Polymer Characterization

Polymer
Characterization
Physical techniques

Second Edition

D. Campbell
School of Engineering
University of Northumbria, UK

R.A. Pethrick
Department of Pure and Applied Chemistry
University of Strathclyde, UK

and

J.R. White
Department of Mechanical, Materials and Manufacturing
Engineering
University of Newcastle-upon-Tyne, UK

Taylor & Francis

Taylor & Francis Group
Boca Raton London New York

A CRC title, part of the Taylor & Francis imprint, a member of the
Taylor & Francis Group, the academic division of T&F Informa pic.
Published in 2000 by
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742

© 2000 by D. Campbell, R.A. Pethrick and J. R. White

CRC Press is an imprint of Taylor & Francis Group
First edition published by Chapman & Hall 1989
H

No claim to original U.S. Government works

International Standard Book Number-10: 0-7487-4005-8 (Softcover)

International Standard Book Number-13: 978-0-7487-4005-5 (Softcover)
This book contains information obtained from authentic and highly regarded sources. Reprinted material is
quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts
have been made to publish reliable data and information, but the author and the publisher cannot assume
responsibility for the validity of all materials or for the consequences of their use.
No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic,
mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and
recording, or in any information storage or retrieval system, without written permission from the publishers.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only
for identification and explanation without intent to infringe.

Library of Congress Cataloging-in-Publication Data

Catalog record is available from the Library of Congress

informa Taylor & Francis Group

is the Academic Division of Informa pic.
Visit the Taylor & Francis Web site at
http ://[Link]

and the CRC Press Web site at

[Link]

Publisher’s Note
The publisher has gone to great lengths to ensure the quality of this book but
points out that some imperfections from the original may be apparent.
Contents

Preface ix

1 Introduction: Preliminary survey 1

1.1 The purpose of characterization 1
1.2 Molecular architecture 6
1.3 Amorphous polymers 7
1.4 Crystallizing polymers 8
1.5 Survey of characterization techniques 11
References 14
Further reading 14

2 Molecular mass determination 15

2.1 Introduction 15
2.2 Primary methods 19
2.3 Secondary methods 34
2.4 Other methods 48
2.5 Conclusions 49
Further reading 49
Exercises 50

3 Introduction to spectroscopic techniques 51

3.1 General remarks 51
3.2 Strategy in spectroscopic studies 52
3.3 Energy level calculations 53
3.4 Properties of electromagnetic radiation 53
3.5 Double beam optics 54
3.6 Fourier transform spectroscopy 55
References 57
vi Contents

4 Ultraviolet-visible spectroscopy 58
4.1 Introduction 58
4.2 Instrumentation 59
4.3 Theoretical estimation of electronic energy levels 61
4.4 Application to polymers 64
Reference 66

5 Vibrational spectroscopy: infrared and Raman

spectroscopy 67
5.1 Fundamentals of vibrational spectroscopy 67
5.2 Experimental techniques 75
5.3 Applications to polymers 85
5.4 Raman analysis 96
5.5 Near-infrared (MR) 106
Conclusions 106
References 107

6 Nuclear magnetic resonance spectroscopy 108

6.1 Introduction 108
6.2 Principles of magnetic resonance 109
6.3 Experimental techniques 124
6.4 Applications of NMR to polymers 133
6.5 NMR of polymers in the solid state 145
6.6 Two-dimensional NMR 150
6.7 NMR imaging 153
6.8 Conclusions 153
References 153
Exercises 154

7 Electron spin resonance spectroscopy 155

7.1 Introduction 155
7.2 Basic theory 156
7.3 Experimental considerations 168
7.4 Polymerization studies 174
7.5 Radiation effects in polymers 178
7.6 Mechanical degradation 186
7.7 Thermal degradation 188
7.8 Relaxation studies 188
7.9 ESR imaging 191
7.10 Conclusions 191
Further reading 192
Exercises 192
 Contents vii

8 X-ray diffraction 194

8.1 Introduction 194
8.2 Generation and properties of X-rays 194
8.3 Diffraction theory 195
8.4 Practical aspects 210
8.5 Applications to polymer characterization 217
8.6 Conclusions 235
References 235
Further reading 236
Exercises 236
9 Transmission electron microscopy and diffraction 237
9.1 Introduction 237
9.2 Layout of the TEM 238
9.3 Resolution limitation 242
9.4 Electron diffraction 244
9.5 Image contrast 248
9.6 Operating conditions and procedures for
beam-sensitive materials 262
9.7 Polymer studies in TEM 266
9.8 Multiple dark field imaging 287
9.9 Scanning transmission electron microscopy (STEM) 289
References 291
Further reading 291
Exercises 292
10 Scanning electron microscopy 293
10.1 Introduction 293
10.2 Design and operation of the SEM 293
10.3 Observation of polymers in the SEM 301
10.4 Examples of SEM applications with polymers 313
10.5 Conclusions 323
Further reading 325
Exercise 325
11 Light optical techniques 326
11.1 Introduction 326
11.2 Refractive index 326
11.3 Birefringence 328
11.4 Light microscopy 344
11.5 Small angle light scattering 357
Reference 360
Further reading 360
Advanced reading 360
Exercises 360
viii Contents

12 Thermal analysis 362

12.1 Introduction 362
12.2 Differential thermal analysis (DTA) and
differential scanning calorimetry (DSC) 362
12.3 Thermomechanical analysis 376
12.4 Thermogravimetric analysis 381
12.5 Dynamic mechanical thermal analysis 382
12.6 Dielectric thermal analysis (DETA) 399
References 407

13 Density, surface properties and neutron diffraction 408

13.1 Density 408
13.2 Surface analysis of polymeric materials 412
13.3 Chemical analysis of the polymer surface 416
13.4 Surface probe or scanning tip microscopy 427
13.5 Neutron scattering studies 432
References 435
Further reading 436
Exercise 436

14 Multiple characterization applications 438

14.1 The structure of a spherulite 438
14.2 Skin-core characterization in moulded polymers 440
14.3 Weathering of polymers 442
14.4 Particular studies 443
14.5 Conclusions 450
References 451

15 Theoretical predictions of the physical properties of

polymeric materials 452
15.1 Introduction 452
15.2 Types of approach available for modelling
polymer physical properties 452
15.3 Molecular based structural additivity calculations 453
15.4 Ab initio calculations of polymeric properties 455
References 464

Appendix: Solutions to exercises 465

Index 472
Preface

The spectacular growth in the number and variety of polymers, has

generated materials with a wide range of physical properties and appli-
cations. Being able to identify and characterize these materials is not just
of academic interest but also of commercial and environmental concern.
Many of the important structural and microstructural features of poly-
mers have no counterparts in other classes of materials. For instance,
molecular mass has little or no significance for metals or ceramics
whereas it has a major influence on the properties of polymers.
Furthermore a characterization method may need to be considerably
modified when it is applied to polymers. An example of this is trans-
mission electron microscopy, where the electron beam sensitivity of poly-
mers has necessitated the development of procedures that are
unnecessary with metals and ceramics. This book concentrates on those
techniques peculiar to polymers and on those aspects which are of partic-
ular value in polymer studies. In recent years there has been an
increasing interest in the characterization of surfaces which can have a
significant effect on the physical properties and technological applica-
tions of these materials.
The book is intended for use as the basis of a course offered to second
or third year undergraduate students in Materials Science, Polymer
Technology and Chemistry, and it is hoped that it will be of value to
many scientists and technologists working in the polymer industry.
Although the book is meant to be self-contained, a certain knowledge
of physics, chemistry and of polymeric materials is assumed. The
emphasis of this book is on illustrating applications of the methods and
the basic theory behind the various techniques is introduced only spar-
ingly. A bibliography is provided at the end of each chapter to help the
interested reader to find out more about both the background theory
and practical aspects of the techniques. The scope of this book has been
enlarged to include polymer composites whenever there are good exam-
ples of applications with these methods.
D. Campbell
R.A. Pethrick
J.R. White
Introduction:
Preliminary survey 1
1.1 THE PURPOSE OF CHARACTERIZATION
There are many reasons why it is necessary to characterize a polymeric
material. Forensic analysis, recycling and reclamation and failure
analysis of a component all require identification of the polymer type
although they may not be concerned with a precise measurement of all
of its physical properties. Design of polymers for specific applications
will often also require a knowledge of the physical properties of a
material and the techniques used may in some instances also be those
used for characterization.

1.1.1 IDENTIFICATION OF A POLYMER MATERIAL

The typical polymeric material is often not a single component and hence
simple chemical analysis will rarely provide all the information required.
For example, knowing that a component having a particularly poor
outdoor weathering performance is made of poly (vinyl chloride) is of
little use since this polymer, when formulated with a suitable ultravi-
olet stabilizer, can have excellent properties. Thus in determining its
stability to light it might be more pertinent to analyse the additive
content than the base polymer. A polymer formulation may contain
significant amounts of additives which can include dyes and pigments;
inorganic or organic fillers which are usually added to reduce cost but
may also have the effect of improving some aspect of the mechanical or
physical properties; and processing aids which increase the thermal
stability or aid flow by segregating to interfaces. All these additives can
have effects other than changing the aesthetic appearance of the mate-
rial. Some consideration is given to the analysis of additives but this
topic is not really within the scope of this book.
2 Introduction: Preliminary survey

1.1.2 CLASSIFICATION OF POLYMERIC MATERIALS

Whilst most polymers are synthesized from carbon containing molecules,
the precise method can have a profound influence on the physical prop-
erties which are developed. Most polymeric materials can be subdivided
into two classes of materials. Thermoplastics - indicating that the
material has the ability to be reprocessed by heating into the molten
state - and thermosets - implying that the process of formation gener-
ates a structure which may not be reprocessed by heating to a high
temperature. If a polymer is formed by the joining of one or more
monomers by no more than two chemical bonds, then the resultant
polymer is a linear chain and has thermoplastic characteristics. If,
however, more than two linkages are possible then a three dimensional
network can be generated and the material is a thermoset. Polymeric
chains may exist as highly organized - crystalline - or as highly disor-
dered - amorphous - materials. Polymer solids are rarely totally crys-
talline or amorphous, and partially crystalline materials are often
observed. The molecular organization has a very significant effect on the
physical properties and study of the morphology is often an essential
part of the characterization of any polymer system. There are a variety
of methods used for the synthesis of polymer chains, however two
distinctly different routes are used to generate commercial polymers and
hence are worthy of note. If the reaction involves addition to an ethene
or vinyl group then a polymer with carbon-carbon bonds may be
produced and the synthetic methods allow polymers with well
controlled chain lengths - molar masses - to be generated. The alterna-
tive approach is to link functional groups by a ring opening or conden-
sation type of reaction. The chains generated are produced by stepwise
growth and in a reaction mixture many different lengths of chain will
exist at any time. The molar mass of such a polymeric material will reflect
a spread of chain lengths and this often has significant implications for
the physical properties. Being able to associate a particular material with
one of these groupings will often immediately give an indication of its
physical properties and possible use. It is convenient to schematically
link the classification of polymer materials. We shall refer to this classi-
fication later in the book to illustrate how various techniques can provide
specific information on particular types of polymer.
Chemical analysis is very important to aid the specific identification
of the atomic species present and can often point directly to a particular
class of polymer. This is a standard method used by organic chemists
and will not be discussed in detail. Routine chemical analysis will allow
the amount of hydrogen, carbon, sulphur, halogens, nitrogen and oxygen
to be determined. Addition polymers often contain little or no oxygen
or nitrogen and hence are easily recognized for their chemical simplicity.
 The purpose of characterization 3

CONDENSATION
POLYMERS

AMORPHOUS-
ELASTOMERS

THERMOSETS -
[rarely crystalline]

Condensation polymers in contrast will often contain oxygen and or

nitrogen in addition to carbon and hydrogen and similarly are easily
recognized. There are, however, many polymers which break these
simple guidelines and can confuse the analysis of the data. The chem-
ical composition of a polymer will often be determined by surface
analysis and spectroscopic techniques in conjunction with simple chem-
ical analysis.
A number of excellent textbooks exist entirely devoted to the synthesis
and chemical structure of polymers and a selection of these is listed at
the end of this chapter. The above classification does not consider the
detailed topography of the polymer chain which can have a profound
effect on the physical properties. A very flexible polymer may be able
to adopt a tightly coiled structure whereas the same molar mass in a
more rigid polymer will appear more like a rod. The chemistry can lead
to polymers of the same molar mass being either liquids or solids.
Depending on the flexibility of the chain, the resultant material may
have at ambient temperatures elastomeric, amorphous or crystalline
properties. The above scheme is an attempt to rationalize what is almost
an infinite variety of structures and properties.

1.1.3 MATERIAL DESIGN AT THE MOLECULAR LEVEL

When a chemist is asked to design a polymer for a particular applica-
tion, a number of factors have to be considered. When a chemist prepares
a polymer it is essential that a number of characteristics are determined.
Some of the properties which are often required to define a polymer are:
4 Introduction: Preliminary survey

• the chemical composition of the monomer

• the detailed sequence structure of the polymer skeleton
• the melting or glass to rubber transition temperature
• the molar mass of the individual polymer chains
• the microstructure of the solid phase
• the dynamic mechanical properties of the solid
• the thermal stability of the polymer chain
Accurate characterization of a range of polymers permits correlation to
be made between features of the molecular structures and (macroscopic)
physical properties.
Having established such a correlation the chemist may then 'design7
a molecule with a structure favouring a particular property or combi-
nation of properties. After devising and executing a method of prepa-
ration, the product must be checked to confirm that it has the required
target structure.

1.1.4 MICROSTRUCTURE AND INFLUENCE OF PROCESSING

The arrangement of molecules in a solid polymer can be modified consid-
erably by the presence of additives and processing conditions used
during moulding, fibre drawing, etc. In the case of crystallizing poly-
mers, the spherulite size can be influenced considerably by small
amounts of deliberately added nucleating agents and the cooling rate.
Thus the microstructure is partly controlled by the formulation and
partly by the fabricator. Crystallinity may vary widely and can be influ-
enced by moulding and post-moulding conditions. Molecular flow
during moulding creates anisotropy and may influence the crystal
morphology. Fibre orientation in short fibre reinforced thermoplastics is
similarly sensitive to moulding conditions. Molecular orientation may
also be present in mouldings made from glassy polymers, causing phys-
ical properties to be anisotropic. Thus methods of characterization are
needed in order to be able to determine how formulation and processing
conditions influence the microstructure. In many applications the surface
structure is crucial and this can be influenced by the method of
processing.

1.1.5 FAILURE ANALYSIS

When a component fails it is frequently important to investigate why.
This will assist avoiding failure in the future. Was the failure caused by
the wrong choice of material? Were there flaws present that were caused
by the fabrication conditions? Was it the result of an accidental event,
such as an overload, and is a similar problem likely to recur? Did the
 The purpose of characterization 5

material degrade in service due to weathering or some other environ-

mental factor? Some of the methods of characterization described in the
following chapters can help to answer these questions. One of the most
common ways in which a component fails is by fracture. Inspection of
the fracture surface by one or more of the microscopic techniques can
lead to identification of the mechanism of failure (often by comparison
with fracture surfaces produced under strictly controlled conditions in
the laboratory) and hence determine whether failure was a fault of the
material, component design, the fabrication process (e.g. if a stress
concentrating void or agglomerate of filler is present and has nucleated
the crack) or service conditions.

1.1.6 WHAT IS A POLYMER SOLID?

The term polymer is used for a variety of different materials. Moulded
components are often produced from a moulding compound which
contains:
• organic polymer - homo or co-polymer
• inorganic or organic filler particles which are often cheaper or can
improve the mechanical properties
• a moulding compound (or processing aids) may be added to avoid
sticking of the polymer to the extruder or mould. These compounds
are often low molecular mass and also lower the surface energy of
the article when they segregate to the surface
• pigments and dyes which produce the required cosmetic effect; dyes
will be dispersed in the polymer, pigments which do not dissolve and
act like filler, and include a brightener such as titanium dioxide which
will enhance the colour
• antioxidants and ultraviolet stabilizers which suppress degradation
during processing at elevated temperatures and enhance the dura-
bility of the article in use.
More specialized polymer materials may have antistatic agents, surface
modifiers, mechanical property modifiers, including plasticizers,
crosslinking agents, tribo modifiers, etc. All these agents can significantly
modify the properties of the final material and many of the components
may only be present in the final article at concentrations which are below
1% w/w. This book is concerned with the study of pure polymers;
however reference will be made to formulated plastics which are to be
found commercially and form the main bulk of materials which are
analysed.
6 Introduction: Preliminary survey

1.2 MOLECULAR ARCHITECTURE

The properties of a polymer are determined not only by the atomic
constituents of each monomer unit but also the stereochemistry of the
polymer. There are several types of structure which can give physically
different properties from chemically identical materials. Amongst these
variations are:
• stereochemical variations; gauche and trans isomers can have quite
different properties.
• chiral centres can generate polymers with different tacticity; and the
stereochemical ordering leads to significantly different physical
properties.
The isotactic conformation may be a rigid rod-like polymer whereas
the atactic form may be a flexible mobile chain. The precise nature of
the polymer depends on the chemical structure of the R and X groups
and the ability or otherwise of the backbone to form a helical structure.
The constituent monomers and shape that they make within the
polymer molecule determine to a large extent the packing characteris-
tics of the molecules. In an atactic polymer the disruption of periodicity
along the polymer molecule inhibits packing and will influence the bulk
structure. If the polymer backbone is regular then there will be a
tendency for the chains to pack close together and a crystalline struc-
ture is observed. Introduction of a bulky side group which prevents the
close approach of parallel molecule segments will lead to a disorderd,
or amorphous, solid structure. The side chain bulky groups can also
inhibit conf ormational change around the backbone and will lead to rigid
properties being observed even at relatively high temperatures. In most
amorphous materials heating will allow the disordered glassy state to
be converted into a mobile flexible rubbery state on passing through the
glass transition temperature T . In fact all conformational changes which
occur are influenced by the chemical structure and will ultimately deter-

H
H H H H /
H
I I I I N^N^ H H H
YrW^
t i l l
i l l 1
H HXXVi^S^\H
H H H H y y
H H
trans - trans conformation gauche - trans conformation
 Molecular architecture 7

R X R X R X R X R X R X R X
Isotactic conformation

R X X R R X X R R X X R R
Syndiotactic conformation

R X R X X R R X X R R X R X
Atactic - random conformation

mine such properties as the stiffness, strength and gas/liquid perme-

ability. Optical properties also depend on molecular structure and strong
birefringence can be observed if highly polarizable bonds (multiple
bonds) are present and if, during processing, they adopt a preferred
orientation.

1.3 AMORPHOUS POLYMERS

A large range of polymers are unable to form an ordered structure in the
solid state and resemble glasses or rubbers. The transition from glass to
rubber occurs at a temperature, T , which is determined not only by the
chemical structure but also by the free volume available for the chains
to undergo rotational motion. The free volume is an ambiguous concept
which may be defined as the difference between the projected slope of
the glass state expansion of the solid and that observed for the rubber
state. It represents the averaged volume available for the collective
motion of the chain backbone and has a critical value which defines the
value of Tg. The free volume is a function of the chemistry of the poly-
mer; below Tg the volume between chains will, on the average, be below
this critical value. However, the fact that creep occurs in polymers below
Tg indicates that a very small number of these 'holes' may still exist or
be formed as a consequence of chain rearrangement. The free volume is
closely associated with disorder. It is possible for a material to be frozen
as an amorphous solid and then to slowly transform into a crystalline
state, a process which is occasionally observed with polymers.
8 Introduction: Preliminary survey

Volume of solid
Free volume

Rubber

Temperature
Fig. 1.1 Volume-temperature relationship for an amorphous polymer.

One of the simplest ways of determining the value of T is by the

direct measurement of the temperature dependence of the density. For
amorphous materials, Tg is probably the most important characteristic
and the wide range of values which can be observed differentiates
organic polymers from other materials. The T is a 'second order' thermo-
dynamic process and is therefore kinetically controlled and the precise
value which is observed depends on the rate or frequency of the applied
stimulus (e.g. deformation rate, cyclic loading frequency, temperature
ramp, etc.). The second order character of the process is a reflection of
the fact that it is associated with the motion of the chain backbone, which
in turn is influenced by both the available energy and volume for the
motion to occur.

1.4 CRYSTALLIZING POLYMERS

1.4.1 CRYSTALLIZATION CONDITIONS

Many polymers crystallize and others do not. Since crystals have well
ordered periodic structures we would not expect to be able to crystal-
lize those polymers which do not possess a periodic molecular structure
(e.g. random copolymers). On the other hand, linear polymer molecules
with regular periodic structures arrange themselves side by side to form
crystals. In nature, many polymers form helical structures and these
are able to align themselves in a parallel orientation and form crystal
structures. Long thin molecules, small or polymeric, can form disordered
partially oriented states which have a significant degree of mobility
and these are called liquid crystalline materials. Crystals formed
by alignment of neighbouring molecules will only have weak secondary
bonding holding them together. The internal energy is usually small
 Crystallizing polymers 9

and the melting temperatures tend to be low, sometimes below room

temperature. Side groups can induce steric effects as in the case of
some isotactic polymers and force the molecules to adopt a helical
conformation which can pack efficiently. If the side group is too large
to permit close packing, the internal energy can never be sufficient to
produce a stable structure and the polymer does not crystallize and is
amorphous. Sometimes a polymer that shows little or no tendency to
crystallize at a particular temperature may be made to crystallize by
orientating it (e.g. shearing it in solution or by extending in the rubbery
state). Polymer molecules that take up conformations that are more
nearly straight have a high trans population and can readily form
crystals.
For crystallization to occur the polymer molecule must have a regular,
relatively symmetric structure about the backbone and sufficient freedom
of motion to permit the molecule or segments of it to achieve the con-
formation required for crystal growth. The formation of an organized
state can be assisted by hydrogen bonds between neighbouring mole-
cules, as often happens in biological materials. These systems are often
very sensitive to the distance and frequency of hydrogen bonding groups
(e.g. variation in the properties of nylons with changes in the structure
are dominated by hydrogen bonding). The temperature must be below
that at which spontaneous melting occurs, but not so far below it to
inhibit molecular motion. Nucleating agents may act as templates to
encourage molecule segments to arrange in the correct manner to form
a crystal nucleus. These agents may be included specifically for this
purpose but nucleation may often take place on a reinforcing fibre or
other additive included in the formulation for purposes other than
crystal nucleation.
Cooling rate influences crystallization, and since large temperature
gradients may exist during fabrication, it is not unusual to find different
crystalline forms within a single article. Injection moulding involves the
hot melt entering a cold mould cavity and the material near the cavity
wall cools and solidifies rapidly to form a skin. This region contains
molecules which have been rapidly cooled and are subject to high shear
forces which can produce a high population of the trans conformation.
Both the cooling rate and flow conditions will influence the orientation
of the polymer chains and hence the crystalline morphology which is
formed. The material in the interior of the moulding cools much more
slowly as a result of the low thermal conductivity and often has a quite
different crystal morphology to the surface.
The pressure exerted during moulding operations (up to 250 MPa in
injection moulding) has an influence on crystallization. Some experi-
ments at much higher pressures have produced quite different
morphologies [1] from those produced at normal pressure.
10 Introduction: Preliminary survey

1.4.2 CRYSTAL STRUCTURE

It is relatively straightforward to determine the crystal structure
but more difficult to determine the types of defects and their abundance.
Dislocations can occur in polymer crystals but they are less easy to
observe and characterize than in metals (see Chapter 9). There are
types of defect that do not have counterparts in other materials asso-
ciated with the chain ends, chain folds at surfaces, and where buried
loops occur within the crystal instead of at the surface. The general levels
of disorder can be deduced from X-ray diffraction measurements
(Chapter 8).

1.4.3 CRYSTAL MORPHOLOGY

When the polymer crystals are grown from solution they are normally
formed as chain folded lamellae [1]. The concept of chain folded crys-
tallization raises many questions. Does the molecule double back on
itself in the minimum possible number of chain bonds? Does it form a
loose loop at the crystal surface? Does it re-enter the crystal at the nearest
('adjacent') site, or somewhere else ('switchboard')? What determines the
fold lengths? These questions are not easily answered, but many careful
laboratory investigations involving several of the characterization tech-
niques described in this book have helped in the study of polymer struc-
ture. The results have led to a deeper understanding of the structure of
bulk samples, solidified from the melt.
When polymers are crystallized from the melt they usually form
lamellae, frequently arranged into spherulites [1]. The internal structure
of the spherulite and its size influence other physical properties of the
solid. Some molecules or parts of molecules are always rejected from
the crystal phase and fractional crystallization (i.e. the fraction of mate-
rial that is actually present as crystals) has important effects on the phys-
ical properties. Crystalline and non-crystalline regions have very
different physical properties and the overall bulk properties will there-
fore depend on the fraction of each present. Hence a number of methods
have been devised to measure the fractional crystallinity and most of
them are unique to polymer studies.
When crystallization takes place in a flowing system molecular orien-
tation is induced which will influence the crystal morphology. In extreme
cases crystallization occurs without significant chain folding occurring
('extended chain crystallization'). Sometimes chain folded crystals grow
out from the main fibre to form a 'shish-kebab' structure. This can be pro-
moted by simply rotating a rod about its axis in a solution in a coaxial
cylindrical beaker. The fibrous growth attaches itself to the rod. Similar
structures sometimes form from the melt during moulding operations that
involve flow.
 Survey of characterization techniques 11

Crystal morphology can be altered by deformation in the solid state.

Fibre drawing causes pronounced orientation so that the molecular axis
and the corresponding crystal axis tend to lie parallel to the draw direc-
tion. Crystal orientation produces anisotropic properties and occurs
mainly during the fabrication process. In fibre drawing a large enhance-
ment of stiffness and strength is produced parallel to the fibre axis, with
little reduction in properties in the transverse directions. Thus methods
of measuring orientation are of importance.
Many natural polymers, e.g. cellulose, form a highly organized crys-
talline structure in which the individual molecules are in helical or
ribbon like forms and the surface structure which is observed reflects
the internal orientation of the individual polymers. In inorganic solids,
the crystalline structure can be related directly to the atomic spatial
distribution. However, in polymers the organization observed depends
on the scale of observation. At a molecular level a high degree of order
may exist, but disorder increases as the scale of observation is increased.
Crystalline regions pack as lamellae, but these elements may cluster to
form spherulites in which the overall orientation is random or aligned
to form fibre structures. The microfibres themselves may be clustered to
form structures which may be recognized as biological structures.

1.5 SURVEY OF CHARACTERIZATION TECHNIQUES

One of the most important characteristics of a polymer is its molecular
mass. Measurement techniques have been developed for polymers which
are different from those used for low molar mass materials. The tech-
niques for absolute molecular mass determination are not particularly
versatile. The exception is gel permeation chromatography in which
different molecular mass fractions are separated as part of the proce-
dure to determine molar mass distribution and this therefore enables
the preparation of (tiny) samples of selected molecular size with a very
narrow molar mass distribution ('monodisperse'). Molecular mass deter-
mination is dealt with in Chapter 2.
Spectroscopic techniques are used in many branches of chemistry for
molecular structure analysis. Many different structural features can be
examined and each technique has a range of applications. Further
explanation is given in Chapter 3, then Chapters 4-7 are devoted to
individual techniques. Ultraviolet - visible spectroscopy is placed in
Chapter 4 because it uses electromagnetic radiation of the lowest wave-
length among those techniques discussed. It has wide application in
organic chemistry for identifying molecules with conjugated molecular
structures. Another use of this method is the measurement of the num-
ber of carbonyl groups and this has been effectively used to monitor
degradation caused by oxidation/weathering of polymeric materials.
12 Introduction: Preliminary survey

Chapter 5 deals with vibrational spectroscopy: both infrared (IR) and

Raman are often described as 'fingerprinting7 techniques and are the most
popular methods for identifying a polymer. It is possible to determine
information about orientation and crystallinity by appropriate experi-
mental arrangements and analysis.
Nuclear magnetic resonance (NMR), Chapter 6, is also used for mole-
cular structure analysis. Again details concerning the precise three
dimensional arrangement of atoms in the molecules (tacticity, cis-trans
isomer discrimination, etc.) can be deduced from the measurements. The
technique allows selection of a particular nucleus for examination and
can provide very detailed structural information in both the liquid and
solid state. The response is very different in the liquid and solid states
and changes in the line shape can be observed on passing through the
glass transition temperature T . The use of NMR for the study of the
dynamics of polymer systems has allowed a detailed understanding of
the types of motion which can occur in polymers and is a valuable addi-
tion to other methods of observation [2], The electron spin resonance
technique measures the paramagnetism associated with the unpaired
electron spin and can be used to determine the concentration of free
radicals (and the structure of the reaction site). The technique has been
applied to the study of the kinetics of polymer formation and also the
chain scission products generated after fracture of the polymeric mate-
rial (Chapter 7).
Whereas the techniques outlined so far are concerned primarily with
the size of molecules and the arrrangement of atoms within them,
Chapters 8-11 deal with the arrangement of molecules and structural
features on a larger scale. Many polymers form crystals and methods of
determining crystal structure are of great value. X-ray diffraction
(Chapter 8) is foremost in this field and electron diffraction (Chapter 9)
is also important. Orientation and fractional crystallinity can also be
measured using X-ray diffraction. Details of crystal morphology can often
be inferred from continuum analytical measurements, but the most pos-
itive morphological determinations are made by microscopy (Chapters
9-11). Microscopy can provide detailed and unambiguous morphologi-
cal information, though great care must be taken to avoid artefacts that
might lead to serious errors. These techniques can sometimes be used to
measure lamellar thickness and other details of crystal morphology
directly, though the accuracy of the method must always be carefully
assessed. X-ray line broadening can also be used to estimate lamellar
thickness, though in this case an average value is obtained whereas a
microscopic measurement is based on a feature selected by the observer.
Raman spectroscopy provides yet another technique for lamellar thick-
ness measurement. The average spherulite size can be obtained by small
angle X-ray scattering (Chapter 8) or small angle light scattering (Chapter
 Survey of characterization techniques 13

11), or can be obtained directly by microscopy, most commonly light

microscopy (Chapter 11) or scanning electron microscopy (Chapter 10).
X-ray diffraction can be used to measure crystallinity whereas trans-
mission electron microscopy can be used to show details of the arrange-
ment of the crystal lamellae and non-crystalline regions. Crystal
orientation in the bulk is usually determined using X-ray diffraction,
though crystal habit analysis is most elegantly performed in the trans-
mission electron microscope using a combination of microscopy and
selected area electron diffraction (Chapter 9). Molecular orientation in
the amorphous state is often measured using birefringence (Chapter 11).
Fracture mechanisms can be determined with the help of microscopic
examination of fracture surfaces using either light or scanning electron
microscopy. These techniques are also suitable for studying the distrib-
ution of fillers, including reinforcing fibres.
Methods of thermal analysis (Chapter 12) provide information on the
bulk thermal properties. Information can be obtained about transitions;
the crystal melting temperature, the glass to rubber transition, T , and
sub-Tg processes as well as defining the onset of degradation. In addi-
tion calorimetric measurements leading to determination of the heat of
melting and enthalpies of reaction in thermoset systems can be made
using differential scanning calorimetry (DSC). Thermogravimetric
analysis (TGA) is regularly used to determine residual monomer content,
onset of thermal degradation and absorbed moisture levels. Thermal
expansion measurements will be anisotropic if the structure is
anisotropic (orientated). Other thermal techniques involve making a
measurement of some physical property as the temperature is scanned
through a wide range and are often able to provide information on mole-
cular relaxation processes (e.g. dynamic mechanical thermal analysis
[DMTA] and dielectric relaxation thermal analysis [DETA]) [2].
When choosing topics for inclusion in this text it is intended to provide
the reader with an overall view of methods of polymer characterization.
However the subject is so broad that it is impossible to be comprehen-
sive and several valuable specialist techniques have been given only a
passing reference. Chapter 13 contains some of these techniques and
includes ESC A (electron spectroscopy for chemical analysis). This
method can produce unique information about the surface of a material
(namely the composition and the bonding state) and for some applica-
tions is indispensable. It is, however, rather specialized and not as widely
available as the other methods described here. Neutron diffraction is
even less accessible, but can provide unique information about the
conformational state of polymer molecules as well as the phase struc-
ture. Conversely, density measurements are quite common (and the
equipment is relatively cheap) yet is sufficiently straightforward to be
dealt with in a few paragraphs and does not merit expanding into a
14 Introduction: Preliminary survey

chapter. Measurements of density can be related to crystallinity or, in

the case of glassy polymers, to the thermomechanical history.
It is frequently found that several different techniques are required to
provide information to enable a particular characterization to be
executed and examples are given in Chapter 14. In some cases it is
possible to conduct characterization with two or more techniques simul-
taneously. It is often possible to examine a specimen using a light micro-
scope during the performance of an experiment. A microscope
heating/cooling stage in which the specimen is located in a DSC cell
with optical windows can be purchased permitting calorimetric data
to be gathered while observing a phase change in liquid crystalline
materials. Similarly, commercial instruments are available that can
perform differential thermal analysis and thermogravimetric analysis
simultaneously.
A short introduction (Chapter 15) to the semi-empirical calculation of
polymer properties is included to aid the experimentalist to connect the
chemical structure characteristics to those of the bulk physical proper-
ties. Software now exists which, when used with atomistic modelling,
can be used in the interpretation of various forms of spectroscopy and
allows polymer scientists with a limited chemical background to assign
with confidence many of the observed features.
Finally we note that, although our emphasis is firmly on characteri-
zation of polymers, many of the techniques described here are of value
in detecting and identifying the additives present in commercial plastics
and we take the opportunity of pointing out examples whenever they
arise.

REFERENCES
1. Bassett, D.C. (1981) Principles of Polymer Morphology, Cambridge University
Press, Cambridge.
2. Bailey, R.T, North, A.M. and Pethrick, RA. (1981) Molecular Motion in High
Polymers, Oxford University Press, Oxford.

FURTHER READING
Lovell, PA. and Young, R.J. (1991) Introduction to Polymers, 2nd edn, Chapman
and Hall, London.
Billmeyer, F.W. Jr. (1984) Textbook of Polymer Science, 3rd edn, Wiley, New York.
Cowie, J.M.G. (1991) Polymers: Chemistry and Physics of Modern Materials, 2nd
edn, Blackie, Glasgow and Chapman & Hall, New York.
Ward, I.M. and Hadley, D.W. (1993) Mechanical Properties of Solid Polymers, Wiley,
Chichester.
Molecular mass
determination 2
2.1 INTRODUCTION
The essential distinguishing characteristic of polymeric materials is their
molecular size. The properties which have enabled polymers to be used
in a diversity of applications derive almost entirely from their long chain
macro-molecular nature. The molecular size of a monodisperse molecu-
lar substance is expressed in terms of its relative molar mass (RMM) (or
molecular mass) which is the mass in grams of 1 mole of the substance
relative to the mass of 1 mole (12 g) of 12C. 1 mole of a substance con-
tains Avogadro's number (6.02205 X lO^mol"1) of units (atoms, ions or
molecules). Actual molar masses are given by the RMM multiplied by
the atomic mass unit (amu) where 1 amu = 1.6606 X 10~27 kg. Thus in
practice the RMM is expressed in units of g mol"1 or amu, although the
units are normally omitted. The relative molar masses of polymers are
related to the degree of polymerization X (the number of repeat units in
the chain) and the relative molecular mass of the repeat unit, Mr by M
= X Mr. Non-network polymers, thermoplastics, uncured elastomers and
thermosetting resins, consist of an assembly of molecules having a dis-
tribution of molecular sizes, i.e. they are polydisperse. In order to char-
acterize fully these materials it is essential to have some means of defining
and determining their molar masses and molar mass distributions. For
crosslinked network structures the molecular size is considered to be
essentially infinite so that the concept of molar mass is less useful and
their properties are determined largely by the crosslink density.
Condensation polymers will generally have a broader molecular mass
distribution compared with vinyl monomers. The latter are easily
obtained in a wide variety of relative molecular masses varying from
103-107 amu whereas condensation polymers usually are obtained with
values of 105 amu or lower.
It is not always practicable to determine absolute relative molecular
mass distributions, although as we shall see in Section 2.3.2, this is
16 Molecular mass determination

H,
Mv

HP

Mz
Number of
Molecules

Molecular Mass
Fig. 2.1 Typical molecular mass distribution curve.

becoming more common practice using chromatographic techniques

usually coupled with other characterization methods.

2.1.1 MOLECULAR MASS AVERAGES

A simple molecular mass distribution is shown in Fig. 2.1 and represents
the number of molecules (Nz) having relative molar masses (Mz). It is use-
ful to define several averages of molar mass, the most common being
Mn the number, and Mw the mass average values. Averaging carried out
on the basis of the number of molecules (Nt) of a particular size (Mf) gives
the Number Average Molar Mass Mn defined by

M=*M
" 2N,
(2.1)

The corresponding number average degree of polymerization is

expressed as Xn such that Mn = Xn Mr
An important consequence of this definition is that the Number
Average Molar Mass in grams contains Avogadro's number of mole-
cules. This definition of molar mass is consistent with that of a monodis-
perse molecular species, i.e. all the molecules having the same molecular
mass. Of more significance is the recognition that, if the number of mole-
cules in a given mass of a poly disperse polymer can be determined in
some way, then Mn can be calculated directly. This is the basis of colliga-
 Introduction 17

tive property measurements (Section 2.2.1 (b)) and in principle allows

absolute determination of the molar mass.
Averaging on the basis of the mass fractions (w{) of molecules of a
given mass (Mf) leads to the definition of Mass Average Molar Mass
(MJ

- = 2n; f M | . = SN|.A^2
(2.2)
2wf 2NZ-MZ.

The_corresponding mass average degree of polymerization is expressed

by Xw. Mwis probably a more useful means of expressing polymer molar
masses than Mn since it reflects more accurately properties such as melt
viscosity and tensile strength of polymers. It should be emphasized that
a knowledge of the molecular mass distribution is essential to explain
fully the physical properties of polymers. There are two other molecular
mass averages which are useful in describing polymer properties. The
first is the z-average Mz defined as

— SN- M 3
(23)
*-!fi£
which can be determined experimentally using sedimentation/diffusion
measurements, but will not be discussed further in this text since its
applications are limited and very specialized. The second is the Viscosity
Average Mv defined as

M
—
--fed
rSN-M-1+al1/a
<">
where a is an empirical constant for a given polymer, solvent and
temperature. Mv is not a widely quoted molecular mass average but
solution viscosity measurements are used extensively for determining
molar masses (Section 2.3.1).
The molecular mass or molar mass and its distribution have a
profound influence on the properties of polymers. In most situations
engineers or scientists will make objects out of or perform experiments
on samples which have been prepared using a variety of different
synthetic methods. The simplest methods of synthesis, radical and
condensation polymerization, lead to materials with broad distributions
of molar mass. In general, condensation will produce materials with
lower average molar mass than addition processes. Catalysed and ionic
polymerization will produce narrower molar mass distributions than
other routes. Ring opening and group transfer polymerization can yield
relatively narrow distribution materials but can also produce broad
18 Molecular mass determination

distributions under other conditions. Measurement of the molar mass

distribution can therefore be a useful indication of the route used for
polymerization. Monodispersed samples with ratios close to unity are
usually an indication of an ionic initiated process. Step condensation
polymerization in which_statistically the most probable distribution is
generated has M z : Mw : Mn =3:2:1. With increasingly broader molecular
mass distributions, such as those produced by chain addition polymer-
ization at particularly high levels of conversion, the ratio Mw: Mn can
increase to >25. Commercial polymers will often have a distribution
which not only reflects the synthesis route but also the fact that the final
sample is obtained by blending materials with the same chemical struc-
ture but different molecular masses. Adhesives used in pressure-sensi-
tive applications will often have two or three peaks in their distributions
reflecting the fact that they are produced by blending high, medium and
low molecular mass batches of polymer.
Why are we concerned about the average molar mass and the molar
mass distribution? Many physical properties of polymers are influenced
by the length of the polymer chain. For instance:
• viscosity_of the melt depends on the molar mass below a critical value
called Mc and above this value increases with molar mass with an
approximately 3.5 power law. The consequence in processing is that,
whereas low molar mass materials form free flowing Newtonian low
viscosity fluids on melting, higher molar mass materials form highly
viscous difficult to process materials which reflect the entangled
nature of the chains in the melt and which are difficult to process;
• the glass to rubber temperature which is associated with the onset
of long range molecular motion of the chain backbone is a function
of molar mass. For very short chains the ends move at a very different
rate to the backbone and this process lowers the observed T . The
value of T varies with molar mass up to a limiting value which occurs
at chain lengths of about 105 and then becomes constant.
These two very important physical parameters have a profound effect
on other mechanical properties and consequently on the use of poly-
mers in various applications.
When considering instrumental methods for determining polymer
molecular masses, we should distinguish those primary or absolute
methods which are capable of determining molecular masses from first
principles (although in practice comparative methods may be employed)
and secondary methods which require calibration with known molec-
ular mass standards. Secondary methods usually offer additional advan-
tages including experimental convenience.
 Primary methods 19

2.2 PRIMARY METHODS

Primary methods include colligative property measurements to deter-
mine Mn and light scattering techniques to determine M^. Secondary
methods include solution viscosity and gel permeation chromatography
measurements. The latter method is particularly useful in providing a
direct measure of molecular mass distributions.

2.2.1 NUMBER AVERAGE MOLECULAR MASSES

When we considered the definition of Mn it was pointed out that where
it is possible effectively to count the number of molecules in a known
mass of a polymer, then Mn could be readily determined. This relation-
ship provides the basis of the primary methods for determining Mn. The
experimental methods employed differ in the ways in which the
numbers of molecules are determined. The simplest approach may be
applied to those polymers possessing characteristic chain end groups
which can be analysed quantitatively. The method, termed end group
analysis, will only be considered briefly as it is only useful for low mole-
cular mass polymers. Other primary methods are based on the colliga-
tive properties of solutions.

(a) End group analysis

In simple step condensation polymerizations of polyesters
n HO-R-OH + n HOOC-R'-COOH ->
HO-R-O-IOC-R'-COO-R-OJ^OC-R'COOH + (n - 1)H2O
and it can be presumed that on average each polymer chain possesses
one -COOH and one -OH group. Spectroscopic techniques such as
infrared or nuclear magnetic resonance enable the concentration of the
end groups to be determined relative to the in-chain ester links. It is
also possible to determine the concentration of these end groups by
direct titration using acidic or basic conditions. The same general proce-
dures may be applied to other condensation polymers such as
poly amides. In the case of the -COOH groups simple chemical analyt-
ical procedures such as acid-base titration or pH measurements have
been used successfully. The essential requirement in every case is that
the structure of the chain end groups be known.
For chain addition polymerizations which proceed with no measurable
chain transfer the chain end initiator fragment may be sufficiently dis-
tinct to enable identification and quantitative analysis to be carried out.
Provided that the mode of termination is known, i.e. combination or dis-
proportionation, Mn can be determined. For example, polymerization of
20 Molecular mass determination

styrene initiated with AZBN (azobisisobutyronitrile) would yield the fol-

lowing chain structure when the mode of termination is by combination:

(CH 3 >zC-C- -CH—CH2 ]—[—CH2 CH ]—C ~C(CH3)2

I m | n |
CN CN

It is possible to distinguish the in-chain phenyl protons and the

aliphatic protons including the chain end f-butyl groups by means of
proton NMR (Chapter 6). Absorption peaks at 7.2 ppm can be assigned
to phenyl protons (intensity = A) and absorptions in the range 0.6-
2.8 ppm can be assigned to the aliphatic protons (total integrated inten-
sity = B). Assuming that
A __ 5(n + m)
B ~~ 12 + 3(n + m)
(n + m) may be calculated, and hence, from the molecular masses of the
respective groups, Mn =W4(n + ra) + 136. Chain transfer can however
complicate the analysis and lead to ambiguities in the values obtained.
The major problem encountered in end group analysis is the sensi-
tivity of the method which is limited to about 0.1% and this limits the
minimum molar mass which can be measured. The method relies on
measurement of the concentration of only one or two groups per mole-
cule and the sensitivities needed to determine polymer molecular masses
of 104 and 105 amu with an accuracy of 10% are respectively 10~5 and
lO"6 moles. Spectroscopic techniques such as IR, NMR and tracer tech-
niques can be enhanced by labelling with, for example, 14C to give high
accuracy, but nevertheless end group analysis is restricted to molecular
mass averages < 20,000 amu. Since useful vinyl thermoplastics normally
have molecular masses greater than this range, other methods for deter-
mining molecular masses are preferred for routine measurements. The
technique is most useful for the analysis of condensation polymers where
reaction of specific labels can enhance the sensitivity.
Identification of the nature of an end group using Spectroscopic tech-
niques may also be important in order to determine the method used
for synthesis or the susceptibility of the polymer to undergo further
polymerization or degradation. NMR techniques have been developed
to identify very small concentrations of end groups but these are usually
not sufficiently quantitative to allow their use in molar mass deter-
mination.
 Primary methods 21

(b) Colligative properties

Solution properties such as lowering of vapour pressure, elevation of
boiling point, depression of freezing point and osmotic pressure are
termed colligative properties. The name derives from the fact that, at
infinite dilution, the properties depend on the number of independent
kinetic units in the solution and hence on the number of solute polymer
molecules. The colligative properties may all be expressed in terms of
the free energy of dilution of the solution AG/ which is the total Gibb's
Free Energy when 1 mole of a solvent is added to a very large volume
of solution:
AGa = G! - Ga° (2.5)
where Gl and G^ are the partial molar free energies of a solution and
pure solvent respectively. For ideal solutions at low concentration the
colligative property relationships may be expressed by

AG^RTln-^ (2.6)
Pi
where plf and p^ are the vapour pressure of solution and solvent respec-
tively and R is the ideal gas constant,
AGX = -HV (2.7)

where II is the osmotic pressure and V the molar volume of the solvent,

AG^^AK (2.8)
h
AGi^AT, (2.9)
*e

where Lf and Le are the latent heats of fusion and evaporation, T^ and
Te are the freezing point and boiling point of the solvent, and AT^ and
ATe are the depression of the freezing point and elevation of the boiling
point of the solution respectively.
Since for ideal solutions the free energy of dilution can be expressed
by
RTV
45,--^ (2.10,

where M2 is the molecular mass of solute and c its concentration, it

follows that
22 Molecular mass determination

ln^=--£-c (2.11)
M
Pi 2

RT
I, = -c (2.12)

Ai>-sae
1
LfM2 (2.13)
and
PT^T/
AT = c (214)
' r1sr
L,M 2

so that measurement of colligative properties permits the solute molec-

ular mass to be determined. As with end group analysis, the accuracy
with which the various measurements can be made imposes limitations
on their use. In addition it is not expected that polymer solutions would
behave ideally so that the above expressions require modification to take
into account specific polymer-solvent interactions. This is done by
expressing the colligative property, such as osmotic pressure, in the form
of a virial equation i.e.
property - A-^c + A2c2 + A3c3+ ... (2.15)
where the higher power terms express the departure from ideality and may
be used to provide additional information on polymer-solvent interactions.
Applications of colligative property methods will be considered by
specific reference to ebulliometry and osmotic pressure measurements
since these have been used most widely and are more likely to be
encountered in the laboratory. Some reference will also be made to
vapour pressure osmometry which is a practical means of determining
vapour pressure lowering.

(i) Ebulliometry
As we have already seen, the elevation in boiling point (AT) of a solu-
tion of known concentration (c) of solute in a particular solvent is related
to the number average molecular mass Mn. For a dilute, ideal solution
the molecular mass is given by

Mn = &£ (2-16)

where KE is the ebullioscopic constant - RT^Ml/Lep, Ml the relative

molecular mass of the solvent and p its density.
 Primary methods 23

For non-ideal solutions, such as for polymers, additional terms need

to be introduced to account for the departure from ideality. An expres-
sion of the form
AT = A^c + A2c2 + A3c3+ ... (2.17)
where Alf A2 and A3 are thejirst, second and third virial coefficients
respectively and Al = KE/Mn, is found to express the behaviour
adequately. For ideal solutions the second and third terms in the expres-
sion, which take into account heat and entropy of mixing, vanish and
the expression is then identical to equation (2.16). Mn for polymers can
then be determined by measuring AT as a function of c at low concen-
trations. A plot of AT/c against c will be approximately linear as c—> 0
provided that A2 and A3 are not too great. The intercept when c = 0 is
then equal to KE /Mn so that, knowing KE , Mn can be obtained.
In practice, errors will arise in applying this procedure if there is a
'zero error'. The zero error can be expressed by modifying the equation
by the inclusion of an additional term AT0
AT - AT0 + A$ + A2c2 + A3c3+ ... (2.18)
A plot of AT against c should be linear with intercept equal to AT0 when
the higher terms are negligible (Fig. 2.2). The source of the zero error is
not completely understood but probably arises from differences in the
nature of the boiling process for dilute solutions due to surface tension
effects.
When zero error effects are noticeable, treatment of the data needs to
be modified in order to yield Mn. The recommended procedure is the
divided difference method in which the ratio of differences in boiling
points, (Tn+1 — Tn) due to successive small changes in concentration
(cn+i ~ cn)' are described in terms of the virial equation in the following
manner:

ir

Concentration

Fig. 2.2 Ebulliometry; zero error plot.

24 Molecular mass determination

( n+l
l _^=A 1 + A2(cn+1 + cn) (2.19)
\Ln + l c
n)

A plot of the left hand side of equation (2.19) versus (cn+1+cn) should
be linear with intercept equal to A x which ought to be independent of
any zero error. Extrapolation of the data to zero concentration differ-
ence should be carried out using linear regression analysis to ensure
accuracy.
Although it is possible in principle to determine KE from known
solvent properties, i.e. Te/, Le/, Mlf and p (equation (2.16)) it is more
normal practice to employ comparative methods using polymers of
known molecular masses as reference standards. In this way problems
which may arise from instrumental variables can be minimized. The
latter may include factors, such as solvent held in the reflux system,
effectively changing the solution concentration, and superheating due to
the effects of hydrostatic pressure. In addition, there is some doubt about
the constancy of KE with changes in molecular mass. Determination of
the boiling point elevations produced by a series of dilute solutions
of a solute of known molecular mass enables a calibration curve to be
constructed by plotting the data in the form of equation (2.19). From the
calibration curve an effective value of KE may be obtained which takes
into account some of the instrumental variables that may exist.
Measurements of boiling point elevations for polymer solutions using
the same experimental procedure enable M n to be determined using the
previously determined KE value.
Inspection of equation (2.16) shows that AT is directly proportional to
solute concentration and inversely proportional to polymer molecular
mass. Since it is necessary to work with dilute solutions, the magnitude
of AT is small and the practical difficulties in measuring small temper-
ature differences places limits on the range of applicability of the tech-
nique. For example, polymer molecular masses of 104 and 5 X 104 amu
give AT values of approximately 2 X10~3 and 5 XlO~ 4 °C respectively in
1% solutions. Measurement of such small temperature differences is best
achieved using bridge circuits employing thermocouples or thermistors.
Some highly sensitive temperature sensing devices have been reported
in the literature with accuracies down to 2 X 10~6 °C and have enabled
polymer molecular masses up to 105 amu to be determined. For most
systems, however, the practical limit for molecular mass determination
by ebullioscopic techniques is considered to be approximately 3-4 X 104
amu and even then, great care is needed to ensure accuracy. Thus the
method is most suitable for low molecular mass materials such as
novolak resins, condensation polymers and other intermediates.
Measurement of boiling point elevation is best carried out by means
of a differential ebullioscope. That is to say that the boiling point of the
 Primary methods 25

solution is directly compared with the boiling point of the refluxing

solvent. There are two particular experimental arrangements currently
used. The first utilizes a single refluxing column in which the tempera-
ture of the boiling solution is compared to the temperature of the
refluxing vapour. The second arrangement employs two identical
columns placed side by side with one containing pure solvent and the
other the solution of known concentration. For very high sensitivity,
the former arrangement is considered to be more accurate since instru-
mental variables are reduced but for many routine applications the twin
column system is adequate and is likely to be simpler to construct
and operate.
A typical arrangement is shown in Fig. 2.3. A particular feature is the
Cottrell pump which causes the boiling solution to flow over the ther-
mistor or thermocouple pocket so reducing effects of possible super-
heating of the boiling vapour. The experimental procedure involves
placing pure solvent in the flask and balancing the Wheatstone bridge
network to give a null reading on the galvanometer. Accurately weighed
samples of solute are added to one column and the system allowed to
reach equilibrium when the bridge is again balanced by means of a cali-
brated variable resistor or potentiometer. The magnitude of the change
in the variable resistor is directly proportional to the boiling point eleva-
tion. The results may be plotted directly as potentiometer dial reading
against sample mass or as the divided difference discussed earlier and
then compared with results obtained from the reference standard.
Errors in temperature measurement can arise due to excessive foaming
of the solution. For high accuracy, steps need to be taken to reduce the

Condenser

Thermistor

Cottrell pump

Heating
element

Fig. 2.3 Ebulliometry column showing Cottrell pump.

26 Molecular mass determination

effects of foaming including the use of mechanical devices to disrupt

the foam.

(ii) Osmometry
The chemical potential of a solvent in solution is lower than that in the
pure solvent in proportion to the concentration of the solute species. If
it is so arranged that the solution is separated from pure solvent by
means of a semipermeable membrane, i.e. one which only permits
passage of solvent and not solute molecules, then solvent will diffuse
through the membrane in order to equalize the chemical potentials and
so establish equilibrium. The result of the flow of solvent is to generate
a pressure on the solution side of the membrane which at equilibrium
is just sufficient to prevent any further net flow of solvent. The pressure
difference is termed the osmotic pressure II. This is illustrated schemat-
ically in Fig. 2.4, where the osmotic pressure is indicated by the differ-
ence in heights of the liquids in the side arms.
The Knauer membrane osmometer will serve to show the general
procedures employed. The essential part of the instrument is the steel
measuring cell which is divided into two halves by the semipermeable
membrane. In operation the lower half of the cell is filled with degassed
solvent and sealed by the semipermeable membrane which forms a seal
when the upper half of the cell is placed in position (Fig. 2.5).
The membrane materials are typically regenerated cellulosics, such as
cellophane, cellulose nitrate and cellulose acetate, which are required to
be conditioned in the solvent prior to use. Other membrane materials
include polytetrafluoroethylene and polyvinylidene fluoride. The
membranes are supplied in a swollen, wet state and should not be
allowed to dry out. Conditioning of the membranes involves successive
immersions in a series of intermediate solutions containing increasing
amounts of the required solvent until 100% solvent is reached.
Recommended immersion times can vary significantly, depending on
the membrane and solvent.

Solution Solvent

x
Membrane

Fig. 2.4 Schematic representation of osmotic pressure.

 Primary methods 27

Fig. 2.5 Measuring cell of Knauer membrane osmometer (courtesy of Owens

Polyscience Ltd,). 1. Capacitive pressure measuring system. 2. Mushroom over
which the membrane is placed. 3. Semi-permeable two layer membrane.
4. Extractor device for rinsing.

Sample solutions of known concentration in the range 0.2-2.0 g/100

ml are injected successively into the upper half of the cell using a syringe.
Diffusion of solvent through the membrane into the solution results in
a negative pressure being developed on the solvent side of the
membrane. A steel membrane in the lower half of the cell is caused to
deflect under the action of the negative pressure and this in turn is
detected as a change in capacitance of an electronic circuit which is there-
fore related directly to IT.
The volume of the solvent side is kept as small as possible (about
120 jjil) so as to reduce effects due to temperature variations. Likewise
only small volumes of sample solution are necessary (about 1.5 ml) for
each reading. Membrane osmometers are capable of operating up to
130°C which enables polymer of low solubility to be examined.
Reference to the virial equation
O/c = RT/Mn+A2 c+ A3 c2 + ... (2.20)
shows that a plot of IT/c versus c should be linear for dilute solutions,
when the third term vanishes, with the intercept being equal to RT/Mn
from which Mn is readily determined.
As with all colligative property measurements, the accuracy of the
technique is dependent on the linearity of the curve at low concentra-
tion in order to permit extrapolation to zero concentration. This in turn
depends on the magnitude of the deviations from ideal solute behav-
iour as indicated by the magnitude of the second and higher order virial
coefficients. For most systems the third and higher virial coefficients are
28 Molecular mass determination

negligibly small and may usually be ignored. The second virial coeffi-
cient, which takes into account specific polymer-solvent interactions, is
more important in determining how good particular solvents are.
Thermodynamically good solvents correspond to high values of A2 (typi-
cally of the order of 10~4 m3 mol kg~2), and as A2 decreases in value, the
solvent becomes increasingly poorer. Theta solvent conditions arise
when A2 = 0 corresponding to the situation when an infinitely high
molecular mass polymer just swells but does not dissolve in the solvent.
In this case, a plot of H/c versus c has zero slope. When A3 becomes
significantly large, a plot of n/c versus c is no longer linear but shows
some curvature with increasing concentration. Extrapolation to zero
concentration is then more uncertain. However, if the data are plotted
in the form of (TI/c)1/2 versus c, the graph should be linear, permitting
extrapolation to give A2 as the intercept.
Membrane osmometry is particularly useful in that absolute molecular
mass averages can be determined over a much wider molecular mass
range than is possible for the other colligative property measurements.
Molecular masses up to one million can be determined, although
accuracy diminishes with increasing molecular masses. Unfortunately
the technique has limitations for polymers having low molecular
mass averages due to the possibility of diffusion of low molecular mass
molecules through the membrane. When this occurs, the measured
osmotic pressure is not the true equilibrium value and so constitutes
a major source of error. In addition, it is necessary to completely replace
the contaminated solvent which may involve dismantling and re-
assembling the osmometer and replacing the membrane which is very
inconvenient.
Osmotic pressure measurements are also very temperature dependent
so that careful thermostatting of the equipment is essential. This is
common practice and standards specify the temperature at which
measurements must be made.

(Hi) Vapour phase osmometry

The limitations placed on osmotic pressure measurements due to the
presence of low molecular mass fractions are not encountered in methods
based on lowering of vapour pressure. Consequently vapour pressure
measurements are more accurate for low molecular masses. Practical
methods for determination of molecular masses are based on lowering
of vapour pressure caused by the preferential condensation of solvent
from the vapour by a dilute polymer solution. The condensation process
is monitored by small temperature changes. Transport of solvent to solu-
tion proceeds via the vapour phase and thus the method is termed
Vapour Phase Osmometry.
 Primary methods 29

A typical vapour phase osmometer (Fig. 2.6) consists of a thermostat-

ted sample chamber containing two matched thermistors connected to a
bridge circuit capable of measuring small temperature differences with
high accuracy. The chamber is initially saturated with solvent vapour at
the given temperature. A small amount of solvent is placed by means of
a syringe on to one of the thermistors and a small amount of a dilute
polymer solution on to the other. Differences in vapour pressure between
the solvent and solution cause solvent to condense on to the solution
with a resultant rise in temperature of the solvent due to the evolution
of the latent heat of vaporization. The measured temperature difference
AT is directly proportional to the lowering of the vapour pressure and
hence to the polymer molecular mass.
In practice, the equipment is calibrated using known molecular mass
standards and recording differences in electrical resistance AR rather
than AT directly so that the molecular mass is obtained from
AR 1
(2.21)
Wc Wn+A^C
=

where K' is a calibration constant.

Fig. 2.6 Schematic of vapour phase osmometer (courtesy Owens Polyscience

Ltd). 1. Measuring probe. 2. Aluminium block for thermostatic control of
syringes. 3. Seal. 4. Wick. 5. Thermistors. 6. Windows. 7. Glass container for
solvent. 8. Measuring cell (aluminium).
30 Molecular mass determination

As with other methods, R values are obtained for a series of dilute

solutions of concentration c and AR/K'c is plotted against c with extrap-
olation to zero concentration to give Mn.

2.2.2 DETERMINATION OF MASS AVERAGE MOLECULAR MASSES BY

LIGHT SCATTERING
When a parallel beam of light passes through a transparent system, a
small part of the light is scattered elastically (Rayleigh Scattering). The
scattering arises due to fluctuations in the density of the medium which
in gases and liquids can be due simply to the molecules themselves. At
a molecular level scattering arises from interactions between the electric
vector of the incident electromagnetic radiation and the electrons in the
molecules inducing dipoles in the molecules oscillating at the same
frequency as the incident radiation. The oscillating dipoles act as scat-
tering centres by emitting light in every direction at the same wave-
length as the incident radiation. For solutions of molecules that are small
relative to the wavelength of the incident radiation (<AV20 where A' is
the wavelength in the solution) the scattering envelope is symmetrical.
For solutions of polymer molecules, additional scattering arises from the
presence of the solute molecules and this may be shown to be a func-
tion of the concentration of the polymer molecules, as well as their size
and shape. From measurements of the differences in intensity of the scat-
tered light between the solvent and a series of dilute solutions, it is
possible to determine the average size of the solute scattering centres,
and hence their mass average molecular masses.
There are a number of experimental arrangements available for deter-
mining light scattering of polymer solutions and these have been
described in some detail in the literature so that only an outline descrip-
tion will be given here. Light scattering photometers (Fig. 2.7) consist
essentially of a high intensity monochromatic light source (S) (either a

C P R.C S.C

zJ 1V U 11
\
i V
i \
— 1
•tiT:
s
i (PM)

[Link]. PM
s

Fig. 2.7 Schematic of typical light scattering photometer. S = source, F filter,

C= condenser, P = polarizer, RC = reference chamber. RPM = reference photo-
multiplier, SC = sample chamber, PM = rotatable photomultiplier.
 Primary methods 31

mercury arc lamp or, increasingly, laser light sources) and suitable filters
(F), a plane polarizer (P), a sample cell (SO, so constructed as to permit
passage of the incident beam and to enable the scattered light to be
detected, and a rotatable photomultiplier tube (PM) to permit measure-
ment of the scattered light intensity over a range of scattering angles.
The arrangement of the photomultipliers in the sample cell is designed
to reduce or eliminate entirely scattering from the container walls which
causes spurious results. Modern machines use lasers as the light source
and diodes to replace the photomultipliers.
The scattering of light by polymer solutions may be expressed in terms
of the Debye equation
H K <2 22)
r i=mv+2A* -
where

_ 32Tj2n2 (dn/dc)2
4
3N0A
(2-23)
and r is the turbidity (from I = I0 e T / ; Beer Lambert law), / is the path
length, n is the refractive index; dn/dc is the concentration dependence
of the refractive index of the polymer solution (a constant); N0 is
Avogadro's number; A is the wavelength of light in air; P(0) the particle
scattering factor, describes the angular dependence of the scattered light
and is related to the dimensions of the scattering entities.
K is given by the equation

27jV (dn/dc)2 (2.24)

~ N0A4

and Re = R0 (solution) - Re (solvent) where Re, the Rayleigh Factor, is

the ratio of the scattered light intensity per unit volume of the scattering
liquid, per unit solid angle of the detector, to the incident light inten-
sity (I0), i.e.

Re=*oljT (2'25)
where r is the distance from the scattering centre to the detector and ie
is the amount of light scattered per unit volume at angle 0 to the inci-
dent beam by one scattering centre.
dn/dc is first determined by measuring the refractive indices of a
range of solutions of known concentrations. Second, the scattered light
intensity is measured at different angles to the incident beam for a
hh

similar range of solution concentrations. In principle, for small mole-

cules a plot of Kc/Re versus c is linear with intercept equal to 1/MW so
that determination of molecular mass is relatively straightforward.
However, as the molecular size increases destructive interference effects
arise from light scattered from different parts of the molecule and the
scattering envelope is unsymmetrical. These interference effects are taken
into account by inclusion of the function P(ff) in equation (2.22)*. P(6)
is the ratio of the measured value of Re at a scattering angle 6 to the
intensity in the absence of interference. For small angles:
1 16772

m = Ws*(e/2}
1+ (2.26)

where s| is the z-average mean square radius of gyration and \s = \/n

is the_wavelength of light in the solution. When 6 = 0 then P(6) = 1
and Mw may be determined from the experimental data.
The experimental procedure followed for determining Mw for poly-
mers involves measuring the intensity of light scattered over a range of
angles relative to the incident beam (in the range 10-160°). Calculation
of Mw makes use of the Zimm plot which is a method of plotting the
data in such a way that the extent of light scattering at 6 = 0 may be
obtained by extrapolation so that P(6) = 1. From the equation
1 + =_
Kc = =_
=- 16V2} i 27 / _
1 __ 0W ,„„
S n S 2 +Ac 2.27)
R e Mw M W \ 3 A 2 / \2/ z 2

double plots are constructed of ICc/R^ against sin2 (9/2) +k'c where k' is
a scaling factor to give a reasonable spread of the lines on the graph.
These are a series of plots obtained by first keeping c constant and vary-
ing Of then keeping 0 constant whilst varying c, as shown in Fig. 2.8.
Extrapolation for c = 0 and 6 = 0 gives 1/MW as the intercept on the
Kc/Re axis.
In addition to Mw, the graphs yield a value of A2 which is given by
the slope of the 0 = 0 curve,
A2 =(k'/2) [slope] c^0
and the radius of gyration s| which is given by the slope of the c = 0
curve:
s | = (3A 2 M ro /16^n 2 )[slope] e _ 0

* The rotation of a solid body about an axis may be represented by a point mass M rotating
at a radius known as the radius of gyration R such that the moment of inertia of the body
is the product of the mass and the square of the radius of gyration I = MR2. Polymer
molecules in solution can assume different conformations so that the mean square radius
of gyration is used.
 Primary methods 33

f,
c
o c,
?3 C4
I

'0c
/ -

'Te,
I
.Ik e3
*e
9
2

S
1
e
o

sin2 (j)*k'c

Fig. 2.8 Zimm plot showing double extrapolation technique; extrapolated

points •. k' is a scaling factor.

Both of these quantities are useful in describing the behaviour of

polymer molecules in solution. The value of sj- obtained in this way
does not depend on the shape of the polymer molecule. These ideal
curves are not always obtained and the effect of polymer-polymer inter-
actions and aggregates can lead to deviations from linearity at high
values of the concentration.
The most serious problem that arises in determining molecular masses
by light scattering techniques is that due to the presence of impurities
such as dust particles which contribute to the overall scattering inten-
sity. Great care is required in preparing the solutions to remove any
extraneous matter. The recommended procedures require filtration of
the solvent and solutions through ultrafine sintered glass filters and
ensuring that there is no further contamination during the transfer of
the samples to the light scattering cell. The concentration of the solu-
tions is best determined gravimetrically by evaporating a known volume
of the filtered solution and weighing the residues to give the mass of
dissolved polymer.
Recent developments in light scattering photometers have involved
laser light sources. Because of the much higher light intensity, it is
possible to detect light scattered at small angles (< 10°) even for
dilute polymer solutions. Since P(0)-»l as 6—>0 we can rewrite
equation (2.22):
Kc 1
s:'^*2^
(2.28)
34 Molecular mass determination

Measurement of Re for a single angle then permits Mwio be determined

from the intercept of the graph of Kc/Re versus c. This markedly reduces
the number of measurements required, so greatly simplifying the proce-
dure. With the increased availability of laser light sources, such instru-
ments are becoming the norm and, when allied to on-line data
processing, lead to more rapid molecular mass determination.

2.2.3 MOLAR MASS DETERMINATION OF COPOLYMERS

We will not specifically discuss the problems of determination of the
molar mass of copolymers under the separate techniques. The main
differences between a homopolymer and a copolymer is that, whereas
the effect of change of solvent or temperature is determined relatively
simply by solution thermodynamics in the case of a single polymer/
solvent combination, in the case of a copolymer this is more complex.
It is not uncommon to find that a particular solvent may be a poor
solvent for one element of the copolymer and yet be a good solvent for
another. This effect is particularly evident in the case of a block
copolymer where the segments in the good solvent will be expanded
and those in the poor solvent contracted or collapsed. Change in temper-
ature in this case can have a dramatic effect on the collapsed regions
and relatively small effects on the expanded regions. It is also not
uncommon to find that only mixtures of solvents can be used to dissolve
copolymers and is a reflection of the preferential solvation of particular
parts of the polymer chain by a specific solvent type. Copolymers there-
fore exhibit second virial coefficients and radii of gyration which are not
simple sums or averages of those of the component homopolymer
values. Analysis of copolymers is very complex and invariably requires
application of a number of different methods and conditions in order to
achieve a definitive description of the chain dimensions.

2.3 SECONDARY METHODS

2.3.1 SOLUTION VISCOMETRY

A secondary method widely used for the determination of polymer mole-
cular masses is that based on solution viscosity measurements. It is
reasonable to presume that the viscosity of a polymer solution would
depend both on the concentration and on the average molecular size of
the molecules. Although there is no complete theoretical treatment of
the dependence of viscosity on molecular masses, useful empirical rela-
tions have been established. For instance, the viscosity average molec-
ular mass (MJ may be expressed by equation (2.4).
Assuming that polymer molecules in solution behave as impermeable
 Secondary methods 35

coils, i.e. the polymer molecule plus solvent move together, it_may be
shown that the hydrodynamic volume Vh is proportional to [77] Mv where
[17] is the intrinsic viscosity for the polymer-solvent system. The exper-
imental procedure employed requires determination of the intrinsic
viscosity for the polymer/solvent system [77] which is related to Mv/
through the Mark-Houwink equation:
— (2.29)
[77] = KM;

where K and a are empirical constants for the particular system. We

have noted that solution viscosities are dependent in part on the size of
the polymer molecules but this statement requires some amplification.
In the present context the effective size or hydrodynamic volume of the
molecules depends on the extent of interactions with the solvent. In good
solvents the molecules assume expanded conformations but, as the
solvent becomes poorer, intramolecular interactions become more signif-
icant until for theta solvents the chains assume their unperturbed dimen-
sions. Consequently the solution viscosity is expected to show quite
different dependence on molecular mass in different solvents. In addi-
tion, even for a particular polymer/solvent combination, effects such as
chain branching will influence the variation of [r/] with M^ so that the
empirical parameters K and a may not be constant for the same nominal
polymer type.
The intrinsic viscosity may be determined experimentally by measure-
ment of flow times of solvent (t) and a series of dilute polymer solu-
tions of known concentration (c) in a standard suspended level capillary
viscometer. The specific viscosity is calculated from the equation
(2 30)
•*, = ^^
^o
= ^~
o f -
where tc is the flow time for the polymer solution and t0 that for pure
solvent, and from the Huggins equation

^-M + nni'c <2-31)

where k'is a positive constant; the intrinsic viscosity is obtained by plot-
ting j]sp/c versus c and extrapolating to zero concentration (Fig. 2.9).
The Ubbelohde viscometer (Fig. 2.10) offers a convenient method for
determining the flow times under essentially constant conditions. The
experimental procedure involves introducing filtered solvent in the lower
bulb of the viscometer via the filling tube. The viscometer is then
mounted vertically in a water bath thermostatted to 25 ± 0.02°C and left
to equilibrate at that temperature. It is advisable to immerse the solvent
36 Molecular mass determination

«sp
15 -

1
04 08 12
Concentration gdl

Fig. 2.9 Dilute solution viscometry; Muggins plot of rjsp/c against concentra-
tion c.

container in the water bath itself to reduce the time required to reach
temperature equilibrium. The solvent is drawn up the central capillary
tube by suction using a hand aspirator until it is above the upper etched
mark. The time for a constant volume of the solvent to flow between the
two etched marks is determined to high accuracy using a stopwatch. This
value is t0. The viscometer is emptied, cleaned, dried and a known vol-
ume of a filtered solution just sufficient to be drawn up the capillary
without introducing air bubbles, and containing approximately 1% w/v
polymer in the same solvent, is introduced. The flow time for the solu-
tion is determined in the same manner as for the solvent. The stock
solution is then progressively diluted by the addition of known amounts
of the solvent to give at least 5 different concentrations, care being taken
at each stage to mix thoroughly. This can be achieved by repeatedly low-
ering and raising the solution up the capillary tube using the hand aspi-
rator. The flow times for each dilution are determined (tc) and then [77]
and Mv, can be calculated with known values of K and a. In general,
Mv, will be close to, but somewhat lower, than Mw (Fig. 2.1). In practice
the errors involved in the viscosity measurement increase with dilution
and data are usually restricted to concentrations in the range 5-0.5 g/dm3.
Values of the empirical constants K and a have been determined exper-
imentally by measurements of [17] for polymer samples of known mole-
cular masses which themselves have been determined using one of the
absolute methods. For most polymer/solvent systems reference can be
made to tabulated values. It should be emphasized that K and a values
vary with polymer type, molecular mass range, solvent and tempera-
ture so that care must be exercised in interpreting values of molecular
mass from solution viscosity measurements. A polymer in a poor solvent
will have a value of a of ~ 0.5; as the quality of the solvent improves
 Secondary methods 37

(7

Fig. 2.10 Ubbelohde viscometer.

the value of a moves towards unity. Some typical values of K and a

taken from [7] are given in Table 2.1.

Table 2.1

Polymer Solvent Temp. K X 106 a Molecular mass

°C [dm3g-1] range x10~4

Polystyrene Tetrahydrofuran 25 11 0.725 100

Polystyrene Tetrahydrofuran 25 13.6 0.714 4000
LDPE p-xylene 75 135 0.63 0.2-7.6
LDPE p-xylene 81 105 0.63 1-10
HOPE Decalin 135 67.7 0.67 3-100
HOPE Decalin 135 46 0.73 3-64
HOPE Trichlorobenzene 135 95.4 0.64 45
HOPE Trichlorobenzene 135 51 0.706 123
Nylon 6 Formic acid -10 26.8 0.82 12
Nylon 6 Formic acid 0 24.8 0.82 12
Nylon 6 Formic acid 10 23.4 0.82 12
Nylon 6 Formic acid 20 75 0.7 0.45-1.6

For many practical situations manually operated viscometers such as the

Ubbelohde capillary viscometer are adequate and are relatively simple to
operate. It is, of course, quite feasible to measure the flow times automat-
ically using photoelectric techniques, and to employ automatic filling and
diluting of the solution, and practical systems have been reported.
Automatic on-line viscometers have also been reported in which the
viscosity 77 is determined by accurately monitoring the pressure drop
AP across a capillary of known length, L and radius, R, and applying
Poiseuille's formula
38 Molecular mass determination

dV _ [7rR4AP]
dt ~ 87?L
where dV/dt is the output flow rate.
Instruments based on this principle find use particularly in gel perme-
ation chromatography (Section 2.3.2) for continuously monitoring the
eluent, enabling the molecular mass distribution to be determined.

2.3.2 GEL PERMEATION CHROMATOGRAPHY - SIZE EXCLUSION

CHROMATOGRAPHY
Although not a primary method for determining molecular masses, gel
permeation chromatography (GPC) or size exclusion chromatography
(SEC) has developed into one of the most useful methods for routine
determination of average molecular masses and molecular mass distrib-
utions of polymers. It is the latter facility which makes the technique so
attractive to polymer scientists and technologists and, allied to the devel-
opment of on-line microprocessor based data handling, has facilitated the
calculation of the molecular mass averages from the size distribution
chromatograms. In addition, only very small sample sizes are required.
One application that exploits this advantage is the study of the effects of
weathering on the surface of polymers. Samples can be taken every 100
microns depth from the surface, permitting examination of molecular
changes due to weathering at different depths. Figure 2.11 shows the
changes in molecular mass distribution occurring as a result of exposure
of polystyrene to UV radiation. Photo-oxidation of polymers causes severe
polymer degradation in sunny climates and the reactions can be followed
using ultraviolet exposure in the laboratory. In most cases photo-oxida-
tion leads to chain scission and this causes the observed deterioration in
mechanical properties. The GPC distributions for polystyrene given in Fig.
2.11 clearly show the reduction in molecular mass with exposure but also
indicate the development of a high molecular mass tail at some depths.
This is attributed to crosslinking, and detailed analysis of the shape of the
molecular mass distribution permits estimation of the ratio of chain scis-
sion to crosslinking reactions.
GPC is a form of liquid chromatography in which the molecules are
separated according to their molecular size or, more correctly, their
hydrodynamic volume in solution. When the solvent has a constant
composition the method is termed isocratic chromatography. The proce-
dure involves injecting a dilute solution of a polydisperse polymer into
a continuous flow of solvent passing through a column containing tightly
packed microporous gel particles. The gel has particle sizes in the range
5-10 jjim in order to give efficient packing and typically possesses a
range of pore sizes from 0.5-105 nm, which correspond to the effective
 Secondary methods 39

1.2

1.0 Polystyrene
Virgin sample
6 weeks UV
_ °'8
^
.2 0.6
c
£
0.4

0.2

3 4 5 6 7

log M

Fig. 2.11 Molecular-size distributions for virgin PS (broken line) and a

degraded sample removed from the surface (0.33 mm) of a bar that was UV-
irradiated for 6 weeks (solid line).

size range of polymer molecules. Separation of the molecules occurs by

preferential penetration of the different sized molecules into the pores;
small molecules are able to permeate more easily through the pores
compared to the larger sized molecules so that their rate of passage
through the column is correspondingly slower. The continuous flow of
solvent leads to separation of the molecules according to size with the
larger molecules being eluted first and the smaller molecules, which have
penetrated more deeply into the pores, requiring longer elution times.
It follows that the time or, more usually volume of elution (Ve), is
inversely proportional to the molecular size. If the pore size is too small
to permit penetration by any of the molecules, or if the pore size is so
large that all of the molecules can penetrate with the same relative ease,
there would be little or no separation of the molecules. Consequently,
selection of the column packing material to have the appropriate pore
size distribution is crucial and different columns are usually required
for polymers having widely different molecular mass distributions. The
availability of gels of mixed pore sizes which can operate over four
decades of molecular mass has made this a less demanding requirement.
The ability of the polymer molecule to enter a pore will depend upon
its size or hydrodynamic volume in solution. Change of solvent and