded urea in comparison to that observed in a flexible copoly(urethane-
urea) foam. In a very short time frame (30–60 sec depending on the
formulation) chemical gelation of the polymer network and vitrification of
the polyurea kinetically “trap” the evolving morphology and freeze it in place.
The result is a very irregular, phase-separated morphology that exhibits
fractal symmetry.

7.5 Physical Properties

7.5.1 Thermal Conductivity
7.5.1.1 K-factor Contributions
It is generally accepted that thermal insulation performance of rigid polyure-
thane foam arises from three heat transfer controlling phenomena: polymer
conduction, gas conduction, and radiation. The overall thermal conductivity,
indicated as l or K-factor, is determined by the sum of these contributions, as
well as by the density of the foam and the material distribution over struts

© 2004 by CRC Press LLC

www.ketabdownload.com
and windows.5,136–139 The example136 in Figure 7.21 shows that, typically, the
K-factor versus density profile exhibits a minimum at about 30–50 kg/m3,
which can be explained as the result of a balance between radiation and
polymer conduction. At constant cell size, radiation is particularly important
at very low densities where the number of radiation scattering cell windows
per unit volume is smallest. An increase in density results in an increasing
number of cell windows per unit volume, hence the radiation contribution
decreases as does the K-factor. At higher densities, this effect is counteracted
by an obvious increase in polymer conduction, which explains the upswing
of the K-factor profile. Similarly, at constant density and cell gas composition,
the radiation transfer contribution to thermal conductivity decreases with
decreasing cell size,139,140 down to values below 1 mW/mK for cells smaller
than 100 mm.141
Figure 7.21 also suggests that, at 30 kg/m3 density, the radiation heat trans-
fer contributes about 20% to the K-factor. This number should be compared
with quotes of several older studies which range from 15 to 20%137 to 30%139,140
for CFC-11 foams, whereas estimates of 10–20% and 6% have been made in
more recent work for 141B142 and air-filled open celled foams,143 respectively.
Values reported for the contribution of polymer conduction also range
between 17 and 30%.136, 139,140, 142,143 Table 7.4 gives a summary of preferred
mean values. This table clearly shows that, by far, the most important K-factor
contribution arises from gas conduction, i.e., roughly 50–70%, depending on
the type of foam. Optimization of blowing agents is thus an obvious choice
to improve product performance, and it is not surprising that significant
efforts have been undertaken in recent years to develop the CFC-11 alterna-

K-factor (mW/M.K) Contribution (%)

23 100%
Radiation
22
80%
21 Polymer conduction
60%
20

19
40%

18
Gas conduction
20%
17

16 0%
0 10 20 30 40 50 60 70 80 90 100 10 20 30 40 50 60 70 80 90 100
Foam density (kg/m3) Foam density (kg/m3)

FIGURE 7.21
Contributions to foam K-factor as a function of foam density. (Reprinted with permission from
Grünbauer, H.J.M., Rigid polyurethane foams, in Polymeric Materials Encyclopedia, vol. 10, ed.
J.C. Salamone, CRC Press, Boca Raton, FL, 1996, 7504–7512.)

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 7.4
Contributions to Rigid Polyurethane Foam K-factor
Weight
Parameter Influencing Factor (%)
Radiative heat transfer—Radiation Opacity of the foam—Cell structure 13
through the foam
Solid heat transfer—Conduction Thermal conductivity of the polymer 20
through the polymer matrix
Gaseous heat transfer through gas Gas K-factor of the blowing agent 67
mixture contained in cells mixture (conductive)

tives of Table 7.1. Once the blowing agent (combination) is fixed, further K-
factor improvements could be made by reduction of either the polymer con-
duction or the radiative heat transfer component.
Heat transfer from conduction through the solid matrix is determined by
the chemical structure of the polymer, by the density of the foam and the
distribution of mass between the struts and windows at the cellular level of
the foam.136,139,140 The minimum foam density permitted is limited by the
long-term dimensional stability of the foam; see Section 7.5.2.2. For this
reason, reduction of radiative heat transfer is the most attractive option for
K-factor improvement. In polymeric foam, the radiative contribution to heat
transfer is influenced by the ability of the polymer to absorb radiative
energy.139,140 Polyurethane foams can be thought of as an aggregate of opaque
struts and thin transparent windows. Reducing the cell size is equivalent to
increasing the number of opaque barriers.139,140 An alternative approach to
reduce the radiative thermal conductivity is to increase the ability of the
polymeric material to interact with the radiative energy modification of the
polymeric structure or by addition of additive that would absorb, reflect, or
scatter the radiative energy.144

7.5.1.2 Condensation

Inspection of Table 7.2 shows that several currently used blowing agents are
in the liquid state at ambient conditions, with boiling points ranging from
15˚C for HFC-245 fa to 50˚C for cyclopentane. Accordingly, their vapor pres-
sures are below 1 bar under ambient conditions with a minimum value for
cyclopentane of 0.34 bar at 20˚C. Classical thermodynamics dictates that con-
densation of a vapor always occurs at pressures equal to or higher than its
vapor pressure at a given temperature.71,136 This implies that, at 20˚C, conden-
sation of, for example, cyclopentane will occur when the gas phase of a closed
celled foam contains cyclopentane at a mole fraction of 0.34 or higher.145,146 A
similar reasoning holds for other blowing agents, such as CFC-11136 and HCFC-
141b.147 Condensation will also take place at lower mole fractions when the
temperature is decreased until all blowing agent is eventually condensed. This

© 2004 by CRC Press LLC

www.ketabdownload.com
phenomenon is of significant practical importance for two main reasons. First
of all, the thermal conductivity of the foam is usually affected appreciably as
a result of several partially balanced effects. Liquified blowing agent no longer
contributes to gas phase thermal conductivity as the gas phase will be enriched
in gases with higher thermal conductivity, such as N2 , O2 from ingressed air,
and (residual) CO2. A second reinforcing effect could be due to the thermal
conductivity contribution of liquified blowing agent which is higher than that
of gaseous blowing agent. Third, the overall reduction of cell gas pressure may
have an opposite influence which is, however, small at the cell gas pressures
considered. As a result, overall thermal conductivity of the foam is usually
increased significantly.
Reduction of total cell gas pressure is also at the origin of the second
practical impact of blowing agent condensation. Due to the high porosity of
95–97% of rigid polyurethane foams, their dimensional stability can be char-
acterized as a subtle balance between polymer strength and cell gas pres-
sure.136,145,146 Partial loss of cell gas pressure may hence cause significant
problems, especially at low service temperatures in freezers and refrigerators
or when construction foams are exposed to cold climates. Both effects are
nicely illustrated by Figure 7.22, which shows gas thermal conductivity and
cell gas pressure of a free-rise foam containing 0.09 mol of total blowing
agent per 100 g of formulation.146 The total cell gas pressure is about 0.7 bar
due to the cooldown of the foam from its maximum exotherm to ambient
temperature.136,146 At 20˚C, this pressure is maintained at cyclopentane mole
fractions below 0.5. Above this limiting concentration, cyclopentane starts
to condense and the total cell gas pressure decreases with increasing cyclo-
pentane content until, at 100% cyclopentane, the vapor pressure (0.34 bar)
of cyclopentane at 20˚C is reached.

Gas thermal conductivity Total pressure,

@10°C, mW/(mk) 105 Pa
18 0.8
17 0.7
16 20°C 0.6
10°C
15 0.5
14 0.4
13 condensation 0.3
12 0.2
no condensation
11 0.1
10 0.0
0.0 0.2 0.4 0.6 0.8 1.0
Mole fraction of cyclopentane in the formulation
FIGURE 7.22
Gas thermal conductivity and cell gas pressure as a function of blowing agent composition in
a CO2/cyclopentane blown free-rise foam. (Reprinted with permission of Sage Publications
Ltd., from Fleurent, H. and Thijs, S., J. Cell. Plast, 31, 580, 1995.)

© 2004 by CRC Press LLC

www.ketabdownload.com
 Figure 7.22 also shows this effect at 10˚C, together with the concomitant
increase in K-factor. At this lower temperature, condensation starts earlier
due to the lower value (0.22 bar) of the vapor pressure of cyclopentane.
Needless to say, the strength of the polymer matrix should be designed to
counteract this pressure drop in order to maintain dimensional stability,
especially in refrigerators and freezers. Alternatively, low boiling blowing
agents are frequently considered in these cases as they allow for the use of
more cost-effective lower foam densities while maintaining K-factor perfor-
mance due to the absence of condensation effects. Not surprisingly, Table 7.1
shows several examples, which are currently in use.
Foam dimensional stability is further affected by penetration of blowing
agent into the foam polymer matrix, especially in the longer term as well as
by the exchange of cell gases with external gases which takes place during
the foam aging process, which is discussed below. The matrix “solubility”
effect has been extensively investigated148–150 for cyclopentane, pentane iso-
mers, HCFC-141b, and CFC-11. In all cases, matrix loadings up to 10% by
weight have been found, particularly in foams that have been in use for a
prolonged period of time, i.e., for 10 years or longer.149

7.5.2 Aging
7.5.2.1 Gas Diffusion
The most important aging phenomenon observed with rigid foams is
undoubtedly the loss of thermal insulation performance, and sometimes
dimensional stability, due to the exchange of blowing agent and atmospheric
gases between the foam core and its environment. The thermodynamic force
driving this exchange is the difference in concentration across the foam
boundary of the gases involved. This process can be approximated as a
classical sorption/desorption process that can be quantified in terms of effec-
tive diffusion coefficients, indicated by Deff, by applying Fick’s law, once the
concentrations of the various gases in the foam are known as a function of
time.151 In practice, concentrations are obtained, for example, via internal gas
pressure recording,152 weight loss measurement,153,154 and GC analysis of
gases evolved from vacuum exposed foam surfaces155–157 or from crushed
cylinders, sampled on various location of a foam slab.158–163
Essential in this treatment is that a foam slab is considered as a continuum.
This approach has the advantage that details of the foam structure, such as
open cell content and material distribution, need not be known. Also, Deff
data from relatively fast aging thin foam slices can be scaled up to predict
long-term aging performance of real live slabs of greater thickness,151 an
approach which is frequently referred to as the “slicing method.”162,164 Its
major disadvantage is its dependence on local material properties at the
sampling point of forthcoming Deff values. This needs to be taken into account
when computing predictions of actual long-term aging behavior.163 In addi-
tion, variations among Deff data derived from different foam types may be
expected due to their dependence on the actual foam sample considered.

© 2004 by CRC Press LLC

www.ketabdownload.com
 TABLE 7.5
Effective Diffusion Coefficients of Various Blowing Agents and
Atmospheric Gases
Blowing Agent or Gas Deff ¥ 1013 (m2/s) Ref.
CFC-11 2.7 80, 81, 84, 87
HCFC-141b 1.5 84, 87
HCFC 22 32 79, 87
Isomeric pentanes 2.5 86
N2 57 80, 81, 83, 86, 87
O2 647 80, 81, 83, 86, 87
CO2 1540 79, 83, 84, 86, 87, 90

Although not negligible, these variations are nevertheless smaller than those
between blowing agents, so that meaningful conclusions can be drawn about
diffusivity trends for various atmospheric gases and blowing agents from the
Deff compilation given by Table 7.5. This table gives average Deff values
obtained under ambient conditions by various authors. Despite the spread
in the original data, the conclusion seems warranted that Deff values of phys-
ical blowing agents are of the order of 10–13 m2/s, except perhaps for HCFC-
22, which is however gaseous at ambient conditions. Deff of air, 80% N2 + 20%
O2 , is of the order of 10–11 m2/s, whereas Deff of CO2 ranks highest with an
average Deff value of the order of 10–10 m2/s. The relative ranking of Table 7.5
is clearly manifest in aging studies of unfaced foam slabs where CO2 diffusion
is completed within days, air exchange takes from several months to years,
and full exchange of the blowing agent itself is difficult to measure during
the service life of the foam.165,166 This is illustrated in Figure 7.23 which shows
the time dependence of partial pressures of N2, O2, CO2 and cyclopentane in
unfaced aging foam.

60
nitrogen
Partial pressure (kPa)

40

oxygen
20
c-pentane

carbon dioxide
0
0 20 40 60
Time (days)
FIGURE 7.23
Time dependence of partial pressures for a cyclopentane blown polyurethane foam (30 mm slab
= 45 kg m–3). (Reprinted with permission from Svanström, M. et al., Cell. Polym., 16, 182, 1997.)

© 2004 by CRC Press LLC

www.ketabdownload.com
 The outward diffusion of blowing agent being very slow, several
approaches to predict aging from so-called accelerated aging tests have been
proposed. One approach, the slicing method, has been mentioned already. It
has the advantage of simplicity, speed, and ease of scaling,162 but it is disad-
vantaged by the need to cut thin slices. This introduces a layer of damaged
cells at the surface, which is difficult to quantify. An alternative approach is
the so-called ACERMI method,167,168 which comprises acceleration of gas
exchange by increasing the temperature of the foam. Despite its wide accep-
tance, this method has its shortcomings as well. Detailed examination of the
temperature dependence of the diffusion of various gases and blowing agents
has clearly demonstrated significant variations among thermally induced
acceleration factors of various gases and blowing agents.169,170 As a result, the
balance between the different diffusion processes is shifted and predicted
aged properties may be, and are actually found to be,171 in error.

7.5.2.2 Dimensional Stability

Dimensional stability is almost as important as K-factor performance to the

success or failure of a given foam formulation. It is frequently measured by
the compressive strength at 10% compression. This is a reliable means of
assessing dimensional stability because it is related to the pressure inside
the foam cells and its evolution with time and temperature, as well as the
strength of cell walls of the foam. It is influenced by several factors, such as:

1. The polymer formulation, which can give higher or lower strength,

and the foam density.
2. The boiling point of the blowing agent. A blowing agent with a lower
boiling point will give a higher initial cell pressure and therefore a
higher compressive strength and will also tend to condense less at
lower temperatures (which occur during the operation of a refriger-
ator or in a cold climate) than a product with higher boiling point.
3. The flow properties of the foam. A foam with good flowability will
have a better density and compressive strength distribution through-
out the cabinet and a more homogeneous cell structure, i.e., fewer
elongated cells.
4. The ratio between the rate of outward CO2 diffusion and inward air
diffusion.
5. The penetration of physical blowing agent into cell walls.

Except for the first, all factors are mediated by the choice of the blowing
agent and the exchange of gas and blowing agent between the foam interior
and the outside world. It follows that it is hardly possible to overemphasize
the importance of these phenomena for the performance of polyurethane
insulation foams in industrial practice. Assuming typical lifetimes of 20 to
50 years for construction materials and 15 years for refrigerators, it is thus

© 2004 by CRC Press LLC

www.ketabdownload.com
of obvious importance to assess the impact of diffusion processes on long-
term thermal conductivity and stability after the short-term depletion of CO2
has been completed. In that respect, real life studies over periods of 30 years
for CFC-11 and up to 9 years for HCFC-141b and pentane are rather reas-
suring.172 In full agreement with expectations based on Deff of the type dis-
played by Table 7.5, K-factor aging profiles of these foams show a tendency
to flatten after an initial upswing with plateau values ranging from 25 to 30
mW/mK, depending on the blowing agent, sample thickness, and presence
or absence of a facer material.172
The position of CO2 as a blowing agent deserves some special attention in
this context. In principle, CO2 comprises a very attractive option for
expansion of insulation foams due to its moderate gas K-factor, its low global
warming potential versus physical blowing agents, and the convenience of
the reactive route given by Equation 7.2a or one of its counterparts. Apart
from the relatively high cost of generating CO2, ultimately via phosgenation
of amines, its high diffusivity is the only drawback preventing its widespread
use as a blowing agent. Several attempts have been reported in the literature
to counteract or circumvent this unfortunate feature. The most obvious route
is to increase the foam density in order to balance the pressure drop that
occurs within weeks after a fully CO2 blown foam is manufactured.136 In that
case, the plateau K-factor will be around 32 mW/mK,173 i.e., the value for
air. Lower plateau values can be attained by applying suitable diffusion
barriers,174 which are capable of retaining a significant portion of the initial
CO2 concentration over a prolonged period of time (or forever in the case of
metal facings). Other alternatives that have been considered are reduction
of CO2 diffusion by means of optimizing the urethane-urea copolymer
morphology,120,121 polyurethane polymer structure,175 via relief of the density
requirement by using open cells80,81,176 or via improved crosslinking of the
polyurethane resin involved.177–179

7.6 Conclusions
Inherent to the very nature of rigid polyurethane foam formation is the
requirement of a proper balance between properties and actions of all
ingredients present in the starting formulation. This review has clearly
shown that several years of intense research have greatly enhanced our
understanding of the fundamentally complex science of foam formation
and physical property development. As a result, today’s polyurethane
industry is much better positioned to implement product improvements
required by customers and/or regulatory bodies than 15 years ago when
the first attempts to replace CFC-11 technology were necessarily guided
by “Edisonian” trial and error approaches. In hindsight, one can only
admire the historical development of CFC-11-based polyurethane insula-
tion foams and their early non-CFC-based counterparts “from scratch!”

© 2004 by CRC Press LLC

www.ketabdownload.com
Meanwhile, rigid polyurethane foam technology has now matured to an
environmentally responsible position.
Contrasting with the implied optimism of this statement however is the
general feeling that responding to ever more stringent upcoming environ-
mental regulations on, e.g., flammability and energy consumption remains
a challenge to the industry unless dramatic performance improvements can
be realized. It is, for example, unlikely that current developments in blowing
agent technology will give rise to quantum-leap improvements of, say, 30%
or more in thermal conductivity. To achieve these levels of performance,
entirely new product and processing concepts need to be invented and
implemented. A good example in this context is the use of open-celled
foam79–81 in combination with vacuum insulation179 for refrigerators. It is
from this type of innovative thinking that new concepts will emerge to serve
tomorrow’s customer and environmental needs. Rigid polyurethane foam
technology is sufficiently broad and versatile to face this challenge as posi-
tively as it is currently responding to the challenge of implementing zero
ODP, low GWP technology on a global basis.

Acknowledgments
The authors would like to thank the following colleagues: Professor Anthony
J. Ryan, University of Sheffield, U.K., and Dr. Puteri S. Megat Yusoff, Uni-
versity of Technology, Kuala Lumpur, Malaysia. They are also grateful to
The Dow Chemical Company for granting permission to publish this chapter.

References
1. Oertel, G., ed., Polyurethanes Handbook, 1st ed., Carl Hanser Verlag, Munich, 1985.
2. Woods, G., The ICI Polyurethanes Book, John Wiley & Sons, New York, 1987.
3. Gum, W.F., Riese, W. and Ulrich, H. eds., Reaction Polymers Chemistry: Technol-
ogy, Applications, Markets, 1st ed., Carl Hanser Verlag, Munich, 1992.
4. Grünbauer, H.J.M., Rigid polyurethane foams, in Polymeric Materials Encyclo-
paedia, vol. 10, ed. J.C. Salamone, CRC Press, Boca Raton, FL, 1996, 7504–7512.
5. Smits, G.F. and Thoen, J.A., J. Therm. Insul., 14, 81, 1990.
6. Lewis, K.M., Kijak, I. and Reuter, K.B., J. Cell. Plast., 32, 235, 1996.
7. Britton, D.J., Cell. Polym., 8, 125, 1989.
8. Dämmstoffe im Hochbau (Insulation Materials in Buildings), Information for
Builders, Architects and Engineers, Wirschaftsministerium Baden-Würtenberg,
Stuttgart, 1999.
9. Steildachdämmung mit Polyurethan-Hartschaumstoff (Pitched Roof Insulation
with Rigid Polyurethane Foam), IVPU, Stuttgart, 1995.
10. European Thermal Insulation Markets: The Current and Future Prospects for
Polyurethane, UTECH 2000 (Urethanes Technology), Crain Communications
Ltd, London, 2000.

© 2004 by CRC Press LLC

www.ketabdownload.com
 11. Rowland, F.S., Annu. Rev. Phys. Chem., 42, 731, 1991.
12. Tominaga, T., Pure Appl. Chem., 64, 529, 1992.
13. Graber, M., Mutisya, G., and Mulumba, M.A., eds., Production and Consumption
of Ozone Depleting Substances under the Montreal Protocol 1986–2000, UNEP,
Nairobi, 2002.
14. Skochdopole, R.E., Chem. Eng. Prog., 57, 55, 1961.
15. Collishaw, P.G. and Evans J.R.G., J. Mat. Sci., 29, 486, 1994.
16. Glicksman, L.R. and Torpey, M., Polyurethanes 87, Proceedings of the SPI, 30th
Annual Conference, 80, 1987.
17. King, J.A., Latham, D.D., and Martin, C.A., Polyurethane Expo 2001, Proceed-
ings of the Alliance for the Polyurethane Industry, 473, 2001.
18. Barton, A.F.M., Handbook of Solubility Parameters and Other Cohesion Parameters,
CRC Press, Boca Raton, FL, 1983.
19. Koschade, R., Die Sandwichbauweise (How to Build a Steel Sandwich), Ernst &
Sohn, Berlin, 2000.
20. Troitzsch, J., International Plastics Flammability Handbook, 2nd ed., Hanser Verlag,
Munich, 1990.
21. SPE 13823 (en): Reaction to fire tests for building products - Building products
excluding floorings exposed to the thermal attack by a single burning item,
ICS 13.220.50, Nederlands Normalisatie Instituut (NEN), Delft, 2000.
22. Ward, E.R., Chem. Britain, 15, 297, 1979.
23. Ingold, C.K., Structure and Mechanism in Organic Chemistry, 2nd ed., Cornell
University Press, Ithaca, NY, 1969.
24. Haber, F., Z. Elektrochem., 22, 506, 1898.
25. Janssen, H.J., Kruithof, A.J., Steghuis, G.J., and Westerterp, K.R., Ind. Eng. Chem.
Res., 29, 754, 1990.
26. Wagner, E.C., J. Org. Chem., 19, 1862, 1954.
27. Richter, R.H. and Priester, Jr., R.D., Isocyanates, organic, in Kirk-Othmer Ency-
clopedia of Chemical Technology, vol. 14, 4th ed., Wiley-Interscience, New York,
1995, 902-934.
28. Iben, H., Kattanek, S., Backhauss, L., Welschinger, B., Gansera, B., Loeschau,
S., and Meyer, A., French Patent 2,325,637, 1997.
29. Zaby, G., Judat, H., Boonstra, R., De Vos, S., Eckermann, R.W., and Humburger,
S., German Patent 3,744,001, 1998.
30. Zaby, G., Judat, H., Humburger, S., De Vos, S., and Eckermann, R.W., German
Patent 3,736,988, 1989.
31. Ohlinger, R., Schnez, H., Pfannenstiehl, L., Blumberg, B., and Raabe, H.J., U.S.
Patent 4,581,174, 1986.
32. Richter, R. and Ulrich, H., Tetrahedron Lett., 22, 1875, 1974.
33. Thiele, L. and Becker, R., Adv. Urethane Sci. Technol., 12, 59, 1993.
34. Kanner, B. and Decker, T.G., J. Cell. Plast., 5, 32, 1969.
35. Vincent, B., in Surfactants, ed. T.F. Tadros, Academic Press, London, 1984.
36. Grüning, B. and Koerner, G., Tenside Surf. Det., 26, 313, 1989.
37. Hostettler, F., German patent 1,091,324, 1960.
38. Sandridge, R.L., Morecroft, A.S., Hardy, E.E., and Saunders, J.H., J. Chem. Eng.
Data, 5, 495, 1960.
39. Hill, R.M., ed., Silicone Surfactants, Surfactant Science Series, vol. 86, Marcel
Dekker, New York, 1999.
40. Owen, M.J., in Surfactants in Solution, vol. 6, ed. K.L. Mittal and P. Bothorel,
Plenum Press, New York, 1986.

© 2004 by CRC Press LLC

www.ketabdownload.com
 41. Owen, M.J. and Kendrick, T.C., Macromolecules, 3, 458, 1970.
42. Kendrick, T.C. et al., J. Colloid Interface Sci., 24, 135, 1967.
43. Kendrick, T.C. and Owen, M.J., Chem. Phys. Appl. Praat. Surface (Spain) 5th ed.,
2, 57, 1968.
44. Schmidt, G., Tenside Surf. Det., 27, 324, 1990.
45. Haluska, L.A., U.S. patent 2,868,824, 1956.
46. Boudreau, R.J., Mod. Plast., 44, 133, 1967.
47. Lee, L.J., Ottino, J.M., Ranz, W.E., and Macosko, C.W., Polym. Eng. Sci., 20, 868,
1980.
48. Kolodziej, P., Macosko, C.W., and Ranz, W.E., Polym. Eng. Sci., 22, 388, 1982.
49. Kolodziej, P., Yang, W.P., Macosko, C.W., and Welinghoff, S.T., J. Polymer Sci.
Part B: Polym. Phys., 24, 2359, 1986.
50. Molnar, J.A. Jr. and Lee, L.J., J. Appl. Polym. Sci., 37, 2295, 1989.
51. Nelson, M. and Lee, L.J., J. Appl. Polym. Sci., 37, 251, 1989.
52. Manas-Zloczower, I. and Macosko, C.W., Polym. Eng. Sci., 28, 1219, 1988.
53. Richter, E.B. and Macosko, C.W., Polym. Eng. Sci., 18, 1012, 1978.
54. Tucker, C.L. and Suh, N.P., Polym. Eng. Sci., 20, 875, 1980.
55. Baser, S.A., Shetty, D.G., and Khakhar, D.V., Polym. Eng. Sci., 33, 1611, 1993.
56. Fields, S.D., Thomas, E.L., and Ottino, J.M., Polymer, 27, 1423, 1986.
57. Wickert, P.D., Macosko, C.W., and Ranz, W.E., Polymer, 28, 1105, 1987.
58. Machuga, S.C., Midje, H.L., Peanasky, J.S., Macosko, C.W., and Ranz, W.E.,
AIChE. J., 34, 1057, 1988.
59. Rojas, A.J., Marciano, J.H., and Williams, R.J., Polym. Eng. Sci., 22, 840, 1982.
60. Baser, S.A. and Khakhar, D.V., Polym. Eng. Sci., 34, 632, 1994.
61. Baser, S.A. and Khakhar, D.V., Polym. Eng. Sci., 34, 642, 1994.
62. Basile, G., Musso, E., Tesser, R., and Di Serio, M., Cell. Polym., 13, 98, 1994.
63. Williams, D.J., Bogdan, M.C., Parker, R.C., and Knopeck, G.M., J. Cell. Plast.,
33, 238, 1997.
64. Singh, S.N., Burns, S.B., and Costa, J.S., Cell. Polym., 16, 444, 1997.
65. Biesmans, G., Colman, L., and Vandensande, R., J. Coll. Int. Sci., 140, 199, 1998.
66. Hager, S.L., Craig, T.A., Jorgenson, M.W., Artevia, L.D., and Macosko, C.W., J.
Cell. Plast., 30, 44, 1994.
67. Grünbauer, H.J.M., Thoen, J.A., Folmer, J.C.W., and van Lieshout, H.C., J. Cell.
Plast., 28, 36, 1992.
68. Han, J.H. and Han, C.D., J. Polym. Sci., 28, 743, 1990.
69. Grünbauer, H.J.M., Broos, J.A.F., Thoen, J.A., Smits, G.F., Lehnert, A.B., and
Hoenke, M.S., Polyurethanes 92, Proceedings of SPI 34th Annual Conference,
1990, 36, 1992.
70. Grünbauer, H.J.M., Cell. Polym., 14, 366, 1995.
71. Pitzer, K.S. and Brewer, L., Thermodynamics, McGraw-Hill, New York, 1961.
72. Volkert, O., Polyurethanes 91, Proceedings of SPI/ISOPA World Congress, 1991,
740, 1991.
73. Modesti, M., Adriani, V., and Simioni, F., Polym. Eng. Sci., 40, 2046, 2000.
74. Vaze, N. and Khakhar, D.V., in Macromolecules: Current Trends, vol.2, ed. S.
Venkatachalam, Allied Publishers, New Delhi, 1995; Niyogi, D., Kumar, R., and
Gandhi, K.S., AIChE J., 38, 1170, 1992; Niyogi, D., Kumar, R., and Gandhi, K.S.,
Polym. Eng. Sci., 39, 199, 1999.
75. Latham, D.D., J. Cell. Plast., 34, 124, 1998.
76. Smits, G.F. and Grünbauer, H.J.M., U.S. Patent 4,945,119, 1990.
77. Grünbauer, H.J.M. and Smits, G.F., U.S. Patent 4,972,003, 1990.

© 2004 by CRC Press LLC

www.ketabdownload.com
 78. van Stralen, S. and Cole, R., Boiling Phenomena, vol. 1, McGraw-Hill, New York,
1979.
79. Smits, G.F., Proceedings of UTECH 1994 Conference, The Hague, paper 19,
1994.
80. Smits, G.F. and Thoen, J.A., U.S. Patent 5,250,579, 1993.
81. Smits, G.F., Cini, G., Grünbauer, H.J.M., and Broos, J.A.F., U.S. Patent 5,721,284,
1998.
82. Owen, M.J., Kendrick, T.C., Kingston, B.M., and Lloyd, N.C., J. Coll. Int. Sci.,
24, 141, 1967.
83. Wurtz, A., C.R. Hebd. Seances Acad. SCI, 27, 241, 1848.
84. Bayer, O., Angew. Chemie, A59, 257, 1947.
85. Van Gheluwe, P. and Leroux, J., J. Appl. Polym. Sci., 28, 2053, 1983.
86. Hébrard, M.J. and Laroux, J., Anal. Instrum., 15, 201, 1986.
87. Lovering, E.G. and Laidler, K.J., Can. J. Chem., 40, 26, 1962.
88. Pinner, S.H., Plastics (London), 11, 206, 215, 1947.
89. Cooper, W., Pearson, R.W., and Darke, S., Ind. Chem., 3, 121, 1960.
90. Sayigh, A.A.R., Ulrich, H., and Farissey, W.J., Diisocyanates, in Condensation
Monomers, ed. J.K. Stille and T.W. Campbell, Wiley-Interscience, New York,
1972.
91. Richter, R. and Ulrich, H., Synthesis and preparative applications of isocyan-
ates, in The Chemistry of Cyanates and Their Thio Derivatives, Part 2, ed. S. Patai,
John Wiley & Sons, New York, 1977.
92. Dillon, J.G., Infrared Spectroscopic Atlas of Polyurethanes, Technomic, Lancaster,
PA, 1989.
93. Jennings, R., J. Cell. Plast., 5, 159, 1969.
94. Rowton, R.L., J. Cell. Plast., 16, 287, 1980.
95. van Thuyne, A. and Zeegers, B., J. Cell. Plast., 14, 150, 1978.
96. Jianqiu, Y., Jianyuan, Z., Dening, W., Chunpu, H., Shengkang, Y., Yiou, C., Yufu,
C., Ziqian, X., Jin, S., and Yin, W.J., J. Cell. Plast., 26, 39, 1990.
97. Artevia, L.D., Macosko, C.W., Priester, R.D., Jr., Schrock, A.K., and Turner, R.B.,
Proceedings of SPI/ISOPA World Congress 1991, 509, 1991.
98. Baser, S.A. and Khakhar, D.V., J. Cell. Plast., 29, 280, 1993.
99. Mora, E., Artavia, L.D., and Macosko, C.W., J. Rheol., 35, 921, 1991.
100. Carrierre, C.J., Bank, D.H., and Christenson, C.P., Polym. Eng. Sci., 32, 426, 1992.
101. Carrierre, C.J., Bank, D.H., and Christenson, C.P., Mater. Eng., 9, 743, 1995.
102. Lipshitz, S.D. and Macosko, C.W., Polym. Eng. Sci., 16, 803, 1976.
103. Bessette, M.D. and Sundstrom, D.W., Polym. Process Eng., 3, 25, 1985.
104. Nabata, Y., Mamada, A., and Yamasaki, H., J. Appl. Polym. Sc., 35, 155, 1988.
105. Kostrzewski, W., J. Cell. Plast., 21, 424, 1985.
106. Kostrzewski, W. and Lindt, J.T., J. Polym. Eng., 16, 187, 1986.
107. Corfield, M., UTECH Asia ‘ 99 Conference Book of Papers, science and testing
paper1/11, 1999.
108. Jones, S.A., Scott, K.W., Willoughby, B.G., and Sheard, E.A. J. Cell. Plast., 38,
285, 2002.
109. Webb, D.D., Proceedings of SPI 28th Annual Conference, 2, 1984.
110. Matusik, F.J. and Nelson, D.W., U.S. patent 4,488,427, 1984.
111. Fitzgerald, J.V., Matusik, F.J., Nelson, D.W., and Schrag, J.L., U.S. Patent
4,754,640, 1988.
112. McClusky, J.V., O’Neill, R.E., Priester, R.D., Jr., and Ramsey, W.A., J. Cell. Plast.,
30, 224, 1994.

© 2004 by CRC Press LLC

www.ketabdownload.com
113. Zhang, X.D., Giles, D.W., Barocas, V.H., Yasunaga, K., and Macosko, C.W., J.
Rheol., 42, 871, 1998.
114. Ferry, J.D., Viscoelastic Properties of Polymers, John Wiley & Sons, New York, 1980.
115. Marciano, J.H., Rojas, A, and Williams, R.J.J., Polymer, 23, 1489, 1982.
116. McClusky, J.V., Priester, R.D., Jr., O'Neill, R.E., Turner, R., Harthcock, M.A., and
Davis, B.L., J. Cell. Plast., 26, 346, 1990.
117. Peng, S., Jackson, P., Sendijarevic, V., Frisch, K.C., Prentice, G.A., and Fuchs,
A., J. Appl. Polym. Sci., 77, 374, 2000.
118. Grünbauer, H.J.M. and Thoen, J.A., Proceedings of SPI 33rd Annual Technical
and Marketing Conference, 318, 1990.
119. Grünbauer, H.J.M., Folmer, J.C.W., van Lieshout, H.C., Lidy, W., and Thoen,
J.A. Polym. Prep., 32, 517, 1991.
120. Grünbauer, H.J.M. and Folmer, J.C.W., J. Appl. Polym. Sci., 54, 935, 1994.
121. Megat-Yusoff, P.S., Ph.D. thesis, Victoria University of Manchester, 1996.
122. Turner, R.B., Spell, H.L., and Wilkes, G.L., Proceedings of SPI 28th Annual
Technical and Marketing Conference, 244, 1988.
123. Armistead, J. P., Turner, R.B., and Wilkes, G.L., J. Appl. Polym. Sci., 35, 601, 1988.
124. Creswick, M.L., Lee, K.D., Turner, R.B., and Huber, L.M., Proceedings of SPI
28th Annual Technical and Marketing Conference, 11, 1985.
125. Artavia, L.D., Ph.D. thesis, University of Minnesota, 1991.
126. Mora, E., M.Sc. dissertation, University of Minnesota, 1991.
127. Elwell, M.J.A., Ph.D. thesis, Victoria University of Manchester, 1993.
128. Neff, R., Ph.D. thesis, University of Minnesota, 1995.
129. Li, W., Ph.D. thesis, University of Sheffield, 2002.
130. Wilkes, G.L., Abouzhar, S., and Radovich, D., J. Cell. Plast., 19, 248, 1983.
131. Avnir, D., ed., The Fractal Approach to Heterogeneous Chemistry, John Wiley &
Sons, New York, 1989.
132. Schaefer, D.W., Bunker, B.C., and Wilcoxon, J.P., Proc. R. Soc. Lond. A., 423, 35,
1989.
133. Feder, J., Fractals, Plenum Press, New York, 1989.
134. Harrison, A., Fractals in Chemistry, Oxford Science Publications, Oxford, 1995.
135. Kaiser, T., Polym. Sci., 373, 14, 1989.
136. Smits, G.F. and Thoen, J.A., J. Cell. Plast., 29, 57, 1993.
137. Buist, J.M., Doherty, D.J., and Hurd, R., Urethane rigid foams: factors affecting
their behaviour as thermal insulants, in Progress in Refrigerator Science and
Technology, Pergamon Press, Oxford, 1963.
138. Schuetz, M.A. and Glicksman, L.R., J. Cell. Plast., 20, 114, 1984.
139. Glicksman, L.R., Cell. Polym., 10, 276, 1991.
140. Cunningham, A., Jeffs, G.M.F., Rosbotham, I.D., and Sparrow, D.J., Cell. Polym.,
7, 1, 1988.
141. Eeckhaut, G. and Cunningham, A., J. Cell. Plast., 32, 528, 1996.
142. Chung-jen Tseng, Yamaguchi, M., and Ohmori, T., Cryogenics, 37, 305, 1997.
143. Jhy-Wen Wu, Wen-Fa Sung, and Hsin-Sen Chu, Int. J. Heat Mass Transfer, 42,
2211, 1999.
144. Gluck, D.G., Soukup, T.G., and Moore, W.J., U.S. patent 4,795,763, 1989.
145. Fleurent, H., Cell. Polym., 13, 419, 1994.
146. Fleurent, H. and Thijs, S., J. Cell. Plast., 31, 580, 1995.
147. Yourd, R.A., J. Cell. Plast., 32, 601, 1996.
148. Daems, D., Rosbotham, D., Singh, S.N., and Franco, M.V., J. Cell. Plast., 32, 485,
1996.

© 2004 by CRC Press LLC

www.ketabdownload.com
149. Svanström, M. and Ramnäs, O., J. Cell. Plast., 32, 159, 1996.
150. Singh, S.N., Biesmans, G., Karramans, M., and Randall, D., J. Cell. Plast., 34, 75,
1998.
151. Crank, J., The Mathematics of Diffusion, 2nd ed., Oxford University Press, Oxford,
1993.
152. Cuddihy, E.F., J. Cell. Plast., 12, 114, 1976.
153. Cuddihy, E.F. and Moacanin, J., J. Cell. Plast., 3, 73, 1967.
154. Bhattacharjee, D. and Booth, J.R., J. Cell. Plast., 31, 244, 1995.
155. Norton, F.J., J. Cell. Plast., 3, 23, 1967.
156. Brandreth, D.A. and Ingersoll, H.G., J. Cell. Plast., 16, 134, 1980.
157. Page, M.C. and Glicksman, L.R., J. Cell. Plast., 28, 268, 1992.
158. du Cauze de Nazelle, G.M.R., Ph.D. thesis, Delft University of Technology, the
Netherlands, 1995.
159. Cunningham, A., Rosbotham, I.D., and Sparrow, D.J., Cell. Polym., 9, 367, 1990.
160. Svanström, M. and Ramnäs, O., J. Cell. Plast., 31, 375, 1995.
161. Svanström, M., Ramnäs, O., Olsson, M., and Jarfelt, U., Cell. Polym., 16, 182,
1997.
162. Bart, G.C.J. and du Cauze de Nazelle, G.M.R., J. Cell. Plast., 29, 29, 1993.
163. Boscoletto, A.B. and Cellarosi, B., Cell. Polym., 10, 263, 1991.
164. Isberg, J., Ph.D. thesis, Chalmers University of Technology, Götenburg, Sweden,
1988.
165. Ball, G.W., Healy, W.G., and Partington, J.B., Eur. J. Cell. Plast., 1, 50, 1978.
166. Ball, G.W., Cell. Polym., 13, 434, 1994.
167. Roux, G., Tramblay, G., and Venuti, G., Cell. Polym., 9, 278, 1990.
168. Ball, G.W., Cell. Polym., 16, 110, 1997.
169. Biesmans, G., De Vos, R., and Rosbotham, I.D., Cell. Polym., 13, 292, 1994.
170. Biesmans, G., Karremans, M., Randall, D., and Singh, S.N., J. Cell. Plast., 34,
349, 1998.
171. Albrecht, W. and Zehendner, H., Cell. Polym., 16, 35, 1997.
172. Albrecht, W., Cell. Polym., 19, 319, 2000.
173. IVPU Joint Project, Cell. Polym., 10, 470, 1991.
174. Grünbauer, H.J.M., Cikut, L., Haworth, G.J., and Beerwart, A., J. Cell. Plast., 33,
140, 1997.
175. Kaplan, W.A. and Tabor, R.L., J. Cell. Plast., 30, 243, 1994.
176. Christman, D.L. and Reichel, C.J., J. Cell. Plast., 32, 82, 1996.
177. Moore, S.E., J. Cell. Plast., 30, 494, 1994.
178. Nichols, J.B., Bhattacharjee, D., Moreno, O., and Mirasol, S., J., Cell. Plast., 32,
139, 1996.
179. Kucukpinar, E. Proceedings of UTECH 2003 Conference, The Hague, 2003.

© 2004 by CRC Press LLC

www.ketabdownload.com