Advances in Polymer Science 276

Finizia Auriemma
Giovanni Carlo Alfonso
Claudio De Rosa Editors

Polymer
Crystallization I
From Chain Microstructure to Processing
276
Advances in Polymer Science

Editorial Board:

A. Abe, Yokohama, Kanagawa, Japan

A.-C. Albertsson, Stockholm, Sweden
G.W. Coates, Ithaca, NY, USA
J. Genzer, Raleigh, NC, USA
S. Kobayashi, Kyoto, Japan
K.-S. Lee, Daejeon, South Korea
L. Leibler, Paris, France
T.E. Long, Blacksburg, VA, USA
M. M€ oller, Aachen, Germany
O. Okay, Istanbul, Turkey
V. Percec, Philadelphia, PA, USA
B.Z. Tang, Hong Kong, China
E.M. Terentjev, Cambridge, UK
P. Theato, Hamburg, Germany
M.J. Vicent, Valencia, Spain
B. Voit, Dresden, Germany
U. Wiesner, Ithaca, NY, USA
X. Zhang, Beijing, China
Aims and Scope
The series Advances in Polymer Science presents critical reviews of the present and
future trends in polymer and biopolymer science. It covers all areas of research in
polymer and biopolymer science including chemistry, physical chemistry, physics,
material science.
The thematic volumes are addressed to scientists, whether at universities or in
industry, who wish to keep abreast of the important advances in the covered topics.
Advances in Polymer Science enjoys a longstanding tradition and good reputa-
tion in its community. Each volume is dedicated to a current topic, and each review
critically surveys one aspect of that topic, to place it within the context of the
volume. The volumes typically summarize the significant developments of the last
5 to 10 years and discuss them critically, presenting selected examples, explaining
and illustrating the important principles, and bringing together many important
references of primary literature. On that basis, future research directions in the area
can be discussed. Advances in Polymer Science volumes thus are important refer-
ences for every polymer scientist, as well as for other scientists interested in
polymer science - as an introduction to a neighboring field, or as a compilation of
detailed information for the specialist.
Review articles for the individual volumes are invited by the volume editors.
Single contributions can be specially commissioned.
Readership: Polymer scientists, or scientists in related fields interested in poly-
mer and biopolymer science, at universities or in industry, graduate students.
Special offer:
For all clients with a standing order we offer the electronic form of Advances in
Polymer Science free of charge.

More information about this series at http://www.springer.com/series/12

Finizia Auriemma • Giovanni Carlo Alfonso •
Claudio De Rosa
Editors

Polymer Crystallization I
From Chain Microstructure to Processing

With contributions by
R.G. Alamo
R. Androsch
F. Auriemma
X. Chen

S.Z.D. Cheng
C. Cioce
X.-H. Dong
H. Gao

R. Di Girolamo
C.-H. Hsu
M. Huang
W. Hu
Y. Li

H. Liu
B. Lotz
A. Malafronte
V.B.F. Mathot

R.M. Michell
A. Mugica
A.J. Müller
R.A. Pérez-Camargo

C. De Rosa
L. Santonja-Blasco
C. Schick
M. Scoti

H.-J. Sun
J. Wang
C.-L. Wang
X. Yu
K. Yue
X. Zhang

W.-B. Zhang
M. Zubitur
Editors
Finizia Auriemma Giovanni Carlo Alfonso
Department of Chemical Sciences Department of Chemistry and Industrial
University of Naples Federico II Chemistry
Napoli, Italy University of Genova
Genova, Italy

Claudio De Rosa
Department of Chemical Sciences
University of Naples Federico II
Napoli, Italy

ISSN 0065-3195 ISSN 1436-5030 (electronic)

Advances in Polymer Science
ISBN 978-3-319-49201-8 ISBN 978-3-319-49203-2 (eBook)
DOI 10.1007/978-3-319-49203-2

Library of Congress Control Number: 2016961289

© Springer International Publishing AG 2017

This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of
the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or
dissimilar methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt
from the relevant protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publisher nor the
authors or the editors give a warranty, express or implied, with respect to the material contained
herein or for any errors or omissions that may have been made.

Printed on acid-free paper

This Springer imprint is published by Springer Nature

The registered company is Springer International Publishing AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Preface

The APS volumes “Polymer Crystallization: From Chain Microstructure to

Processing” appear about 10 years after the three APS volumes (180, 181, and
191) “Interphases and Mesophases in Polymer Crystallization” edited by Giuseppe
Allegra. The volumes follow a series of workshops on polymer crystallization held
in Genova in 2010, 2012, and 2014, which were triggered by the need to stimulate
debate and share new ideas among leading scientists from academia and industry on
emerging topics related to the crystallization of polymers. We decided to collect
some of these contributions into two APS volumes, eventually including the
contributions of additional authors to fix the new concepts, ideas, and findings
into a unified project reflecting the state of art.
With the development of new theoretical and experimental tools for investigat-
ing matter at the atomic level, significant advances in the understanding of phe-
nomena associated with polymer crystallization have been achieved. However,
elucidating the fundamental physical and chemical issues that govern the crystal-
lization process in a polymer, by which chain molecules move from the melt state to
a semicrystalline state with formation of lamellar crystals, is still a challenge.
The volumes include a wide range of different topics. The first section of
Volume I is related to molecular aspects of polymer crystallization, with chapters
on polymorphism (“Crystallization of Statistical Copolymers”), properties of sta-
tistical copolymers (“Molecular View of Properties of Random Copolymers of
Isotactic Polypropylene”), the crystallization of cyclic polymers (“Crystallization
of Cyclic Polymers”), and precision ethylene copolymers (“Crystallization of
Precision Ethylene Copolymers”). This section ends with a chapter devoted to the
crystallization of giant molecules (“Supramolecular Crystals and Crystallization
with Nanosized Motifs of Giant Molecules”). The second section of Volume I deals
with two different basic aspects of the nucleation process that are also important in
industrial processes: self-nucleation (“Self-nucleation of Crystalline Phases Within
Homopolymers, Polymer Blends, Copolymers and Nanocomposites”) and nucle-
ation at high supercooling (“Crystal Nucleation of Polymers at High Supercooling
of the Melt”).

v
vi Preface

Volume II begins with a section concerning aspects of polymer crystallization

that have often been overlooked in the literature and are related to concomitant
crystallization and cross-nucleation (“Concomitant Crystallization and Cross-
Nucleation in Polymorphic Polymers”), surface-induced epitaxial crystallization
(“Epitaxial Effects on Polymer Crystallization”), and study of the origin of banded
spherulites with nanofocus X-ray diffraction (“Microstructure of Banded Polymer
Spherulites: New Insights from Synchrotron Nanofocus X-Ray Scattering”). The
two latter chapters are illustrative examples of modern investigation of crystal
morphology at the molecular level. The second section of Volume II collects
important issues in industrial application and processing. Topics includes the use
of synchrotron light for studying phase transformation during processing or defor-
mation in real time (“Real-Time Fast Structuring of Polymers Using Synchrotron
WAXD/SAXS Techniques”), the role of amorphous phase in stress-induced crys-
tallization of natural rubber (“Strain-Induced Crystallization in Natural Rubber”),
the influence of cooling rate and pressure on polymer crystallization
(“Non-isothermal Crystallization of Semicrystalline Polymers: The Influence of
Cooling Rate and Pressure”), and the modeling of flow-induced crystallization
(“Modeling Flow-Induced Crystallization”).
We are thankful to all contributors to the project for their high quality work.
These two volumes cover only a few aspects of polymer crystallization, and final
solutions to the big problems in the field have not been assessed. Several topics
covered in the volumes are still under development and need additional in-depth
analyses, checks, and improvements. Nonetheless we hope that the selected topics
will stimulate new discussions, inspire new theories and experiments, intrigue new
followers, and initiate new research in this fascinating world.

Napoli, Italy Finizia Auriemma

Claudio De Rosa
Genova, Italy Giovanni Carlo Alfonso
04 July 2016
Contents

Crystallization of Statistical Copolymers . . . . . . . . . . . . . . . . . . . . . . . 1

Wenbing Hu, Vincent B.F. Mathot, Rufina G. Alamo,
Huanhuan Gao, and Xuejian Chen
Molecular View of Properties of Random Copolymers of Isotactic
Polypropylene . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Finizia Auriemma, Claudio De Rosa, Rocco Di Girolamo,
Anna Malafronte, Miriam Scoti, and Claudia Cioce
Crystallization of Cyclic Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93
Ricardo A. Pérez-Camargo, Agurtzane Mugica,
Manuela Zubitur, and Alejandro J. Müller
Crystallization of Precision Ethylene Copolymers . . . . . . . . . . . . . . . . 133
Laura Santonja-Blasco, Xiaoshi Zhang, and Rufina G. Alamo
Supramolecular Crystals and Crystallization with Nanosized Motifs
of Giant Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183
Xue-Hui Dong, Chih-Hao Hsu, Yiwen Li, Hao Liu, Jing Wang,
Mingjun Huang, Kan Yue, Hao-Jan Sun, Chien-Lung Wang, Xinfei Yu,
Wen-Bin Zhang, Bernard Lotz, and Stephen Z.D. Cheng
Self-Nucleation of Crystalline Phases Within Homopolymers,
Polymer Blends, Copolymers, and Nanocomposites . . . . . . . . . . . . . . . 215
R.M. Michell, A. Mugica, M. Zubitur, and A.J. Müller
Crystal Nucleation of Polymers at High Supercooling of the Melt . . . . 257
René Androsch and Christoph Schick

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 289

vii
Adv Polym Sci (2017) 276: 1–44
DOI: 10.1007/12_2016_349
© Springer International Publishing Switzerland 2016
Published online: 3 September 2016

Crystallization of Statistical Copolymers

Wenbing Hu, Vincent B.F. Mathot, Rufina G. Alamo, Huanhuan Gao,

and Xuejian Chen

Abstract Conventional polymers contain various chemical, geometrical, and

stereo-optical sequence irregularities along the backbone chain, which can be
treated as noncrystallizable comonomers in statistical copolymers. For statistical
copolymers, the link between chemistry (copolymerization to characterize statisti-
cal copolymers) and physics (crystallization to determine structures and properties)
has recently been enhanced. This review discusses how the crystallization behavior
and resulting semicrystalline structure of statistical copolymers are affected by the
various microstructure parameters of their comonomers, such as content, distribu-
tion along or even among polymer chains, and size (determining their inclusion in
or exclusion from the crystallites). The discussion of crystallization is focused on its
interplay with component segregation at three different length scales: monomer,
monomer sequence, and macromolecule. The first two mainly occur in homoge-
neous copolymers, whereas the last one is only operative for heterogeneous copol-
ymers. In addition, some unique phenomena such as strong memory effects and
(cross)fractionation are discussed briefly.

Keywords Component segregation • Crystallization • Heterogeneous •

Homogeneous • Statistical copolymers

W. Hu (*) and H. Gao

Department of Polymer Science and Engineering, State Key Lab of Coordination Chemistry,
School of Chemistry and Chemical Engineering, Nanjing University, 210023 Nanjing, P.R.
China
e-mail: [email protected]
V.B.F. Mathot
SciTe B.V., Ridder Vosstraat 6, 6162 AX Geleen, The Netherlands
R.G. Alamo and X. Chen
Department of Chemical and Biomedical Engineering, FAMU-FSU College of Engineering,
2525 Pottsdamer St., Tallahassee, FL 32310-6046, USA
2 W. Hu et al.

Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
2 Chain Microstructure of Statistical Copolymers: The (Missing) Link to Crystallization
Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.1 Markov Modeling of Copolymerization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2 Determination of Chain-Growth Parameters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.3 Homogeneous and Heterogeneous Copolymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3 Crystallization with Monomer Segregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
4 Crystallization with Monomer-Sequence Segregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
5 Crystallization with Macromolecular Segregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

1 Introduction

Synthetic chemistry provides an economic way to produce polymers with the same
monomer units and regular sequence structures, both of which favor crystallization.
However, a polymer with fully crystalline structure has only hard and brittle
properties, which limit its practical application. Introducing intramolecular defects,
mainly on the backbone chain, is the conventional path for obtaining a useful
semicrystalline product. Sequence irregularities on the backbone can be classified
according to three sources in the detailed polymerization processes: chemical
modifications, geometric isomers, and stereoisomers. For instance, various kinds
of chemical modifications on the backbone have led to important new classes of
commercial polyolefins, such as high impact polypropylene (HIPP), high density
polyethylene (HDPE), low density polyethylene (LDPE), heterogeneous and homo-
geneous linear low density polyethylene (LLDPE), and very low density polyeth-
ylene (VLDPE).
Over the years, every part of the ‘chain of knowledge’ for polymer crystalliza-
tion (meaning the successive steps from polymerization via processing to proper-
ties) has been well studied, but the links between these parts are still weak or even
absent [1]. For instance, studies of the connection between chain microstructure and
crystallization behavior often stopped at the content of sequence defects and
overlooked their detailed distribution, probably as a result of technical difficulties
in obtaining a clear characterization of sequence distribution. However, defect
distribution has a huge impact on the crystallization behavior of polymers, which
is why determination of the distribution should have the full attention of researchers
studying the crystallizability of polymers, especially because crystallization has
appreciable effects on the end properties of the product. During crystallization, the
sequence segments containing defects do not usually match the geometric or spatial
requirements of compact packing in the crystalline ordered regions composed of
crystallizable units. Thus, they behave like noncrystallizable entities and hinder the
development of crystallinity by the polymer main chains. In this sense, they bring
Crystallization of Statistical Copolymers 3

an intramolecular chemical confinement, being inclusion in the case of mobile

defects or exclusion in the case of immobile defects in the crystalline phase,
which depends on the size and rigidity of the defect-containing segments relative
to the crystallizable units. The sequence defects, on the one hand, bring more chain
connections between crystallites to improve the durability of plastic products, but,
on the other hand, suppress crystallinity and in extreme cases give rise to an
amorphous polymer such as atactic polystyrene or atactic poly(methyl methacry-
late) (PMMA). Driven by strong industrial interests, along with modern develop-
ments in the use of well-designed catalysts, NMR characterization, and molecular
simulations, research has stepped forward to examine how the detailed distribution
of sequence defects affects crystallization behavior and the resulting semicrystal-
line structure of synthetic polymers.
Many synthetic polymers show sequence defect distributions that resemble some
statistical features; therefore, their crystallization behaviors can be discussed by
treating them as statistical copolymers produced by typical addition polymerization
following a statistical mechanism. Statistical homogeneous and heterogeneous
copolymers, although distinguished by their differing homogeneity of comonomer
distribution among macromolecules, both exhibit an intramolecular
multicomponent behavior. Upon crystallization, two levels of component
microphase segregation (from small to large scale) may be of relevance: monomer
segregation (according to different chemical species of the chain units) and
monomer-sequence segregation (according to different consecutive lengths of
monomer units along the chain). Heterogeneous copolymers have, in addition, an
intermolecular multicomponent phase-separation behavior (caused by different
comonomer contents and distributions between macromolecules), which could
lead to macromolecular segregation. Various levels of liquid–liquid phase separa-
tion are likely to occur prior to liquid–solid crystallization and, thus, change the
course of the latter. Therefore, the interplay between crystallization and component
segregation is an interesting issue [2] and of importance in understanding better the
complex phase transition behavior of statistical copolymers.
In this review, we survey current understanding of different levels of component
segregation and the resulting crystallization phenomena of statistical copolymers.
After a description of statistical copolymers, the discussion focuses on the occur-
rence of monomer segregation, monomer-sequence segregation, and macromolec-
ular segregation during the process of crystallization of statistical copolymers. For
each type of segregation, we discuss factors of chemical structure such as como-
nomer content, comonomer distribution, and molar mass, as well as the thermody-
namic conditions (mainly temperature). More specifically, we compare various
copolymers of different comonomer mobility in the crystalline phase, which rep-
resent different extents of intramolecular confinement in the crystallization of
statistical copolymers. We hope that with this strategy, the complex crystallization
behavior of statistical copolymers can clearly unfold in front of the reader’s eyes.
4 W. Hu et al.

2 Chain Microstructure of Statistical Copolymers: The

(Missing) Link to Crystallization Behavior

2.1 Markov Modeling of Copolymerization

We have taken commercial polyethylene-based linear statistical copolymers as

typical examples for our discussion. Figure 1 provides an overview of the classifi-
cation of polyethylene products [3–6]. These are not just grades of polyethylene,
but intrinsically new polymers because of their unique properties. Thus, polymer
chemistry plays a crucial starting role, leading to additional large-scale production
for globally demanding markets. In addition, the chain of knowledge from polymer
microstructure to macroscopic properties is continuously being improved by the
sophisticated, state-of-the-art techniques of polymer characterization. The weakest
(and in fact often the ‘missing’) link is usually the relation between the detailed
sequence distribution and its influence on crystallization. In the case of copolymers,
the missing link is knowledge of comonomer distribution in and between chains and
its translation to crystallization behavior, morphology, and so on.
Chemical modifications of polymer chains are often realized through addition or
condensation copolymerizations. A typical example of addition copolymerization

Fig. 1 Classification of polyethylene products. LPE linear polyethylene, UHMWPE ultra-high

molecular weight polyethylene, HDPE high density polyethylene, LDPE low density polyethyl-
ene, SCB short chain branching, LCB long chain branching, LLDPE linear low density polyeth-
ylene, VLDPE very low density polyethylene, EP ethylene–propylene, EB ethylene–1-butene, EO
ethylene–1-octene, EPDM ethylene propylene diene monomer. Adapted from [3]. SciTe ©.
Courtesy of VBF Mathot
Crystallization of Statistical Copolymers 5

Fig. 2 Schematics of
homo- and cross-
propagation reactions in
addition copolymerization

resulting in a product of large market share is LLDPE (‘linear’ indicates the absence
of long-chain branching). Historically, although production started with heteroge-
neous types of LLDPE (resulting, for example, from Ziegler–Natta polymeriza-
tion), homogeneous types of LLDPE were later produced by means of single-site
catalysts, including metallocene-based catalysts. By varying the type of comono-
mer and the number and distribution of noncrystallizable short-branched comono-
mers, LLDPEs can show behaviors ranging from thermoplastic elastomers to hard
plastics [3, 4, 7].
In principle, the sequence distributions of the (co)monomers of LLDPE as
produced by one (or each) catalytic site can be fairly described by a statistical
model on the basis of a simplified stochastic mechanism of addition copolymeri-
zation, which provides a unified classification of copolymers under the name of
statistical copolymers [8]. In addition to copolymerization, linear polymer chains
can be produced by stepwise addition of free monomers or comonomers, while the
choice made depends upon the type of unit at the chain end, the concentrations of
free monomers and comonomers, and their relative reactivity. In such a case, the
chain-propagation process can be treated as a stochastic Markov process. If prop-
agation at polymer chain ends quickly loses the memory of previously added units,
only short-range interactions with the active chain ends are relevant. A first-order
Markov process (called the terminal model) corresponds to the case where the
active ends can be characterized by the last added (co)monomer. A second-order
Markov process (called the penultimate model) corresponds to the case where the
last two (co)monomers have influence, which in specific cases gives a more
accurate description of the complex kinetic process [9], as discussed in more detail
below.
First, we look at the terminal model of copolymerization of monomers and
comonomers on a single-site catalyst. The chain propagation process can be
modeled as shown in Fig. 2, where monomers and comonomers are separately
marked as M1 and M2, and the reaction rate constants are kij (with i, j elements of
{1, 2} to characterize the preference of reaction of monomers and comonomers,
respectively). The asterisks marked at the ends of propagating chains represent the
active unit of polymerization. Then, the propagating probabilities for the formation
of the two successive units are:
6 W. Hu et al.

k11 ½M1 ½M1 * r1 F

P11 ¼ ¼ ; ð1Þ
k11 ½M1 ½M1 * þ k12 ½M2 ½M1 * r 1 F þ 1
P12 ¼ 1  P11 ; ð2Þ
k22 ½M2 ½M2 * r2
P22 ¼ ¼ ; ð3Þ
k22 ½M2 ½M2 * þ k21 ½M1 ½M2 * r 2 þ F
P21 ¼ 1  P22 ; ð4Þ

where r1  k11/k12 and r2  k22/k21 are the reactivity ratios related to monomer and
comonomer incorporation at the chain ends, and F  [M1]/[M2] is the ratio of mole
fractions of monomer and comonomer as representative of the feed composition in
the reaction mixture. In this sense, the probability of adding the consecutive
sequences of monomers with length n (n > 0) is given by P21P11n1P12. When
n ¼ 0, the probability of having two successive comonomers is given by P22.
One could wonder whether the chain-growth preferences can also be described
by the reactivity ratios, as well as by the probabilities, especially when conversion
between the two is simple. Indeed, expressions regarding content and distribution in
a statistical copolymer sample can be described by either using a specific P-set and
the specific mole fraction of monomer x1 (or comonomer x2 ¼ 1  x1) or, alterna-
tively, using a specific r-set and specific F, and the one can be calculated from the
other. The advantage of the P description is that it can be done on a sample without
any additional information and is thus very useful for analysis of a competitor
sample. For a series of homogeneous copolymers (well-polymerized under constant
external conditions and with knowledge of the feeds used), the advantage of the
description with an r-set is that it is a constant, intrinsic property of the catalyst,
implying that is useful for describing the whole series by one common r-set, in
addition to characterizing specific samples by a specific r-set and a specific F. In
practice, this turns out to be feasible and very useful information becomes
available.
The product of two reactivity ratios, r1r2, is decisive for the type of sequence
length distribution of both monomer and comonomer. Values of r1r2 approaching
zero (the alternative description is P11 << x1) and infinity (the alternative is
P11 >> x1) correspond to the two extreme situations of forming alternating copol-
ymers and blocky copolymers, respectively. Between these extremes, a character-
istic value for the r-set is achieved in the case of statistically random copolymers for
which different r1 and r2 values give the product r1r2 ¼ 1 (since r1 6¼ 1, according to
Eq. (1), P11 6¼ x1). Usually, in a first-order Markov (terminal model) process, r1 and
r2 have quite different values that lead to intramolecular heterogeneity during chain
growth. However, in the special case of a zero-order Markov (Bernoullian-like)
process, random copolymers also result for the reason that r1r2 ¼ 1 because r1 ¼ 1
and r2 ¼ 1, and P11 ¼ x1. In the latter case, the distribution of (co)monomer units in
the chains is the same random distribution as the (co)monomer distribution in the
reactor before the copolymerization. Figure 3 shows the distribution of monomer
and comonomer units for a fictive copolymer [1]. One can see that at the same
Crystallization of Statistical Copolymers 7

Fig. 3 Schematics (to be read like a book) of (fictive) ethylene (E)–propylene (P) copolymers
with (a) alternating (rEP  rPE < 1), (b) random (rEP  rPE ¼ 1), and (c) blocky (rEP  rPE > 1)
distributions of the monomer (open squares) and comonomer ( filled squares) units. The values rij,
Pij, and xi are the reactivity ratios, the propagation probabilities, and mole fractions of (co)
monomer incorporated in the chain, respectively. Note that the propylene content is 50% for
each of the three copolymers. Also note that PEE is lower than xE for alternating, equal to xE for
random, and higher than xE for blocky copolymers. Schematics taken from [1]. SciTe_TU/e-F.G.
Karssenberg ©. Courtesy of VBF Mathot

comonomer mole fractions of 50%, the chain structure can be quite different
depending upon whether distribution is alternating, random, or blocky. Clearly,
knowledge of the comonomer content alone is not enough for a copolymer descrip-
tion nor for evaluation of its crystallization behavior. This implies that if the
copolymerization process is not random, one has to use P11 rather than x1, because
the latter could lead to totally wrong results.
8 W. Hu et al.

In Fig. 3, it is assumed that the feed composition F is constant during the

copolymerization process of a specific grade, similar to the steady-state continuous
reaction processes used in industry. In such a case, the sequence distributions are
homogeneous among all the copolymer chains. This leads to a definition of a
homogeneous copolymer [10]: for each sample polymerized, all chains have the
same average (co)monomer content and have the same statistics with regard to the
(co)monomer distribution. To this end, the statistics in each sample are character-
ized by one specific set of chain propagation probabilities during (co)monomer
incorporation in the chains (a P-set) in combination with one mole fraction of the
monomer or comonomer actually incorporated in the chains, xi. Alternatively, a
specific sample can also be characterized by the combination of one specific set of
reactivity ratios (an r-set) and one specific monomer/comonomer ratio F during
copolymerization. For a series of samples, a common set of reactivity ratios should
be applied and then various samples are characterized by various values of F. All
other copolymers, by definition, are heterogeneous, and some possibilities are
discussed later.

2.2 Determination of Chain-Growth Parameters

Before discussing the crucial influence of the statistics of (co)monomer incorpora-

tion on the crystallization and melting behavior, we need to explain how the
reactivity ratios (and propagation probabilities) can be determined from experi-
mental data [11–14].
In early studies of copolymerization, the determination of (co)monomer distri-
bution lacked rigorous treatment. There were three main reasons for this: First, the
best technique available, NMR, was not good enough at the time to provide the
information needed. Second, the limited information (mainly from signals related to
the copolymer) available from NMR (comonomer content, diads, triads, methylene
sequences) was not fully used. These two items frustrated determination not only of
the short sequences but also of longer sequences [15]. Third, the lack of well-
copolymerized and characterized series of homogeneous copolymers led to com-
parison of copolymers of different origins [16], in which copolymers were not
homogeneous or had different statistics such as random, blocky, or alternating. The
result could be best described as restricted to information concerning the smallest
sequences and was therefore not good enough to serve as an input for realistic
crystallization studies.
To improve this situation, the direct 13C-NMR peak method (DPM) [11–13] was
developed to analyze the results of high field, quantitative 13C-NMR [17]. This
method makes use of the total information available in the NMR spectrum. All
relevant peaks are calculated using a Markovian model of the relevant order, an r-
set, and an F. After minimization of the differences between the experimental and
calculated intensities, the chi-square value determines the reliability of the fit. The
Markovian order of the system is also found in this way. An important aspect of
Crystallization of Statistical Copolymers 9

DPM is that it is possible to model a series of samples as a whole. For copolymers of

ethylene and a higher α-olefin with long relaxation times of the atoms of the short
chain branches (e.g., ethylene–1-octene copolymers), quantitative, high signal-to-
noise 13C-NMR spectra were acquired by adding an optimized amount of
chromium(III)-triacetylacetonate as relaxation agent [17] in order to enable DPM.
So far, DPM has been applied successfully to various series of homogenous
copolymers from ethylene–propylene (EP) copolymers to ethylene–1-octene
(EO) copolymers, as polymerized by vanadium and different metallocene catalysts.
This method is not limited to ethylene–1-alkene copolymers but can be used for all
olefin copolymers (e.g., propylene–1-pentene copolymer) [1, 18]. With adjust-
ments, DPM could be generalized for other types of copolymers. It was also
found that DPM is able to unravel chain structures resulting from a mixture of
two homogeneous catalysts [19]. This has opened the way towards understanding
and modeling the chain microstructures of heterogeneous copolymers having
bimodal (co)monomer distributions.
Next, examples are given of detailed and realistic chain microstructures of
copolymers that have been analyzed by using the methods discussed above.

2.3 Homogeneous and Heterogeneous Copolymers

As the first example of the application of DPM in the determination of the structure
of a mainstream commercial polymer, EPDM (ethylene propylene diene monomer)
rubber is discussed (the minor diene is tentatively not included in the following
consideration). For the purpose of mimicking the essence of the molecular structure
of the terpolymer EPDM, a series of 19 EP copolymers with propylene content (xp)
ranging from 0 (linear polyethylene) to 35.3% were prepared using the same
promoted catalyst system, consisting of an aluminum alkyl combined with a
vanadium component. This system has been known to act as a single-site catalyst,
and the copolymers were produced under steady-state conditions and by stepwise
changing of the feed in the reactor. In this case, the occurrence of inversion of the
comonomer complicates the modeling. Normally, a propene unit is incorporated via
1,2 (or primary) insertion. With some catalysts of industrial importance (e.g.,
constraint geometry catalysts), inverted 2,1 (or secondary) insertion is also possible,
which is relatively slow. Thus, in fact, one deals with a ‘degenerated’ terpolymer. In
this case, a first-order Markov description was successfully used [11–13], based on
the methylene sequence length method (MSLM), even though only part of the 13C
NMR data was used. As a result, the following parameters were found (where
subscript 1 stands for ethylene, 2 for inverted propylene, and 3 for normal
propylene):
r12 ¼ 19.6; r13 ¼ 120; r21 ¼ 0.032; r23 ¼ 2.8; r31 ¼ 0.005; r32 ¼ 1 (meaning that
head–head linking of propylenes did not occur: P32 ¼ 0). The reliability was
<90.0%.
10 W. Hu et al.

In this model, if the difference between ‘normal’ and ‘inverted’ propylenes is

removed and everywhere is changed into ‘propylenes’ a kind of product of reac-
tivity ratio, rErP, can be calculated for xP between 15 and 40% [11, 12]:
rE ¼ 17; rP ~ 0.029; rErP ~ 0.50
The value of rErP characterizes the copolymer as slightly alternating.
Later, the analysis using DPM was improved [20] such that all relevant data
were utilized, leading to the following reactivity ratios:
r12 ¼ 20.0 (2.0%); r13 ¼ 162 (16.2%); r21 ¼ 0.015 (6.7%); r23 ¼ 1.3 (7.7%);
r31 ¼ 0.060 (5.0%); r32 ¼ 1.
The reliability was > 99.5 %, illustrating the improvement obtained using DPM.
The second example is a series of EO copolymers, produced using the same
vanadium-based catalyst system with concomitant occurrence of inversion. The
crystallization and melting behavior by simulation are discussed later (see, for
example, Fig. 12). Extensive experimental research results on both EP and EO
series are found in the literature ([4, 6] and related publications). To facilitate a
better understanding of the interaction between chain microstructure and the crys-
tallization behavior of samples of the EO copolymer series, the crystallization and
melting behaviors of one of these (EO-J) were studied using differential scanning
calorimetry (DSC) and were discussed in detail with respect to chain microstructure.
Figure 4 shows the structure of part of a chain of the homogeneous EO copol-
ymer EO-J, as determined using MSLM for first-order Markov modeling of 13C
NMR data [4, 11], taking inversion into account (which is needed to get the full

Fig. 4 Left: Schematics (to be read like a book) of the EO-J (vanadium-based) ethylene-1-octene
copolymer with the same types of comonomer units (either normal or inverted 1-octene; filled
squares). Symbols as in Fig. 3. sn is the number-average sequence length. Right: Normalized
sequence length distributions of EO-J for ethylene and 1-octene on the basis of a degenerated
terpolymer model. There is a wide distribution in the case of ethylene, up to approximately
50 units, and a very narrow distribution for 1-octene, up to sequences of approximately 4 units.
SciTe_DSM ©. Courtesy of VBF Mathot
Crystallization of Statistical Copolymers 11

picture). Subsequently, the difference between ‘normal’ and ‘inverted’ 1-octenes

can be removed and everywhere is changed into ‘1-octenes’, because EO-J has long
sequences and inversion does not influence the crystallization behavior. The value
of rErO is 0.41, so the ‘copolymer’ is between alternating and random. The
inversion percentage is approximately 19%, which is not constant but varies with
the fraction of 1-octene, reaching 100% for poly(1-octene) alone. Figure 5 gives an
overview of DSC cooling and heating curves and the corresponding crystallinity
curves [6].
EO-J clearly shows broad crystallization and melting peaks (see Fig. 5), caused
by the broadness of the ethylene-sequence-length distribution, as calculated by
Monte Carlo simulation (see Sect. 3). DSC studies of a series of 11 samples show
cooling and heating (see Fig. 5) results for samples ranging from amorphous to
highly crystalline. In Sect. 3, Fig. 11 shows the crystallinity curves for both a series
consisting of EO-A with xO ¼ 0.44 to EO-J with xO ¼ 0.115 and the series JW-1121
with xO ¼ 0.80 to the linear polyethylene JW-1114 with xO ¼ 0.
Several more examples on EP copolymers can be found in the literature [21–
23]. As mentioned before, metallocene-catalyst-based propylene–1-pentene copol-
ymers have also been studied successfully, showing that the statistics of the
copolymerization process could be denoted as random because the products rPrrPe
of the first-order Markov reactivity ratios were found to be 1.06 and 1.09 for two
different metallocene catalysts denoted EI and MBI, respectively [18]. In conclu-
sion, the modeling discussed here, including the application of DPM, can be
generalized for all the crystallizable copolymers of interest.

Fig. 5 Left: DSC cooling curves at 10 C/min and related crystallinity curves of ethylene-1-
octene copolymers. The curves of EO-J are indicated [6]. Right: Ditto in subsequent heating at
10 C/min [6]. SciTe_KU Leuven-Sofie Vanden Eynde ©. Courtesy of VBF Mathot
12 W. Hu et al.

In contrast to copolymerization at constant feed, in the case of a batch reaction

process when the monomer reactivity ratios are usually not the same, one compo-
nent is used up more quickly than the other and, thus, the feed composition F drifts
with time. In this case, the comonomer content differs from one chain to another.
Such a product is not homogeneous but is an example of a so-called heterogeneous
copolymer.
Most commercially available heterogeneous copolymers are produced with
multisite Ziegler–Natta catalysts in continuous steady-state processes. The multiple
reaction sites of the catalyst produce copolymers of different comonomer contents
and different sequence length distributions [7–10, 24–26]. It is also characteristic
that the comonomers can influence chain length by increasing the chance of ending
the polymerization process, such that the higher the comonomer content, the shorter
the chain. For example, two different catalytic sites could lead to two (very)
different distributions and, as a result, such heterogeneous copolymers can be
regarded as binary blends. This situation was mimicked by a 1:5 mixture of
metallocene catalysts 1 and 2, resulting from a semibatch polymerization during
which interaction between the two metallocenes was prohibited [19]. Thus, the two
catalysts operate independently and simultaneously in dual-site experiments, cre-
ating the same copolymers as in single-site experiments. Values of 2.02 and 0.03,
respectively, were found for the classical rErP products, demonstrating that catalyst
1 produces an almost random sequence of copolymers and catalyst 2 alternating
copolymers. Melting temperatures of ethylene-type and propylene-type crystallin-
ity could be attributed to one of the two catalysts, indicating the use of this kind of
calculation for analysis of properties.
Comonomer distributions can still be regarded as heterogeneous inside each
chain, even for homogeneous copolymers, except for truly random copolymers
(zero-order Markov statistics) or alternating copolymers. The extreme cases are
gradient copolymers and block copolymers. Their crystallization behaviors com-
bine some features from both random copolymers and heterogeneous copolymers.
Besides the chemical modifications described above, other defects may also be
generated during monomer addition, such as regio- and stereo-defects that confer
geometric or stereo-isomerism to the copolymer chains. For example, a few head–
head additions may join the chain propagation of the dominant head–tail additions
of propene units during polymerization, as in the inversion case of EPDMs
(vanadium-based). Other causes of sequence defects in polypropylene are
stereoenantiomers, leading to isotactic, syndiotactic, or atactic polypropylenes
having totally different crystallizabilities that range from high to zero and result
in different properties. Both isotactic and syndiotactic sequences can be regarded as
crystallizable, depending on the temperature range and time provided for their
crystallization. The hydrogenated polybutadienes (HPBDs) can be regarded as a
random copolymer of ethylene and 1-butene, because a few 1,2 additions accom-
pany the dominant 1,4 addition during the polymerization of polybutadienes. In the
next sections, the outcomes obtained by applying DPM are used as a key input for
the simulations of crystallization and morphology.
Crystallization of Statistical Copolymers 13

3 Crystallization with Monomer Segregation

The phase transition from the liquid state to the crystalline state of copolymers can
be thermodynamically described on the basis of a two-component system com-
posed of crystallizable A and noncrystallizable B building units of the copolymer.
Any theoretical approach to formulation of the equilibrium line as a function of
comonomer composition must assume, a priori, the distribution of the components
between the two phases. The best-known equilibrium theory is Flory’s classical
approach, dating back to 1947. In his consideration, comonomer B was completely
excluded from the crystalline phase of monomer A. In other words, comonomers
are completely segregated from monomers upon crystallization. He treated the
crystalline phase of random copolymers as limited by the monomer sequence
lengths [27–29]. A terminal-type (first-order Markovian type) of chain growth
was considered, and the monomer sequence length was characterized by the
crystallizable sequence propagation probability PAA. For this type of chain, when
the crystalline phase remains pure, the equivalence of chemical potentials between
the crystalline phase and the amorphous phase gives an equilibrium melting
temperature of the copolymer, Tm, expressed as:

1 1 R
 0¼ lnPAA ; ð5Þ
Tm Tm ΔHu

where T0m is the melting temperature of the pure parent homopolymer, R is the gas
constant, and ΔHu is the enthalpy of fusion per repeating unit. This equation gives
the very interesting result that the melting temperature of a copolymer does not
depend directly on its composition, but rather on the nature of the sequence
distribution. The reason for this unique result lies in the chain-like character of
polymers. The chemical potential of a unit in the chain, in either state, depends on
the sequence distribution rather than on the composition [28–30]. Copolymers are
more complex and cannot be treated as simple monomer molecules that contain
isolated impurities. We can analyze three different types of sequence distributions
in terms of xA, the mole fraction of crystallizable units. For an ordered or block
copolymer, PAA >> xA. For such copolymers, there is at most a slight decrease in
the melting temperature compared with that of the corresponding homopolymer.
For an alternating copolymer, PAA << xA and a drastic reduction in the melting
temperature occurs. The prediction that copolymers having the same composition,
but constituted in different ways (see Fig. 3), will have drastically different melting
temperatures has been amply demonstrated by experimental observation of a wide
variety of copolymer types [31–34]. For a truly random copolymer (following zero-
order Markov statistics), PAA ¼ xA and Eq. (5) becomes:
14 W. Hu et al.

1 1 R
 ¼ lnxA ; ð6Þ
Tm Tm0 ΔHu

Coleman found that this treatment for random copolymers is also applicable to
stereoisomer-based copolymers [35], and Allegra et al. re-examined this treatment
based on conditional probabilities [36]. Many experimental observations follow
Eq. (6), without recognition of chemical differences between comonomers, as seen
in Fig. 6 for data extracted from early work on ethylene–1-alkene copolymers

Fig. 6 Melting peak temperatures for rapidly crystallized hydrogenated polybutadienes and
metallocene random ethylene copolymers as a function of branching composition. The weight-
average molar mass for all is 90,000  20,000 g/mol. Data extracted from similar plots in
[7, 37]. The arrow indicates the line calculated on the basis of Tm0 ¼ 418.5 K and
ΔHu ¼ 970 cal/mol in Eq. (6). Courtesy of RG Alamo
Crystallization of Statistical Copolymers 15

[7, 37, 38]. When the comonomer content is small, Eq. (6) applies and ln(xA) ¼ ln
(1-xB)  xB. In Fig. 6, the melting temperatures of rapidly crystallized samples are
plotted versus branching content (half comonomer content) for HPBDs. These are
models for random ethylene–1-butene copolymers as their synthesis ensures very
narrow molar mass, random distribution of ethyl branches, and uniform
intermolecular branching composition. Also plotted are copolymers synthesized
with a metallocene catalyst, having the most probable molar mass distribution,
random sequence distribution, and very narrow comonomer composition distribu-
tion. The equilibrium line is also plotted as reference using T0m ¼ 418.5 K and
ΔHu ¼ 970 cal/mol (ΔHu ¼ 290 J/g), drawn from the homopolymer. The deviation
of this line from the data points implies the nonideality (for instance, the surface
tension) of ethylene-sequence crystals. Fortunately, the data points at small como-
nomer fractions maintain the linear relationship as expected from Eq. (6). Clearly,
there is no significant deviation in the experimental melting temperatures between
the different types of copolymers. Even ethylene–norbornenes with a very bulky
side group fall on the melting versus composition line of the HPBDs. The conclu-
sion can be made that both types of copolymers must display the same behavior
with respect to branch partitioning between crystalline and noncrystalline regions.
If the norbornene cannot be accommodated into the crystalline lattice, the ethyl
branch must also be rejected from the crystal.
Thus, Flory’s treatment is reasonable for those ethylene-based copolymers with
relatively ‘large’ short-chain branches (1-butene, 1-hexene, 1-heptene, 1-octene,
etc.) that cannot enter the compact-packing region composed of crystallizable
ethylene monomers (denoted as ’exclusion’). This phenomenon is not only sub-
stantiated experimentally in many other works [3, 38–43] but also recently via
molecular simulations [44]. An example of the latter is given in Fig. 7, where Tm
values for different copolymers were obtained from the onset of template crystal-
lization on cooling, and the parameter Ep/kTm is plotted versus ln xA. A linear
relation is obtained, as predicted by Eq. (6).
Industrial processing often orients semicrystalline polymers via a strain-induced
crystallization process, such as plastic molding, thin-film stretching, or fiber spin-
ning, in order to gain a high mechanical performance. Flory derived the melting
point of strained polymers [45], which can be combined with Eq. (6) to predict the
melting point of strained random copolymers [46], as given by:

1 1 R
¼  ½lnxA þ f ðsÞ; ð7Þ
T msco T 0m ΔH u

where

1=2

6 1 s2 þ 2s þ 1 1
f ðsÞ ¼ ð s þ 1Þ  þ ;
πN N 2 sþ1

and s is the strain of polymers, and N is the chain length. The combination has been
validated by the onset strains in the simulations of strain-induced copolymer
crystallization, as demonstrated in Fig. 8 [46].
16 W. Hu et al.

0.5

0.4
Ep/kTm

0.3
Equation y = a + b*x
Weight No Weighting
Residual Sum of 5.98262E-5
Squares
0.2 -0.99915
Pearson's r
Adj. R-Square 0.99788
Value Standard Error
Intercept 0.20917 0.00246
Slope -0.84571 0.01744
0.1
-0.3 -0.2 -0.1 0.0
Ln x A

Fig. 7 Reciprocal of melting points versus logarithmic mole fraction of monomers, with the data
points adopted from the equilibrium melting points in dynamic Monte Carlo simulations of
random copolymers [44]. Comonomers were not allowed to slide into the crystalline regions
made up of crystallizable monomers. Straight line is linear fitting of the data points according to
Eq. (6). Courtesy of W-B Hu

0.40
Comonomer fraction
0
0.35 0.06
0.12
( msco)

0.24
0.30
Linear fit
Ec /(kT

0.25

0.20
y=a+b*x Adj. R-Square Intercept a Slope b
All samples 0.96859 0.316 ±
0.001 -0.336 ±
0.009
Homopolymer 0.99575


0.283 ±
0.001 -0.215 ±
0.004
0.15
-0.2 -0.1 0.0 0.1 0.2 0.3 0.4
lnxA+f (s)
Fig. 8 Inverse melting points (Ec is the bending energy of bond connection on the chain) versus
the sum of logarithmic monomer fraction and strain function for homopolymer and random
copolymers with various comonomer mole fractions, as labeled, according to Eq. (7). The long
straight line results from linear regression with the fitting parameters shown in the inset table. The
homopolymer data were also fitted for comparison [46]. Courtesy of W-B Hu

Conversely, the methyl branch in ethylene–propylene copolymers can be par-

tially accommodated in the crystallites, as found in early works [47–50]. If como-
nomers can be included in the crystalline ordered regions composed of
crystallizable monomers (denoted as ‘inclusion’), Colson and Eby suggested that
Crystallization of Statistical Copolymers 17

the comonomers homogeneously distributed in the ordered region can be regarded

as crystalline lattice defects with excess free energy ε [51]. They derived the
expression:


ε
Tm ¼ Tm 0
1 xB : ð8Þ
ΔH u

Following the initial derivations of Helfand and Lauritzen [52], a more elaborated
equation was derived by Sanchez and Eby [53, 54]. The melting temperature of an
infinitely thick crystal of a copolymer with an overall mole fraction of B units, xB,
and a mole fraction of B units in the crystal xC, is given as:
  

1 1 R xC ð1  x C Þ xC
 ¼  ε þ ð1  xC Þ ln þ xC ln : ð9Þ
Tm0 Tm ΔH u RT m ð1  xB Þ xB

When xC reaches the equilibrium value, it becomes:

xB eε=RT
xeq ¼ : ð10Þ
1  xB þ xB eε=RT

For these conditions, the equilibrium melting point of the copolymer becomes:

1 1 R  
ε=RT
 ¼ ln 1  x B þ x B e : ð11Þ
T 0m Tm ΔH u

For the case of exclusion, when ε ¼ 1, Eq. (11) reverts to Eq. (6). For uniform
inclusion, when xB ¼ xC, Eq. (10) reverts to Eq. (8). This situation implies that the
concentration of comonomer inside the crystallites is identical to the comonomer
concentration in the copolymer chain, which is a case rarely encountered in
ethylene-based random copolymers. The results of Monte Carlo simulations of a
hypothetical case of uniform inclusion are consistent with the linearity of Eq. (8)
and are shown in Fig. 9 [55]. Moreover, we need to consider that both equilibrium
approaches [Eqs. (6) and (8)] predict a linear correlation between Tm and xB for
relatively low total xB, as those of interest for LLDPE. Therefore, from experimen-
tal data showing a decrease in Tm with an increase in comonomer content, one
cannot predict whether the B units are included or rejected from the crystallites.
Additional data are required to make a conclusion regarding comonomer
partitioning [56–58].
The melting behaviors of a series of random propylene–1-alkene copolymers,
synthesized with a single-site metallocene catalyst to ensure narrow intermolecular
composition distributions, are also excellent examples of the effect of branch
partitioning and of correspondence with the premise of thermodynamic derivations
based on exclusion or inclusion of the comonomer units.
18 W. Hu et al.

4.5
4.0
3.5
kT m /Ec
3.0
2.5
2.0
1.5
1.0
0.0 0.1 0.2 0.3 0.4 0.5
xB

Fig. 9 Reduced melting point versus the mole fraction of comonomer B obtained from dynamic
Monte Carlo simulations of bulk random copolymers of 16-mers in a 163 cubic lattice box.
Comonomers are mobile in the crystalline region of monomers. Ep/Ec ¼ 1, where Ep is the
compact-packing energy for two parallel-packed bonds, and Ec is the bending energy for two
consecutively connected bonds on the chain; k is the Boltzmann constant. The straight line is
drawn according to Eq. (8) to guide the eye [55]. Courtesy of W-B Hu

A plot of melting temperature versus defect composition is shown in Fig. 10 for

sets of propylene-based copolymers with ethylene (PE), 1-butene (PB), 1-hexene
(PH) or 1-octene (PO) as co-units [59]. Defects for these copolymers are the
comonomers and other stereo- and regio-irregularities in the chain. The groups of
both Baer and Galeski investigated the structure–property relationship of homoge-
neous PH copolymers [60]. De Rosa et al. systematically investigated the influence
of chemical and stereo- and regio-defects on the physical properties of propylene-
based copolymers [61–63]. For these data, the final melting temperature of the
endotherm (Tmf) was plotted instead of the melting peak because the composition of
the melt at Tmf most closely resembles the defect chain composition for any
copolymer. Furthermore, plotting Tmf for copolymers with <15 mol% defects
also ensures that the melting behavior of the same alpha crystals is comparatively
analyzed for these copolymers. The relationship between melting temperature and
composition of these copolymers falls into three groups. For participation of the
1-butene comonomer in the crystalline regions at the highest level [64], their
crystallizable sequence lengths are the longest among the series for any given
comonomer content; hence, they display the highest melting temperatures. The
ethylene comonomer also participates in the crystallization [56–58] but at a lesser
extent than 1-butene; consequently, lower melting temperatures are observed. PH
and PO copolymers with up to ~13 mol% comonomer that are rejected from the
crystallites display the lowest melting temperatures, with no difference between
these two comonomers [59, 65].
The trends in melting points of propylene–1-alkene copolymers are explained on
the basis of phase equilibrium between crystallites of the same crystallographic
Crystallization of Statistical Copolymers 19

160
PE
140 PB
PH
PO
Tmf (oC) 120

100

80

60

40
0 5 10 15 20 25
Defect Content (mol %)

Fig. 10 Final melting temperature (Tmf), as determined by DSC for propylene–ethylene (PE),
propylene–1-butene (PB), propylene–1-hexene (PH), and propylene–1-octene (PO) copoly-
mers melt-crystallized at 23 C and kept at room temperature for 2 weeks. Lines are drawn to
guide interpretation of the experimental data. Re-plotted from published data [59]. Courtesy of
RG Alamo

nature and a composition-dependent melt. As a result of rejection from the crys-

talline lattice, the comonomer composition of the melt coexisting with crystals of
PH or PO is richer than the corresponding compositions of PE or PB. Therefore, on
equilibrium basis, and in reference to the homopolymer, the melting of PO and PH
crystallites is more depressed than the melting of PE or PB.
Thus, DSC measurements are regularly used to observe the crystallization and
melting behaviors of statistical copolymers. Figures 5 [6] and 11 [66] demonstrate
the experimental observations of cooling and heating curves for the crystallinity of
EO copolymers with various comonomer contents. Higher comonomer content
results in lower crystallinity as well as lower crystallization and melting tempera-
tures, demonstrating the depression of crystallization as a result of intramolecular
chemical confinement of sequence defects.
Dynamic Monte Carlo simulations of lattice-model polymers have reproduced
the phenomena above [67]. Three series of statistical copolymers separately
representing random copolymers, slightly alternating copolymers, and heteroge-
neous copolymers with various comonomer contents were considered. In analogy to
the copolymerization of ethylene and 1-octene with a vanadium-based Ziegler–
Natta catalyst (r1 ¼ 24.7 and r2 ¼ 0.017) [68], sequence distributions of homoge-
neous alternating-to-random copolymers resulted from copolymerization via a
continuous reaction with constant feed compositions, whereas that of heteroge-
neous copolymers resulted from a batch reaction with shifting feed compositions.
The crystallinity curves upon cooling and heating of random copolymers with
20 W. Hu et al.

Fig. 11 Crystallinity curves of a set of ethylene-based copolymers with different comonomer

contents from low (JW series) to high (EO series) during cooling and heating processes of DSC
scanning at 10 C/min. Re-plotted from published data [66]. SciTe_KU Leuven-Sofie Vanden
Eynde ©. Courtesy of VBF Mathot

various comonomer contents are shown in Fig. 12, demonstrating the reproduction
of experimental observations.
In the temperature-scanning curves shown in Fig. 12, the crystallinity is relative to
the fraction of crystalline bonds (characterized by the release of latent heat during
melting) in the total amount of monomers (characterized by the monomer mole
fractions multiplied by the total sample mass). In principle, the estimation of
crystallinity should be based on the total amount of crystallizable bonds of mono-
mers, which is different from the total amount of monomers because of their
connections with comonomers. However, the former cannot be accurately estimated
experimentally. Thus, in simulations, the absolutely crystallinity was defined by the
fraction of crystalline bonds in the total amount of crystallizable monomer bonds.
Figure 13 shows the parallel simulation results of absolute crystallinity for compar-
ison with Figs. 11 and 12. An interesting observation is that, although the experi-
mentally defined crystallinity is suppressed by comonomers present, the absolute
crystallinity of the copolymers is hardly influenced by the comonomer content.
Nearly 80% of crystallizable bonds are eventually transferred into crystalline states
at low temperatures. This result implies that crystallization exhausts the local
crystallizable bonds, and that the observed depression of crystallinity is mainly a
result of our improper but pragmatic definition of relative crystallinity.
The significant hysteresis between heating and cooling curves reflects the
nucleation-controlled process of crystallization. If the copolymers are reheated
directly from a middle temperature of the cooling process, as demonstrated in
Fig. 13, a temporally continuing increase in crystallinity is observed at the begin-
ning of reheating. This is because the crystals can still grow once the nucleation
process has been initiated. The homogeneous slightly alternating copolymers
Crystallization of Statistical Copolymers 21

Random copolymers
1.0
first cooling
0.0
0.8 then heating
Relative crystallinity 0.06
0.12
0.6
0.24
0.36
0.4
0.44

0.2

0.0
1 2 3 4 5 6
kBT/Ep

Fig. 12 Crystallinity curves of a set of homogeneous random copolymers during cooling (solid
lines) and reheating (dashed lines) processes of molecular simulations. The crystalline bonds were
defined as those monomer bonds containing more than five parallel neighbors of the same species
(which also occurs in the melt, giving positive responses). The comonomer mole fractions are
marked below the corresponding curves. The arrows show the direction of temperature evolution
[67]. Courtesy of W-B Hu

Random copolymers
1.0

0.8
Absolute Crystallinity

0.6
0.0
0.06
0.4 0.12
0.24
0.36

0.2 0.44

0.0
1 2 3 4 5 6
kBT/Ep

Fig. 13 Absolute crystallinity curves of a set of homogeneous random copolymers during cooling
(solid lines) and reheating (dashed lines) processes in molecular simulations. The comonomer
mole fractions are marked below the corresponding curves. The arrows show the direction of
temperature evolution [67]. Courtesy of W-B Hu

exhibit similar behaviors to random copolymers (as shown in Fig. 14), indicating
the common features of homogeneous copolymers.
With an increase in comonomer content, crystallite morphologies of homoge-
neous copolymers fade out dramatically, as schematically illustrated in Fig. 15
[1, 6]. The crystallites decay from spherulites in homopolymers to random stacking
22 W. Hu et al.

Slightly-alternating copolymers
1.0

Absolute crystallinity
0.8

0.6
0.0
0.06
0.4 0.12
0.24
0.36
0.2 0.44

0.0
1 2 3 4 5 6
kBT/Ep

Fig. 14 Absolute crystallinity curves of a set of slightly alternating copolymers during cooling
(solid lines) and heating (dashed and dotted lines) processes in molecular simulations. The
comonomer mole fractions are marked below the corresponding curves. The arrows show the
direction of temperature evolution [67]. Courtesy of W-B Hu

Fig. 15 Morphological degeneration of polymer crystals in statistical copolymers with the

increase in comonomer content [69]. SciTe_TU/e-F.G. Karssenberg_KU Leuven-S. Vanden
Eynde ©. Courtesy of VBF Mathot
Crystallization of Statistical Copolymers 23

Fig. 16 Snapshots of a set of homogeneous random copolymers when cooled to a low temper-
ature, as shown in Fig. 12 in molecular simulations. The comonomer mole fractions are (a) 0, (b)
0.06, (c) 0.12, (d) 0.24, (e) 0.36, and (f) 0.44. The amorphous bonds are represented by blue
cylinders, and the crystalline bonds are represented by yellow cylinders [67]. Courtesy of W-B Hu

of lamellar crystals, granular crystals (probably progressing to fringed micelles),

and eventually unstable clusters of crystalline bonds in an amorphous matrix.
Figure 16 shows snapshots of homogeneous random copolymers obtained after
cooling to T ¼ 1.0 in the unit of E_p/k_B, where the crystallization has been
saturated in molecular simulations [67]. Even with a comonomer content as low
as 0.06, the lamellar thickness of polymer crystals is significantly depressed,
demonstrating the effects of chemical confinement of comonomers along the
direction of polymer chains. Similar observations also exist for the series of
homogeneous slightly alternating copolymers with various comonomer contents.
Closer observation of the comonomer distribution in the copolymer matrix reveals
comonomers preferring to accumulate at the peripheries of crystals, as demon-
strated in Fig. 17. With an increase in comonomer content, the morphology of
polymer crystals decays from lamella to granular crystallites. Because the crystal-
lizable bonds are almost completely used by the crystallites at low temperatures,
polymer chains trespass several crystallites to make a network in 3D space. In this
sense, the crystallites act as physical crosslinking points, which enables homoge-
neous LLDPE samples with high enough comonomer content to perform as ther-
moplastic elastomers [70].
In ethylene-based homogeneous copolymers, relatively small comonomers such
as propylene residues still have the possibility of entering the crystalline region of
ethylene sequences [50]. The ability of comonomer units to slide into the crystalline
region determines the effect of chemical confinement along polymer chains, and
thus influences the crystallization behaviors as well as crystal morphologies of
homogeneous copolymers. In molecular simulations, a comparison of the crystal-
lization morphology of homogeneous random copolymers with or without
24 W. Hu et al.

Fig. 17 Zoom-in snapshot of the semicrystalline texture of a slightly alternating copolymer with
comonomer content of 0.12 at T ¼ 1, as shown in Fig. 18. The monomer bonds are drawn as thin
cylinders, and those bonds containing comonomers are drawn as double thicker cylinders
[67]. Courtesy of W-B Hu

Fig. 18 Snapshots of crystallites in homogeneous random copolymers (a) with and (b) without
comonomer sliding restrictions in the crystalline regions of monomers prepared by cooling to
T ¼ 1. Only the crystalline bonds are drawn as small cylinders. The amorphous part is omitted for
clarity. More details can be found in [71]. Courtesy of W-B Hu

comonomer sliding diffusion restrictions has been made, as shown in Fig. 18

[71]. As a result of removal of chain-sliding restrictions, which weaken the chem-
ical confinement along polymer chains, lamellar crystals can be well developed in
Crystallization of Statistical Copolymers 25

the vertical directions, but crystal growth in lateral directions appears less domi-
nant, as demonstrated in Fig. 18b. If comonomers are able to slide into monomer
crystals, the morphology change appears mainly on the thicker lamellar crystals,
although not reaching fully extended sequences. The extended-chain crystals of
polyethylene can mainly be prepared by crystallization via the hexagonal phase
under high pressures [72, 73].

4 Crystallization with Monomer-Sequence Segregation

As shown in Figs. 11, 12, 13, and 14, homogeneous copolymers with higher
comonomer contents exhibit broader temperature ranges of crystallization and
melting. An interesting phenomenon in Figs. 13 and 14 is that, irrespective of the
temperature chosen to stop cooling and start the reheating process, the crystallinity
curves soon converge to the heating curves starting from the lower temperatures.
Similar behavior also exists in the cooling curves for various periods of annealing at
a certain crystallization temperature, as demonstrated in Fig. 19. The annealing
process raises the crystallinity, the extent depending upon the annealing period
(Fig. 19). However, the annealing process disturbs the crystallinity curves on both
cooling and heating in the local temperature range. The crystallinity curves merge
into each other away from the annealing temperature. Similar annealing experi-
ments were performed by Androsch and Wunderlich in their DSC measurements of
EO copolymers, as demonstrated in Fig. 20 [74]. The observations imply that there
exists a pair of master curves separately for crystallization on cooling and for

Slightly-alternating copolymer with comonomer content 0.36

Annealing time /1000 MC cycles

0.8
5
Absolute crystallinity

4
0.6 3
2
0.4 1
0

0.2

0.0
1.0 1.5 2.0 2.5 3.0 3.5
kBT/Ep

Fig. 19 Crystallinity curves of slightly alternating copolymers with comonomer content of 0.36
upon cooling and reheating after they had been cooled to T ¼ 2.0 and then annealed at various time
periods, as denoted. The arrows show the direction of temperature evolution [67]. Courtesy of
W-B Hu
26 W. Hu et al.

Fig. 20 DSC scanning curves of (a) cooling and (b) reheating of ethylene-1-octene copolymers
after annealing at 296 K for different time periods, as labeled in (a) [74]. Reprinted with the
permission of ACS Publisher

melting on reheating of homogeneous copolymers. In a given temperature range,

crystallization of homogeneous copolymers generates crystals with corresponding
thermal stability. The amount of generated crystals is limited by the amount of
monomer bonds available for crystallization in this temperature range.
The master-curve phenomenon for crystallization and melting of homogeneous
copolymers can be attributed to the sequence-length segregation in combination
with crystallization at different temperatures. On cooling, longer monomer
sequences hold stronger crystallization capability and crystallize in a higher prior-
ity. They form thicker crystals at higher temperatures. During reheating, the crystals
of longer monomer sequences with larger lamellar thickness formed at higher
temperatures also melt at higher temperatures. Therefore, any monomer sequences
with specific lengths crystallize and melt in a narrow temperature range, and a wide
distribution of sequence lengths results in the master crystallization and melting
curves over a broad temperature range. The average lengths of monomer sequences
in crystals reflect this kind of partitioning or monomer-sequence segregation during
crystallization and melting processes [40, 41]. A simulation example of a slightly
alternating EO copolymer with comonomer content of 0.24 during cooling is shown
in Fig. 21 [75], demonstrating lower average sequence lengths joining the crystal-
line phase at lower temperatures on cooling.
The local crystallization and on-the-spot melting for separation sequence lengths
can also be observed from the snapshots shown in Fig. 22 [75]. One can see that the
crystallinity obtained at lower temperatures can be mainly attributed to the edges of
existing lamellar crystals, instead of forming new individual lamellae. This kind of
inserting crystal growth in copolymers has been related to the sequence-length
selection in random copolymers [76, 77]. When the temperature is modulated
instead of linearly decreasing or increasing, the formation and melting of the
crystals at the lamellar edge are reversible, which lead to a high value for the
Crystallization of Statistical Copolymers 27

1.0 6

Average length of crystallized sequences

0.8

Absolute crystallinity
4
0.6

0.4
2
Cooling
0.2
Heating

0.0 0
1.0 1.5 2.0 2.5 3.0 3.5
kBT/Ep

Fig. 21 The average sequence length in crystals for a slightly alternating ethylene-1-octene
copolymer with comonomer content of 0.24 and a hard sliding diffusion restriction during the
cooling and reheating process. The sequence length was calculated when more than half of the
crystallizable bonds become crystalline (each containing more than five parallel neighbors of the
same species). The absolute crystallinity curves on cooling and heating are also shown for parallel
simulations [75]. Courtesy of W-B Hu

Fig. 22 Snapshots of an alternating copolymer with comonomer content of 0.24 during cooling
when (a) T ¼ 2.3 and (b) T ¼ 1; and followed by heating back to (c) T ¼ 2.8. The red circles denote
the places of crystal edges for the inserting crystal growth and melting of short sequences
[75]. Courtesy of W-B Hu

reversing heat capacity of random copolymers over a wide temperature range. The
high reversing heat capacity has been treated as evidence of reversible melting at
the lateral surface of copolymer lamellar crystals by Wunderlich and collaborators
[74, 78, 79]. By contrast, for linear polyethylene (LPE) reversible melting was
found at the longitudinal surface of the lamellar crystallites, which is facilitated by
sliding diffusion occurring smoothly in LPE [80, 81].
For homogeneous copolymers, a sliding diffusion restriction of comonomers
stabilizes the crystal interface slightly, resulting in a slight increase in the melting
point of crystallites. By decreasing or removing the comonomer sliding diffusion
28 W. Hu et al.

1.0 10

Average length of crystallized sequences

Without restriction
With restriction
0.8 8

Crystallinity 0.6 6

0.4 4

0.2 2

0.0 0
1.0 1.5 2.0 2.5 3.0
Temperature /Ep/kB

Fig. 23 Crystallinity and average sequence lengths in crystals as a function of temperature on

cooling for homogeneous random copolymers with/without (solid/dashed lines) comonomer
sliding restrictions in the crystalline regions made up of crystallizable monomers. More details
can be found in [71]. Courtesy of W-B Hu

restriction and allowing comonomer inclusion into the crystalline phase, the
monomer-sequence segregation upon crystallization on cooling can be enhanced,
as demonstrated in Fig. 23 [71]. However, at the same time, the crystallization
temperature and the obtained crystallinity both decrease. This result is not obvious
for the weakened chemical confinement along polymer chains in this case. Proba-
bly, allowing the comonomers to slide into crystallites significantly increases the
amount of defects in the crystalline region, but it is a relatively time-consuming
process that delays crystallization.
Based on the nature of sequence-length segregation for the master-curve phe-
nomenon of random copolymer crystallization, various fractionation DSC methods
have been proposed for characterization of the sequence-length distribution of
copolymers [82–84]. The step crystallization method (SC) [85, 86] divides the
sample into fractions obtained by isothermal crystallization during stepwise cooling
from high to low temperatures, as illustrated in Fig. 24. The fractions differ
according to their crystallizability and thus chains are separated according to the
presence of segments with those monomer sequence lengths that are capable of
crystallization at the isothermal temperature set. The only way to produce high
quality thermal fractions by SC is to employ a very long fractionation time (over
1 week), which can induce sample degradation. A much faster and more efficient
way to obtain thermally fractionated samples is provided by the successive self-
nucleation and annealing method (SSA) [83]. In this thermal fractionation method,
successive heating and cooling cycles are applied to promote faster molecular
segregation and annealing at each selected fractionation temperature. As a result,
ethylene–α-olefin copolymers, for instance, can be fractionated in minutes,
depending on the heating/cooling rates employed, and the fractions have better
resolution than those obtained by SC (see Fig. 25 and a recent review on SSA [84]).
Speeding up of measurement has also been explored and up to 50 C/min have been
Crystallization of Statistical Copolymers 29

Fig. 24 Illustration of step-crystallization fractionation method that collects the particular

sequence length fractions by isothermal crystallization at a series of step-down temperatures. (a)
The temperature program for step-crystallization fractionation. (b) Comparison of DSC heating
curves for ethylene-1-octene copolymers (4.2% comonomer mole fraction) resulting from step
crystallization (2) and from conventional continuous cooling at 10 C/min (1) [83]. Reprinted with
permission from Wiley Publisher

obtained, and even higher rates with HPer DSC and Flash DSC. These fractionation
methods are based on the crystallization capabilities of sequence lengths and can
also be applied to compare differences in comonomer distributions of heteroge-
neous copolymers.
Crystallization temperatures of polymers on cooling usually depend on their
thermal history, such as the presence of crystalline residues, foreign surfaces for
nucleation, oriented segments, or less entangled melt [87–91]. Memory of crystal-
lization in the melt can raise the subsequent crystallization temperature, increase
the crystallization rate, decrease the crystal sizes, and even change the crystalline
lattice [92–95]. Such memory effects are typically erased by annealing of polymers
above their equilibrium melting points. Recently, the strong memory effects of
crystallization in homogeneous copolymers with relatively low comonomer content
were observed, even when the annealing temperature was above the equilibrium
melting point of the copolymers [96]. Illustrated in Fig. 26 are data for HPBDs that
are models for random ethylene–1-butene (EB) copolymers. A series of HPBDs
with a fixed content of ethyl branches of ~2.2 mol% and molar masses ranging from
800 to 420,000 g/mol were studied [96]. These samples were first cooled from
200 C to 40 C at a rate of 10 C/min to develop a standard crystalline state. Then,
the samples were heated to different annealing temperatures (Tmelt) for 5 min. The
second-round crystallization temperatures (Tc_peak) on DSC cooling curves were
compared to test for the memory effect of crystallization (see Fig. 26a). For low
molar mass samples, Tc peak is independent of Tmelt; hence, no memory effect can be
observed above the equilibrium melting point (Fig. 26b). However, as shown in
Fig. 26c, a strong memory effect is observed for copolymers of high molar mass,
30 W. Hu et al.

Fig. 25 Illustration of successive self-nucleation and annealing (SSA) fractionation method that
collects the particular sequence-length fractions by avoiding primary nucleation upon isothermal
crystallization of that sequence length. (a) The temperature program for successive self-nucleation
and annealing fractionations. (b) Comparison of DSC heating curves for hydrogenated
1,4-butadiene polymer resulting from the SSA method and conventional continuous cooling
[83]. Reprinted with permission from Wiley Publisher

even when the annealing temperatures are above the equilibrium melting point. The
onset of melt memory (Tonset) is the highest annealing temperature for which a
subsequent crystallization is enhanced. The strength of the melt memory of ethyl-
ene copolymers is measured by the difference between Tonset and the equilibrium
melting temperature of the copolymer. The latter can be calculated according to
Eq. (6). Tonset is plotted versus the weight-average molar mass in Fig. 26d, giving a
critical molar mass of about 1,311 g/mol for such a strong memory effect. This
threshold length for preventing strong melt memory is coincidentally the critical
entanglement molecular weight for polyethylene [97–99].
The strong crystallization memory in the melt observed in model HPBDs is a
general feature of random ethylene copolymers. Ethylene–1-alkene copolymers
synthesized with a single-site metallocene catalyst display the same self-nucleation
features as observed in HPBDs. The increase in nucleation density as a result of
melt memory is dramatic, as illustrated by the polarized optical micrographs in
Fig. 27. As seen in Fig. 28, a plot of the difference between Tonset and the
equilibrium melting point against branching content of the copolymer shows a
parabolic-shaped curve with a maximum at ~2 mol%. The seeds that speed up the
crystallization disappear at temperatures about or below the equilibrium melting
point for copolymers with low branching and for copolymers with comonomer
contents above ~4.5 mol% [96, 100].
Crystallization of Statistical Copolymers 31

Fig. 26 (a) Temperature program for study of the memory effect of crystallization in homoge-
neous random copolymers. (b) Crystallization temperature (Tc peak) versus annealing temperature
(Tmelt) for low molar mass hydrogenated polybutadienes with 2.2 mol% ethyl branches in the
second-time crystallization. (c) Same as (b) for high molar mass samples. (d) Plot of upper-limit
annealing temperatures (Tonset) for the memory effect of crystallization in hydrogenated poly-
butadienes of differing molar mass (weight-averaged). The dashed line indicates the equilibrium
melting point of these copolymers following Eq. (6) (137.8 C) [96]. Courtesy of RG Alamo

The strong molar mass effect for melt memory, and the weakening of melt
memory with increasing comonomer content, eventually vanishing for copolymers
that develop low levels of crystallinity, point to a self-nucleation phenomenon, as
shown schematically in Fig. 29. For annealing temperatures between the equilib-
rium melting point and Tonset, the copolymer melt is heterogeneous. Here, seeds
remain as a memory of the crystalline sequence partitioning to form the initial
crystallites. Because residues of ordered structures are not observed by polarized
optical microscopy or in the X-ray diffraction patterns of melts between Tonset and
Tm0copo, the seeds are associated with long ethylene sequences that, albeit molten,
remain in close proximity and are unable to diffuse quickly to the randomized melt.
These seeds speed up subsequent crystallization from such heterogeneous melts
because the critical initial step of monomer sequence partitioning is partially
bypassed.
32 W. Hu et al.

Fig. 27 Polarized optical micrographs of model ethylene 1-butene copolymer P108: (a) Mor-
phology from a homogeneous melt, Tmelt ¼ 200 C, (b) morphology from a heterogeneous melt,
Tmelt ¼ 150 C. The scale bars represent 20 μm [100]. Courtesy of RG Alamo

Fig. 28 Plot of difference between onset temperature for melt memory and equilibrium melting
temperature against branching content of narrow ethylene–1-alkene copolymers and hydrogenated
polybutadienes (HPBDs) [100]. Courtesy of RG Alamo

Conversely, the melts above Tonset are homogeneous because all crystalline
sequences randomize quickly when reaching these annealing temperatures. Crys-
tallization from homogeneous melts is independent of the annealing temperature
and is relatively slow as a result of the much lower number of nuclei. Tonset, the
critical temperature for melt memory, is well defined and easy to obtain experi-
mentally, as shown in Fig. 26.
The stability and extent of melt memory above the equilibrium melting point has
been proven to be a function of the initial level of crystallinity, and is explained by a
Crystallization of Statistical Copolymers 33

Fig. 29 Schematics of the copolymer crystalline structure (left), the heterogeneous melt with self-
seeds (center), and homogeneous melt free of self-seeds (right) [96]. Courtesy of RG Alamo

progressive increase in the entangled topology of the intercrystalline regions of the

copolymer during evolution of the semicrystalline structure [101]. There is a
threshold level of crystallinity for which melt memory can be observed. This is
explained as the need to develop sufficient constraints in the intercrystalline regions
in the process of dragging crystalline sequences to the crystal front. These con-
straints may slow down diffusion of long ethylene sequences, thus preventing
dissolution of self-seeds [101].
The strong memory effects of crystallization in homogeneous random copoly-
mers were confirmed by dynamic Monte Carlo simulations under parallel temper-
ature programs [44]. De Gennes once pointed out that homogeneous random
copolymers may undergo weak segregation in the case of very large Flory–Huggins
interaction parameters [102]. In principle, owning to the sequence-length segrega-
tion, the melting of large crystals of homogeneous random copolymers generates
local regions rich in long monomer sequences, as evidenced in molecular simula-
tions [44]. In consequence, the local equilibrium melting point is increased, which
brings a larger supercooling for the initiation of crystallization on cooling,
appearing as a strong memory effect above the equilibrium melting points of the
original homogeneous melts. Short copolymer chains do not hold sufficient long
monomer sequences for such a crystallization-enhanced segregation as well as the
strong memory effect. The same is true for high comonomer contents.

5 Crystallization with Macromolecular Segregation

In comparison with homogeneous random copolymers, where only the molar mass
distribution is a factor in changing properties at a fixed comonomer content,
heterogeneous copolymers exhibit two parallel distributions: molar mass distribu-
tion and comonomer content across the molar mass distribution (termed bivariate
distribution). The crystallization of heterogeneous copolymers depends on the
34 W. Hu et al.

Fig. 30 (a) Three-dimensional bivariate distribution for commercial broad ethylene–1-hexene

copolymer with average branching content of 1.66 mol%. (b) Temperature of the initial melt
(Tmelt) versus crystallization peak temperature (Tc peak). Closed circles correspond to Tmelt
approached from below. Open circles, Tmelt is approached from above. Regions A, B, C, and
D correspond to regions of different melt structures. The melt undergoes liquid–liquid phase
separation in region C [100]. Courtesy of RG Alamo

breadth of the comonomer distribution and the nature of the bivariate (i.e., whether
mono- or bimodal) [25, 26]. It has been experimentally demonstrated that ethylene
copolymers with broad bimodal distributions and with contents of 1-hexene como-
nomer changing from 1 to 14 mol% across the molar mass distribution display an
interplay between the strong memory effect of crystallization and liquid–liquid
phase separation (LLPS) at relatively high temperatures [100, 103]. This phenom-
enon is illustrated in Fig. 30 for a commercial broad ethylene–1-hexene copolymer.
Bringing the reference crystalline structure to temperatures of 170 C or higher,
subsequent crystallization gives constant Tc peak because the crystallization takes
place from a one-phase homogeneous melt that is free of self-seeds. For Tmelt of
170 C to ~150 C, the crystallization peak increases as a result of remaining self-
seeds, akin to the behavior of narrowly distributed copolymers. For Tmelt of 150 C
down to 125 C, Tc peak decreases, denoting an inversion in the crystallization rate.
The inversion in the rate demarcates the onset of a self-seed-assisted LLPS between
comonomer-rich and comonomer-poor macromolecules. The crystallization rate
decreases in this range of Tmelt because the number of self-seeds decreases. A
fraction of the seeds dissolve as a result of the strong thermodynamic driving force
to diffuse macromolecules to each liquid phase domain in order to equilibrate the
composition of the two phases. The overall crystalline morphology changes accord-
ingly, as documented by polarized optical microscopy in samples cooled from
different melt structures [100]. The comonomer-rich domains that develop when
cooling from a two-phase melt structure were identified by TEM as having a
diameter of about 0.5 μm [100].
The low branching of the high molar mass molecules present in classical
Ziegler–Natta heterogeneous copolymers means that their crystallization and
Crystallization of Statistical Copolymers 35

Heterogeneous copolymers
1.0

0.8
Absolute crystallinity
0.6
0.0
0.06
0.4 0.12
0.24
0.36
0.2 0.44

0.0
2 3 4 5 6
kBT/Ep

Fig. 31 Absolute crystallinity curves for a set of heterogeneous copolymers with various como-
nomer mole fractions, as labeled, during cooling (solid lines) and reheating (dashed lines)
processes. The arrows show the direction of temperature scanning [67]. Courtesy of W-B Hu

melting temperatures do not appear to be very sensitive to the average comonomer

content [24, 104]. This insensitivity of crystallization rate to the average comono-
mer content in heterogeneous copolymers has been observed in both experiments
[105–108] and simulations [67], as demonstrated in Fig. 31 for the simulation
results of temperature scans in comparison to the parallel results for homogeneous
copolymers in Fig. 14. In addition, one can see that with the increase in comonomer
content, heterogeneous copolymers display lower absolute crystallinity.
These features of heterogeneous copolymers – in contrast to those of homoge-
neous random copolymers – can be attributed to macromolecular segregation prior
to crystallization. In simulations, the preparation of heterogeneous copolymers
results in extreme diversity of comonomer content among the molecules present,
much like a binary blend with one component containing few comonomers with
high crystallization capability and another containing many comonomers with very
limited crystallization capability. Figure 32 demonstrates the prior phase separation
that results in crystals accumulated in a local region [67]. More clearly, the
temperature evolution of demixing parameters of comonomers (average occupation
fraction of monomers around each comonomer) demonstrates prior phase separa-
tion at high temperatures before crystallization, as shown in Fig. 33. Because
crystallization of heterogeneous copolymer takes place in the separated phase, the
crystallization and melting temperatures as well as the crystal morphologies of a
heterogeneous copolymer are less sensitive to comonomer content than those of
homogeneous random and slightly alternating copolymers. Prior phase separation
has also been observed by small-angle neutron-scattering experiments on hetero-
geneous LLDPE samples and blends [109–111]. Transmission electron microscopy
(TEM) has made distinct semicrystalline domains (CSDs) visible [112], composed
36 W. Hu et al.

Fig. 32 Snapshots of a set of heterogeneous random copolymers with various comonomer

contents when cooled to T ¼ 2 in Fig. 31 of simulations. The comonomer mole fractions are
separately (a) 0, (b) 0.06, (c) 0.12, (d) 0.24, (e) 0.36, and (f) 0.44. The amorphous bonds are
represented by blue cylinders, and the crystalline bonds are represented by yellow cylinders
[67]. Courtesy of W-B Hu

Fig. 33 Cooling curves of absolute crystallinity and demixing parameters of comonomers for a
heterogeneous copolymer with comonomer mole fraction of 0.36 with (solid lines) and without
(dotted lines) the hard restriction of sliding diffusion of comonomers in the crystalline regions
[67]. Courtesy of W-B Hu

of the least branched molecules in heterogeneous VLDPEs, irrespective of whether

the most branched molecules form a matrix (Fig. 34, right) or are dispersed (Fig. 34,
left). This implies a (massive) segregation on a macromolecular scale. It is striking
that when the CSDs are made up of the least branched molecules (see Fig. 34 right),
Crystallization of Statistical Copolymers 37

Fig. 34 Bottom left: SEM image of a VLDPE (5.7 mol% of 1-octene) after extraction with xylene
at room temperature to remove the most branched molecules, revealing their dispersed,
low-crystalline or amorphous nature within the matrix of mainly the least branched (well-
crystallized, up to HDPE type) molecules. Top left: TEM image of (bulked-fixed/stained by
means of chlorosulfonation in the vapor phase) ultrathin sections of the same VLDPE showing
the dispersed phase and surrounding long and thick lamellae with shorter and thinner lamellae
crystallized on the longer ones. Bottom right: Compact semicrystalline domains (CSDs) of another
VLDPE (8.5 mol% of 1-octene) made up of the least-branched molecules (up to HDPE-type) in a
low-crystalline matrix of the most-branched molecules. Top right: CSDs made visible by a more
severe fixation/staining treatment are connected by single lamellae. From [112]. SciTe_DSM-R.
Deblieck ©. Courtesy of VBF Mathot

they are connected by single lamellae – segregation on a much smaller scale – and
thus a network is established. Heterogeneous copolymers therefore provide us with
a good model of polymer blends to study the interplay between polymer crystalli-
zation and LLPS [113].
Even if the mixing interactions between monomers and comonomers are
athermal, the different crystallizability of comonomer-rich and comonomer-poor
macromolecules alone results in phase separation prior to crystallization. This
conclusion was achieved by the development of classical lattice thermodynamic
theory of polymer solutions to include the consideration of polymer crystallization
[114]. The mixing free energy of binary polymer blends is given by [115]:
38 W. Hu et al.

ΔFmix
¼ n1 lnφ1 þ n2 lnφ2
kT
"

 #
B12 2 1 2 Ep
þ φ 1 r 2 n2 ð q  2Þ þ 1 1 ; ð12Þ
kT q r 2 kT

where n1 and n2 are the number of moles of two component polymers, φ1 and φ2 are
their corresponding volume fractions, r2 is the chain length of crystallizable poly-
mers, q is the coordination number of the regular lattice, and B12 is the mixing
interaction of two component monomers. One can see that because the mixing
entropy of the two polymers in the first two terms is relatively small, even with
B12 ¼ 0, the crystallizability of the second component holding Ep still makes the
mixing free energy remain positive (i.e., the polymers are immiscible). In practice,
this principle explains the incompatibility of isotactic polypropylene (PP) with
either atactic or syndiotactic PP in the melt. Near the equilibrium melting point,
the strong thermal fluctuations of crystalline ordering also enhance LLPS in the
binary blends [116]. On the other hand, the interfaces of two incompatible polymers
facilitate crystal nucleation [117]. The interplay between LLPS and polymer
crystallization brings us rich information to understand better the complex phase
transition behaviors of multicomponent polymer systems [2].
In heterogeneous copolymers, besides the two components with extreme high
and low comonomer contents, the intermediate component with intermediate
comonomer content also plays an important role in tuning the mechanical perfor-
mance of copolymer materials [101]. One commercial example is high-impact
polypropylene, as PP toughened by in-situ blending with ethylene–propylene
copolymers and HDPE [118–120]. The intermediate component can be analyzed
and extracted by methods such as temperature rising elution fractionation (TREF)
and crystallization analysis fractionation (CRYSTAF) [121, 122] but is best
achieved through cross-fractionation schemes [123, 124]. These fractionation
methods are based on the differences in molar mass and/or crystallizability in
solvents of the various components present. Interactive liquid chromatography
has been recently developed to characterize the crystalline and noncrystalline
components of homogeneous and heterogeneous copolymers [124–127].
In simulations [128], the reactivity ratios can be changed to generate heteroge-
neous copolymers containing a significant amount of intermediate components. The
intermediate component mainly acts as an amphiphile between the two extreme
components, which enhances their compatibility at the interfaces, because of its
richness in both monomers and comonomers in sequences. With prior phase
separation, crystallization of copolymers takes place in the monomer-rich domain
surrounded by the matrix of more comonomers. The presence of intermediate
components allows networking of crystallites in the amorphous matrix, which
provides toughness for such types of heterogeneous copolymers [119, 121]. Unlike
traditional amphiphiles, that contain only two blocks, the intermediate component
contains many short sequences of monomers or comonomers, which is more similar
to multiblock amphiphilic copolymers.
Crystallization of Statistical Copolymers 39

For the extreme case of intramolecular-heterogeneous but homogeneous copol-

ymers (e.g., diblock copolymers), microphase separation on the length scale of
monomer sequences competes with crystallization on cooling [129]. The
pre-existing microdomains of diblock copolymers bring a spatial nanoconfinement
effect to the crystallization behavior of monomer sequences, which is another
interesting issue, and whose simulation progress has been recently reviewed [130].

6 Summary

An overview is presented of our current understanding of the complex crystalliza-

tion and melting behaviors of both homogeneous and heterogeneous statistical
copolymers, in combination with possible segregation of components. The statisti-
cal nature of copolymerization and related distributions of monomers and como-
nomers and their sequence lengths (as determined by modeling of the complete 13C
NMR spectrum) is the starting point of the discussion. These parameters are
missing links in the understanding of crystallization behavior, which is why they
are a necessary input for subsequent extensive Monte Carlo simulations of the
crystallization, melting, and component-segregation phenomena of statistical
copolymers. The impact of (co)monomer segregation on copolymer crystallization
can be classified into three length scales for chain microstructures: monomer
segregation, monomer-sequence segregation, and macromolecular segregation.
The former two mainly occur in homogeneous copolymers and the latter mainly
in heterogeneous copolymers. Such segregations are of direct relevance for crys-
tallization because they can give rise to unique phenomena such as strong memory
effects on copolymer crystallization. In addition, we have briefly discussed (cross)
fractionation using methods such as TREF, CRYSTAF, and (TREF-)SEC-HPer
DSC to characterize the microstructure features of heterogeneous copolymers, and
thermal fractionation methods such as step-crystallization and SSA to characterize
the microstructure features of homogeneous copolymers. The interplay between
polymer crystallization and component segregation at various length scales pro-
vides us with a clear clue to understanding the complex phase transition behaviors
of both intramolecular and intermolecular multicomponent copolymer systems.

Acknowledgement We appreciated the helpful discussion with Professor Alejandro J. M€ uller at

the University of the Basque Country UPV/EHU, Spain. Financial support from the National
Natural Science Foundation of China (NO. 21274061 and 21474050) and the Program for
Changjiang Scholars and Innovative Research Team in University is appreciated. RGA acknowl-
edges funding from the USA National Science Foundation under grant No. DMR1105129.
40 W. Hu et al.

References

1. Karssenberg FG (2005) Chain microstructures of homogeneous olefin copolymers and

characteristics of single site catalysis. Thesis Technische Universiteit Eindhoven. doi:10.
6100/IR591592
2. Hu W-B (2013) Polymer physics: a molecular approach. Springer, Wien
3. Mathot VBF (1994) Chapter 5: Thermal characterization of states of matter, pp. 105-167;
Chapter 9: The crystallization and melting region, pp. 231-299. In: Mathot VBF (ed)
Calorimetry and thermal analysis of polymers. Hanser, Munich, pp 231–299
4. Mathot VBF, Scherrenberg RL, Pijpers TFJ, Engelen YMT (1996) New trends in polyolefin
science and technology. Research Signpost, Trivandrum, pp 71–95
5. Mathot VBF, Scherrenberg RL, Pijpers TFJ (1998) Polymer 39:4541–4559
6. Vanden Eynde S. (1999) Homogeneous ethylene-1-alkene copolymers: a study of the crys-
tallization and melting behavior at ambient and elevated pressures. Thesis Katholieke
Universiteit Leuven, pp 1–178
7. Alamo RG, Mandelkern L (1994) Thermochim Acta 238:155–201
8. Ring W, Mita I, Jenkins AD, Bikales NM (1985) Pure Appl Chem 57:1427–1440
9. Coote ML, Davis TP (1999) Prog Polym Sci 24:1217–1251
10. Mathot VBF (1984) Molecular structure of LLDPE. In: Polycon ’84-LDPE Papers and
Proceedings, The Plastics and Rubber Institute, Chameleon, London, pp 1–15
11. Mathot VBF, Fabrie CM, Tiemersma-Thoone GPJM, van der Velden GPM (1985) In: Pro-
ceedings of the international rubber conference, Kyoto, Japan, pp 334–340
12. Mathot VBF, Fabrie CM (1990) J Polym Sci Part B Polym Phys 28:2487–2507
13. Mathot VBF, Fabrie CM, Tiemersma-Thoone GPJM, van der Velden GPM (1990) J Polym
Sci Part B Polym Phys 28:1–15
14. Karssenberg FG, Mathot VBF, Zwartkruis TJG (2006) J Polym Sci Part B Polym Phys
44:722–737
15. Randall JC (1977) Polymer sequence determination, carbon-13 NMR method. Academic,
New York
16. Wunderlich B (1980) Macromolecular physics, vol. 3: crystal melting. Academic, New York,
p 225
17. Adriaensens PJ, Gelan FG, Karssenberg JM, Mathot VBF (2003) Polymer 44:3483–3489
18. Karssenberg FG, Karssenberg FG, Joubert DJ, Assumption HJ, Mathot VBF (2005) Reac-
tivity ratios of metallocene catalysts in copolymerizations of propene and 1-pentene, Chapter
8. In: Karssenberg FG (ed) Chain microstructures of homogeneous olefin copolymers and
characteristics of single site catalysis. Thesis Technische Universiteit Eind-hoven. doi: 10.
6100/IR591592
19. Piel C, Karssenberg FG, Kaminski W, Mathot VBF (2005) Macromolecules 38:6789–6795
20. Karssenberg FG, Mathot VBF (2006) J Polym Sci Part B Polym Phys 44:738–746
21. Hopf A, Kaminsky W, Karssenberg FG, Mathot VBF, Piel C (2005) Macromol Theory Simul
14:289–290
22. Karssenberg FG, Piel C, Hopf A, Mathot VBF, Kaminsky W (2005) Macromol Theory Simul
14:295–299
23. Karssenberg FG, Piel C, Hopf A, Mathot VBF, Kaminsky W (2006) J Polym Sci Part B
Polym Phys 44:747–755
24. Alamo RG, Blanco JA, Agarwal P, Randall JC (2003) Macromolecules 36:1559–1571
25. Vadlamudi M, Subramanian G, Shanbhag S, Alamo RG, Varma-Nair M, Fiscus DM, Brown
GM, Lu C, Ruff CJ (2009) Macromol Symp 282:1–13
26. Vadlamudi M, Alamo RG, Fiscus DM, Varma-Nair M (2009) J Therm Anal Calorim
96:697–704
27. Flory PJ (1947) J Chem Phys 15:684
28. Flory PJ (1949) J Chem Phys 17:223–240
29. Flory PJ (1955) Trans Faraday Soc 51:848–857
Crystallization of Statistical Copolymers 41

30. Baker CH, Mandelkern L (1966) Polymer 7:7–21

31. Casey HK, Elston CT, Phibbs MK (1964) J Polym Sci Polym Lett Ed 2:1053–1056
32. Bastien IJ, Ford RW, Mark HD (1966) J Polym Sci Polym Lett Ed 4:147–150
33. Alamo RG, Domszy K, Mandelkern L (1984) J Phys Chern 88:6587–6595
34. Hosoda KS (1988) Polym J 20:383–397
35. Coleman BD (1958) J Polym Sci 31:155–164
36. Allegra LG, Marchessault RH, Bloemberger S (1992) J Polym Sci Polym Phys Ed
30:809–815
37. Isasi JR, Haigh JA, Graham JT, Mandelkern L, Alamo RG (2000) Polymer 41:8813–8823
38. Chen H-Y, Chum SP, Hiltner A, Baer E (2001) J Polym Sci Polym Phys Ed 39:1578–1593
39. Alamo RG, Mandelkern L (1989) Macromolecules 22:1273–1277
40. Alamo RG, Chan EKM, Mandelkern L, Voigt-Martin IG (1992) Macromolecules 25:6381
41. Alamo RG, Viers BD, Mandelkern L (1993) Macromolecules 26:5740–5747
42. Bensason S, Minick J, Moet A, Chum SP, Hiltner A, Baer E (1996) J Polym 34:1301–1315
43. Alizadeh A, Richardson L, Xu J, McCartney S, Marand H, Cheung YW, Chum S (1999)
Macromolecules 32:6221–6235
44. Gao H-H, Vadlamudi M, Alamo RG, Hu W-B (2013) Macromolecules 46:6498–6506
45. Flory PJ (1947) J Chem Phys 15:397–408
46. Nie Y-J, Gao H-H, Wu Y-X, Hu W-B (2014) Soft Matter 10:343–347
47. Richardson MJ, Flory PJ, Jackson JB (1963) Polymer 4:221–236
48. Baker CH, Mandelkern L (1966) Polymer 7:71–83
49. Ver Strate G, Wilchinsky ZW (1971) J Polym Sci Part A-2 9:127–142
50. Ruiz de Ballesteros O, Auriemma F, Guerra G, Corradini P (1996) Macromolecules
29:7141–7148
51. Colson JP, Eby RK (1966) J Appl Phys 37:3511–3514
52. Helfand E, Lauritzen JI (1973) Macromolecules 6:631–638
53. Sanchez IC, Eby RK (1973) J Res Nat Bur Stand Sec A 73:353
54. Sanchez IC, Eby RK (1975) Macromolecules 8:638–641
55. Hu W-B (2000) J Chem Phys 113:3901–3908
56. Vander Hart DL, Alamo RG, Nyden MR, Kim MH, Mandelkern L (2000) Macromolecules
33:6078–6093
57. Alamo RG, VanderHart DL, Nyden MR, Mandelkern L (2000) Macromolecules
33:6094–6105
58. Clas S-D, McFaddin DC, Russell KE, Scammel-Bullock MV, Peat IR (1987) J Polym Sci
Polym Chem 25:3105
59. Jeon K, Palza H, Quijada R, Alamo RG (2009) Polymer 50:832–844
60. Poon B, Rogunova M, Hiltner A, Baer E, Chum SP, Galeski A, Piorkowska E (2005)
Macromolecules 38:1232–1243
61. De Rosa C, Dello Iacono S, Auriemma F (2006) Macromolecules 39:6098–6109
62. De Rosa C, Auriemma F, Ruiz de Ballesteros O, Resconi L, Camurati I (2007) Macromol-
ecules 40:6600–6616
63. De Rosa C, Auriemma F, Ruiz de Ballesteros O, Resconi L, Camurati I (2007) Chem Mater
19:5122–5130
64. Ruiz-Orta C, Alamo RG (2012) Polymer 53:810–822
65. Ruiz-Orta C, Fernandez-Blazquez JP, Pereira EJ, Alamo RG (2011) Polymer 52:2856–2868
66. Vanden Eynde S, Mathot VBF, Koch MHJ, Reynaers H (2000) Polymer 41:4889–4900
67. Hu W-B, Mathot VBF, Frenkel D (2003) Macromolecules 36:2165–2175
68. Madkour TM, Goderis B, Mathot VBF, Reynaers H (2002) Polymer 43:2897–2908
69. Litvinov VM, Mathot VBF (2002) Solid State Nucl Magn Reson 22:218–234
70. Bensason S, Stepanov EV, Chum S, Hiltner A, Baer E (1997) Macromolecules 30:2436–2444
71. Hu W-B, Karssenberg FG, Mathot VBF (2006) Polymer 47:5582–5587
72. Vanden Eynde S, Rastogi S, Mathot VBF (2000) Macromolecules 33:9696–9704
42 W. Hu et al.

73. Vanden Eynde S, Mathot VBF, Hoehne GWH, Schawe JWK, Reynaers H (2000) Polymer
41:3411–3423
74. Androsch R, Wunderlich B (1999) Macromolecules 32:7238–7247
75. Hu W-B, Mathot VBF (2004) Macromolecules 37:673–675
76. Strobl GR, Engelke T, Meier H, Urban G, Zachmann HG, Hosemann R, Mathot VBF (1982)
Colloid Polym Sci 260:394–403
77. Strobl G (2007) The physics of polymers. Springer, Berlin/Hedelberg, p 208
78. Androsch R, Wunderlich B (2000) Macromolecules 33:9076–9089
79. Wunderlich B (2003) Thermochim Acta 396:33–41
80. Hu W-B, Albrecht T, Strobl G (1999) Macromolecules 32:7548–7554
81. Goderis B, Reynaers H, Scherrenberg R, Mathot VBF, Koch MHJ (2001) Macromolecules
34:1779–1787
82. Starck P (1996) Polym Int 40:111–122
83. M€uller AJ, Arnal ML (2005) Prog Polym Sci 30:559–603
84. M€uller AJ, Michell RM, Pérez RA, Lorenzo AT (2015) Eur Polym J 65:132–154
85. Chen F, Shanks RA, Amarasinghe G (2004) Polym Int 53:1795–1805
86. Keating MY, McCord EF (1994) Thermochim Acta 243:129–145
87. Mamun A, Umemoto S, Okui N, Ishihara N (2007) Macromolecules 40:6296–6303
88. Xu J-J, Ma Y, Hu W-B, Rehahn M, Reiter G (2009) Nat Mater 8:348–353
89. Martins JA, Zhang W, Brito AM (2010) Polymer 51:4185–4194
90. Lorenzo AT, Arnal ML, Sanchez JJ, Muller AJ (2006) J Polym Sci Part B Polym Phys
44:1738–1750
91. Zhang Y-S, Zhong L-W, Yang S, Liang D-H, Chen E-Q (2012) Polymer 53:3621–3628
92. Maus A, Hempel E, Thurn-Albrecht T, Saalwaechter K (2007) Eur Phys J E Soft Matter Biol
Phys 23:91–101
93. Khanna YP, Kumar R, Reimschuessel AC (1988) Polym Eng Sci 28:1607–1611
94. Cho K, Saheb DN, Choi J, Yang H (2002) Polymer 43:1407–1416
95. Cho K, Saheb DN, Yang HC, Kang BI, Kim J, Lee SS (2003) Polymer 44:4053–4059
96. Reid BO, Vadlamudi M, Mamun A, Janani H, Gao H-H, Hu W-B, Alamo RG (2013)
Macromolecules 46:6485–6497
97. Litvinov VM, Ries ME, Baughman TW, Henke A, Matloka PP (2013) Macromolecules
46:541–547
98. Ries ME, Brereton MG, Ward IM, Cail JI, Stepto RFT (2002) Macromolecules
35:5665–5669
99. Graessley WW (1982) Adv Polym Sci 47:67–117
100. Mamun A, Chen X, Alamo RG (2014) Macromolecules 47:7958–7970
101. Chen X, Mamun A, Alamo RG (2015) Macromol Chem Phys 216:1220–1226
102. de Gennes PG (2003) Macromol Symp 191:7–10
103. Ren M, Chen X, Sang Y, Alamo RG (2015) Macromolecular Symposia 356:131-141
104. Alamo RG (2004) Macromol Symp 213:303
105. Krigas T, Carella J, Struglinski M, Crist B, Graessley WW, Schilling FC (1985) J Polym Sci
Polym Phys Ed 23:509–520
106. Mirabella FM, Westphal SP, Fernando PL, Ford EA, Williams JG (2005) J Polym Sci Part B
Polym Phys 26:1995–2005
107. Kim MH, Phillips PJ (1998) J Appl Polym Sci 70:1893–1905
108. Gelfer MY, Winter HH (1999) Macromolecules 32:8974–8981
109. Wignall GD, Alamo RG, Ritchson EJ, Mandelkern L, Schwahn D (2001) Macromolecules
34:8160–8165
110. Alamo RG, Londono JD, Mandelkern L, Stehling FC, Wignall GD (1994) Macromolecules
27:411–417
111. Alamo RG, Graessley WW, Krishnamoorti R, Lohse DJ, Londono JD, Mandelkern L,
Stehling FC, Wignall GD (1997) Macromolecules 30:561–566
112. Deblieck RAC, Mathot VBF (2001) J Mater Sci Lett 34:1779–1787
Crystallization of Statistical Copolymers 43

113. Wang H, Shimizu K, Kim H, Hobbie EK, Wang Z-G, Han CC (2002) J Chem Phys
116:7311–7315
114. Hu W-B, Frenkel D (2005) Adv Polym Sci 191:1–35
115. Hu W-B, Mathot VBF (2003) J Chem Phys 119:10953
116. Ma Y, Hu W-B, Wang H (2007) Phys Rev E 76:031801
117. Ma Y, Zha L-Y, Hu W-B, Reiter G, Han CC (2008) Phys Rev E 77:061801
118. Fu Z, Fan Z, Zhang Y, Feng L (2003) Eur Polym J 39:795–804
119. Chen R-F, Shangguan Y-G, Zhang C-H, Chen F, Harkin-Jone E, Zheng Q (2011) Polymer
52:2956–2963
120. Zhang CH, Chen RF, Shangguan YG, Zheng Q (2011) Chinese J Polym Sci 29:497–505
121. Wild L (1990) Adv Polym Sci 98:1–47
122. Xu J-T, Feng L-X (2000) Eur Polym J 36:867–878
123. Cheruthazhekatt S, Pijpers TFJ, Harding GW, Mathot VBF, Pasch H (2012) Macromolecules
45:5866–5880
124. Cheruthazhekatt S, Pijpers TFJ, Harding GW, Mathot VBF, Pasch H (2012) Macromolecules
45:2025–2034
125. Macko T, Br€ull R, Alamo RG, Thomann Y, Grumel V (2009) Polymer 50:5443–5448
126. Macko T, Br€ull R, Alamo RG, Stadler FJ, Losio S (2011) Anal Bioanal Chem 399:1547–1556
127. Lee D, Shan CLP, Meunier DM, Lyons JW, Cong R, de Groot AW (2014) Anal Chem
86:8649–8656
128. Yang F, Gao H-H, Hu W-B (2012) J Mat Res 27:1383–1388
129. Ma Y, Li C, Cai T, Li J, Hu W-B (2011) J Phys Chem B 115:8853–8857
130. Zha L-Y, Hu W-B (2016) Prog Polym Sci 54–55:232–258
Adv Polym Sci (2017) 276: 45–92
DOI: 10.1007/12_2016_350
© Springer International Publishing Switzerland 2016
Published online: 24 September 2016

Molecular View of Properties of Random

Copolymers of Isotactic Polypropylene

Finizia Auriemma, Claudio De Rosa, Rocco Di Girolamo, Anna Malafronte,

Miriam Scoti, and Claudia Cioce

Abstract The yield behavior during uniaxial drawing of isotactic random copol-
ymers of propene with ethylene (iPPEt), 1-butene (iPPBu), 1-pentene (iPPPe),
1-hexene (iPPHe), and 1-octadecene (iPPOc) is analyzed within the framework of
our current understanding of deformation properties of semicrystalline polymers,
that is, the intrinsic stability of lamellar crystals and related polymorphism phe-
nomena, along with the ability of entangled amorphous chains to transmit stress.
The samples selected for analysis were synthesized using single-site metalorganic
catalysts, are highly stereoregular, and contain small amounts of regiodefects
caused by secondary 2,1 erythro units. Moreover, the interchain and intrachain
distribution of comonomeric units is uniform. In the case of iPPEt copolymers,
samples containing
3.5 mol% stereodefects were also studied. The yield behavior
of these samples depends on the kind and concentration of defects, and is directly
related to the level of inclusion in and exclusion from crystals of the comonomeric
units. Apart from iPPBu copolymer samples with high butene content, the yield
stress of all samples increases with the thickness of lamellar crystals according to a
common trend, regardless of comonomer. In the case of iPPBu copolymers
containing a high concentration of butene units, the yield stress decreases with
increasing lamellar thickness. The increase in yield stress with lamellar thickness is
rationalized in terms of the micromechanical model of crystallographic slips, based
on thermal activation of screw dislocations. The parameters of the model describing
the yield behavior are the critical free energy required to form a screw dislocation
and the shear modulus associated with the slip planes of the dislocations. These
were set as identical to those deduced for isotactic polypropylene homopolymer
samples (iPP) crystallized under different conditions. Study of the yield behavior of

F. Auriemma (*), C. De Rosa, R. Di Girolamo, A. Malafronte, M. Scoti, and C. Cioce

Dipartimento di Science Chimiche, Complesso Monte Sant’ Angelo, via Cintia, 80126 Naples,
Italy
e-mail: [email protected]
46 F. Auriemma et al.

these copolymers extends the use of the dislocation model to a set of samples
crystallized under similar conditions but characterized by differences in
comonomeric unit, degree of crystallinity, lamellar thickness, polymorphism, and
intrinsic flexibility of the chain backbone. The results indicate that for a homoge-
neous class of propene-based copolymers, namely crystallized in the α-form of iPP
under similar conditions, lamellar thickness controls the level of plastic resistance
provided that the concentration of structural irregularities in the crystals is not too
high. iPPBu copolymers with high comonomer concentration do not obey this rule
because of the high level of inclusion of comonomers in the crystals, which induces
an increase in lamellar thickness but also a decrease in crystal stability.

Keywords Random copolymers • Yield behavior • Crystallographic slip process •

Dislocation model • Polymorphism

Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
2 Experimental Details . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53
3 Structural Analysis and Thermal Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4 Mechanical Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
5 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85

1 Introduction

The uniaxial deformation of semicrystalline polymers having spherulitic or lamel-

lar morphology transforms an initially isotropic material into oriented fibers. In
semicrystalline polymers, the complex interwoven structure of amorphous and
crystalline phases that are tightly intermingled [1] entails that the deformation
process is complex [2, 3, 4, 5–9]. Deformation involves movement of atoms or
groups of atoms, both in the crystalline and amorphous phase, in a cascade of events
over different length scales. These range from the length scale of monomeric units,
unit cells, and coils to the thickness of lamellar crystals, assembly of lamellae in
stacks separated by amorphous phase, and spherulites [5, 7, 10–19]. The move-
ments follow a common scheme during uniaxial stretching and involve both elastic
(reversible) deformation and plastic (permanent) deformation as a result of viscous
flow activated by the stress field [3, 5–7, 20, 21].
At low strains, the stress increases linearly and the polymer sample responds to
the applied strain elastically, obeying Hooke’s law [22]. The main mechanisms
behind the Hooke regime of deformation occurring at the subnanometric length
scale are the deformation of covalent bonds and valence angles, librational motion
of internal rotation angles, and reversible deformation of unit cell axes [22, 25]. By
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 47

releasing the tension, the sample recovers its initial shape and dimensions. The
proportionality constant between stress and strain is the Young’s modulus, which
depends mainly on the intrinsic flexibility of the chains and the volume fraction of
the crystalline phase [22–24].
With further increase in strain, the stress within the polymer increases to reach the
yield point. The high mobility achieved at the yield point initiates plastic (irreversible)
deformation in the material, at a rate equal to the applied strain rate. After yielding,
strain softening takes place, leading to strain localization and subsequent necking.
From now on, plastic deformation dominates until break [5, 7, 26, 27].
The molecular mechanisms involved at yield and during plastic deformation up
to the break point depend on the deformation rate, temperature, and crystallization
conditions of the sample [5, 7, 27, 28]. Possible mechanisms associated with
yielding behavior and plastic deformation of a polymer are partial melting and
recrystallization [29, 30], thermal activation of screw dislocations with the Burgers
vector parallel to the chain axis [6, 8, 14–19, 25, 31, 32], cavitation, and
micronecking [3, 20, 21].
The concept that yielding and successive plastic deformation are the result of
strain-induced melting of crystals followed by recrystallization into new crystals in
adiabatic conditions (mechanical melting) was suggested by Yoon and Flory [29]
on the basis of speculative considerations and not on experiments. According to this
concept, the melting of initially unoriented crystals followed by recrystallization of
the molten material into new crystals with a predominant chain orientation induced
by strain is the thermodynamic driving force for plastic deformation, because it
would allow reduction of the local stress level during drawing. Although this
mechanism can explain the decrease in thickness and lateral dimensions of lamellar
crystals during drawing, and the high degree of orientation achieved in a fiber, it
does not account for the yielding process. Indeed, only after yielding does the
stress-induced melting–recrystallization mechanism play some role in the plastic
flow of a polymer [5–7]. The new crystals may either correspond to the same
polymorph initially present in the sample or to a completely different polymorph
[33–44]. It has been shown that occurrence of stress-induced polymorphic transi-
tions during stretching produces a neat increase in ductility, because this provides a
mechanism for conversion of mechanical energy into latent heat of fusion, which
induces local melting of the crystals followed by recrystallization into a new phase
[40–43].
By contrast, the crystallographic model based on thermal activation of screw
dislocations is the most general and important micromechanical model of plastic
deformation in polymers [23, 24, 31, 32, 35–37]. The mechanism of thermal
activation of screw dislocations becomes effective from the yield point through
occurrence of crystal slip processes, assisted by interlamellar slip in the amorphous
layers. Moreover, for some polymers, additional twinning modes or stress-induced
polymorphic transformations can also occur at high deformations [45–52].
Stress-induced phase transition, in turn, may occur not only through mechanical
melting followed by recrystallization [29, 40, 41, 43, 44, 53], but can also be first-
order as in martensitic (displacive) processes [35–45, 54–59]. The crystallographic
48 F. Auriemma et al.

approach correctly predicts the dependence of the yield stress on the stem length of
lamellae, temperature, and strain-rate [16, 60–67].
More precisely, the crystallographic approach to plastic deformation of semi-
crystalline polymers originates from basic ideas borrowed from the classical theory
of crystal plasticity, that is, that the yield stress is governed by the energy required
to nucleate a dislocation within a lamellar crystal [68–72]. Bowden and Young [25]
adopted this idea and demonstrated that the picture based on classical concepts of
nucleation of dislocations and their glide along the crystal lattice agrees well with
the behavior of semicrystalline polymers [6, 8, 12–19, 25, 31, 32, 73–76].
According to this approach, the plastic deformation of polymer crystals is, in
essence, of crystallographic origin and takes place without destroying the crystal-
line order. It occurs by crystallographic slips in the planes of closest packing (slip
planes), generally corresponding to large interplanar distances, in directions coin-
ciding with the direction of the closest packing located in the slip plane [25]. The
slip mechanism is produced by the glide of a linear defect, namely a screw
dislocation, along the slip plane. It can therefore accommodate plastic strains
much more easily than other mechanisms such as twinning or martensitic transi-
tions. The slip begins when the shear stress in the slip direction τ reaches a value
higher than a threshold level that is critical for the given slip system. Such a stress
level τ0 corresponds to the critical resolved shear stress [5–7, 14–19, 25]. Thus,
yielding starts when the critical resolved shear stress is reached in any family of
lattice planes with low τ0.
For polymer crystals, the slip systems need to operate in planes parallel to the
chain axes. The most typical modes are chain slip involving a glide parallel to
chain axes and transverse slip involving a glide perpendicular to chain axes
(Fig. 1A, B). Additional constraints to the crystallographic deformation process
are imposed by chain folds [5]. Chain folds should not be destroyed during
deformation. Therefore, slip processes parallel to planes containing chain folds
are generally preferred.
The slip process during deformation may occur in two different ways, producing
either fine slips (Fig. 1D) or coarse slips (Fig. 1E) [7, 14–19, 49–52, 77–79]. Fine
slips consist of displacements by one or two lattice vectors on every other lattice
plane of a crystal [77–79]. With increased slip processes, the global effect results in
a progressive increase in chain tilting with respect to the lamellar normal and a
decrease in lamellar thickness (Fig. 1D). Coarse slips consist of significant shear
displacements of crystal blocks on well-separated crystal planes. In general, coarse
slips take place in lamellae containing a high concentration of defects or having a
block fine-structure and in the late stages of deformation, when the crystals are
already thinned as a result of advanced fine slip processes [77–79]. Eventually, at
this stage of deformation, lamellae become so weak that they undergo slip insta-
bilities, that is, complete fragmentation, orientation, destruction [32], and recrys-
tallization in oriented crystals of fibrillar morphology.
The third micromechanical model of plastic deformation in semicrystalline
polymers is based on the role of cavitation and micronecking [3, 4, 20, 21,
80]. This model was elaborated by Peterlin and coworkers and assumes that plastic
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 49

Fig. 1 (A, B) Two types of crystallographic slip in macromolecular crystals: chain slip (longitu-
dinal) (A) and transverse slip (B) [6, 25]. Regular chain folds connect adjacent stems in the crystal.
Arrows denote the direction of the chain translation. Slip occurs parallel to chain fold. (C–E)
Deformation modes of lamellar crystals [32]: undeformed crystal (C), models of fine chain slip
(D), and coarse chain slip (E). The orientation of chain axes and the vector normal to lamellar
surface are indicated. (F–H) Deformation modes of the amorphous phase: undeformed stack (F),
interlamellar separation mode (G), and interlamellar slip (shear) (H). In G, H only the chains
bridging adjacent lamellar crystals are drawn

yielding is a result of the shearing of crystalline lamellae followed by their

simultaneous fragmentation into crystal blocks [20, 21, 81]. In practice, the initial
lamellar stacks inside the initial isotropic spherulites transform into fibrillar entities
during stretching through formation of micronecks, which are generated at
microcrack boundaries. Upon fragmentation of the lamellar blocks through chain
unfolding, the blocks become oriented with chain axes in the stretching direction,
originating microfibrils that are characterized by alternation of crystalline and
amorphous regions [20, 21]. The basic mechanism for formation of micronecks is
cavitation, because cavities remove mechanical constraints on block rotations.
Therefore, cavitation and micronecking constitute the basic steps of Peterlin’s
micromechanical model [20, 21]. Successive studies have shown that morpholog-
ical transformation from initial isotropic structures into microfibrils by plastic
deformation can also take place without formation of cavities or microvoids
[7, 32, 82, 83]. This, for instance, occurs in the regime of plane-strain or uniaxial
50 F. Auriemma et al.

compression [84]. Therefore, cavitation is merely a side effect produced by partic-

ular deformation modes and is not essential for the plastic deformation process and
related transformation of polymer morphology [84]. Moreover, Peterlin’s model
completely neglects crystallographic slip processes, and although it describes well
the effect of tensile drawing, it fails completely for other deformation modes such
as compression or plane-strain [7, 84]. By contrast, the crystallographic approach is
more general and can explain the full deformation sequence in any deformation
mode, without invoking any catastrophic events such as micronecking, melting–
recrystallization transformation, or cavitation [1, 2, 5–8].
According to a generalized view, the mechanisms that govern the process of
tensile deformation of semicrystalline polymers at low and moderate deformations
appear strain controlled, rather than stress controlled [33, 34, 66, 67, 85–88]. Within
this scheme, the amorphous phase also plays a key role because it participates in the
plastic flow of a polymer at any deformation, starting from the yield point, as a
result of the high degree of interconnection between crystals and amorphous phase.
This connectivity is ensured by chains crossing the crystal–amorphous interphase
and bridging adjacent lamellae, either through tie chains or entanglements created
by chains emanating from a crystal that re-enter into the same crystal, after passage
through a portion of the adjoining amorphous layer (Fig. 1F) [46, 85]. The principal
deformation modes of the amorphous phase are interlamellar shear and
interlamellar separation (Fig. 1G, H).
In general, the contribution of the amorphous phase to plastic deformation at
yield is small and the contribution of the amorphous phase becomes predominant
only at large deformations, that is, at deformations corresponding to almost com-
plete lamellar fragmentation and consequent transformation of the spherulitic
morphology into fibrillar morphology. In principle, deformation of the intralamellar
amorphous regions at temperatures higher than the glass transition is largely
recoverable, especially at low deformation. This is a result of the rubbery state of
the amorphous phase and the high degree of connectivity of the amorphous phase
with the crystalline scaffold, which hampers viscous flow. Moreover, because of
this connectivity and the intrinsic incompressibility of the amorphous phase, there
is an intrinsic difficulty for the amorphous phase to compensate the deformation
along a given direction with shape distortions in the transversal section, as required
for volume conservation of the rubbery state [5–7, 14–19]. A direct consequence of
this difficulty is that, after yielding, there can be formation of microvoids, lamellar
bending (kinking), and consequent stack rotation because these modes cause
relaxation of the local strain and prevent scission of the tie chains [85–88]. In all
cases, at both large and small deformations, the chains involved in the bridges
between adjacent crystals act as efficient stress transmitters [5, 7, 10, 89, 90] that
facilitate macroscopic deformation of the sample, up to breaking at large
deformations.
More precisely, for deformation temperatures Tdef higher than the glass transi-
tion Tg and immediately after the elastic regime, the plastic deformation of a
semicrystalline polymer starts with small distortions of the amorphous portions of
the chains located between crystals. The compliant amorphous regions are expected
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 51

to deform more easily than the crystals, according to the modes depicted in Fig. 1D,
E [2–7]. However, this deformation is quickly exhausted because of the high
increase in local stress. This stress is transferred to adjacent crystals. Yield starts
as soon as local stress reaches the level of critical resolved shear stress for the
easiest slip system, so that the crystals become involved in plastic flow [2–7, 25–
28]. From this point on, the plastic deformation of crystals begins to control the
whole deformation kinetics of the sample, whereas the amorphous layers respond
trough continuous adjustments of their conformation [2–7, 25, 27, 62–65]. With
increased deformation, the conformation of amorphous chains eventually becomes
so taut that collective movements of the crystals are induced, including fragmen-
tation of lamellae into blocks, complete destruction of the initial morphology, and
rotation of the stacks [62–65, 85–88]. Therefore, the entire deformation process
involves the simultaneous, combined deformation of amorphous and crystalline
components. Crystallographic control dominates until the breakdown of crystallites
[32–34, 66, 67, 83, 91]. Afterwards, strain hardening may intervene at large
deformations prior to breaking [5–7, 26]. Strain hardening is related to the orien-
tational hardening of the amorphous phase and, to a lesser extent, to reorientation of
crystals as a result of crystal slip in the late stages of the deformation process. Along
the true stress–true strain curves of polymeric materials, the compliance changes at
well-defined points corresponding to changes in crystalline morphology and in the
relative response of a material in terms of plastic versus elastic deformation [33, 34,
66, 67, 85–87]. These critical points correspond to:
(A) The onset of isolated inter- and intralamellar slip processes after the initial
Hooke’s elastic range
(B) Change into a collective activity of slip motions of crystal blocks at the point of
maximum curvature of the true stress–strain curve
(C) The beginning of destruction of crystal blocks followed by re-crystallization
with formation of fibrils
(D) The beginning of disentanglement of the amorphous network or strain harden-
ing as a result of stretching of the amorphous entangled network at high
deformations
The values of the strains at critical points A, B, and C are invariant, for each class
of polymer, with variation in crystallinity, temperature, strain rate, and crystal
thickness [33, 34, 66, 67, 85–87]. In contrast to the strain, stresses at the critical
points vary with deformation rate, and present larger values for higher crystallin-
ities and lower values for higher temperatures. These observations comply well
with the general assumption that the strain is homogeneously distributed in semi-
crystalline polymers, whereas the stress is not [48–52, 68, 69]. At low stresses or
strains, the forces transmitted by the interconnected crystallites dominate, whereas
at high strains the rubber-like network forces are superior [68, 69].
The yield point in engineering stretching experiments is always located a little
above point B [85–87]. The position of the critical strain at point C, at which the
critical stress that starts destruction of the crystal blocks is achieved, depends on
the interplay between the entanglement density of the amorphous phase and the
52 F. Auriemma et al.

intrinsic stability of crystals [33, 34, 49–52, 66, 67, 85–87]. A higher entanglement
density implies that a higher stress is generated when the sample is stretched. The
more stable the crystallites, the higher the stress needed for their destruction [49–
52].
Within the framework of the crystallographic approach, models for quantitative
predictions of the yield stress of semicrystalline polymers have been developed,
based on the assumption that yielding involves thermal activation of screw dislo-
cation with the Burgers vector parallel to the chain axis (vide infra) [25, 31, 68, 69,
82, 92–94]. In the resultant model of thermal activation of dislocation, the free
energy required to nucleate a dislocation within the crystalline region has been
correlated with considerable success to the measured yield stress of various samples
of polyethylene (PE) and isotactic polypropylene (iPP) at temperatures higher than
Tg [14–19, 60, 67, 71, 92–96]. In particular, it has been found that at a given
temperature the stress required to initiate these dislocations depends on the thick-
ness of the crystals, which accounts quite well for the observed dependence of yield
stress on crystal thickness for various samples of PE and iPP, regardless of
crystallization conditions, degree of crystallinity, and molecular mass. In particular,
the critical free energy required to form a screw dislocation and the shear modulus
associated with the slip planes of the dislocation can be extracted from this analysis
and correlated with crystallographic features of the material [14–19, 60, 67, 71, 92–
96].
In this chapter, the yield behaviors of isotactic copolymers of propene with
ethylene, 1-butene, 1-pentene, 1-hexene, and 1-octadecene, prepared with different
metallocene catalysts [97–106], are analyzed in terms of the crystallographic
approach using the dislocation model. The catalysts allow synthesis of copolymers
with compositionally uniform chains, uniform distribution of comonomers
along the chain, and tailored microstructure [107–110]. Samples with a very
small concentration of stereodefects or regiodefects and variable amounts of
comonomeric units, or similar concentration of comonomeric units but different
concentration of stereo- and regiodefects, have been prepared. Stereodefects
(namely, isolated rr triads), regiodefects, and different types and concentrations
of comonomeric units have different effects on the crystallization of α- and γ-forms
of iPP, and on crystallization properties in general. The differences in polymor-
phism and crystallization properties, in turn, induce differences in mechanical
properties. The polymorphism and crystallization properties of these systems
depend not only on the concentration of comonomers, which in a random copoly-
mer regulates the average length of the fully (crystallizable) propylene sequences,
but also on the different degree of inclusion of these defects (stereo- and regio-
irregularities, comonomeric units) in the crystals of α- and γ-forms of iPP. The
inclusion of stereo- and regio-irregularities and comonomeric units in the crystal
produces point-like defects and an increase in entropy and/or decrease in internal
energy and, consequently, influences the relative stability of the crystals [97–106,
111–117]. Therefore, study of the yield behavior of these copolymers allows use of
the dislocation model to be extended to a set of samples crystallized under similar
conditions but characterized by differences in the degree of crystallinity, lamellar
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 53

thickness, polymorphism, and intrinsic flexibility of the chain backbone. The values
of these parameters can be finely tailored independently of each other by the type
and concentration of defects by simply selecting a different catalyst system. In
particular, the present investigation aims at establishing the influence of different
degrees of inclusion of point-like defects inside crystals on the parameters of the
dislocation model, namely the critical free energy required to form a screw dislo-
cation and the shear modulus associated with the slip planes of the dislocation
[25, 92–94]. The final goal is to understand the macroscopic properties of materials
at the molecular scale.

2 Experimental Details

The samples selected for this study were iPP homopolymers and propylene-
ethylene (iPPEt) [97, 98], propylene-(1-butene) (iPPBu) [97–99], and propylene-
(1-hexene) (iPPHe) [98, 105, 106] copolymers prepared at temperatures between
60 C and 70 C with the metallocene catalysts A–C shown in Scheme 1, activated
with methylalumoxane (MAO). Samples of propylene-(1-pentene) (iPPPe)
[103, 104] and propylene-(1-octadecene) (iPPOc) [102] were prepared at 25 C
with catalyst D (see Scheme 1). The three C2-symmetric metallocenes A, C, and
D are not completely regioselective, but highly isoselective [107, 108, 110]. The
C1-symmetric metallocene B is fully regioselective but not perfectly isoselective
[109]. The MAO-activated metallocenes A and B for the synthesis of iPPEt and
iPPBu copolymers were supported on spherical SiO2 particles, or on porous poly-
ethylene or polypropylene particles, following a Basell proprietary
technology [118].
All samples are listed in Table 1. The copolymers are designated YZx, where Y
is the catalyst (A, B, C, or D) and x is the concentration of the comonomeric unit Z
(where Z¼E, B, P, H, and O stands for ethylene, 1-butene, 1-pentene, 1-hexene, and
1-octadecene, respectively).
The microstructural data of all samples were obtained from 13C NMR analysis
(see [97–106] for details). The samples of iPP homopolymer prepared with the
catalysts A (iPPA), C, and D (iPPD) are similar. They are highly stereoregular and
contain only small amounts of stereoerrors (0.2 and <0.1 mol% of rr triad defects in
iPPA and iPPD, respectively) and regiodefects caused by secondary 2,1 erythro
units (2,1e) (0.8 and 0.2 mol% of 2,1e units in iPPA and iPPD, respectively). The
iPP sample prepared with catalyst B is highly regioregular (no 2,1 regiodefects
detectable) but less stereoregular, and contains 3.5 mol% rr triads [119].
All catalysts produce copolymer samples with microstructures similar to those of
the homopolymer samples prepared with the same catalyst, with small oscillations
in the concentration of rr stereoerrors and 2,1e regiodefects around those of the
corresponding iPP (Table 1). In particular, for iPPBu copolymers prepared with the
catalyst A the concentration of 2,1e regiodefects decreases with increasing butene
54 F. Auriemma et al.

Scheme 1 Structures of metallocene catalysts A–D used for synthesis of the samples listed in
Table 1

content (Table 1) [97, 98]. For iPPEt and iPPHe copolymers prepared with
catalyst A, the content of stereoerrors is not determinable and is assumed to be
the same as that found in the corresponding homopolymer iPPA (Table 1) [97, 98,
100, 101]. All the copolymers have a random distribution of comonomers and
narrow molecular mass distributions. Details of the NMR analysis are described in
the literature [97–106].
The films used for structural and thermal characterization and for mechanical
tests were prepared by compression molding. Powder samples were heated at
temperatures 20–30 C higher than the melting temperatures between flat brass
plates under a press at low pressure and slowly cooled to room temperature by
fluxing water in the refrigerating circuit of the press plates. Special care was taken
to obtain films of uniform thickness (0.3 mm) and to minimize surface roughness,
according to the recommendation of the standard ASTM D-2292-85.
Calorimetric data were collected with a differential scanning calorimeter (DSC)
Mettler DSC-30 in a flowing N2 atmosphere at heating rate of 10 C/min. All
samples showed a Tg lower than
0 C, which decreased with increasing comono-
mer concentration and length of the side chains [97–106].
Table 1 Details of iPP homopolymers and various copolymers prepared using different MAO-activated metallocene catalysts
Catalyst/cocatalyst/ Mw Mve Comonomer conc.g [rr]h [2,1e]i εtot j
a
Sample carrierb (kg/mol)c Mw/Mnd (kg/mol) Tmf ( C) Comonomer (mol%) (mol%) (mol%) (mol%)
Homopolymers iPP [97, 103]
iPPA A/MAO/PE 237 2.2 – 151 – 0 0.2 0.8 1.00
iPPB B//MAO/PP 247 2.3 – 135 – 0 3.5 0 3.5
iPPD D//MAO/PP – – 680 151 – 0 <0.1 0.2 0.2
(iPPC)
Propene-ethylene copolymers iPPEt [97, 98]
AE0.6 A/MAO/PE – – – 146 Ethylene 0.6 0.2k 0.7 1.5
AE4.0 A/MAO/PE 293 2.1 – 130 Ethylene 4.0 0.2k 0.6 4.8
AE7.4 A/MAO/PE 289 2.1 – 117 Ethylene 7.4 0.2k 0.4 8.0
BE3.6 B/MAO – – – 121 Ethylene 3.6 3.6 0 7.2
BE13.1 B/MAO 193 2.0 – 47, (113)l Ethylene 13.1 3.2 0 16.3
Propene-1-butene copolymers iPPBu [97–99]
AB1.9 A/MAO/SiO2 316 2.2 – 144 1-Butene 1.9 <0.1 0.5 2.4
AB4.3 A/MAO/PE 229 2.1 – 137 1-Butene 4.3 <0.1 0.4 4.7
AB8.3 A/MAO/SiO2 200 2.1 – 130 1-Butene 8.3 <0.1 0.3 8.6
AB13.6 A/MAO/SiO2 161 2.4 – 115 1-Butene 13.6 <0.1 0.1 13.7
CB12 C/MAO – – 165 119 1-Butene 12.0 0.2 0.8 13.0
CB27.6 C/MAO – – 182 88 1-Butene 27.6 0.2 0.8 28.6
CB37.3 C/MAO – – 177 76 1-Butene 37.3 0.2 0.8 38.3
(continued)
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene
55
Table 1 (continued)
56

Catalyst/cocatalyst/ Mw Mve Comonomer conc.g [rr]h [2,1e]i εtot j

a b c d
Sample carrier (kg/mol) Mw/Mn (kg/mol) Tmf ( C) Comonomer (mol%) (mol%) (mol%) (mol%)
Propene-1-pentene copolymers iPPPe [103, 104]
DP3.2 D/MAO 428 1.8 – 126 1-Pentene 3.2 <0.1 0.2 3.4
DP5.3 D/MAO 347 2 – 105, (51)l 1-Pentene 5.3 <0.1 0.2 5.5
DP8.8 D/MAO 279 2 – 85, (51)l 1-Pentene 8.8 <0.1 0.2 9.0
DP11.0 D/MAO – – – 68, (50)l 1-Pentene 11.0 <0.1 0.2 11.2
Propene-1-hexene copolymers iPPHe [100, 101, 105, 106]
AH2.0 A/MAO 122 2.4 – 127 1-Hexene 2.0 0.2 0.2 2.4
AH3.7 A/MAO 333 2.0 – 116 1-Hexene 3.7 0.2 0.2 4.1
AH6.8 A/MAO 239 2.2 – 96 1-Hexene 6.8 0.2 0.1 7.1
AH11.2 A/MAO 266 1.9 – 70 1-Hexene 11.2 0.2 0.1 11.5
Propene-1-octadecene copolymers iPPOc [102]m
DO1 D/MAO – – – 139 1- 1 <0.1 0.2 1.2
Octadecene
DO2.2 D/MAO – – – 117 1- 2.2 <0.1 0.2 2.4
Octadecene
DO4.8 D/MAO – – – 93 1- 4.8 <0.1 0.2 5.0
Octadecene
F. Auriemma et al.
DO6.0 D/MAO – – – 85 1- 6.0 <0.1 0.2 6.2
Octadecene
DO7.5 D/MAO – – – 55, 69n 1- 7.5 <0.1 0.2 7.7
Octadecene
a
iPP homopolymers (iPPA, iPPB, and iPPD) and iPPEt, iPPBu, iPPPe, iPPHe, and iPPOc copolymers were prepared with the MAO-activated metallocenes A–
D shown in Scheme 1. Sample designations are of the form YZx, where Y is the catalyst (A, B, C, or D) and x is the concentration of the comonomeric unit Z
(where Z¼E, B, P, H, and O stands for ethylene, 1-butene, 1-pentene, 1-hexene, and 1-octadecene, respectively)
b
PE polyethylene, PP polypropylene
c
Mass-average molecular mass (Mw) was obtained by GPC
d
Polydispersity Mw/Mn. SEC curves show molecular mass distributions for all samples of around 2.0
e
Values of viscosity average-molecular masses (Mv) were obtained from values of the intrinsic viscosities [η]
f
Melting temperature (Tm) of compression-molded samples crystallized from the melt by slow cooling to room temperature while fluxing cold water in the
refrigerating circuit of the press plates
g
Concentration of comonomeric units (mol%) determined from solution 13C NMR analysis
h
Percentage content of primary stereoerrors over all monomer units, [rr] ¼ [mrrm] + [mrrr]. For copolymers samples prepared with A, C, or D where the
content of stereoerrors is not determinable, the quantity [rr] is assumed to be the same as that found in the homopolymers iPPA and iPPD
i
Concentration of secondary 2,1-erythro units [2,1e]. Secondary insertions 2,1 are only of the erythro type, and their amount is normalized over all monomer
units. For iPPEt copolymers, [2,1e] is the sole concentration of isolated secondary 2,1-erythro units (PSP, P ¼ propylene, S ¼ secondary 2,1 propylene unit)
j
Total concentration of defects (εtot)
k
Content of stereoerrors was not determinable and was assumed to be the same as that found in the corresponding homopolymer iPPA
l
Temperature of a low melting endothermic peak
m
Samples were fractioned with boiling ethoxyethane to remove unreacted 1-octadecene comonomer
n
Sample shows a double melting endotherm
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene
57
58 F. Auriemma et al.

X-ray diffraction patterns (WAXS) were obtained at room temperature with

Ni-filtered CuKα radiation (λ ¼ 1.5418 Å). The powder profiles were obtained
using a Philips diffractometer with continuous scans of the 2θ angle and scanning
rate of 0.02 /s.
The indices of crystallinity (xc, relative error 10%) were evaluated from the
X-ray powder diffraction profiles from the ratio between the crystalline diffraction
area (Ac) and the total area of the diffraction profile (At), xc ¼ Ac/At. The crystalline
diffraction area was obtained from the total area of the diffraction profile by
subtracting the amorphous halo. The procedure used for evaluation of the amor-
phous halo for each sample and for the subtraction is the same as previously
published [97–106].
Small angle X-ray scattering (SAXS) data for compression-molded films were
collected at room temperature using a Kratky compact camera SAXSess (Anton
Paar, Graz, Austria) in the slit collimation configuration, attached to a conventional
X-ray source (CuKα, wavelength 1.5418 Å). The scattered radiation was recorded
on a BAS-MS imaging plate (Fujifilm) and processed with a digital imaging reader
(Fujibas 1800). The range of scattering vector modulus, 0.1 nm1  q  2 nm1,
where q ¼ (4πsinθ/λ) and 2θ is the scattering angle, was analyzed. After subtraction
for dark current, the empty sample holder, and a constant background caused by
thermal density fluctuations, the slit smeared data were de-convoluted with the
primary-beam intensity distribution using the SAXSquant 2.0 software to obtain the
corresponding pinhole scattering (desmeared) intensity distribution. The constant
value of intensity approximating the background Iback was found by fitting the
smeared SAXS intensity curve in the range 2 < q < 4 nm1 [I(qhigh)] with the
function [120, 121]:


I qhigh ¼ I back þ bq3 ð1Þ

where Iback and b are fitting parameters. The average value of the long period L was
calculated as L
2π/q*, where q* is the q value corresponding to the maximum in
the Lorentz-corrected intensity (i.e., the SAXS intensity multiplied by q2/π). Crystal
thickness lc was then calculated by lc
xc L, where xc corresponds to the degree of
crystallinity, as evaluated from WAXS profiles. The thickness of amorphous layers,
la, was evaluated as la ¼ Llc. In practice, we used the mass fraction of the
crystalline phase derived from WAXS analysis instead of the volume fraction,
because the density of amorphous copolymers was not directly determined. There-
fore, even though the calculated values of la and lc are affected by an absolute error,
they are of significance in comparing the properties of the different samples. It is
worth noting that the average values of the long period and lamellar thickness (and
thickness of amorphous layers) evaluated using the one-dimensional correlation
function [120] or the interface distribution function [121] would give similar results
to those evaluated directly from the q values at the maximum of the Lorentz-
corrected scattering intensity.
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 59

Mechanical tests were performed at room temperature on compression-molded

films with a universal testing machine Zwicki (Zwick/Roell), following the stan-
dard test method for tensile properties of thin plastic sheeting (ASTM D882-83).
Rectangular specimens 10 mm long, 5 mm wide, and 0.3 mm thick were stretched
to the break point or to a given deformation ε ¼ [(Lf  L0)/L0]100, where L0 and Lf
are the initial and final lengths of the specimen, respectively. Two benchmarks were
placed on the test specimens and used to check the local elongation versus the
nominal elongation measured from the grip-to-grip distance. In mechanical tests,
the ratio between the drawing rate and the initial length was fixed at 10 mm/
(mm min). The stress–strain curves and the reported values of the mechanical
parameters were averaged over at least five independent experiments.

3 Structural Analysis and Thermal Behavior

The crystallization behavior of isotactic copolymers of propene with ethylene,

butene, pentene, hexene, and octadecene synthesized with single-center
metallocene catalysts has been extensively studied [97–106, 111–117], and com-
pared with that of the corresponding homopolymer iPP produced with the same
catalyst system. Single-center metallocene catalysts allow perfect control over the
chain microstructure [122]. Thus, iPP-based homo- and copolymer samples char-
acterized by different kinds and amounts of defects along the chain can be produced
while maintaining tight control over molecular mass, molecular mass distribution
(with polydispersity index close to two), and uniform inter- and intrachain distri-
butions of the defects. Study of these systems has allowed isolation of the different
influences of each kind of defect, namely stereodefects (isolated rr triads),
regiodefects (e.g., secondary 2,1 insertions of monomeric units), and comonomeric
units, on the crystallization of α- and γ-forms of iPP.
In particular, it has been shown that chain microstructure strongly influences the
polymorphic behavior and physical properties of iPP [40, 41, 97, 100, 111, 114,
119, 123–130]. Samples characterized by chains containing microstructural defects
(stereodefects and regiodefects) and/or comonomeric units, generated by different
catalysts, crystallize as a mixture of the α- and γ-forms (Fig. 2A, B) [40, 41, 97, 100,
111, 114, 119, 123–130]. In general, formation of the γ-form seems favored by the
presence of these defects [40, 41, 97, 100, 111, 114, 119, 123–130]. However, each
kind of defect influences the crystallization of α- and/or γ-forms in a different way
according to different mechanisms. A rational and unified picture of the complex
polymorphism of these systems [97, 100, 102] has been achieved and the general
rules controlling the polymorphism of iPP in defective samples have been
identified.
The first important parameter that influences the crystallization of α- and
γ-forms of iPP corresponds to the average length of the regular isotactic propylene
sequences [40, 41, 119, 123–130]. Short regular isotactic sequences generally favor
60 F. Auriemma et al.

Fig. 2 Structural models of α-form (A) and γ-form (B) of isotactic polypropylene. (C) Trigonal
form of isotactic copolymers of propene with 1-pentene and 1-hexene containing pentene con-
centrations higher than 10 mol% and hexene concentrations higher than 15–16 mol% [100, 101,
103, 104]. The structural model of iPPHe copolymers in the trigonal form is shown as an example.
The lateral butyl groups of 1-hexene units are statistically included in the unit cell with occupancy
factor close to the average content of comonomers in the copolymer chain. (D) New mesomorphic
form of isotactic copolymers of propene with long 1-alkene [102].
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 61

crystallization of the γ-form. For metallocene-made homo- and copolymers of

propene characterized by random distribution of defects along the chains, the
average length of propene crystallizable sequences scales with the reciprocal
concentration of defects. Therefore, even a small number of defects shortens the
average length of the regular isotactic sequences, reducing the melting temperature
and favoring crystallization of the γ-form [40, 41, 119, 123–130]. Because of the
non-parallel arrangement of chains in crystals of the γ-form (Fig. 2B), defects are
easily accommodated at the lamellar boundaries, with no need for chain folding. By
contrast, in the α-form (Fig. 2A) defects are rejected at the fold surface because
chain folding is a necessary requisite for crystallization in order to avoid
overcrowding at the lamellar surface.
A second remarkable effect that drives the crystallization of α- and/or γ-forms is
the possible inclusion of defects in crystals of the two polymorphs [97, 100]. The
inclusion effect favors crystallization of the form that better tolerates the defect
within its crystalline lattice. Therefore, the two effects can either act synergistically
in favoring crystallization in a polymorph, or in competition [97, 100]. In principle,
the interruption effect is common to any defect (stereo- and regiodefects and
comonomeric units) and always favors crystallization of the γ-form. The inclusion
effect produces point-like defects inside the crystals, which may influence the
conformational and packing energy of α- and γ-forms to equal or different extents.
The final crystalline form obtained upon crystallization depends on the effective
level of disturbance of the defects inside the crystalline lattice, which, in the case of
copolymers, is related to the size of the comonomeric units [97, 100, 131].
A third peculiar effect has also been demonstrated in the case of some copoly-
mers that, above a threshold concentration of comonomeric units, tend to crystallize
in a polymorphic form that is different to both α- and γ-forms [100–106]. Crystal-
lization into the new polymorph is driven by the easy inclusion of comonomers
inside crystals of the new form. The process is driven by an increase in entropy and
density [103–106, 132–140] or by kinetic factors [102]. In other words, crystalli-
zation of the new polymorph is competitive with crystallization of the α-form
(expected on the basis of the inclusion effect) and/or the γ-form (expected on the
basis of the interruption effect). We call this effect the “competitive crystallization
effect.”
It has been shown that iPP homopolymer samples with different concentrations
of rr defects [40, 41, 119, 123–130] and samples of iPPEt and iPPBu copolymers
[97, 111] crystallize from the melt as mixtures of the α- and γ-forms. The amount of
γ-form increases with increasing crystallization temperature, ethylene concentra-
tion, and content of rr stereodefects. By contrast, in iPPBu copolymers, the amount
of the γ-form first increases, then decreases for concentrations of butene units higher
than 10–14 mol% and is always lower than that crystallized in stereodefective
iPP and iPPEt copolymers [97, 99]. Therefore, rr stereodefects and ethylene units
favor the crystallization of the γ-form, whereas butene units favor crystallization of
the α- and γ-forms at high and low concentrations, respectively.
These data have been rationalized by resorting to the combined effect of
interruption and inclusion [97, 119]. First, it has been shown that in stereodefective
62 F. Auriemma et al.

iPP, iPPEt, and iPPBu copolymers, different proportions of rr defects, ethylene, and
butene are included in crystals of the α- and γ-forms [112, 113]. On the one hand,
the interruption effect favors crystallization of the γ-form. On the other hand, the
inclusion effect also comes into play and the two effects act simultaneously, with
one prevailing over the other depending on the compatibility of the different defects
within the crystalline lattices of the different polymorphs.
Ethylene and rr stereodefects are included in crystals of both α- and γ-forms but
are more easily included in crystals of the γ-form [97]. In iPPEt copolymers and in
stereodefective iPP samples, the effects of crystal inclusion and shortening the
regular propylene sequences produce the same result of favoring crystallization of
the γ-form [97, 119].
In the case of iPPBu copolymers, butene units are included without differenti-
ation between crystals of the α- and γ-forms, but are more easily included in the
α-form at high concentrations [97]. At low butene concentrations (< 10 mol%), the
effect of shortening the length of regular isotactic propylene sequences prevails and
induces crystallization of the γ-form. Hence, at low concentration of butene units
(for average propene sequences of 10–100 monomeric units), the relative amount of
γ-form increases with increasing butene concentration [97]. For butene concentra-
tions higher than 10 mol%, the effect of inclusion of butene units in crystals of the
α-form prevails over the interruption effect [97]. As a consequence, the relative
amount of γ-form decreases and iPPBu samples with butene concentrations higher
than 20–30 mol% always crystallize in the pure α-form, crystallization of the
γ-form being completely inhibited.
It has been shown that the crystallization of iPPPe [103, 104] and iPPHe
[100, 105, 106] copolymers from the melt produces mixtures of α- and γ-forms at
low pentene or hexene concentrations. For comonomer concentrations higher than a
threshold, they crystallize almost completely into the α-form. Further increase in
comonomer content produces crystallization into the trigonal form of iPP [103, 104,
132–140] (Fig. 2C). This is a result of the high inclusion, at high concentrations, of
pentene and/or hexene units into crystals of the α-form, driven by density increase,
favoring crystallization of the α-form instead of the γ-form [100, 103–106]. There-
fore, the inclusion effect prevails at these comonomer concentrations. The inter-
ruption effect becomes efficient in promoting crystallization of the γ-form only at
very low concentrations of pentene and/or hexene (2–3 mol%) [100, 103–
106]. The trigonal form does not crystallize by cooling the melt but crystallizes
from the amorphous state by cold-crystallization or, for samples with high pentene
or hexene concentration, by aging amorphous samples at room temperature
[100, 103, 104, 132–140]. The hexene or pentene units are included in crystals of
the trigonal form and, at low concentration, also in crystals of the α-form, producing
an increase in the unit cell dimensions. The change in crystallization habit from
monoclinic into trigonal, for pentene concentrations higher than 10 mol% and
hexene concentrations higher than 15–16 mol%, allows incorporation of higher
amounts of monomer in crystals of the trigonal form than in the α-form, and
produces an increase in entropy. Therefore, at high pentene/hexene concentration,
competitive crystallization of the trigonal form prevails.
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 63

In the case of iPPOc copolymers, octadecene units are completely excluded from
the crystals of α- and γ-forms, and only the interruption effect plays a role
[102]. These systems crystallize as mixtures of α- and γ-forms for low octadecene
concentrations, even though the relative amount of γ-form is low, probably because
of kinetics factors and/or the intrinsic tendency of long side chains to favor
formation of chain-folded lamellae of the α-form to alleviate steric hindrance. At
octadecene concentrations above 7–8 mol%, the α-form is also destabilized and
samples crystallize from the melt into a new mesomorphic form [102] (Fig. 2D).
This mesophase is different from the quenched mesomorphic form of iPP homo-
polymer [141]. It is characterized by parallel chains in 3/1 helical conformation
packed at average interchain distances of about 6 Å, defined by self-organization of
the flexible side groups and high degree of disorder in the lateral packing of chains
[102] (Fig. 2D). Therefore, at high octadecene concentrations, the competitive
crystallization effect prevails, leading to formation of the new mesophase instead
of α- and/or γ-forms, probably because the crystallization kinetics of the normal α-
and/or γ-forms become too slow.
It is worth noting that the crystallization conditions (cooling rate, maximum
temperature achieved in the melt, and maximum time that the sample is left at that
temperature) can have a strong impact on the resulting structural and morphological
features and on the optical and mechanical properties of isotactic copolymers of
propene [142–148]. This paper focuses on isotactic copolymers of propene crys-
tallizing from the melt essentially in α- and/or γ-forms, using identical crystalliza-
tion conditions, as described in “Experimental Details.” Samples (Table 1) were
selected to probe the effect on yield behavior of inclusion/exclusion of
comonomeric units in crystals of the two forms. The X-ray powder diffraction
profiles of the compression-molded samples are reported in Fig. 3.
All iPP homopolymer samples crystallize from melt in the α-form. This is
indicated by presence of the (130)α- reflection at 2θ
18.6 of the α-form
[149, 150] and absence or negligible intensity of the (117)γ reflection at
2θ
20.1 of the γ-form [151, 152], as shown by the X-ray powder diffraction
profiles in Fig. 3A (curves a, e) and Fig. 3C (curve a).
The iPPEt copolymers crystallize from the melt as mixtures of α- and γ-forms of
iPP (Fig. 3A). The relative amount of crystals of the γ-form increases with increas-
ing concentration of comonomers, as indicated by the increase in intensity of the
(117)γ reflection of the γ-form in the diffraction profiles shown in Fig. 3A.
iPPBu copolymers also crystallize from the melt as mixtures of α- and γ-forms
(Fig. 3B) but, in contrast to iPPEt systems, the relative amount of γ-form first
increases with butene concentrations up to 10–15 mol% (Fig. 3C, curves a–e) then
decreases. For butene concentrations of 26–40 mol%, the pure α-form is obtained
(Fig. 3B, curves f, g).
In the case of iPPPe and iPPHe samples, small amounts of crystals in the γ-form
are obtained only at low concentrations (2–3 mol%) of comonomeric units (Fig. 3C,
D, curves b, c). At higher comonomer concentrations, the pure α-form is obtained
(Fig. 3C, D, curve d). Partial inclusion of pentene and hexene units is indicated by
64 F. Auriemma et al.

Fig. 3 X-ray powder diffraction profiles of compression-molded samples of the copolymers iPPEt
(A), iPPBu (B), iPPPe (C), iPPHe (D), and iPPOc (E), and of homopolymers iPPA, IPPB, and
iPPD prepared with catalysts A–D of Scheme 1. The (130)α and (117)γ reflections at 2θ
18.6 and
20 of the α- and γ-forms of iPP, respectively, and the (110)t reflection at 2θ
10 of the trigonal
form of iPP are indicated

the progressive shift of diffraction peaks toward lower 2θ values with higher
comonomer concentrations. With further increase in pentene units, the iPPPe
samples crystallize as mixtures of trigonal (Fig. 2C) and α-forms (Fig. 2A) of iPP
[100, 101, 103–106]. The relative amount of crystals in the trigonal form is small at
a pentene concentration of 8.8 mol% (sample DP8.8; Fig. 3C, curve d), as indicated
by the low intensity of the diffraction peak at 2θ
10 , corresponding to (110)t
reflection of the trigonal form of iPP [103, 104], and increases at higher pentene
concentrations (sample DP11; Fig. 3C, curve e).
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 65

Samples of iPPOc crystallize from the melt in the pure α-form (Fig. 3E, curves a,
c–e). Crystals of both α- and γ-forms are obtained only for sample DO2.2. How-
ever, presence of the new mesophase (Fig. 2D) cannot be excluded [102].
In all cases, the degree of crystallinity (Fig. 4) decreases with comonomer
concentration. The decrease is low in the case of iPPBu copolymers and becomes
progressively steeper with increasing the size of the comonomeric unit from
pentene to octadecene. In particular, in the case of copolymers crystallized with
the highly stereoselective but not completely regioselective catalysts A, C, and D it
is possible to discern the effect of the presence of comonomeric units on crystal-
linity and melting temperature. More precisely, the degrees of crystallinity of the
copolymers of propene with ethylene (AEx) and pentene (DPx) decrease with the
concentration of comonomeric units according to a nearly common trend, in
agreement with the fact that ethylene is partially included in the crystals of α-
and γ-forms, and pentene in the crystals the of α-form (Fig. 4A). In the case of
iPPHe copolymers prepared using catalyst A (AHx), hexene units are partially
included in the crystals of the α-form of iPP, and the decrease in crystallinity is
only slightly steeper than in iPPPe systems formed using catalyst D (DPx). How-
ever, in the case of iPPOc copolymers prepared using catalyst D (DOx), octadecene
units are not included at all in the crystals of α- and/or γ-forms because of the high
the size of the lateral side chains. As a consequence, the degree of crystallinity
rapidly decreases with the concentration of octadecene units and is always lower
than for other copolymers of identical concentration (Fig. 4A). In the case of iPPBu
copolymers prepared using catalyst A (ABx), which are crystalline in the whole
composition range [97–99], the decrease in degree of crystallinity with butene
concentration is low (Fig. 4B), in agreement with the good inclusion of the units
in the α- and γ-forms of iPP.
In the case of the highly regioregular, less stereoregular, iPPEt copolymers
synthesized with catalyst B (BEx), the degree of crystallinity decreases with
increasing concentration of ethylene units more rapidly than in the stereoregular
and slightly regiodefective samples prepared using catalyst A (AEx) of identical
composition, in agreement with the presence of a concentration of rr stereodefects
of about 3.5 mol%. However, because stereodefects rr are also partially included in
the α- and γ-forms of iPP and these defects play the same role as ethylene co-units, a
plot of degree of crystallinity as a function of the total concentration of defects
identifies a common trend for iPPEt copolymers formed using catalysts A or B
(Fig. 4B0 , inset).
The melting temperature Tm also decreases with increasing concentration of
comonomeric units (Fig. 5) but the slope varies according to the comonomer and/or
catalyst. Once again, in the case of the regiodefective copolymer samples produced
with the catalysts A, C, and D (Fig. 5A), the decrease is least for iPPBu copolymers.
This diminution increases for iPPEt copolymers and becomes steeper with increas-
ing size of comonomeric units, following a common trend for iPPPe and iPPHe
copolymers, but then drops rapidly for iPPOc systems. Furthermore, in the case of
the highly regioregular but stereodefective iPPEt copolymers synthesized with
catalyst B, the decrease in melting temperature with increasing total concentration
66 F. Auriemma et al.

Fig. 4 Degree of crystallinity of iPP homopolymers and iPPEt, iPPBu, iPPPe, iPPHe, and iPPOc
copolymer samples prepared with catalysts A–D of Scheme 1, as a function of the concentration of
comonomeric units (A, B) and the total concentration of defects (B0 , inset)

of defects is similar to that for the slightly regioirregular iPPEt copolymers syn-
thesized with catalyst A at similar defect concentration (Fig. 5B).
The Lorentz-corrected SAXS intensity for the compression-molded homo- and
copolymer samples listed in Table 1 is shown in Fig. 6. All samples show a broad
correlation peak around q*
0.5 nm1 as a result of lamellar stacking. The
broadness of the peak increases with increasing comonomer concentration.
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 67

Fig. 5 Melting temperature of iPP homopolymers iPPA, iPPB, iPPC, and iPPD, and of iPPEt,
iPPBu, iPPPe, iPPHe, and iPPOc copolymer samples prepared with the catalysts A–D of Scheme 1,
as a function of concentration of comonomeric units (A) and, in the case of iPPA, iPPB and iPPEt
samples, as a function of the total concentration of defects (B)

Simultaneously, the SAXS intensity in the background and at low q regions

increases, especially for the highly defective samples AB13.6 ( Fig. 6B, curve d),
DP5.3, DP8.8, and DP11 (Fig. 6C), AH11.2 (Fig. 6D, curve d), and DO6.0 (Fig. 6E,
68 F. Auriemma et al.

Fig. 6 Lorentz-corrected SAXS intensity of iPP homopolymers and iPPEt (A), iPPBu (B), iPPPe
(C), iPPHe (D), and iPPOc (E) copolymer samples prepared with catalysts A–D of Scheme 1

curve d). The SAXS profiles shown in Fig. 6 can be interpreted in terms of a
lamellar morphology, which becomes highly imperfect with increasing comonomer
content. Imperfections typically correspond to the formation of distorted lamellae
having small lateral dimensions, large distributions of the thicknesses of the
crystalline and amorphous layers in the lamellar stacks, formation of short lamellar
stacks and/or of more than one population of lamellar stacks with different thick-
nesses, and, especially for copolymers with a higher concentration of comonomeric
units, the presence of single lamellar entities together with a population of periodic
arrays of parallel lamellae [153].
The average values of the long spacing L, thickness of the crystalline layer lc
(lamellar thickness), and thickness of amorphous la layers for the most representa-
tive stacks formed in the compression-molded samples, calculated from the posi-
tion q* of the main correlation peak in the SAXS profiles shown in Fig. 5, are
compared in Table 2 and Fig. 7. With the exception of iPPBu samples (Fig. 7D), in
all cases the lamellar periodicity L is around 11–13 nm. With increasing concen-
tration, L first shows a slight decrease up to a monomer content of 6–8 mol%, then
tends to increase at higher comonomer concentration (Fig. 7A–C). The thickness of
lamellar crystals tends to decrease, whereas that of amorphous layers tends to
increase with increasing comonomer concentration.
Three kinds of behavior can be identified, depending on the degree of inclusion
of the comonomeric units in α- and/or γ-forms of iPP. The first kind of behavior
corresponds to the case of the isotactic copolymers iPPEt (samples AEx), iPPPe
(samples DPx), and iPPHe (samples AHx), produced with the highly stereoselective
catalysts A and D, and containing only small amounts of regiodefects. Samples with
identical concentrations of comonomeric units develop a lamellar morphology
characterized by identical values of the parameters L, lc, and la (Fig. 7A). In these
samples, the ethylene, pentene, and hexene units are partially included in the
crystals, the decrease in lamellar thickness lc and increase in thickness of amor-
phous layers la with comonomer concentration are monotonous, whereas the long
spacing L first decreases and, then, at higher comonomer concentration, tends to
increase slightly. In the case of copolymers iPPEt synthesized with catalyst B
(samples BEx) and containing 3.3–3.6 mol% of rr stereodefects, the lamellar
morphology is characterized by parameters identical to those of samples
Table 2 Comparison of lamellar parameters for iPP homopolymers and various copolymers prepared using different MAO-activated metallocene catalysts
Samplea q* b (nm1) xcc (%) Ld (nm) lce (nm) laf (nm) Tmg ( C) Comonomer conc.h (mol%) εtot i (mol%) σ yj (MPa) εyk
Homopolymers iPP
iPPA 0.51 69 12.32 8.50 3.82 151 0 1.00 – –
iPPB 0.55 55.3 11.42 6.32 5.11 135 0 3.5 17  2 49  12
iPPD (iPPC) 0.51 69 12.32 8.50 3.82 151 0 0.2
Propene-ethylene copolymers iPPEt
AE0.6 0.55 61.4 11.4 7.47 3.95 146 0.6 1.5 – –
AE4.0 0.57 59.4 11.0 6.55 4.48 130 4.0 4.8 19  5 38  10
AE7.4 0.60 53.1 10.5 5.56 4.91 117 7.4 8.0 17  5 47  10
BE3.6 0.60 53.4 10.5 5.59 4.88 121 3.6 7.2 15  4 49  8
BE13.1 0.56 34.5 11.2 3.87 7.35 47 13.1 16.3 3.9  1 45  9
Propene-1-butene copolymers iPPBu
AB1.9 0.50 64 12.6 8.04 4.52 144 1.9 2.4 – –
AB4.3 0.52 62 12.1 7.49 4.59 137 4.3 4.7 18  2 39  4
AB8.3 0.52 63 12.1 7.61 4.47 130 8.3 8.6 17  2 28  3
AB13.6 0.47 55 13.4 7.37 6.03 115 13.6 13.7 14  1 21  2
CB12 0.49 59 12.8 7.56 5.26 119 12.0 13.0 14  1 33  3
CB27.6 0.41 57 15.3 8.74 6.59 88 27.6 28.6 11  1 34  3
CB37.3 0.34 57 18.5 10.5 7.95 76 37.3 38.3 81 22  2
Propene-1-pentene copolymers iPPPe
DP3.2 0.60 60 10.5 6.28 4.19 126 3.2 3.4 22  1 16  2
DP5.3 0.58 55 10.8 5.96 4.88 105 5.3 5.5 13  1 23  2
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene

DP8.8 0.53 40 11.9 4.74 7.11 86 8.8 9.0 8.6  0.5 20  1

DP11 0.52 21 12.1 2.54 9.54 68 11.0 11.2 8.2  0.2 19  1
(continued)
69
Table 2 (continued)
70

Samplea q* b (nm1) xcc (%) Ld (nm) lce (nm) laf (nm) Tmg ( C) Comonomer conc.h (mol%) εtot i (mol%) σ yj (MPa) εyk
Propene-1-hexene copolymers iPPHe
AH2.0c 0.57 57 11.0 6.28 4.74 127 2.0 2.4 27  3 16  2
AH3.7c 0.58 53 10.8 5.74 5.09 116 3.7 4.1 19  1 25  6
AH6.8c 0.61 48 10.3 4.94 5.36 96 6.8 7.1 11  1 30  2
AH11.2c 0.45 31 14.0 4.33 9.63 70 11.2 11.5 31 32  4
Propene-1-octadecene copolymers iPPOc
DO1 0.55 54 11.4 6.17 5.25 139 1 1.2 20  2 13  2
DO2.2 0.60 51 10.5 5.34 5.13 117 2.2 2.4 13  1 19  2
DO4.8 0.63 28 9.97 2.79 7.18 93 4.8 5.0 71 18  2
DO6.0 0.54 25 11.6 2.91 8.73 85 6.0 6.2 5  0.4 18  2
DO7.5 0.57 24 11.0 2.65 8.38 69 7.5 7.7 2.7  0.2 23  2
a
iPP homopolymers (iPPA, iPPB, iPPD) and iPPEt, iPPBu, iPPPe, iPPHe, and iPPOc copolymers were prepared with the MAO-activated metallocenes A–D of
Scheme 1. Sample designations as in Table 1
b
Position of the correlation peaks (q*) in the SAXS intensity profiles of Fig. 6
c
Crystallinity index (xc) determined from X-ray powder diffraction profiles of Fig. 3
d
Long spacing (L )
e
Lamellar thickness (lc)
f
Thickness of amorphous layers (la)
g
Peak melting temperature (Tm) of the compression-molded samples crystallized from the melt by slowly cooling to room temperature while fluxing cold
water in the refrigerating circuit of press plates
h
Concentration of comonomeric units (mol%)
i
Total concentration of defects (εtot)
j
Stress (σ y) at yield of compression-molded samples
k
Strain (σ y) at yield of compression-molded samples
F. Auriemma et al.
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 71

Fig. 7 Long spacing L and the thickness of crystalline lc and amorphous la layers, as calculated
from the SAXS profiles shown in Fig. 6, relative to iPPA, iPPB, and iPPD homopolymers and to
iPPEt (A, B), iPPBu (D), iPPPe (A), iPPHe (A) and iPPOc (C) copolymer samples prepared with
catalysts A–D of Scheme 1. In B, the lamellar parameters L, lc, and la of iPPEt copolymers
synthesized with catalyst A and containing only 2,1 regiodefects ([2,1e]
0.4–0.8 mol%) and
iPPEt copolymers synthesized with catalyst B and containing only rr stereodefects ([rr]
3.2–
3.6 mol%) are compared as a function of the total concentration of defects εtot

synthesized with catalyst A (samples AEx) and having an equal concentration of

defects. In particular, as shown in Fig. 7B, plotting L, lc, and la versus the total
concentration of defects εtot ¼ [ethylene] + [rr] + [2,1e] for these copolymer sam-
ples indentifies a unique trend confirming that rr stereodefects play the same role as
ethylene co-units in the crystallization behavior of these systems. Also in this case,
with increasing the comonomer concentration, we observe a monotonous decrease
in lamellar thickness lc, a monotonous increase in the thickness of amorphous layers
la, and a slight decrease in the long spacing L. The second kind of behavior
corresponds to total inclusion and occurs for the samples of isotactic copolymers
iPPBu (Fig. 7D) synthesized with the highly stereoselective catalysts A and C. The
easy inclusion of butene units in the crystals of the α-form of iPP always produces
crystals with lamellar thickness higher than the lamellar thickness of the other
copolymers with identical concentration of units and minor degree of inclusion.
With increasing comonomer concentration, the lamellar thickness lc first decreases,
then increases at high butene content, whereas the thickness of the amorphous
layers la and the long spacing L increase monotonously. The third kind of behavior
corresponds to total exclusion and occurs for the isotactic copolymers iPPOc
72 F. Auriemma et al.

Fig. 8 Melting temperature of iPP homopolymers and iPPEt, iPPBu, iPPPe, iPPHe, and iPPOc
copolymer samples prepared with catalysts A–D of Scheme 1, as a function of lamellar thickness
and concentration of comonomeric units. For the iPPEt samples prepared with catalyst B, the
concentration of rr stereodefects (
3.5 mol%) has been added to the concentration of ethylene
units. Curve a indicates the trend in the decrease in melting temperature for samples characterized
by exclusion of comonomer units from the crystals; curve b partial inclusion; curve c full
inclusion.

(Fig. 7C) synthesized with catalyst D (samples DOx). In this case, the thickness of
the lamellar crystals is smaller than that of the copolymers with ethene, butene,
pentene, and hexene because the bulky side chains are rejected from the crystals.
For the iPPOc copolymers, the decrease in lamellar thickness lc and increase in
thickness of amorphous layers la with comonomer concentration are monotonous,
whereas the long spacing L decreases only slightly.
From the data of Figs. 5 and 7, it is apparent that parallel to the decrease in
melting temperature with comonomer content, the lamellar thickness also
decreases. Indeed, the melting temperatures of our copolymers, and semicrystalline
polymers in general, depend not only the content of comonomeric units but also on
the lamellar thickness [154–158]. A direct correlation between the melting temper-
ature, lamellar thickness, and comonomer content of melt crystallized copolymer
samples obtained by compression molding is depicted in Fig. 8. It is apparent that in
the case of the copolymers iPPEt, iPPPe, iPPHe, and iPPOc there is a concomitant
decrease in melting temperature (Fig. 5) and lamellar thickness (Fig. 7A, B, C and
Fig. 8, curves a, b) with increasing the concentration of comonomeric units. By
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 73

contrast, in the case of iPPBu copolymers, the melting temperature decreases with
increasing butene concentration (Fig. 5), even though the lamellar thickness
increases (Fig. 7D and Fig. 8, curve c).
The differences in melting behavior are also common to other systems [154–
158] and depend on the different degree of inclusion/exclusion of the comonomeric
units in the crystals and/or crystallization into a different polymorph. According to
Flory’s theory [159] of copolymer crystallization, valid in the limit of strict
exclusion, for A/B random copolymers with dilute B units excluded from crystals
of A units, the melting temperature of copolymer crystals is lower than that of the A
homopolymer exhibiting the same crystal thickness. This is a result of different
concentrations of the comonomeric units in the crystals and in the melt in equilib-
rium. Because the melting temperature Tm is the ratio of the melting enthalpy ΔHm
to the melting entropy ΔSm (Tm ¼ ΔHm/ΔSm), the presence of B units in the melt in
equilibrium with crystals, produces a non-zero mixing entropy contribution to the
melting entropy. This contribution increases with increasing concentration of B
units in the copolymer, producing a decrease in melting temperature. On the other
hand, according to the theory of Sanchez and Eby [160, 161], the melting temper-
ature of an A/B random copolymer is lowered, even in the case of inclusion of B
units in the crystals. Even in the limit of uniform inclusion of B units in the
crystalline and amorphous regions, which corresponds to zero mixing entropy at
melting, the enthalpy penalty for incorporation of B units in the crystals produces a
decrease in the melting temperature, whereas lamellar thickness does not decrease.
In our case, the decrease in melting temperature with decrease in lamellar
thickness follows a common trend in the case of iPPEt, iPPPe, and iPPHe copolymer
samples (Fig. 8, curve b), characterized by partial inclusion of the comonomeric
units in the crystals of α- and/or γ-forms of iPP. Moreover, the total exclusion of
comonomers from the crystals of α- and/or γ-forms in the case of iPPOc copolymers
produces melting depression associated with a major decrease in lamellar thickness
(Fig. 8, curve a). Also, in the case of the iPPPe sample with high pentene concen-
tration (DP11.0 containing 11 mol% pentene units), the comonomers are completely
excluded from the crystals of α- and/or γ-forms and are better included into the
trigonal form [105, 106]. The competitive partial crystallization of the trigonal form
causes melting point depression and a decrease in lamellar thickness. In the case of
iPPBu copolymers, instead, the total inclusion of butene units in the crystals pro-
duces melting depression and no decrease in lamellar thickness (Fig. 8, curve c).

4 Mechanical Properties

The stress–strain curves of melt-crystallized samples of iPP homopolymers and

iPPEt, iPPBu, iPPPe, IPPHe, and iPPOc copolymer samples obtained by compres-
sion molding are shown in Fig. 9. Only the first portion of the curves (up to 400%
deformation) is reported, to put into evidence the yield behavior. It is worth noting
that all copolymer samples show high flexibility, toughness, and ductility, with
74 F. Auriemma et al.

Fig. 9 Stress–strain curves of melt-crystallized films prepared by compression molding of iPP

homopolymers iPPA, iPPB, and iPPD and of iPPEt (A, B), iPPBu (C, D), iPPPe (E), iPPHe (F),
and iPPOc (G) copolymer samples synthesized with catalysts A–D of Scheme 1
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 75

deformation at break higher than 300–400% [98, 101, 103]. By contrast, the highly
stereoregular iPP samples prepared with catalyst A, C (data not reported), and D of
Scheme 1, (iPPA, iPPC, and iPPD) are stiff and fragile materials, as shown in the
insets of Fig. 9A, D [40, 41, 127, 128]. Only the highly regioregular, less stereo-
regular, iPPB sample containing 3.5 mol% rr defects shows high flexibility coupled
with high toughness and values of deformation at break of about 350% [40, 41, 127]
(Fig. 9B).
In all cases, the stress at any strain decreases with increasing concentration of
comonomeric units (Fig. 9). Plastic resistance also decreases, as indicated by the
values of the yield stress. The stereodefective homopolymer sample iPPB, and the
copolymer samples iPPEt and iPPBu show uniform deformation and smooth yield
behavior regardless of comonomer concentration (Fig. 9A–D) [98]. By contrast, the
highly stereoregular copolymer samples with side chains longer than the ethyl
group (iPPPe, iPPHe, and iPPOc) show less uniform deformation and sharp yield-
ing behavior at low comonomer concentrations [98, 101–103]. The yielding behav-
ior becomes smoother with increasing comonomer content (Fig. 9E–G), and the
deformation becomes uniform. The copolymer samples iPPEt, iPPHe, and iPPOc
with the highest comonomer concentrations (BE13.1, AH11.2 and DO7.5, respec-
tively) and the iPPBu sample CB12 with 12 mol% butene units show diffuse
yielding behavior and uniform deformation.
The values of stress and deformation at yield are reported in Fig. 10. It is
apparent that the decrease in stress at yield with increasing comonomer concentra-
tion (Fig. 10A, A0 ) is accomplished with an increase in deformation at yield
(Fig. 10B, B0 B00 ). It is also apparent that each kind of comonomer influences the
yield behavior of the copolymers differently. In particular, in the case of highly
stereoregular, slightly regiodefective copolymer samples iPPEt, iPPBu, iPPPe,
iPPHe, and iPPOc prepared with the catalysts A, C, and D, the decrease in σ y
values is smooth and quasilinear for iPPEt and iPPBu systems and becomes steeper
with increasing size of comonomeric units. This decrease is similar for iPPPe and
iPPHe samples (Fig. 10A). Moreover, as shown in Fig. 10A0 (inset), in the case of
the stereodefective iPP homopolymer sample and iPPEt samples prepared with
catalyst B, the decrease in the values of stress at yield as a function of the total
concentration of defects is similar to that of regiodefective iPPEt samples prepared
with catalyst A.
The values of deformation at yield εy for the samples iPPPe, iPPHe, and iPPOc
increase with the concentration of pentene, hexene, and octadecene comonomeric
units (Fig. 10B), but are nearly constant, or increase only slightly, in the case of
iPPEt (Fig. 10B00 ) and iPPBu (Fig. 10B0 ) copolymers.
The data in Fig. 10 indicate that the decrease in plastic resistance is generally
associated with an increase in deformation at yield. The decrease in plastic resis-
tance, in turn, is related to the level of inclusion in and/or exclusion of
comonomeric units from crystals, and to the effective level of disturbance of the
defects included in the crystals in the case of inclusion. The almost complete
inclusion of butene units in the crystals of the α-form of iPP produces a small
decrease in stress at yield (Fig. 10A), and constant values of deformation at yield
76 F. Auriemma et al.

Fig. 10 Values of stress and strain at yield of melt-crystallized films obtained by compression
molding of iPP homopolymer iPPB (A0 , B00 ) and iPPEt (A, A0 , B, B00 ), iPPBu (A, B, B0 ), iPPPe (A,
B), iPPHe (A, B), and iPPOc (A, B) copolymer samples prepared with catalysts A–D of Scheme 1.
Arrows in B indicate the average values of deformation at yield of iPPEt (B00 ) and iPPBu (B0 )
samples. Samples iPPA, iPPC, and iPPD are not included because they break before yielding

(Fig. 10B0 ) that are not dependent on butene concentration. On the other hand,
partial (ethylene, pentene, and hexene) or complete (octadecene) exclusion of
comonomeric units from the crystals of α- or γ-forms of iPP induces a larger
decrease in stress at yield (Fig. 10A), coupled with an increase in yield deformation
(Fig. 10B, B00 ).
As analyzed in the preceding section, melt-crystallized films of these copoly-
mers prepared by compression molding are characterized by different lamellar
thicknesses. For semicrystalline polymers, the values of yield stress generally
increase with lamellar thickness. In fact, thick lamellae generally entail major
crystal stability, and therefore also strong plastic resistance [5–7, 14–19]. The
values of yield stress shown in Fig. 10A for the compression-molded samples of
copolymers crystallized under similar conditions are reported in Fig. 11 as a
function of lamellar thickness, as evaluated from the SAXS profiles shown in
Fig. 7. A unique correlation line can be established, regardless of comonomer
type. In particular, the logarithm of the yield stress increases as a function of the
average values of lamellar thickness according to a sigmoidal master curve. Devi-
ation from this correlation is observed for copolymers of iPPBu with butene
concentrations higher than 12–13 mol%, where the lamellar thickness increases
and the yield stress decreases with increasing butene concentration. For instance,
for samples CB12, CB27.6, and CB37.3 (containing 12, 27.6 and 37.3 mol% of
butene, respectively) with high lamellar thickness, the yield stress decreases with
increasing lamellar thickness. Exceptions occur at low lamellar thickness for the
copolymer iPPPe with pentene content of 11 mol% (sample DP11) and the highly
defective copolymers iPPOc with octadecene content of 4.8–6 mol% (samples
DO4.8 and DO6.0). These samples show yield stress values that are larger than
those expected on the basis of the sigmoidal master curve. However, with further
increase in octadecene content (sample DO7.5 with 7.5 mol% comonomer), the
value of the yield stress suddenly drops, in apparent agreement with expectation.
These results indicate that, on the one hand, for a homogeneous class of propene-
based copolymers (crystallizing in α- and/or γ-forms of iPP under similar condi-
tions) the lamellar thickness controls the level of plastic resistance of the samples.
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 77

Fig. 11 Values of nominal stress (A) and true stress at yield (B) of melt-crystallized films
obtained by compression molding of iPP homopolymer iPPB and iPPEt, iPPBu, iPPPe, iPPHe,
and iPPOc copolymer samples prepared with catalysts A–D of Scheme 1. Inset B0 shows values of
true stress at yield of melt-crystallized films obtained by compression molding of iPPBu copoly-
mers as a function of 2lc + la. Solid lines in B indicate the theoretical predictions of yield stress on
the basis of the crystallographic model [25] based on thermal activation of screw dislocations,
according to Eq. 5, by setting the value of the Burgers vector B ¼ 0.650 nm, the shear modulus of
(040) planes K ¼ 0.84 GPa, and the free energy barrier associated with nucleation of [001]
dislocations ΔG* in the range 40–90 kT, namely ΔG* ¼ 59 kT (curve a) and 90 kT (curve b)
[94]. The shaded area indicates samples with critical dislocation nuclei (calculated using Eq. 3)
too large in size to be acceptable

However, because crystal stability also depends on the concentration and distribu-
tion of the structural irregularities, in the case of iPPBu copolymers with a high
concentration of butene units, the easy inclusion of comonomers in the unit cell of
the α-form of iPP induces formation of lamellar crystals of high thickness but lower
plastic resistance (yield stress) than expected for defect-free crystals of homopol-
ymer having identical thickness. This indicates that butene units act as point-like
defects in the crystals and, therefore, induce a decrease in plastic resistance. On the
other hand, polymorphism and the amorphous phase placed between lamellar
crystals could also play a role [5–7, 14–19]. This is evidenced in the case of
iPPOc copolymers with 4.8 and 6 mol% of octadecene. For these copolymers, the
yield stress is higher than expected on the basis of the low values of lamellar
thickness. Because the long branches are rejected from the crystals, their confine-
ment in the amorphous interlamellar layers close to the fold surfaces produces an
indirect increase in resistance to plastic deformation of the crystals, probably as a
result of topological restraints. Instead, polymorphism is involved for iPPPe and
iPPOc copolymer samples DP11 and DO7.5 (with 11 and 7.5 mol% of pentene and
octadecene units, respectively). As shown in Fig. 3, the sample DP11 crystallizes in
a mixture of crystals of α-form and trigonal form [103, 104] (Fig. 2C), whereas the
sample DO7.5 crystallizes in a mixture of α-form and the second mesophase of iPP
[102] (Fig. 2D). Both the trigonal form and the mesomorphic form that crystallize
along with the α-form are characterized by partial inclusion of comonomers in the
crystalline domains. The trigonal form of the iPPPe copolymer is characterized by
78 F. Auriemma et al.

long-range order in two dimensions for the positioning of chain axes of the 3/1
helices [162] (Fig. 2C); the new mesomorphic form of the iPPOc sample, instead,
presents no lateral order in the position of chain axes, and only an average
periodicity parallel to chain axes of 3/1 helices [162] (Fig. 2D). As a consequence,
the presence of a second polymorph increases the resistance to plastic deformation
in the iPPPe sample partially crystallized in the trigonal form, but decreases the
plastic resistance in the iPPOc samples partially crystallized in the new mesophase.
At temperatures higher than the glass transition, the strong dependence of yield
stress on lamellar thickness, which is generally observed for semicrystalline poly-
mers and in our copolymers in particular (Fig. 11A), entails a yield behavior
possibly controlled by activation of plastic deformation of lamellar crystals through
crystallographic slip processes [5–7, 14–19, 25]. Crystallographic slip processes, in
turn, are facilitated by nucleation and propagation of dislocations and/or defects
[25, 74–76]. According to this mechanism, the stress at yield corresponds to the
point of the stress–strain curve at which local stress reaches the level of the critical
resolved shear stress for the easiest slip system. This level is, in turn, controlled by
nucleation and propagation of dislocations within deforming crystals [5–7, 14–19,
25, 84]. Therefore, the strong dependence of the yield stress of polymer crystals on
lamellar thickness (Fig. 11A) can be explained by the fact that the critical stress for
activation of chain slip is directly related to the strong dependence of the rate of
nucleation of dislocations on the thickness of the crystals.
The minimum stress required for nucleation and activation of a new dislocation
at the edges of lamellar crystals and the relationship between this stress to crystal
thickness can be predicted using the model of Young [25], successively refined by
Shadrake and Guiu [93]. According to this model, any contribution from the chains
in the amorphous phase to the yield stress is neglected in a first approximation. The
role of amorphous chains is merely as force transmitters, because the modulus of
the amorphous phase above Tg is an order of magnitude lower than the modulus of
crystalline phase. The model assumes that deformation occurs by {hk0} <001>
chain direction slip, resulting in the formation of [001] screw dislocations (i.e.,
dislocations parallel to the chain axes) [25]. The Burgers vector B, parallel to the
chain axis (Fig. 12A) at distance r from the dislocation, has magnitude B and
generally spans only a small integer number n (n ¼ 1,2) of chain periodicities c.
The nucleation and activation of [001] dislocations of length lc coincident with the
lamellar thickness is a thermally activated process requiring a critical level τ* of
shear stress. Within the model, only the value of τ* matters because the intrinsic
movement of already formed dislocations occur at Peierls–Nabarro stress (i.e., at a
stress level much lower than τ*) (Fig. 11A). According to the crystallographic
approach, yielding (i.e., the beginning of plastic deformation) starts when the
critical resolved shear stress is reached in any family of planes with low τ*. This
corresponds to slip planes coincident with the lattice planes of maximum packing
and to slip directions parallel to the lattice directions of maximum packing [5–7,
14–19, 25].
In the simplest approach, the free energy associated with the nucleation of such
dislocations can be written as:
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 79

Fig. 12 (A) Model of a [001] screw dislocation, nucleated at the edge of a lamellar crystal of iPP
in the α-form, that has advanced a distance r inside the crystal parallel to the (040) slip planes,
along the slip direction. The length of the dislocation is equal to the crystal thickness lc, whereas
the Burgers vector, parallel to chain axis, has a magnitude equal to the chain periodicity cα
(B ¼ 0.65 nm). Lamellar crystals (A0 ) obtained through deformation in the {0k0}<00l> chain
direction slip leading the dislocation in A to emerge at the opposite edge of the crystal and
consequent formation of a step. (B) The (040) slip planes in two projections parallel (B) and
perpendicular (B0 ) to the chain axis. Symbols L and R stand for left- and right-handed helical
chains, respectively. Rows aαcα of all left-handed helical chains alternate with rows of chains of
opposite chirality forming double layers, delimited by traces of the (040) planes in B0
[149, 150]. The lattice planes (040) are planes of close packing and the slip direction in A is
parallel to the lattice direction of maximum packing [149, 150]. (C) Schemes of chain folding with
adjacent (C) and non-adjacent re-entry (C0 ) for the limit-ordered [150, 162] and limit-disordered
80 F. Auriemma et al.

KB2 lc r
ΔG ¼ ln  lc Brτ ð2Þ
2π r0

In Eq. 2, K is the shear modulus associated with the {hk0}<001> slip process, r0 is
the core radius of the dislocations (generally assumed to be twice the value of
B [26, 94]), and r is the distance of the dislocation from the edge of the crystal
(Fig. 12A). The first term corresponds to the elastic strain energy and the second
term corresponds to the work performed by the external shear stress. In Eq. 2, the
core energy contribution as a result of lattice distortions around the dislocation is
neglected. Dislocations are activated when the distance of a dislocation from the
edge of the crystal reaches a critical value r* (size of the critical dislocation nuclei),
obtained by setting the derivative of ΔG with respect to r as equal to zero [25]. The
obtained value of the critical size of the dislocation r*, the critical nucleus, is given
by:

KB
r* ¼ ð3Þ
2πτ

By combining Eqs. 2 and 3, the activation barrier of free energy ΔG* needed to
nucleate a dislocation of critical size r* is obtained as:
   
KB2 lc r*
ΔG* ¼ ln 1 ð4Þ
2π r0

Finally, introducing Eq. 3 into Eq. 4, the value of tensile stress σ y, which is twice the
critical value of the shear stress τy (i.e., σ y ¼ 2 τy) according to Tresca’s criterion, is
obtained as:
 
K 2πΔG*
σ y ¼ exp  1 ð5Þ
2π lc KB2

Notice that Eq. 5 assumes that the core radius of the dislocation r0 is equal to
twice the length of the Burgers vector B (2r0) [26, 94].
To compare the yield behavior of our samples with the predictions of the
crystallographic model based on thermal activation of dislocations (Eq. 5), the
values of true stress at yield are needed. In general, the transverse strain of
semicrystalline polymers (perpendicular to the stretching direction under uniaxial
elongation) decreases with deformation. For rubbery materials at low deformation,
the transverse section S of the deformed sample is related to the initial section S0 of
the specimen by the relationship S ¼ S0 (l0/l ) where l0 and l are the initial gauge-

Fig. 12 (continued) [149, 162] α-forms of iPP, respectively. Deformations in the {010}<001>
chain direction slip comply well with the chain folding scheme with adjacent re-entry C. (C and C0
are reproduced from [163], with ACS permissions)
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 81

length and the gauge-length in the deformed state, respectively. This relationship
entails that the sample is incompressible, that is, that Poisson’s ratio is equal to 0.5.
However, for semicrystalline polymers, Poisson’s ratio changes during deformation
from the maximum value of 0.5 to values close to zero, as a result of volume
expansion caused by crazing, cracks, and voids [164]. We checked that the defor-
mation of our copolymer samples was largely uniform up to the yield point.
Therefore, we corrected the nominal values of stress at yield by the factor l/l0 to
obtain the true stress, implicitly assuming a Poisson’s ratio close to 0.5. The
obtained values of true stress are reported in Fig. 11B as a function of lamellar
thickness.
Representative curves describing the change in yield stress with lamellar thick-
ness according to the crystallographic model based on thermal activation of screw
dislocations are also given in Fig. 11B. They were generated using Eq. 5 by fixing
the parameters of the model according to the values suggested in the literature for
modeling the yield behavior of iPP [94]. For the sake of simplicity and without loss
of generality, confining attention to the α-form of iPP, it is assumed that plastic
deformation occurs by {0k0}<00l> chain direction slip that is parallel to the (040)
planes of maximum packing of the iPP α-form, with the [001] screw dislocation
parallel to the chain axes (Fig. 12A, B, B0 ). This deformation mechanism complies
well with the chain folding scheme for the α-form of iPP, characterized by adjacent
re-entry (Fig. 12C), as proposed by Corradini [165–167] and confirmed by double
quantum 13C–13C solid state NMR [163]. In particular, the value of the Burgers
vector B is set equal to 0.650 nm, corresponding to the chain periodicity of iPP in
the α-form [149, 150]. The values for the energy barrier ΔG* were selected in the
typical range of 40–90 kT [94] associated with the thermal nucleation of dislocation
at laboratory time scales, namely ΔG* ¼ 59 and 90 kT, respectively. The values of
shear modulus for the (040) lattice planes was set at 0.84 GPa, in agreement with
values suggested in the literature for iPP (in the range 0.84–1.0 GPa) [94].
A screw dislocation scheme similar to that of Fig. 12 for the α-form of iPP can
also be proposed for the γ-form [151, 152], considering that the chain axes in the
γ-form are directed along the diagonal of the C face of the orthorhombic unit cell
(Fig. 2B). This entails that the screw dislocations parallel to the chain axes are
 
parallel to the [110] and/or 110 lattice directions and that plastic deformation
occurs by {00l}<110> chain direction slip (i.e., parallel to the (008) planes of the
γ-form). A further adjustment of the crystallographic model entails that the chain
axes, and therefore the Burgers vector, lie at a tilted angle of
40 to the normal of
the basal face of the lamellar crystals.
With the exception of iPPBu copolymer samples with high butene content, the
results of the model appear to be in good agreement with experimental results in all
cases (Fig. 11B). This indicates that, in spite of the simplicity of the model, the
crystallographic model describes well (without making any fitting attempt) the
yield behavior of our samples, regardless of the kind and concentration of
comonomeric units. This result is noteworthy, especially considering that the
copolymers are characterized by different degrees of inclusion/exclusion of the
82 F. Auriemma et al.

comonomeric units in the crystals. This inclusion generates different concentrations

and kinds of structural disorder and, therefore, the crystals are characterized by
different intrinsic stabilities. Moreover, the spread of experimental data in Fig. 11B
is also the result of measurements being performed in independent experiments, at
ambient conditions (room temperature) subject to significant thermal fluctuations
(5 C) and on samples adopting slightly different deformation rates, although the
ratio between the deformation rate vdef and initial gauge length l0 was fixed at
10 (vdef/l0 ¼ 10). Minor errors also arise from the approximate evaluation of true
stress at yield utilizing a Poisson ratio of 0.5 and the approximate evaluation of
lamellar thickness lc, since the volume fraction of the crystalline phase is slightly
lower than the crystallinity index xc resultant from WAXS analysis. In fact, using
values of the Poisson ratio close to 0.4, typical for semicrystalline polymers at low
deformations, and/or values of lamellar thickness evaluated from the
one-dimensional correlation function [120, 121], the dislocation model can equally
well describe the yield behavior of our samples, using values close to 0.84 GPa for
the shear modulus for the (040) planes K and in the range 59–90 kT for the free
energy barrier associated with the nucleation of [001] dislocations ΔG*.
We also checked that the size of the critical dislocation nucleus r* calculated
using Eq. 2 is in all cases in the range 5–10 nm, that is, less than the typical size of
crystal blocks in lamellar crystals (15–30 nm) [100, 101], as estimated from the
WAXS profiles using the Scherrer formula (the width at mid-height of the equatorial
reflections are all in the range 0.4–0.8 ; see Fig. 2). Only in the case of the samples
with high comonomer content, having low lamellar thickness and low crystallinity,
do the values of r* exceed 15 nm. These samples are indicated in Fig. 10B and
correspond to samples of iPPOc with octene content higher than 4 mol% (samples
DO4.8, DO6.0, and DO7.5), iPPEt with ethylene content of 13 mol% (sample
BE13.1), and iPPPe and iPPHe with
11 mol% comonomer units (samples DP11
and AH11.2, respectively). Because the value of the critical dislocation nuclei r*
cannot be greater than the dimensions of a crystal block, the good agreement of
the model with experimental data for low values of lamellar thickness should be
considered with caution. In fact, the presence of crystals in a different polymorph,
namely the γ-form for sample AH11.2, the trigonal form for samples DP11 and
AH11.2 (Fig. 2C), and the new mesophase for sample DO7.5 (Fig. 2D), could
completely alter the {0k0}<00l> chain direction slip mechanism of [001] disloca-
tion proposed for the α-form (Fig. 12A, B). Furthermore, the rejection of branches
outside the crystals, close to the fold surfaces, increases the plastic resistance of
the amorphous phase and overcomes the role of crystals in the yield mechanism
[14–19]. This indicates that, for samples with low crystallinity in which the lamellar
crystals have a thickness lower than the threshold value of 4–5 nm, the role of the
amorphous phase in the yield behavior cannot be neglected.
A further exception occurs at high lamellar thickness for the iPPBu samples with
butene concentrations higher than 12 mol%. In particular, the decrease in true stress
values with increasing lamellar thickness cannot be explained by resorting to the
crystallographic model. The yield behavior of these systems derives from the easy
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 83

inclusion of butene units in crystals of the α-form. This inclusion, on the one hand,
increases lamellar thickness while decreasing the stability of crystals and, on the
other hand, increases lattice resistance to the gliding of dislocations because point
defects act as Peierls barriers. Therefore, the observed values of yield stress
achieved in these samples is the resultant of two competitive effects, the decrease
in lamellar stability and the simultaneous increment of barriers to the movement of
dislocations, because butene units in the crystals act as localized obstacles. An
alternative or complementary mechanism subtending the yield behavior of these
samples could involve the deformation modes of the amorphous phase, such as
interlamellar separation and interlamellar slip (Fig. 1E, F) [5–7, 14–19]. These
modes become active whenever the easiest movement at the mesoscale is stretching
of the intralamellar amorphous chains connecting adjacent layers instead of crystal
slip. To form an efficient tie, a chain emerging from a lamellar crystal (thickness lc)
should travel across the amorphous layer (thickness la) and enter a new crystalline
lamella (thickness lc). The higher the number density of tie chains, the higher the
yield stress level. Moreover, the higher the degree of separation of adjacent
lamellae, the lower the number density of tie chains. Therefore, the probability of
formation of a tie chain is expected to decrease with the quantity 2lc + la. The
decrease in true stress at yield of these samples with increase in 2lc + la (shown in
Fig. 11B0 , inset), is in agreement with the above argument.
It is worth noting that, using the Eyring formalization of thermally activated
processes [168, 169], the temperature and strain rate dependence of yield stress of
the iPP homopolymer indicate that two basic processes intervene in the stress
response of a semicrystalline polymer [26–28, 62–65]. The first process involves
intralamellar deformations or crystal slips, the mechanisms of which have been
already detailed. The second process involves interlamellar deformations and is
somehow linked to the α-relaxation mechanism of iPP [170]. This relaxation
involves motion of the 3/1 helical chains in the crystals through the combined
effect of 120 rotations around the chain axis and c/3 translation parallel to the
chain axis [171]. These jumps result in chain diffusion through the crystals and
necessarily involve the mobility of repeating units in the constrained amorphous
zone surrounding the crystal [171]. In the case of iPP homopolymer, the
intralamellar deformation contributes to the yield stress at high temperatures or
low strain rates [62–65]. Crystal slip also participates at lower temperatures and/or
high strain rates, but the main process contributing to the observed yield stress is the
interlamellar process [62–65]. Therefore, in the case of iPP homopolymer, the
contribution of α-relaxation and consequent participation of the amorphous phase
to the yield stress should not be neglected at deformation temperatures close to
ambient.
In the case of copolymers, the α-relaxation mechanism is expected to contribute
actively to yield stress only at low comonomer concentration. With increasing
concentration of comonomeric units, this mechanism becomes less important,
regardless of the degree of inclusion/exclusion of comonomers in/from the crystals.
In fact, chain diffusion inside the crystals associated with α-relaxation is prevented
by the large steric hindrance caused by comonomers located inside the crystals in
84 F. Auriemma et al.

the case of inclusion, and by comonomers located in the amorphous regions close to
the fold surfaces in the case of exclusion. The different degrees of participation of
the α-relaxation process to the yielding behavior of our copolymers can explain the
bifurcation of the experimental values of yield stress shown in Fig. 11B. For
lamellar thicknesses higher than 5 nm, samples with lower content of comonomeric
units show values of yield stress close to the curve generated by setting the free
energy barrier for nucleation of dislocations ΔG* ¼ 59 kT (Fig. 11B, curve a),
whereas the samples with higher content of comonomeric units follow the curve
generated by setting ΔG* ¼ 90 kT (Fig. 11B, curve b). We argue that the lower free
energy barrier associated with the thermal nucleation of dislocations observed for
the samples of curve a is a result of participation of the α-relaxation process to their
yield behavior. Lack of this participation, for the samples of the curve b, results in
an increase in the free energy barrier.

5 Concluding Remarks

The yield behavior in tensile experiments of a wide class of semicrystalline poly-

mers is analyzed in the framework of a crystallographic micromechanical model
based on the thermal nucleation of dislocations. The samples are isotactic copoly-
mers of propene with ethylene (iPPEt), 1-butene (iPPBu), 1-pentene (iPPPe),
1-hexene (iPPHe), and 1-octadecene (iPPOc) and possess a random distribution
of comonomeric units and tailored concentrations of stereo- and regiodefects.
It has been shown that the decrease in plastic resistance depends on the level of
inclusion/exclusion of the comonomeric units in/from the crystals and on the
effective level of disturbance of the comonomers included in the crystals. More-
over, a remarkable dependence of yield stress on lamellar thickness has been
demonstrated. In particular, we have shown that, in all cases, the values of yield
stress decrease with lamellar thickness. By contrast, in the case of iPPBu copoly-
mers, the almost complete inclusion of butene units in the crystals of α-form
produces a decrease in stress at yield and simultaneous increase in lamellar
thickness.
According to the crystallographic approach, the phenomenon of yielding marks
the beginning of plastic flow through occurrence of diffuse crystal slip processes,
facilitated by the movement of dislocations, nucleation of new dislocations at the
edge of lamellar crystals, and participation of the amorphous component through
interlamellar slip or interlamellar rotation. Therefore, the thickness and intrinsic
stability of lamellar crystals and the intrinsic mobility of the constrained
interlamellar amorphous phase play key roles. Applying these concepts, we have
shown that, except for iPPBu copolymers with high butene content, in our copol-
ymers the yield behavior is largely controlled by the activation of plastic deforma-
tion of the crystals through crystallographic slip processes, involving, in turn,
nucleation of new dislocations. However, for highly defective copolymers of low
crystallinity, forming lamellar crystals of low thickness, the role of the deformation
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 85

modes of the amorphous phase, such as interlamellar separation and interlamellar

slip, should not be neglected. In the case of iPPBu, the beginning of plastic
deformation is also controlled by the increase in lattice resistance to glide of
dislocations as a result of the butene units in the crystals acting as Peierls barriers.

References

1. Keller A (1968) Polymer crystals. Rep Prog Phys 31:623–704

2. Ward IM, Sweeney J (2004) An introduction to the mechanical properties of solid polymers,
2nd edn. Wiley, Chichester
3. Peterlin A (1977) Plastic deformation of crystalline polymers. Polym Eng Sci 17:183–193
4. Peterlin A (1975) Composite structure of fibrous material. Adv Chem Ser 142:1–13
5. Oleinik EF, Rudnev SN, Salamatina OB (2007) Evolution in concepts concerning the
mechanism of plasticity in solid polymers after the 1950s. Polym Sci Ser A 49:1302–1327
6. Haudin JM (1982) Plastic deformation of semi-crystalline polymers. In: Escaig B, G’Sell C
(eds) (1982) Plastic deformation of amorphous and semi-crystalline materials. Les Editions
de Physique, Paris, pp 291–311
7. Bartczak Z, Galeski A (2010) Plasticity of semicrystalline polymers. Macromol Symp
294:67–90
8. Lin L, Argon AS (1994) Structure and plastic deformation of polyethylene. J Mater Sci
29:294–323
9. Crist B (1993) Plastic deformation in material science and technology. In: Thomas EL (ed) A
comprehensive treatment, vol 12. VCH, New York, pp 427–470
10. Lee BJ, Argon AS, Parks DM, Ahzi S, Bartczak Z (1993) Simulation of large strain plastic
deformation and texture evolution in high density polyethylene. Polymer 34:3555–3575
11. Lee BJ, Argon AS, Ahzi S (1993) Micromechanical modeling of large plastic deformation
and texture evolution in semi-crystalline polymers. J Mech Phys Solids 41:1651–1687
12. Hay IL, Keller A (1965) Polymer deformation in terms of spherulites. Kolloid-Zu Z Polymere
204:43–74
13. Breese DR, Beaucage G (2004) Curr Opinion Solid State Mater Sci 8:439–448
14. Séguéla R (2007) Plasticity of semi-crystalline polymers: crystal slip versus melting-
recrystallization. e-Polymers 032:1–20
15. Seguela R (2002) Dislocation approach to the plastic deformation of semicrystalline poly-
mers: kinetic aspects for polyethylene and polypropylene. J Polym Sci B Polym Phys
40:593–601
16. Séguéla R (2005) Critical review of the molecular topology of semicrystalline polymers: the
origin and assessment of intercrystalline tie molecules and chain entanglements. J Polym Sci
B Polym Phys 43:1729–1748
17. Liu B, Zhang L, Gao H (2006) Poisson ratio can play a crucial role in mechanical properties
of biocomposites. Mech Mater 38:1128–1142
18. Liu B, Feng X, Zhang S-M (2009) The effective Young’s modulus of composites beyond the
Voigt estimation due to the Poisson effect. Compos Sci Tech 69:2198–2204
19. Gorbatikh L, Pingle P (2007) On incompressibility of a matrix in naturally occurring
composites. Appl Phys Lett 91:241913
20. Peterlin A (ed) (1971) Plastic deformation of polymers. Marcel Dekker, New York
21. Peterlin A (1987) In: Mark HF, Bikales NM, Overberger CG, Menges G, Kroschwitz JI (eds)
Encyclopedia of polymer science and engineering. Wiley, New York, pp 72–94
22. Samuels RJ (1974) Structured polymer properties. Wiley, New York
23. Bueche F (1956) Young’s modulus of semicrystalline polymers. J Polym Sci A Polym Chem
22:113–122
86 F. Auriemma et al.

24. Halpin JC, Kardos JL (1972) Moduli of crystalline polymers employing composite theory. J
Appl Phys 43:2235–2241
25. Bowden PB, Young RJ (1974) Deformation mechanisms in crystalline polymers. J Mater Sci
9:2034–2051
26. Schrauwen BAG, Bernard AG, Janssen RPM, Govaert LE, Meijer HEH (2004) Intrinsic
deformation behavior of semicrystalline polymers. Macromolecules 37:6069–6078
27. Govaert LE, Meijer HEH (2005) Mechanical performance of polymer systems: the relation
between structure and properties. Prog Polym Sci 30:915–938
28. Caelers HJM, Govaert LE, Peters GWM (2016) The prediction of mechanical performance of
isotactic polypropylene on the basis of processing conditions. Polymer 83:116–128
29. Flory PJ, Yoon DY (1978) Molecular morphology in semicrystalline polymers. Nature
272:226–229
30. Gent AN, Madan S (1989) Plastic yielding of partially crystalline polymers. J Polym Sci B
Polym Phys 27:1529–1542
31. Young RJ (1974) A dislocation model for yield in polyethylene. Philos Mag 30:85–94
32. Galeski A, Bartczak Z, Argon AS, Cohen RE (1992) Deformation mechanisms and plastic
resistance in single-crystal-textured high-density polyethylene. Macromolecules
25:5705–5718
33. Hiss R, Hobeika S, Lynn C, Strobl G (1999) Network stretching, slip processes, and
fragmentation of crystallites during uniaxial drawing of polyethylene and related copolymers.
A comparative study. Macromolecules 32:4390–4403
34. Men Y, Strobl GJ (2001) Critical strains determining the yield behavior of s-PP. J Macromol
Sci Phys B40:775–796
35. Hughes DJ, Mahendrasingam A, Oatway WB, Heeley EL, Martin C, Fuller W (1997) A
simultaneous SAXS/WAXS and stress-strain study of polyethylene deformation at high strain
rates. Polymer 38:6427–6430
36. Yamada M, Miyasaka K, Ishikawa K (1971) Redrawing of oriented polyethylene film. J
Polym Sci B Polym Phys 9:1083–1096
37. Takahashi Y, Ishida T (1988) J Polym Sci B Polym Phys 26:2267–2277
38. Stoclet G, Seguela R, Lefebvre JM, Elkoun S, Vanmansart C (2010) Strain-induced molec-
ular ordering in polylactide upon uniaxial stretching. Macromolecules 43:1488–1498
39. Seguela R (2005) On the strain‐induced crystalline phase changes in semi‐crystalline poly-
mers: mechanisms and incidence on the mechanical properties. J Macromol Sci C Polym Rev
J 45:263–287
40. De Rosa C, Auriemma F (2007) Stress-induced phase transitions in metallocene-made
isotactic polypropylene. Lect Notes Phys 714:345–371
41. De Rosa C, Auriemma F (2006) Structural–mechanical phase diagram of isotactic polypro-
pylene. J Am Chem Soc 128:11024–11025
42. De Rosa C, Auriemma F, de Ballesteros OR (2006) A microscopic insight into the deforma-
tion behavior of semicrystalline polymers: the role of phase transitions. Phys Rev Lett
96:167801–167804
43. De Rosa C, Auriemma F (2012) The deformability of polymers: the role of disordered
mesomorphic crystals and stress-induced phase transformations. Angew Chem Int Ed
124:1233–1237
44. Liu Y, Kennard CHL, Truss RW, Carlos NJ (1997) Characterization of stress-whitening of
tensile yielded isotactic polypropylene. Polymer 38:2797–2805
45. Hay IL, Keller A (1970) Mechanically induced twinning and phase transformations. J Polym
Sci C Polym Symp 30:289–295
46. Frank FC, Keller A, O’Connor A (1958) Deformation processes in polyethylene interpreted
in terms of crystal plasticity. Philos Mag 3:64–74
47. Kiho H, Peterlin A, Geil PH (1964) Polymer deformation. VI. Twinning and phase transfor-
mation of polyethylene single crystals as a function of stretching direction. J Appl Phys
35:1599–1605
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 87

48. Allan P, Crellin EB, Bevis M (1973) Stress-induced twinning and phase transformations in
polyethylene single crystals. Philos Mag 27:127–145
49. Butler MF, Donald AM, Bras W, Mant GR, Derbyshire GE, Ryan AJ (1995) A real-time
simultaneous small- and wide-angle X-ray scattering study of in-situ deformation of isotropic
polyethylene. Macromolecules 28:6383–6393
50. Butler MF, Donald AM, Ryan AJ (1997) Time resolved simultaneous small- and wide-angle
X-ray scattering during polyethylene deformation: 1. Cold drawing of ethylene-α-olefin
copolymers. Polymer 38:5521–5538
51. Butler MF, Donald AM (1998) A real-time simultaneous small- and wide-angle X-ray
scattering study of in situ polyethylene deformation at elevated temperatures. Macromole-
cules 31:6234–6249
52. Butler MF, Donald AM, Ryan AJ (1998) Time resolved simultaneous small- and wide-angle
X-ray scattering during polyethylene deformation -- II. Cold drawing of linear polyethylene.
Polymer 39:39–52
53. Ran S, Zong X, Fang D, Hsiao BS, Chu B, Phillips RA (2001) Structural and morphological
studies of isotactic polypropylene fibers during heat/draw deformation by in-situ synchrotron
SAXS/WAXD. Macromolecules 34:2569–2578
54. Ciferri A (1963) The α$ transformation in keratin. Trans Faraday Soc 59:562–569
55. Ciferri A, Smith KJ Jr (1964) Phase changes in fibrous macromolecular systems and associ-
ated elasticity. Model phase diagrams. J Polym Sci A 2:731–753
56. Tashiro K, Nakai Y, Kobayashi M, Tadokoro H (1980) Solid-state transition of poly(butylene
terephthalate) induced by mechanical deformation. Macromolecules 13:137–145
57. Auriemma F, De Rosa C (2003) New concepts in thermoplastic elastomers: the case of
syndiotactic polypropylene, an unconventional elastomer with high crystallinity and large
modulus. J Am Chem Soc 125:13143–13147
58. Auriemma F, De Rosa C (2003) Time-resolved study of the martensitic phase transition in
syndiotactic polypropylene. Macromolecules 36:9396–9410
59. Auriemma F, De Rosa C, Esposito S, Mitchell GR (2007) Polymorphic superelasticity in
semicrystalline polymers. Angew Chem Int Ed 46:4325–4328
60. Kazmierczak T, Galeski A, Galeski A, Argon AS (2005) Plastic deformation of polyethylene
crystals as a function of crystal thickness and compression rate. Polymer 46:8926–8936
61. Argon AS, Galeski A, Kazmierczak T (2005) Rate mechanisms of plasticity in semi-
crystalline polyethylene. Polymer 46:11798–11805
62. Schrauwen BAG, van Breemen LCA, Spoelstra AB, Govaert LE, Peters GWM, Meijer HEH
(2004) Structure, deformation, and failure of flow-oriented semicrystalline polymers. Mac-
romolecules 37:8618–8633
63. Sedighiamiri A, Govaert LE, Kanters MJW, van Dommelen JAW (2012) Micromechanics of
semicrystalline polymers: yield kinetics and long-term failure. J Polym Sci B Polym Phys
50:1664–1679
64. van Erp TB, Cavallo D, Peters GWM, Govaert LE (2012) Rate-, temperature-, and structure-
dependent yield kinetics of isotactic polypropylene. J Polym Sci B Polym Phys
50:1438–1451
65. van Breemen LCA, Engels TAP, Klompen ETJ, Senden DJA, Govaert LE (2012) Rate- and
temperature-dependent strain softening in solid polymers. J Polym Sci B Polym Phys
50:1757–1771
66. Hobeika S, Men Y, Strobl G (2000) Temperature and strain rate independence of critical
strains in polyethylene and poly(ethylene-co-vinyl acetate). Macromolecules 33:1827–1833
67. Brooks NWJ, Mukhtar M (2000) Temperature and stem length dependence of the yield stress
of polyethylene. Polymer 41:1475–1480
68. Peterson JM (1966) Thermal initiation of screw dislocations in polymer crystal platelets. J
Appl Phys 37:4047–4050
69. Peterson JM (1968) Peierls stress for screw dislocations in polyethylene. J Appl Phys
39:4920–4928
88 F. Auriemma et al.

70. Crist B, Fisher CJ, Howard P (1989) Mechanical properties of model polyethylenes: tensile
elastic modulus and yield stress. Macromolecules 22:1709–1718
71. Darras O, Seguela R (1993) Tensile yield of polyethylene in relation to crystal thickness. J
Polym Sci B Polym Phys 31:759–766
72. O’Kane WJ, Young RJ, Ryan AJ (1995) The effect of annealing on the structure and
properties of isotactic polypropylene films. J Macromol Sci Phys 34:427–458
73. Bartczak Z, Argon AS, Cohen RE (1992) Deformation mechanisms and plastic resistance in
single-crystal-textured high-density polyethylene. Macromolecules 25:5036–5053
74. Gleiter H, Argon AS (1971) Plastic deformation of polyethylene crystals. Philos Mag
24:71–80
75. Lewis D, Wheeler EJ, Maddams WF, Preedy JE (1972) Comparison of twinning produced by
rolling and annealing in high- and low-density polyethylene. J Polym Sci A-2 Polym Phys
10:369–373
76. Young RJ, Bowden PB, Ritchie J, Rider RJ (1973) Deformation mechanisms in oriented
high-density polyethylene. J Mater Sci 8:23–36
77. Gaucher-Miri V, Seguela R (1997) Tensile yield of polyethylene and related copolymers:
mechanical and structural evidences of two thermally activated processes. Macromolecules
30:1158–1167
78. Seguela R, Gaucher-Miri V, Elkoun S (1998) Plastic deformation of polyethylene and
ethylene copolymers. Part I. Homogeneous crystal slip and molecular mobility. J Mater Sci
33:1273–1279
79. Seguela R, Elkoun S, Gaucher-Miri V (1998) Plastic deformation of polyethylene and
ethylene copolymers. Part II. Heterogeneous crystal slip and strain-induced phase change. J
Mater Sci 33:1801–1807
80. Peterlin A (1971) Molecular model of drawing polyethylene and polypropylene. J Mater Sci
6:490–508
81. Peterlin A (1965) Crystalline character in polymers. J Polym Sci Part C Polym Symp 9:61–89
82. Bartczak Z, Cohen RE, Argon AS (1992) Evolution of the crystalline texture of high-density
polyethylene during uniaxial compression. Macromolecules 25:4692–4704
83. Bartczak Z, Lezak E (2005) Evolution of lamellar orientation and crystalline texture of
various polyethylenes and ethylene-based copolymers in plane-strain compression. Polymer
46:6050–6063
84. Galeski A (2003) Strength and toughness of crystalline polymer systems. Prog Polym Sci
28:1643–1699
85. Men Y, Rieger J, Strobl G (2003) Role of the entangled amorphous network in tensile
deformation of semicrystalline polymers. Phys Lett 91:95502
86. Al-Hussein M, Strobl G (2002) Strain-controlled tensile deformation behavior of isotactic
poly(1-butene) and its ethylene copolymers. Macromolecules 35:8515–8520
87. Men Y, Strobl G (2003) Critical strains in poly(ε-caprolactone) and blends with poly(vinyl
methyl ether) and poly(styrene-co-acrylonitrile). Macromolecules 36:1889–1898
88. Hong K, Rastogi A, Strobl G (2004) A model treating tensile deformation of semicrystalline
polymers: quasi-static stress–strain relationship and viscous stress determined for a sample of
polyethylene. Macromolecules 37:10165–10174
89. Brown N, Ward IM (1983) The influence of morphology and molecular weight on ductile-
brittle transitions in linear polyethylene. J Mater Sci 18:1405–1420
90. Seguela R, Darras O (1994) Phenomenological aspects of the double yield of polyethylene
and related copolymers under tensile loading. J Mater Sci 29:5342–5347
91. Bartczak Z (2005) Influence of molecular parameters on high-strain deformation of polyeth-
ylene in the plane-strain compression. Part II. Strain recovery. Polymer 46:10339–10354
92. Bowden B, Young RJ (1971) Critical resolved shear stress for [001] slip in polyethylene.
Nature 229:23–25
93. Shadrake LG, Guiu F (1976) Dislocations in polyethylene crystals: line energies and defor-
mation modes. Philos Mag 34:565–581
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 89

94. O’Kane WJ, Young RJ (1995) The role of dislocations in the yield of polypropylene. J Mater
Sci Lett 14:433–435
95. Sirotkin RO, Brooks NW (2001) The effects of morphology on the yield behaviour of
polyethylene copolymers. Polymer 42:3791–3797
96. Popli R, Mandelkern L (1987) Influence of structural and morphological factors on the
mechanical properties of the polyethylenes. J Polym Sci B Polym Phys 25:441–483
97. De Rosa C, Auriemma F, de Ballesteros OR, Resconi L, Camurati I (2007) Crystallization
behavior of isotactic propylene-ethylene and propylene-butene copolymers: effect of como-
nomers versus stereodefects on crystallization properties of isotactic polypropylene. Macro-
molecules 40:6600–6616
98. De Rosa C, Auriemma F, de Ballesteros OR, Resconi L, Camurati I (2007) Tailoring the
physical properties of isotactic polypropylene through incorporation of comonomers and the
precise control of stereo and regioregularity by metallocene catalysts. Chem Mater
19:5122–5130
99. De Rosa C, Auriemma F, Vollaro P, Resconi L, Guidotti S, Camurati I (2011) Crystallization
behavior of propylene-butene copolymers: the trigonal form of isotactic polypropylene and
form I of isotactic poly(1-butene). Macromolecules 44:540–549
100. De Rosa C, Auriemma F, Ruiz de Ballesteros O, De Luca D, Resconi L (2008) The double
role of comonomers on the crystallization behavior of isotactic polypropylene: propylene-
hexene copolymers. Macromolecules 41:2172–2177
101. De Rosa C, Auriemma F, Ruiz de Ballesteros O, Dello Iacono S, De Luca D, Resconi L
(2009) Stress-induced polymorphic transformations and mechanical properties of isotactic
propylene-hexene copolymers. Cryst Growth Des 9:165–176
102. De Rosa C, Auriemma F, Di Girolamo R, Romano L, De Luca MR (2010) A new mesophase
of isotactic polypropylene in copolymers of propylene with long branched comonomers.
Macromolecules 43:8559–8569
103. De Rosa C, Auriemma F, Ruiz de Ballesteros O, Di Caprio MR (2012) Crystal structure of the
trigonal form of isotactic propylene–pentene copolymers: an example of the principle of
entropy–density driven phase formation in polymers. Macromolecules 45:2749–2763
104. De Rosa C, Auriemma F, Talarico G, de Ballesteros OR (2007) Structure of isotactic
propylene-pentene copolymers. Macromolecules 40:88531–88532
105. De Rosa C, Auriemma F, Corradini P, Tarallo O, Dello Iacono S, Ciaccia E, Resconi L (2006)
J Am Chem Soc 128:80–81
106. De Rosa C, Dello Iacono S, Auriemma F, Ciaccia E, Resconi L (2006) Macromolecules
39:6098–6109
107. Bingel C, Goeres M, Fraaije V, Winter A (1998) Preparation of preparing substituted
indanones. International patent application WO 1998/040331
108. Resconi L, Ciaccia E, Fait A (2004) Process for preparing porous polymers and polymers
thereof. International patent application WO 2004/092230
109. Resconi L, Guidotti S, Camurati I, Frabetti R, Focante F, Nifant’ev IE, Laishevtsev IP (2005)
C1-symmetric heterocyclic zirconocenes as catalysts for propylene polymerization. 2. ansa-
zirconocenes with linked diethienocyclopentadienyl-substituted indenyl ligands. Macromol
Chem Phys 206:1405–1438
110. Spaleck W, Kueber F, Winter A, Rohrmann J, Bachmann B, Antberg M, Dolle V, Paulus EF
(1994) The influence of aromatic substituents on the polymerization behavior of bridged
zirconocene catalysts. Organometallics 13:954–963
111. Hosier IL, Alamo RG, Esteso P, Isasi JR, Mandelkern L (2003) Formation of the α and γ
polymorphs in random metallocene–propylene copolymers. Effect of concentration and type
of comonomer. Macromolecules 36:623–5636
112. Ruiz-Orta C, Alamo RG (2012) Morphological and kinetic partitioning of comonomer in
random propylene 1-butene copolymers. Polymer 53:810–822
113. Ruiz-Orta C, Fernandez-Blazquez JP, Pereira EJ, Alamo RG (2011) Time-resolved FTIR
spectroscopic study of the evolution of helical structure during isothermal crystallization of
90 F. Auriemma et al.

propylene 1-hexene copolymers. Identification of regularity bands associated with the trigo-
nal polymorph. Polymer 52:2856–2868
114. Jeon K, Palza H, Quijada R, Alamo RG (2009) Effect of comonomer type on the crystalli-
zation kinetics and crystalline structure of random isotactic propylene 1-alkene copolymers.
Polymer 50:832–844
115. Alamo RG, Ghosal A, Chatterjee J, Thompson KL (2005) Linear growth rates of random
propylene ethylene copolymers. The changeover from γ dominated growth to mixed (α + γ)
polymorphic growth. Polymer 46:8774–8789
116. Alamo RG, VanderHart DL, Nyden MR, Mandelkern L (2000) Morphological partitioning of
ethylene defects in random propylene-ethylene copolymers. Macromolecules 33:6094–6105
117. Alamo RG, Mandelkern L (1991) Crystallization kinetics of random ethylene copolymers.
Macromolecules 24:6480–6493
118. Covezzi M, Fait A (2001) Process and apparatus for making supported catalyst systems for
olefin polymerisation. International patent application WO 01/44319
119. De Rosa C, Auriemma F, Di Capua A, Resconi L, Guidotti S, Camurati I, Nifant’ev IE,
Laishevtsev IP (2004) Structure–property correlations in polypropylene from metallocene
catalysts: stereodefective, regioregular isotactic polypropylene. J Am Chem Soc 2004
(126):17040
120. Roe R-J (2000) Methods of X-ray and neutron scattering in polymer science. Oxford
University Press, New York
121. Stribeck N (2007) X-ray scattering of soft-matter. Springer, Heidelberg
122. Resconi L, Cavallo L, Fait A, Piemontesi F (2000) Selectivity in propene polymerization with
metallocene catalysts. Chem Rev 100:1253–1346
123. Fischer D, Mulhaupt R (1994) The influence of regio- and stereoirregularities on the
crystallization behaviour of isotactic poly(propylene)s prepared with homogeneous group
IVa metallocene/methylaluminoxane Ziegler-Natta catalysts. Macromol Chem Phys
195:1433–1441
124. Thomann R, Wang C, Kressler J, Mulhaupt R (1996) On the γ-phase of isotactic polypro-
pylene. Macromolecules 29:8425–8434
125. Alamo RG, Kim MH, Galante MJ, Isasi JR, Mandelkern L (1999) Structural and kinetic
factors governing the formation of the γ polymorph of isotactic polypropylene. Macromol-
ecules 32:4050–4064
126. Auriemma F, De Rosa C (2002) Crystallization of metallocene-made isotactic polypropyl-
ene: disordered modifications intermediate between the α and γ forms. Macromolecules
35:9057
127. De Rosa C, Auriemma F, De Lucia G, Resconi L (2005) From stiff plastic to elastic
polypropylene: polymorphic transformations during plastic deformation of metallocene-
made isotactic polypropylene. Polymer 46:9461–9475
128. De Rosa C, Auriemma F, Paolillo M, Resconi L, Camurati I (2005) Crystallization behavior
and mechanical properties of regiodefective, highly stereoregular isotactic polypropylene:
effect of regiodefects versus stereodefects and influence of the molecular mass. Macromol-
ecules 38:9143–9154
129. VanderHart DL, Alamo RG, Nyden MR, Kim MH, Mandelkern L (2000) Observation of
resonances associated with stereo and regio defects in the crystalline regions of isotactic
polypropylene: toward a determination of morphological partitioning. Macromolecules
33:6078–6093
130. Thomann R, Semke H, Maier RD, Thomann Y, Scherble J, Mulhaupt R, Kressler J (2001)
Influence of stereoirregularities on the formation of the γ-phase in isotactic polypropene.
Polymer 42:4597–4603
131. Nyden MR, Vanderhart DL, Alamo RG (2001) The conformational structures of defect-
containing chains in the crystalline regions of isotactic polypropylene. Comput Theor Polym
Sci 11:175–189
Molecular View of Properties of Random Copolymers of Isotactic Polypropylene 91

132. Poon B, Rogunova M, Hiltner A, Baer E, Chum SP, Galeski A, Piorkowska E (2005)
Structure and properties of homogeneous copolymers of propylene and 1-hexene. Macro-
molecules 38:1232–1243
133. Lotz B, Ruan J, Thierry A, Alfonso GC, Hiltner A, Baer E, Piorkowska E, Galeski A (2006) A
structure of copolymers of propene and hexene isomorphous to isotactic poly(1-butene) form
I. Macromolecules 39:5777–5781
134. Stagnaro P, Boragno L, Canetti M, Forlini F, Azzurri F, Alfonso GC (2009) Crystallization
and morphology of the trigonal form in random propene/1-pentene copolymers. Polymer
50:5242–5249
135. Stagnaro P, Costa G, Trefiletti V, Canetti M, Forlini F, Alfonso GC (2006) Thermal behavior,
structure and morphology of propene/higher 1-olefin copolymers. Macromol Chem Phys
207:2128–2141
136. Garcia-Penas A, Gomez-Elvira JM, Lorenzo V, Perez E, Cerrada ML (2015) Synthesis,
molecular characterization, evaluation of polymorphic behavior and indentation response in
isotactic poly(propylene-co-1-heptene) copolymers. Eur Polym J 64:52–61
137. Polo-Corpa MJ, Benavente R, Velilla T, Quijada R, Perez E, Cerrada ML (2010) Develop-
ment of the mesomorphic phase in isotactic propene/higher alpha-olefin copolymers at
intermediate comonomer content and its effect on properties. Eur Polym J 46:1345–1354
138. Palza H, Lopez-Majada JM, Quijada R, Perena JM, Benavente R, Perez E, Cerrada ML
(2008) Comonomer length influence on the structure and mechanical response of
metallocenic polypropylenic materials. Macromol Chem Phys 209:2259–2267
139. Lopez-Majada JM, Palza H, Guevara JL, Quijada R, Martinez MC, Benavente R, Perena JM,
Perez E, Cerrada ML (2006) Metallocene copolymers of propene and 1-hexene: the influence
of the comonomer content and thermal history on the structure and mechanical properties. J
Polym Sci B Polym Phys 44:1253–1267
140. Garcia-Penas A, Cerrada ML, Gomez-Elvira JM, Perez E (2016) Microstructure and thermal
stability in metallocene iPP-materials: 1-pentene and 1-hexene copolymers. Polym Degrad
Stab 124:77–86
141. Natta G, Peraldo M, Corradini P (1959) Modificazione mesomorfa smettica del polipropilene
isotattico. Rend Accad Naz Lincei 26:14–17
142. Mileva D, Androsch R, Cavallo D, Alfonso GC (2012) Structure formation of random
isotactic copolymers of propylene and 1-hexene or 1-octene at rapid cooling. Eur Polym J
48:1082–1092
143. Mileva D, Androsch R (2012) Effect of co-unit type in random propylene copolymers on the
kinetics of mesophase formation and crystallization. Colloid Polym Sci 290:465–471
144. Mileva D, Androsch R, Zhuravlev E, Schick C, Wunderlich B (2011) Formation and
reorganization of the mesophase of random copolymers of propylene and 1-butene. Polymer
52:1107–1115
145. Mileva D, Cavallo D, Gardella L, Alfonso GC, Portale G, Balzano L, Androsch R (2011) In
situ X-Ray analysis of mesophase formation in random copolymers of propylene and
1-butene. Polym Bull 67:497–510
146. Mileva D, Zia Q, Androsch R (2010) Tensile properties of random copolymers of propylene
with ethylene and 1-butene: effect of crystallinity and crystal habit. Polym Bull 65:623–634
147. Mileva D, Androsch R, Zhuravlev E, Schick C (2009) Critical rate of cooling for suppression
of crystallization in random copolymers of propylene with ethylene and 1-butene.
Thermochim Acta 492:67–72
148. Mileva D, Androsch R, Radusch H-J (2008) Effect of cooling rate on melt-crystallization of
random propylene-ethylene and propylene-1-butene copolymers. Polym Bull 61:643–654
149. Natta G, Corradini P (1960) Structures and properties of isotactic polypropylene. Nuovo
Cimento Suppl 15:40–51
150. Hikosaka M, Seto T (1973) The order of the molecular chains in isotactic polypropylene
crystals. Polym J 5:111–127
92 F. Auriemma et al.

151. Bruckner S, Meille SV (1989) Non-parallel chains in crystalline γ-isotactic polypropylene.

Nature (London) 340:455–457
152. Meille SV, Bruckner S, Porzio W (1990) γ-Isotactic polypropylene. A structure with
nonparallel chain axes. Macromolecules 23:4114–4121
153. Balta-Calleja FJ, Vonk CG (1989) X ray scattering of synthetic polymers. Elsevier,
Amsterdam
154. Wunderlich B (1980) Macromolecular physics, vol 3. Academic, New York
155. Crist B (2003) Thermodynamics of statistical copolymer melting. Polymer 44:4563–4572
156. Crist B, Claudio ES (1999) Isothermal crystallization of random ethylene–butene copoly-
mers: bimodal kinetics. Macromolecules 32:8945–8951
157. Crist B, Williams DN (2000) Crystallization and melting of model ethylene-butene random
copolymers: thermal studies. J Macromol Sci Phys B39:1–13
158. Crist B, Howard PR (1999) Crystallization and melting of model ethylene–butene copoly-
mers. Macromolecules 32:3057–3067
159. Flory PJ (1951) Theory of crystallization in copolymers. Trans Faraday Soc 55:848–857
160. Sanchez IC, Eby RK (1973) Crystallization of random copolymers. J Res Natl Bur St A Phys
Chem 77:353–358
161. Sanchez IC, Eby RK (1975) Thermodynamics and crystallization of random copolymers.
Macromolecules 8:638–641
162. De Rosa C, Auriemma F (2013) Crystals and crystallinity in polymers: diffraction analysis of
ordered and disordered crystals. Wiley, New York
163. Zhen L, You-lee H, Shichen Y, Jia K, Akihiro K, Akihiro O, Toshikazu M (2015) Determi-
nation of chain-folding structure of isotactic polypropylene in melt-grown α crystals by 13C–
13
C double quantum NMR and selective isotopic labeling. Macromolecules 48:5752–5760
164. Nitta KH, Yamana M (2012) Poisson’s ratio and mechanical nonlinearity under tensile
deformation in crystalline polymers, rheology. In: De Vicente J (ed) Rheology. InTech,
Rijeka, Croatia, pp 113–132. Available from: http://www.intechopen.com/books/rheology/
poisson-s-ratio-and-mechanical-nonlinearity-undertensile-deformation
165. Petraccone V, Pirozzi B, Meille SV (1986) Analysis of chain folding in crystalline isotactic
polypropylene. The implications of tacticity and crystallographic symmetry. Polymer
27:1665–1668
166. Corradini P, Giunchi G, Petraccone V, Pirozzi B, Vidal HM (1980) Structural variations in
crystalline isotactic polypropylene (α form) as a function of thermal treatments. Gazz Chim
Ital 110:413–418
167. Auriemma F, de Ballesteros OR, De Rosa C, Corradini P (2000) Structural disorder in the α
form of isotactic polypropylene. Macromolecules 33:8764–8774
168. Ree T, Eyring H (1955) Theory of non-Newtonian flow. I. Solid plastic system. J Appl Phys
26:793–800
169. Ree T, Eyring H (1955) Theory of non‐Newtonian flow. I. Solution system of high polymers.
J Appl Phys 26:800–809
170. Séguéla R, Staniek E, Escaig B, Fillon B (1999) Plastic deformation of polypropylene in
relation to crystalline structure. J Appl Polym Sci 71:1873–1885
171. Schaefer D, Spiess HW, Suter UW, Fleming WW (1990) Two-dimensional solid-state NMR
studies of ultraslow chain motion: glass transition in atactic poly(propylene) versus helical
jumps in isotactic poly(propylene). Macromolecules 23:3431–3439
Adv Polym Sci (2017) 276: 93–132
DOI: 10.1007/12_2015_326
© Springer International Publishing Switzerland 2015
Published online: 11 August 2015

Crystallization of Cyclic Polymers

Ricardo A. Pérez-Camargo, Agurtzane Mugica, Manuela Zubitur,

and Alejandro J. Müller

Abstract The effect of chain topology (ring versus linear polymer chains) on
polymer crystallization is reviewed. Recent advances in ring closure and ring
expansion synthetic techniques have made available a range of well-characterized
samples with higher levels of purity than available decades ago. Cyclic molecules
are fascinating because the structural difference between them and linear chains is
relatively small, yet their behavior can be completely different from that of their
linear analogs of identical chain length. The effect of having no chain ends can
dramatically change the polymer coil conformation and diffusion rate, as well as the

R.A. Pérez-Camargo
POLYMAT and Polymer Science and Technology Department, Faculty of Chemistry,
University of the Basque Country UPV/EHU, Paseo Manuel de Lardizabal 3, 20018 Donostia/
San Sebastián, Spain
Grupo de Polı́meros USB, Departamento de Ciencia de los Materiales, Universidad Sim

on
Bolı́var, Apartado 89000, Caracas 1080, Venezuela
A. Mugica
POLYMAT and Polymer Science and Technology Department, Faculty of Chemistry,
University of the Basque Country UPV/EHU, Paseo Manuel de Lardizabal 3, 20018 Donostia/
San Sebastián, Spain
M. Zubitur
Chemical Engineering Department, Polytechnic College, University of the Basque Country
UPV/EHU, Plaza de Europa, 20080 Donostia/San Sebastián, Spain
A.J. Müller (*)
POLYMAT and Polymer Science and Technology Department, Faculty of Chemistry,
University of the Basque Country UPV/EHU, Paseo Manuel de Lardizabal 3, 20018 Donostia/
San Sebastián, Spain
Grupo de Polı́meros USB, Departamento de Ciencia de los Materiales, Universidad Sim

on
Bolı́var, Apartado 89000, Caracas 1080, Venezuela
IKERBASQUE, Basque Foundation for Science, Bilbao, Spain
e-mail: [email protected]
94 R.A. Pérez-Camargo et al.

chain entanglement density in the melt. These changes are reflected in different
nucleation and crystallization kinetics for cyclic and linear polymeric chains.
However, the results published so far seem to be dependent on the type of polymer
employed. Therefore, a careful look at the literature and evidence reported for each
group of materials has been assembled and compared. The possible reasons for
some of the contradictions in the evidence are also discussed.

Keywords Crystallization
Cyclic and linear polymers
Diffusion
Entropic

factors
Equilibrium melting temperature

Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
2 Types of Cyclic Polymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
3 Crystallization from Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
3.1 Single Crystals Obtained from Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
4 Crystallization from the Melt State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.1 Nucleation and Spherulitic Structures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102
4.2 Spherulitic Growth Rate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.3 Non-isothermal Differential Scanning Calorimetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
4.4 Isothermal Overall Crystallization Kinetics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5 Self-Nucleation and Successive Self-Nucleation and Annealing . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.1 Self-Nucleation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
5.2 Successive Self-Nucleation and Annealing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
6 Reasons for Different Behavior Reported in the Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 122
7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 124
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

1 Introduction

Cyclic polymers are fascinating materials. They only differ from linear analogs in
their lack of chain ends; however, this topological difference can have a large
impact on their molecular behavior.
Cyclic molecules are present in nature. In fact, the interest in cyclic polymers
began in 1958 when Jacob and Wollman [1] concluded that the genetic map of
bacterial chromosomes of Eschericha coli showed circularity. Additionally, DNA
[2–5], peptides [6], and proteins [7] are synthesized in nature with cyclic topolo-
gies. Therefore, an open question naturally arises: Do cyclic topologies provide
enhanced or unique properties compared with linear structures? [8]. According to
Semlyen [9], the absence of chain ends could be an advantage in some applications,
and natural structures (e.g., DNA) with circular topologies are designed to prevent
any possible reactions or interactions through chain ends.
Cyclic or ring polymers develop fewer entanglements and have more compact
molecular conformations than linear chains. As a consequence, properties such as
Crystallization of Cyclic Polymers 95

diffusion rates, melt and solution viscosities, nucleation and crystallization behav-
ior, and degradation kinetics can be dramatically different for chains that have no
ends as compared with linear analogs. In this review, we focus on the effect of this
difference in chain topology (namely cyclic versus linear) on the morphology,
nucleation, and crystallization behavior of polymers.
The crystallization of cyclic polymers is a complex subject that needs investi-
gation in order to understand all relevant factors involved. The synthetic procedures
and purification methods employed in the production of ring polymers with high
purity have been evolving for decades [10]. Nowadays, new synthetic approaches
have allowed the preparation of a wide range of high purity ring polymers, as well
as novel and more complex cyclic-based topologies [11–29]. These new synthetic
approaches have enabled researchers to study differences between the properties of
cyclic and linear polymers, such as the glass transition temperature [9, 30], melt
viscosity and diffusion [9, 30–34], morphology [35–37], and crystallization [38–
52].
Crystallization studies of several cyclic and linear polymers have been
performed and the results obtained have differed substantially from one another.
Therefore, a unique interpretation is not always possible and it seems that the
crystallization of cyclic chains depends on the specific type of material under
study, apart from other factors such as different levels of purity.
Many different aspects of the morphology and crystallization of cyclic polymers
have been reported in the literature, such as solution-grown single crystal morphol-
ogy, nucleation, spherulitic growth, thermal transition, and overall melt crystalli-
zation kinetics. Several arguments have been employed in an effort to explain the
obtained findings, leading to the development of models for chain folding of cyclic
polymers and taking into account cyclic topology in classical thermodynamic
equations. A revision of the most important experimental findings on cyclic and
linear polymer crystallization, as well as their interpretation, is presented in this
review.

2 Types of Cyclic Polymers

The crystallization of several types of cyclic polymers is presented in this work.

These polymers are synthesized by two main types of cyclization synthetic
methodologies:
(a) Ring-closure reaction
(b) Ring-expansion polymerization [53]
Several reviews focus on the synthetic strategies [54–56], which are beyond the
scope of the present work. To illustrate the two main types of cyclization synthetic
methodologies, poly(ε-caprolactone) (PCL) has been selected as an example
because cyclic PCLs have been synthesized by both ring-closure reactions and
ring-expansion polymerizations.
96 R.A. Pérez-Camargo et al.

O
O
L-PCL-OH 1-pentynoic
anhydride
Sn(Oct)2 DMAP O
N3 OH N3 O H N3 O
130oC On DCM/pyr On
O O H

C-PCL
O Cu(I)Br N N
N3 O PMDETA N
On O
O H O
DCM On
O

Scheme 1 Synthetic pathway employed to obtain cyclic PCLs (C-PCLs). The precursor linear
polymer (L-PCL-OH) can be used as a linear analog that has almost the same molecular weight as
the synthesized C-PCL. Reprinted with permission from [48]

Ring-closure reactions are based on traditional coupling routes for cyclization of

linear precursor polymers under conditions of high dilution [54]. It is worth noting
that ring-closure has been improved thanks to “click” reactions in combination with
“living” radical polymerizations and electrostatic self-assembly and covalent fixa-
tion (ESA-CF) [56]. An example of one possible synthetic pathway for obtaining
cyclic PCLs by ring-closure reactions is shown in Scheme 1.
Scheme 1 shows the synthetic pathway employed to obtain a linear PCL
precursor (L-PCL-OH) by ring-opening polymerization of ε-caprolactone. Then,
a bis-functional cyclization reaction is performed by Cu(I)-catalyzed azide–alkyne
cycloaddition, leading to the cyclic polymer (C-PCL in Scheme 1) (more details are
given in [47, 48]). Normally, the chemical structure of the resulting linking unit
(a triazole ring in this case) is different from the polymer repeating unit, although it
is possible in some cases (e.g., for poly(tetrahydrofuran) cyclic materials), to
transform the linking unit into a group identical to the repeating unit (see [40]).
Ring-expansion polymerization is a direct method for forming cyclic polymers.
For ring-expansion polymerization, the monomer inserts into a cyclic initiator to
form larger rings. C-PCL has also been obtained by ring-expansion polymerization.
Scheme 2 illustrates the synthetic pathway used by Shin et al. [43] to obtain C-PCL
by zwitterionic polymerization, which is a process of enchainment whereby the
ionic propagating end and its counterion are contained in the same polymer chain of
ε-caprolactone in the presence of carbene [54].
Tables 1 and 2 show the molecular weight data for selected cyclic polymers
obtained by ring-closure reactions (Table 1) and ring-expansion polymerizations
(Table 2), and their linear counterparts. They have been chosen because the
crystallization of these samples is discussed later in this review (see Sects. 3 and 4).
One of the advantages of ring-closure techniques is that linear analogs have
almost identical number-average molecular weight (Mn) values (and polydispersity
values) as cyclic molecules. This is because the linear molecules are synthesized
first, functionalized with clickable groups, and then reacted to form the cyclic
chains. Once the cycle is closed, a triazole ring remains within the structure of
Crystallization of Cyclic Polymers 97

Scheme 2 Proposed
mechanism for the
zwitterionic ring-opening
polymerization of
ε-caprolactone. Reprinted
with permission from [43]

Table 1 Number-average molecular weight (Mn) and polydispersity index (PDI) data for ring-
closure samples [40, 41, 45, 47–49]
Mn (g/mol) PDIa
Sample GPCb MALDI NMR GPC MALDI SEC References
L-PTHF 3.1k – – 3,100 – – 1.10 [45]
L-PTHF 4.9k – – 4,900 – – 1.14 [40]
L-PTHF 5.7k – – 5,700 – – 1.12 [45]
L-PTHF 9.1k – – 9,100 – – 1.09
C-PTHF 2.9k – – 2,900 – – 1.11
C-PTHF 4.5k – – 4,500 – – 1.15
C-PTHF 5.1k – – 5,100 – – 1.14 [40]
C-PTHF 8.2k – – 8,200 – – 1.10 [45]
L-PCL-Acet 2k 2,190 2,360 – 1.11 1.02 – [41]
L-PCL-OH 2k 2,010 2,040 – 1.17 1.05 –
L-PCL-OH 3k 3,440 3,140 – 1.15 1.03 – [48]
L-PCL-Acet 4.9k 4,730 4,900 – 1.16 1.01 – [47]
L-PCL-OH 7.5k 7,670 7,340 – 1.13 1.02 – [41]
L-PCL-OH 12k 12,000 12,000 – 1.12 1.03 – [48, 49]
L-PCL-Acet 22k 20,070 22,000 – 1.15 1.03 – [48]
C-PCL 2k 1,550 2,320 – 1.17 1.05 – [41]
C-PCL 3k 2,180 3,200 – 1.15 1.03 – [48]
C-PCL 4.9k 3,390 5,040 – 1.08 1.02 – [47]
C-PCL 7.5k 4,800 7,000 – 1.11 1.04 – [41]
C-PCL 12k 10,580 12,000 – 1.15 1.04 – [48]
C-PCL 22k 15,140 22,000 – 1.15 1.05 – [48, 49]
a
PDI ¼ Mw/Mn
b
Corrected value for linear PCL [57] using Mn(PCL) ¼ 0.259  Mn(PS)1.073
GPC gel permeation chromatography, MALDI matrix-assisted laser desorption/ionization, NMR
nuclear magnetic resonance, SEC size-exclusion chromatography

the cyclic chain (as shown in Scheme 1) that is not present in the linear chains;
hence, the slight difference in Mn values obtained by MALDI-TOF [58]. Table 1
shows two types of linear precursors for PCL (L-PCL-OH and L-PCL-Acet).
98 R.A. Pérez-Camargo et al.

Table 2 Weight- and Polymer Mw (g/mol) Mn (g/mol) PDI References

number-average molecular
L-PE 44k 44,000 – – [44]
weights (Mw and Mn) and PDI
data for ring-expansion L-PE 200k 200,000 – 2.0 [31]
polymerization samples [31, C-PE 9k 9,300 – – [44]
42–44] C-PE 43.6k 43,600 – –
C-PE 86.5k 86,500 – –
C-PE 200k 200,000 – 2.0 [31]
L-PCL 50.6k – 50,600 1.6 [42]
L-PCL 63.4k – 63,400 1.6
L-PCL 69.2k – 69,200 1.8
L-PCL 75k 75,000a 103,000 1.38 [43]
L-PCL 77k 77,000 – 2.1 [42]
L-PCL 100k 101,000a 140,000 1.54 [43]
L-PCL 120k 116,000a 168,000 1.29
C-PCL 45k – 45,000 1.6 [42]
C-PCL 57.5k – 57,500 1.6
C-PCL 64k – 64,000 1.8
C-PCL 69k – 69,000 2.1
C-PCL 75k 75,000a 66,000 1.91 [43]
C-PCL 110k 109,000a 79,000 1.83
C-PCL 130k 129,000a 101,000 2.02
C-PCL 140k 142,000a 114,000 2.03
a
Weight-average (absolute) molecular weight determined by gel
permeation chromatography using light scattering

The L-PCL-Acet is the product of functionalization of L-PCL-OH to give an acetate

group at the end of the molecule. This was originally performed to ascertain
whether the OH end groups could affect the diffusion and crystallizability of PCL
chains because of their ability to form hydrogen bonds. In fact, it was demonstrated
that the behavior of both L-PCL-OH and L-PCL-Acet was identical in terms of
crystallization behavior [41]. They can both be considered very similar linear PCL
chains.
Ring-closure techniques normally produce high purity samples for oligomers
and low to intermediate molecular weights (i.e., Mn values in the range 2–22 kg/
mol) with very low polydispersities. Ring-expansion techniques, on the other hand,
can produce high molecular weight cyclic molecules without any chemically
different moiety within the cyclic chain. However, linear analogs must be synthe-
sized separately, with target Mn values that are similar (but not identical) to those of
the cyclic molecules. Also, molecular weight dispersity values can be slightly
different between cyclic and linear molecules. Additionally, the polydispersity of
samples prepared by ring expansion is much larger than for samples obtained by
ring closure (compare Tables 1 and 2).
Crystallization of Cyclic Polymers 99

3 Crystallization from Solution

One of the advantages of crystallization from dilute solution is the minimal

molecular diffusion hindrance and the lack of isothermal thickening during crys-
tallization. Su et al. [47] studied, for the first time, single crystals obtained from
solutions of cyclic PCL samples and their linear counterparts.

3.1 Single Crystals Obtained from Solution

Su et al. [47] prepared single crystals of linear and cyclic PCL samples by
isothermal crystallization in dilute n-hexanol and N,N-dimethylacetamide solutions
at temperatures of 40 C and 26 C, respectively. The samples were obtained by
ring-closure techniques (see Scheme 1). Identical batches of the parent L-PCL-OH
were functionalized with acetic anhydride to provide linear analogs (L-PCL-Acet)
without a hydrogen-bond donating hydroxyl end group. No influence of the end
group was found in the crystallization behavior of the samples.
The most important findings of Su et al. [47] were that the morphology of C-PCL
and L-PCL single crystals was similar [59], but that there were differences in the
degree of truncation and in the average lamellar thickness. Figure 1 presents
selected TEM micrographs and electron diffraction patterns of linear and cyclic
5.1 kg/mol PCLs [47]. Figure 1a, c shows TEM micrographs in which the distorted
hexagonal-shaped morphology of the lamellar crystals is observed.
Figure 1a, c shows multilayered crystals with spiral growths of different hand-
edness. Another feature, such as striations is also observed in the micrographs (see
[47] for details). These morphological characteristics are also found in the atomic
force microscopy (AFM) micrographs shown in Fig. 2. It is worth noting that no
major differences between linear and cyclic PCLs were found.
Figure 1b, d shows electron diffraction patterns from selected crystals
(Mn ¼ 5.1 kg/mol). These patterns are highly regular and show more than 25 inde-
pendent reflections, which can be indexed according to the unit cell reported for
PCL [60–62] (a ¼ 0.747 nm, b ¼ 0.498 nm and c (fiber axis) ¼ 1.705–1.729 nm)
[47]. Su et al. [47] found that the crystal structures of single crystals of linear and
cyclic PCLs are the same. For samples of linear and cyclic PCL [43] and poly
(tetrahydrofuran) (PTHF) [45] crystallized from the melt, wide angle X-ray scat-
tering (WAXS) data indicate that their crystal structures are identical at the unit cell
level.
According to the correlation of bright field TEM micrographs and selected area
electron diffraction patterns (see Fig. 1), single crystals are bounded by four {110}
faces, with two truncated {100} faces of variable dimensions. The degree of
truncation is defined as the ratio between the length of {100} and {110} faces
(see Fig. 1a, c). Su et al. [47] obtained clear differences between the degree of
truncation of cyclic (i.e., 1.8–1.9) and linear (i.e., 2.8–2.9) PCL single crystals.
100 R.A. Pérez-Camargo et al.

Fig. 1 TEM micrographs (a, c) and electron diffraction patterns (b, d) of lamellar crystals for
linear (a, b) and cyclic (c, d) 5.1 kg/mol PCL samples. Reprinted with permission from [47]

C-PCL 50
54.1 nm

5.1k 40.6 nm
40

30
26.0nm
20
13.2 nm
10

560nm 210nm 0
0 100 200 300 400 500

Fig. 2 AFM topographic (left) and height (middle) images and corresponding height profiles
(right) for 5.1 kg/mol C-PCL sample. Crystals were obtained from N,N-dimethyl acetamide at
26 C. Reprinted with permission from [47]
Crystallization of Cyclic Polymers 101

Table 3 Extended chain length values (L ) and their comparison with lamellar thickness values (l )
(data extracted from [47])
Sample L (nm) l (nm) L/l (number of chain folds)
PCL-OH 2ka 14.2 8.7 1.6
PCL-OH 5.1ka 37.4 9.8 3.8
PCL-OH 7.5ka 55.3 11.6 4.8
PCL-Acet 5.1ka 37.1 7.0 5.3
PCL-Acet 7.5kb 55.0 6.73 8.2
PCL-Acet 7.5kc 55.0 7.22 7.6
C-PCL 2ka 6.8 5.4 1.3
C-PCL 5.1ka 18.4 13.2 1.4
C-PCL 7.5ka 27.4 13.22 2.1
C-PCL 7.5kd 27.4 7.17 3.8
a
Values determined by AFM for solution-grown crystals at 26 C
b
Values determined by SAXS for melt-crystallized samples at 20 C
c
Values determined by SAXS for melt-crystallized samples at 30 C
d
Values determined by SAXS for melt-crystallized samples at 25 C

The degree of truncation has been correlated with the supercooling applied during
crystallization by varying the isothermal crystallization temperature [63–66]. The
lower the truncation degree, the higher the supercooling applied during crystal
growth (where supercooling is defined as the difference between the equilibrium
melting temperature and the crystallization temperature). When the same solvent
and the same crystallization temperature are employed, differences in the degree of
truncation indicate that the crystals were grown at different supercoolings. The only
explanation for this result is that cyclic and linear polymer crystals have different
equilibrium melting points. The study of PCL single crystal truncation performed
by Su et al. [47] provides direct experimental evidence for the higher supercooling
applied during cyclic PCL solution crystallization, hence C-PCLs must have a
higher equilibrium melting temperature (T0m ), as corroborated by Su
et al. employing the Gibbs–Thomson extrapolation (see below).
The average lamellar thickness of the 5.1 kg/mol C-PCL sample was determined
from captured AFM images, as shown in Fig. 2 [47]. Table 3 lists the calculated
extended chain values (L ) together with the lamellar thickness (l ) determined by
AFM. The L/l ratio represents the number of times that a chain should fold to be a
part of a single crystal.
The length of the extended chain was calculated by Su et al. [47] using the
following equation:

L ¼ nlfibre ; ð1Þ

where lfibre is the length of two repeating units that are placed in the ideal
intracrystalline chain conformation (i.e., 1.705 nm) according to the literature
[60, 61] and n is the number of such distances along the chain:
102 R.A. Pérez-Camargo et al.

Mn  Mchainends
n¼ : ð2Þ
2  MUR

Because functional groups at the chain ends (see Scheme 1) and at the center of the
C-PCL chains are rejected to the fold surfaces or amorphous zones of the crystals
(they cannot be included in the crystal lattice), they were not taken into account in
the calculations performed by Su et al. [47].
It should be noted that L values for cyclic chains are always about one half of
those corresponding to the extended chain length of their linear analogs, because
they are ring molecules and cannot be completely extended without forming one
fold. According to Table 3, most PCL samples undergo folding during crystalliza-
tion, even cyclic samples. As the molecular weight increases, the number of times
that the PCL chain folds also increases, as expected.
It is interesting to note that cyclic chains form thicker lamellae than linear
chains, in spite of their half extended chain lengths values. In the cases of 2 and
5.1 kg/mol C-PCLs, the chains crystallize in an almost-once-folded chain confor-
mation (the equivalent of completely stretching the cyclic chains). Some values
obtained by small angle X-ray scattering (SAXS) on melt crystallized samples are
included in Table 3 for comparison purposes.
Similar trends were found by Shin et al. [46] for linear and cyclic poly(L-lactide)
(PLLA), in which the lamellar thickness and long period of cyclic PLLA lamellae
are approximately 20% larger than those of linear PLLA. The authors concluded
that their results implied a lack of multiple chain folding in the cyclic PLLA chains,
as a consequence of a topological constraint on lamellar folding [37, 67, 68].
In contrast, in a recent work by Sugai et al. [51], cyclic polylactides were found
to display long spacing and crystal thickness values that were half of those obtained
with their linear counterparts. The authors attributed the results to the more compact
and flat conformation of cyclic chains compared with that of linear PLLA chains.
The authors indicated that these differences in lamellar thickness and long period
could be partly a result of the different thermal history of the samples [69].

4 Crystallization from the Melt State

4.1 Nucleation and Spherulitic Structures

The nucleation and growth stages of crystallization of cyclic and linear polymers
have been studied using polarized light optical microscopy (PLOM) by different
authors. There is consensus in the literature on the faster nucleation of cyclic
molecules as compared with their linear analogs. Several types of polymers have
been used to experimentally determine nucleation kinetics (even with different
synthetic procedures and possibly different purity grades), namely PTHF [40, 45],
polyethylene (PE) [44], and PCL [48, 49].
Crystallization of Cyclic Polymers 103

L–PTHF C–PTHF

100 µm 100 µm

2´30´´ 3´55´´

100 µm 100 µm

4´30´´ 6´48´´
Fig. 3 PLOM micrographs of spherulites of linear and cyclic PTHF at the indicated crystalliza-
tion times at Tc ¼ 11 C. Reprinted with permission from [45]

Takeshita et al. [45] reported that a higher number of nuclei were generated for
cyclic PTHF than for its linear counterpart, as clearly shown in Fig. 3. The authors
attributed this behavior to the possibility that C-PTHF has a higher chain density
than L-PTHF.
Pérez et al. [48] measured the density of nuclei as a function of time (see Fig. 4)
for linear and cyclic 4.9 kg/mol PCL samples. Figure 4 shows a higher nucleation
density for C-PCL than for its L-PCL analog. Pérez et al. [48] estimated that
nucleation density at saturation is approximately 7% higher for C-PCL than for
L-PCL at the same Tc (i.e., 54 C).
On the other hand, from data obtained at the beginning of nucleation (see Fig. 4),
the nucleation rates were determined to be 1  107 and 2  106 cm3 s1 for cyclic
and linear PCL, respectively (4.9 kg/mol samples). Because chains with ring
topology are characterized by their lack of chain ends and by their more collapsed
coil conformation in the melt state [70–73], it is reasonable to expect that cyclic
chains nucleate faster than linear chains.
Tezuka et al. [40] found differences in spherulitic morphology between linear
and cyclic PTHF. Figure 5 shows that L-PTHF exhibits classical spherulites with
Maltese cross extinction patterns, whereas C-PTFH displays not only Maltese cross
extinction but also distorted banding.
104 R.A. Pérez-Camargo et al.

C-PCL 4.9 k PCL-Acet 4.9 k

8
4x10

(N°nuclei/cm )
3
8
3x10

8
ρnuclei 2x10

8
1x10

Tc=47ºC
0
0 100 200 300 400 500
Time (s)
Fig. 4 Nucleation density determined by PLOM at 47 C for linear and cyclic 4.9 kg/mol PCL
samples. Reprinted with permission from [48]

50µm 50µm

Fig. 5 PLOM micrographs of spherulites for linear (left) and ring (right) PTHF crystallized at
10 C from the melt. Reprinted with permission from [40]

Banded spherulites indicate rotation of the optical indicatrix along the radial
direction. This rotation is a result of lamellar twisting. Lamellar twisting is a
complicated phenomenon that is believed to be caused by (a) cumulative
reorientation of lamellae at successive screw dislocation [74, 75] or (b) different
surface stresses on opposite fold surfaces of individual lamellae [76, 77]. Tezuka
et al. [40] speculated that, in C-PTHF, the banded spherulites are caused by the
surface stresses developed by the folded chains, such as uneven fold volume, and
that the morphological differences observed between linear and cyclic PTHFs
might be caused by the distinctive chain folding structures. To illustrate this, they
proposed the model presented in Fig. 6.
The representation of the chain conformation of linear and cyclic PTHF shows
that L-PTHF can crystallize with a monolayer adsorption on the crystal growth face
Crystallization of Cyclic Polymers 105

(a) Linear Polymer Mono-layer adsorption

(b) Cyclic Polymer

Double-Layer Adsorption

Random Coil Conformation Crystalline Growth face

Fig. 6 Chain folding conformation in (a) linear and (b) ring PTHF crystallized from the melt.
Images were extracted from [78]

Fig. 7 Optical micrographs of spherulites for linear (left) and ring (right) PCL of 2 kg/mol
crystallized at 41 C from the melt. Reprinted with permission from [41]

(see Fig. 6a), whereas C-PTHF is nucleated on the crystal growth face with a double
molecular layer (see Fig. 6b) [78].
According to the model, the chain folding surface in C-PTHF is built up by two
different chain folding directions, yielding different surface energies at top and
bottom lamellar surfaces (see Fig. 6a). Hence, this surface difference, according to
Tezuka et al. [40], is the cause of lamellar twisting and banded spherulites in
C-PTHF.
In contrast to PTHFs, banded spherulites were not found in linear or cyclic
PCLs, as shown in Fig. 7. PLOM images show that both types of spherulites are
similar, exhibiting Maltese cross extinction patterns without banding. The spheru-
lites were found to be negative, as expected for PCL [41].
106 R.A. Pérez-Camargo et al.

Table 4 Differences in spherulitic growth rates reported in the literature for cyclic and linear
analogs [39–41, 44, 45, 47, 49]
Material Mn (kg/mol) G (μm/min) (Tc ( C)) References
Case1
L-PTHF 3.1 13.9–0.86 (12–20) [45]
4.9 360 (20) [40]
5.7 8.76–0.64 (11–24) [45]
9.1 6.0–0.21 (11–24)
C-PTHF 2.9 1.5–0.32 (9–17) [45]
4.5 3.64–1.07 (5–13)
5.1 156 (20) [40]
6.0 3–0.64 (10–20) [45]
L-PE 44 2.05–0.8 (121–123) [44]
C-PE 9 1.18–0.36 (117–120) [44]
44 0.25–0.1 (113–115)
87 1.23–0.28 (112–115)
Case 2
L-POE 1.5 44.6 (22) [39]
C-POE 1.5 100 (22)
L-PCL-OH 2 18.8–2.8 (40–44) [41]
C-PCL 2 40.6–14.7 (45–49)
L-PCL-Acet 2 41.9–6.32 (39–45) [47]
4.9 30.2–6.87 (41–47)
C-PCL 2 25.1–4.99 (47–53) [47]
4.9 33.4–4.44 (48–54)
L-PCL-Acet 22 9.88–1.16 (41–50) [49]
C-PCL 22 4.94–0.51 (45–54) [49]

4.2 Spherulitic Growth Rate

Several works have reported the determination of spherulitic growth rates for cyclic
and linear polymers. Two opposite trends have been found:
Case 1: Slower spherulitic growth rate for cyclic polymers than for linear polymer
analogs. This behavior has been found for PTHF [40, 45] and PE [44].
Case 2: Faster spherulitic growth rate for cyclic polymers than for linear polymer
analogs. This behavior has been found for poly(oxyethylene) (POE) [39] and
PCL [41, 47–49].
The experimental data found in the literature are reported in Table 4 and divided
according to the two cases described above. To illustrate these cases, the results of
two groups that have found opposite trends are presented here along with the
interpretations provided by the authors.
Figure 8 clearly shows opposite trends in spherulitic growth rates between cases
1 and 2 (see Fig. 8a, b). The growth rate is determined by competition between the
Crystallization of Cyclic Polymers 107

(a) C-PE 44k L-PE 44k (b) C-PCL 2k L-PCL 2k

LH Fit LH Fit
2.5

40
2.0

30
G ( μm/min)

G ( μm/min)
1.5

20
1.0

0.5 10

0.0 0
105 110 115 120 125 130 40 44 48 52
Temperature (ºC) Temperature (ºC)
Fig. 8 Spherulitic growth rate (G) as a function of isothermal crystallization temperature (Tc) for
(a) C-PE and L-PE and (b) C-PCL and L-PCL. Lines represent fits to Lauritzen–Hoffman
equation. Data points extracted from [44] and [41], respectively. Reprinted with permission
from [41] and [44]

free energies associated with chain transport to the growth front and secondary
nucleation.
In case 1 (see Fig. 8a), the authors [44] argued that entropic factors cause both
transport and secondary nucleation free energies to be higher in cyclic polymers
(i.e., PTHF and PE) and, consequently, they exhibit lower G values than linear
polymers of similar Mn.
According to Tezuka et al. [40], the entropy difference between crystalline and
molten states is larger for cyclic polymer chains than for linear chains. This
assumption implies that both transport and secondary nucleation require higher
free energies in cyclic polymers for the following reasons:
(a) Cyclic polymer chains are conformationally restricted by the lack of chain
ends, especially if the four-folded structure (shown in Fig. 6b) is assumed to be
valid [40, 44, 45, 78]. The double-layer molecular arrangement of the cyclic
polymers requires more secondary nucleation energy than the monolayer
formation of the linear chain [78].
(b) Under the restricted conformation imposed by the lack of chain ends, the
conformational change required in the process of crystallization is hindered in
cyclic molecules. This is related to the transport energy involved in processes
such as adsorption and chain reel-in on the growth front. On the other hand, the
crystal growth rate direction of C-PTHF also changes according to a poisoning
or pinning effect [40, 44, 45, 78].
108 R.A. Pérez-Camargo et al.

In contrast, in case 2 (see Fig. 8b) both transport and secondary nucleation free
energies are claimed to be lower for cyclic polymers than for the linear analogs,
leading to higher G values in cyclic polymers than in linear polymers of similar Mn
values. The authors used molecular diffusion and entropic factors as arguments to
explain the higher G values of cyclic chains:
Diffusion or kinetic arguments: The faster diffusion of cyclic molecules has been
used as an argument to explain faster spherulitic growth displayed by some
cyclic polymers [39, 41]. This faster diffusion has been clearly established in
both experiments and simulations [32, 79–82]. Experimental findings indicate
that cyclic polymers (both entangled and disentangled) have lower melt viscos-
ities and smaller radii of gyration than linear polymers of equivalent chain
lengths [33, 83, 84].
Entropic or thermodynamic arguments: Su et al. [47] found a higher T0m value for
cyclic PCL than for L-PCL. This finding implies a larger supercooling at the
same crystallization temperature for C-PCL as compared with L-PCL (a fact
supported by the experimental evidence of single crystal truncation, as presented
in Sect. 3.1). Hence, different equilibrium melting points can also partly explain
why cyclic PCL molecules crystallize faster than linear molecules. A more
detailed explanation of this point is presented in Sect. 4.4.
The contradictory results presented above (case 1 versus case 2) are difficult to
explain, especially when some materials (belonging to the group of case 2) change
their trends depending on the properties that are being measured, as demonstrated
next for non-isothermal and isothermal overall crystallization.

4.3 Non-isothermal Differential Scanning Calorimetry

Non-isothermal differential scanning calorimetry (DSC) scans provide important

information on the overall crystallization of cyclic and linear polymers. However,
the thermal transitions are sometimes only partially reported. For example, the
crystallization temperatures (Tc) and degrees of crystallinity are not reported in
some references for POEs [38, 39], poly(lactic acid)s (PLAs) [46], PLAs with a
photocleavable linker [51], PTHFs [40, 45], PE [44], and PCL [18]. In these same
sudies, cyclic polymers exhibited lower melting temperatures (Tm) than their linear
analogs. According to Tezuka et al. [40], lower melting points are obtained as a
result of the higher entropy of fusion for cyclic chains than for linear chains.
On the other hand, for a large range of oligomeric and polymeric alkanes, PE
[31, 85], and PCL [41–43, 47–49], higher values for both crystallization and
melting points were found for cyclic molecules compared with their linear analogs.
Su et al. [47] found higher Tm values in C-PCL single crystals mats prepared by
isothermal crystallization from solution than in L-PCL analog samples. Similar
results were obtained from the melt state. Müller et al. [41, 47–49] studied cyclic
Crystallization of Cyclic Polymers 109

Fig. 9 Effect of molecular C-PCL low M n L-PCL low Mn

weight on melting C-PCL high Mn L-PCL high Mn
temperatures for low (Tm, 60
peak) and high (Tm, onset)
Mn. The solid lines were
traced arbitrarily to guide 58
the eye. Data extracted from
[48] (for Mn values lower
56
than 30 kg/mol) and [42]
(for Mn values higher than

Tm (°C)
30 kg/mol) and re-plotted. 54
Reprinted with permission
from [42] and [48]
52

50

48
0 5 10 15 20 40 60 80
Mn (kg/mol)

and linear PCLs of molecular weights ranging from oligomers to intermediate

molecular weight values (i.e., Mn 2–22 kg/mol).
Figure 9 shows a plot of Tm values obtained during standard DSC experiments as
a function of Mn for the PCL case. The peculiar trend found for the dependence of
the melting point on the molecular weight has its origin in competition between
diffusion and nucleation, which has also been reported for PE and explained in
detail [48].
Table 5 presents results from the literature for PCL samples [41, 47–49]. The
general behavior is that C-PCLs always exhibit higher crystallization and melting
points than L-PCLs. Shäler et al. [42] performed NMR, Hahn echo, and advanced
multiquantum measurements, demonstrating that C-PCLs have a higher segmental
mobility in the melt than their linear counterparts.
The enthalpy values determined by DSC and reported in Table 5 indicate that
there are no significant differences in crystallinity between cyclic and linear PCLs
(Table 5), within the experimental errors involved in measuring enthalpies by DSC
(i.e., of the order of 10–15% depending on integration routine, base line fluctuation,
and calibration). The NMR data obtained by Shäler et al. [42], also reported in
Table 5, show values of similar magnitudes to those obtained by DSC. However,
Shäler et al. [42] suggest that, as a result of the enhanced mobility of C-PCL in the
melt, higher crystallinity values compared with L-PCL might be obtained.
 110

Table 5 Non-isothermal differential scanning calorimetry results for L-PCL and C-PCL samples (data extracted from [41, 42, 47–49])
Cooling scans Second heating scans References
Tc Enthalpy change, ΔHc Crystallinity, Xc,c Enthalpy change, ΔHm Crystallinity, Xc,m
Samples ( C) (J/g) (%) Tm ( C) (J/g) (%)
T1m T2m
L-PCL-OH 2k 29.2 78 57 49.2 83 61 [41]
L-PCL-OH 3k 26.7 74 54 48.8 78 57 [48]
L-PCL-Acet 4.9k 34.6 81 60 55.1 94 69 [47]
L-PCL-OH 7.5k 30.4 69 51 54.3 79 58 [41]
L-PCL-OH 12k 24.1 69 51 52.5 77 57 [48]
L-PCL-Acet 22k 19.9 63 46 53.1 56.1 79 58 [48, 49]
L-PCL 35.9k 35.8 – – – – – [42]
L-PCL 50.6k 35a – – 53.7a – 48
L-PCL 77k 36.1a – – 54.1a – 48
C-PCL 2k 36.3 79 58 58.1 84 62 [41]
C-PCL 3k 34.0 68 50 55.2 70 51 [48]
C-PCL 4.9k 41.3 75 55 59.4 61.3 93 68 [47]
C-PCL 7.5k 41.5 61 45 59.4 63.4 69 51 [41]
C-PCL 12k 27.3 60 44 53.5 76 56 [48]
C-PCL 22k 26.9 57 42 57.3 73 54 [48, 49]
C-PCL 28.1k 36.5 – – – – – [42]
C-PCL 45k 36.4a – – 53.3a – 53
C-PCL 69k 36.8a – – 54.2a – 54
a
Tc and Tm are onset values taken from [42] whereas all others are peak values are taken from [41, 47–49]
R.A. Pérez-Camargo et al.
Crystallization of Cyclic Polymers 111

4.4 Isothermal Overall Crystallization Kinetics

The overall isothermal crystallization kinetics has only been investigated by two
groups of researchers in samples of cyclic and linear PCLs with different origins
and Mn ranges. The cyclic samples employed by Müller et al. [41, 47–49] were
synthesized by the group of Scott Grayson at Tulane University (New Orleans, LA),
employing ring-closure techniques derived from click chemistry, with Mn range of
2–22 kg/mol and low dispersity indexes. On the other hand, the samples of Shin
et al. [43] were prepared at Stanford University (Stanford, CA) by the group of
Robert Waymouth using ring-expansion techniques, yielding much higher Mn
values (i.e., 66–114 kg/mol) and larger dispersity indexes (see Tables 1 and 2).
In spite of the differences explained above between the PCL samples obtained
by the two synthetic approaches, the results obtained are in agreement with each
other. Both groups of authors obtained higher overall crystallization rates
(expressed as the inverse of the half-crystallization time, 1/τ50%) for C-PCLs than
for L-PCLs. To illustrate this, Fig. 10a shows the overall crystallization rate as a
function of isothermal crystallization temperature for samples with Mn values of
2 and 7.5 kg/mol, whereas Fig. 10b, c shows the overall crystallization rate at 45 C
as a function of Mn and weight-average molecular weight (Mw) respectively. C-PCL
samples crystallize faster than L-PCL at identical crystallization temperatures.
In Fig. 10a, the solid lines represent linear fits of the Lauritzen–Hoffman theory,
employing the equilibrium melting temperature (T0m ) values reported by Su et al. for
linear and cyclic PCLs of Mn ¼ 7.5 kg/mol [47] and assuming that these values are
independent of molecular weight for the range 2–22 kg/mol. The T0m value has a
direct impact on the supercooling, defined as ΔT ¼ T 0m  T c . In order to establish
the influence of supercooling on the overall crystallization kinetics, Pérez et al. [48]
represented the overall crystallization rate (1/τ50%) as a function of ΔT, employing
the previously determined values of T0m . Interestingly, the representation of 1/τ50%
as a function of ΔT (for details see [48]) brings the curves of C-PCL and L-PCL
closer together. However, the fact that the curves cannot be completely
superimposed indicates that other factors also contribute to the differences in
overall kinetics. In fact, Su et al. [47] also found differences in fold surface free
energies for C-PCL and L-PCL that preclude perfect superposition of the 1/τ50%
versus ΔT curves. The bell-shaped curve presented in Fig. 10b has been discussed
in detail previously and has its origin in the existence of an optimum molecular
weight value for overall isothermal crystallization [48].
The differences in crystallization rates between C-PCLs and L-PCLs have not
only been detected in low and intermediate Mn samples but also in high molecular
weight PCLs. Shin et al. [43] studied the overall crystallization of high molecular
weight cyclic and linear PCLs. These authors obtained two different behaviors,
depending on the molecular weight. For samples with Mw of 75–110 kg/mol, both
linear and cyclic PCLs crystallize at similar overall rates.
However, when Mw values are larger than 110 kg/mol, C-PCLs crystallize faster
than L-PCLs. The authors explained this peculiar behavior by arguing that C-PCLs
112 R.A. Pérez-Camargo et al.

(a) C-PCL 2k L-PCL 2k (b) C-PCL Exp. C-PCL Ext.

C-PCL 7.5k L-PCL 7.5k
L-PCL Exp. L-PCL Ext.
LH fits
1.0
Tc=45°C

0.8

1/ τ50% (min )
1

-1
1/ τ50% (min )
-1

0.6

0.4
0.1

0.2

0.0 0.01
35 40 45 50 55 60 0 4 8 12 16 20 24
Tc (ºC) Mn (kg/mol)

C-PCL L-PCL
(c) 0.09
Tc= 45ºC
1/ τ50% (min )
-1

0.06

0.03

75 90 105 120 135 150

Mw (kg/mol)

Fig. 10 Inverse half-crystallization time (1/τ50%) as a function of (a) crystallization temperature

(Tc) for the indicated PCL samples, (b) Mn for ring-closure samples, solid lines are included to
guide the eye, and (c) Mw for ring-expansion polymerization samples, solid lines are included to
guide the eye. Lines in a represent fits to Lauritzen–Hoffman equation. Reprinted with permission
from [43] and [48]

contain small amounts of L-PCLs as impurities or contaminants. In the lower Mw

range, these impurities might have a dominating influence over the cyclic topology.
Thus, cyclic and linear PCL samples crystallize at similar rates, because the linear
chains can increase the melt viscosity of cyclic polymers and significantly affect
their diffusion to the crystallization front [31, 33, 86–88].
Crystallization of Cyclic Polymers 113

In contrast, for Mw > 110 kg/mol, a dominating influence of cyclic topology on

the crystallization kinetics, irrespective of possible contributions from trace
amounts of linear polymers, was postulated. Hence, cyclic PCLs crystallize faster
than their linear analogs. The reason for the dominating influence of the cyclic
topology, according to Shin et al. [43], is that the melt viscosities of the cyclic PCLs
are lower than those of linear PCLs, as observed for other cyclic polymers [32, 33,
87]. Additionally, the authors claim that disentanglement in the melt is a crucial
feature of polymer crystallization. Therefore, the differences in the entanglement
density between linear and cyclic PCLs could be responsible for faster crystalliza-
tion of the high molecular weight cyclic samples. Evidence for differences in
entanglement density have also been obtained by self-nucleation, as described in
Sect. 5. The results reported by Shin et al. are shown in Fig. 10c.
Shin et al. [43] performed time-dependent SAXS experiments to study overall
isothermal crystallization kinetics of their PCL samples (i.e., samples with Mw
values of 100 kg/mol L-PCL and 130 kg/mol C-PCL). An increase in the scattering
density described by the invariant Q [89] for cyclic PCLs was almost complete
within 10 min, whereas for linear PCL it gradually increased over the course of
120–150 min. The Porod coefficient (P) [90] was determined as a function of time
and used as a means to monitor crystallization kinetics. The Porod coefficient is
proportional to the interface area per unit volume (Oac):

1
P¼ Oac ðΔηÞ2 ; ð3Þ
8π3

where Δη denotes the difference between the electron densities of the lamellar and
amorphous phases. Figure 11 shows that the high molecular weight cyclic PCL
crystallizes faster than its linear counterpart at a crystallization temperature (Tc) of
45 C. The same trend was found for other crystallization temperatures (30 C, 35 C
and 45 C) (for more details see [43]).
Recently, Wang et al. [50] compared the overall crystallization rates of linear
and cyclic oligomeric PCLs with Mn ¼ 2 kg/mol via fast-scanning chip calorimetry
measurements (for further details of the specific techniques employed, see [50]).
Figure 12 shows an inverse bell-shaped curve for the overall crystallization time,
which can be explained by the classical theory of nucleation and growth. The higher
crystallization half-times at temperatures near the glass transition can be assigned to
the limited diffusion of molecules, whereas those near to the melting point can be
assigned to the limited thermodynamic driving force for nucleation.
Figure 12 shows that at Tc higher than 20 C, C-PCL crystallizes faster than
L-PCL, whereas below 20 C the differences in crystallization half-times are not
significant. Qualitatively, both standard and fast-scanning chip calorimetry yield
similar results in the high temperature range of Fig. 12. On the other hand, the
quantitative differences between the standard DSC data and fast-scanning chip
calorimetry are a result of the different equipment involved in the measurements
and also in the ways in which the isothermal crystallization experiments were
performed [50].
114 R.A. Pérez-Camargo et al.

Fig. 11 Crystal growth C-PCL 130k

rates of cyclic (C-PCL L-PCL 100k
130k) and linear (L-PCL 2.0
100k) PCL during Tc=45ºC
isothermal crystallization at
45 C. Reprinted with
permission from [43] 1.5

P (nm )
-7
1.0

0.5

0.0
0 60 120 180 240 300 360
Time (min)

Fig. 12 Crystallization 10
4

half-time as a function of C-PCL 2k

Crystallization half-time (s)

isothermal crystallization 3
L-PCL 2k
10 C-PCL 2k Con.DSC
temperature for both cyclic
and linear PCLs with L-PCL 2k Con. DSC
2
Mn ¼ 2 kg/mol. The 10
crystallization half-times
1
obtained previously [48] by 10
conventional DSC were
added for comparison 10
0

purposes. Reprinted with

permission from [50] -1
10

-2
10
-60 -40 -20 0 20 40 60
Tc (°C)

Figure 13 shows the apparent nucleation half-time derived from the data shown
in Fig. 12 (see [50] for details on how to derive nucleation half-times from fast-
scanning chip calorimetry experiments). The curve of the nucleation half-time as a
function of Tc exhibits two minima: one related to apparent homogeneous nucle-
ation (low-temperature peak) and the other related to heterogeneous nucleation
(high-temperature peak).
At higher Tc values, C-PCL nucleates faster than L-PCL, corroborating the
results obtained by PLOM presented above. In contrast, at very low Tc values,
L-PCL exhibits higher onset temperature (around 10 C) than C-PCL (around
Crystallization of Cyclic Polymers 115

Fig. 13 Nucleation half- 10

2

time as a function of C-PCL 2k

isothermal crystallization

Nucleation half-time (s)

L-PCL 2k
temperature for both cyclic 1
and linear PCLs with 10
Mn ¼ 2 kg/mol. Reprinted
with permission from [50]
0
10

-1
10

-2
10
-60 -40 -20 0 20 40 60
Tc (°C)

25 C) for the occurrence of apparent homogeneous nucleation. This result was
attributed to the free chain ends in L-PCL, which enhance chain mobility and favor
diffusion at lower temperatures [50].
According to the results obtained by different techniques, the overall crystalli-
zation rates of cyclic PCLs are higher than linear PCLs. This difference can be
explained by (a) kinetic factors dominated by diffusion, and (b) thermodynamic
factors given by the differences in supercooling, as a result of the different equi-
librium melting values for linear and cyclic PCLs.

4.4.1 Difference in Equilibrium Melting Temperature Between Cyclic

and Linear Polymers

The equilibrium melting temperature (T0m ) is defined as the transition temperature

between the isotropic melt and the 100% crystalline phase, which is given by the
ratio of the equilibrium enthalpy of fusion (ΔH0) to the equilibrium entropy of
fusion (ΔS0):

ΔH 0
T 0m ¼ : ð4Þ
ΔS0

There is agreement in the literature that both linear and cyclic polymers have a
similar enthalpy of fusion. However, the entropy of fusion is a controversial point,
because two interpretations can be found in the literature:
(a) According to the interpretation of Tezuka et al. [40], the arrangement of the
cyclic polymer chains within a lamella, including fold surfaces, is conforma-
tionally restricted (see Fig. 6). Hence, they consider that the entropy difference
between crystalline and molten states should be larger for cyclic polymers,
116 R.A. Pérez-Camargo et al.

Table 6 Equilibrium melting Sample Mn (kg/mol) T0m ( C) References

temperature (T0m ) values
L-PTHF 9.1 49.9 [45]
reported in the literature for
different polymers C-PTHF 6.0 32.9
L-PE 44 145.2 [44]
C-PE 44 139.4
L-PCL 168 82 [43]
C-PCL 79 84.2
L-PCL-OH 7.5 80 [41]
C-PCL 7.5 81
L-PCL-Acet 7.5 80 [47]
C-PCL 7.5 91.2

even if the conformation in the melt is restricted for rings. If the equilibrium
entropy difference between the melt and the crystalline states is larger for
cyclic molecules, then T0m would be lower for cyclic polymers in comparison
with their linear analogs.
(b) According to Müller et al. [41, 47–49], because cyclic polymers have a lower
configurational entropy in the melt (assuming that the entropy of cyclic and
linear polymers within the crystalline state is the same) [37], a higher T0m is
predicted. In fact, it is considered that cyclic polymers have a “collapsed” coil
conformation in the melt [70–73, 91, 92]. Hence, it is reasonable to expect that
in the melt state cyclic polymers have lower configurational entropies than
those of linear chains.
The T0m values estimated by the Hoffman–Weeks plot or the Gibbs–Thompson
equation are compiled in Table 6. As can be seen, all possible trends have been
reported: similar T0m values for cyclic and linear polymers, higher values for cyclic
than for linear chains, and lower values for cyclic polymers than for linear analogs.
In order to modify the Hoffman–Weeks and Thomson–Gibbs equations, Su
et al. [47] introduced a “cyclization free energy”, as consequence of the difference
in free energy in the melt state between cyclic and linear analogs with the same
molecular weight. Assuming that the enthalpies of cyclic and linear polymers are
the same, the entropy of cyclic chains is smaller than that of linear chains in view of
the absence of chain ends and more collapsed conformation.
The modified Thomson–Gibbs and Hoffman–Weeks equations for cyclic poly-
mers explicitly contain an excess free-energy term contributed by the entropic loss
upon cyclization. For more details on the derivation of these equations, see
[47]. The modified Thomson–Gibbs equation can be expressed as follows:
 
T 0mL 2σ e
Tm ¼   1 0 ; ð5Þ
T 0 Δs
1 þ mL 0 cyc Δh f l
Δh f

where Tm is the melting point of the lamellar crystal formed by the cyclic polymer,
T0mL is the equilibrium melting point of the linear polymer, ΔScyc is the cyclization
Crystallization of Cyclic Polymers 117

entropy difference between cyclic and linear polymers in the melt state, Δh0f is the
equilibrium melting enthalpy difference (being the same for both cyclic and linear
polymers), σ e is the fold surface free energy, and l is the lamellar thickness. In a plot
of Tm versus l1, the intercept can be used to determine the equilibrium melting
point of the cyclic polymer (T0mC ), defined as:

T 0mL
T 0mC ¼  T 0 Δs
: ð6Þ
1 þ mLΔh0 cyc
f

On the other hand, the modified Hoffman–Weeks equation for cyclic polymers
can be expressed as:

T 0mL    
Tm ¼   1 Tc 1 Tc
T Δscyc
0 1  þ ¼ T 0mC 1  þ : ð7Þ
1 þ mL 0 γ γ γ γ
Δh f

Su et al. [47] performed isothermal crystallization experiments from the melt

using DSC and SAXS and applied the above equations. In order to estimate T0m for
7.5 kg/mol cyclic and linear PCLs, a Tm was plotted versus l1 and the modified
Thomson–Gibbs equation employed. As a result (see details in [47]), the value of
T0mC (i.e., 91.2 C) was found to be higher than that of T0mL (i.e., 80 C). This result is
in agreement with morphological observations of single crystals, which indicated
that C-PCLs are more supercooled than L-PCLs at identical crystallization
temperatures.
In Fig. 14, all values reported for the equilibrium melting points of both linear
and cyclic PCLs are plotted as a function of Mn of the samples. The horizontal lines
indicate the values obtained by Su et al. for 7.5 kg/mol samples. In the case of the
L-PCL sample with 7.5 kg/mol, the T0mL value of 80 C is in agreement with other
values reported in the literature, including those of Shin et al. [43], which corre-
spond to much higher Mn values. Many of the values reported in Fig. 14 were
obtained by the Hoffman–Weeks extrapolation, which can be unreliable for PCL,
depending on the range of Tc temperatures employed, as demonstrated by Su
et al. [47]. This is one of the reasons for the wide spread of T0mL values.
In the case of cyclic PCLs, only a few values have been determined. Su et al. [47]
estimated a ΔScyc value of 12 J kg1 K1. Additionally, the values for the surface
free energy, σ e, of linear and cyclic PCLs determined from the slopes of the plots
were 30.3 and 42.8 mJ m2, respectively. The larger σ e observed for cyclic PCL
was attributed to the larger loops formed at the crystal surface when the cyclic
chains fold. Another reason for larger values of σ e in the case of the C-PCLs
employed by Su et al. is the presence of the triazole rings (remaining from the
click ring-closure reactions) in the amorphous fold planes. In view of the different
σ e values for linear and cyclic PCLs, it is not expected that 1/τ50% versus ΔT curves
118 R.A. Pérez-Camargo et al.

Fig. 14 Melting 110

Phillips et al.
equilibrium temperatures
Goulet et al.
reported in the literature. 100 C-PCL 7.5 k Gou y Groeninckx
Horizontal lines indicate Kuo et al.
values obtained by Su 90 Müller et al.
et al. [47] for 7.5 kg/mol Nie et al.
samples. Data points taken 80 Liu et al.

Tm (°C)
from values reported in the Lorenzo et al.
following references [41, 70 L-PCL 7.5k Liu et al.2

0
Heck et al.
43, 47, 93–106]
Wang et al.
60 Chen et al.
Chang et al.
50 Shin et al.L
Su et al. C
40 Su et al. L
7.5 kg/mol Shin et al.C
30
0 30 60 90 120 150 180
Mn (kg/mol)

obtained by Pérez et al. [48] can be completely superimposed (see [48] for more
details).
Recently, Sugai et al. [51] studied PLLA and poly(D-lactide) (PDLA) with a
photocleavable linker and employed an unmodified Thomson–Gibbs equation to
determine the equilibrium melting points of their samples. They obtained a smaller
σ e for cyclic PLLA than for its linear analog. According to the authors, this result
suggested that cyclic PLLA forms a well-ordered looped structure at the crystal
surface that helps stabilize the crystal–amorphous interface, whereas linear PLLA
has chains ends that are probably not as organized as cyclic PLLA. Similar results
were obtained with PDLA.

5 Self-Nucleation and Successive Self-Nucleation

and Annealing

5.1 Self-Nucleation

The self-nucleation (SN) technique was originally devised by Keller et al. [107] and
its use was extended by Fillon et al. [108] by employing DSC [109]. A more
detailed explanation of this technique can be found elsewhere [110]. SN consists
in performing controlled heating of a sample with a standard thermal history up to a
self-seeding temperature called Ts. If Ts is high enough to erase thermal history, the
sample is in domain I (melting domain). If Ts is high enough to melt the polymer
almost completely (or completely) but low enough to leave self-nuclei (crystal
fragments or residual segmental orientation in the melt) capable of nucleating the
Crystallization of Cyclic Polymers 119

(a) L-PCL-OH 7.5k (b) C-PCL 7.5k

D III D II DI D III D II DI

Heat flow endo up (mW)

Heat flow endo up (mW)
2.5 mW

2.5 mW
36 42 48 54 60 66 72 36 42 48 54 60 66 72
Temperature (°C) Temperature (°C)

Fig. 15 Standard DSC heating scans in the temperature range 36–76 C, where the melting of (a)
linear and (b) cyclic PCL samples of 7.5 kg/mol can be observed. The vertical lines indicate the
temperatures at which the materials experience self-nucleation domain transition. Domain II (D II)
is shaded with two shades of grey. The temperature range at which the material experiences
crystalline memory is shaded dark grey. Reprinted with permission from [48]

polymer during cooling, the polymer is in domain II (exclusive self-nucleation

domain). Finally, if Ts is so low that the polymer only partially melts, the sample is
in domain III (self-nucleation and annealing domain), because the unmelted mate-
rial can anneal.
Use of SN for cyclic and linear polymers revealed novel information regarding
crystalline memory, in which a wider domain II was found in L-PCL than in C-PCL
(see Fig. 15). The ability to induce self-nucleation when the sample is completely
molten is directly related to the crystalline memory of the material [111].
The crystalline memory effect has been explained by considering that the
melting process is insufficient to drive the chains back to an isotropic random
coil conformation. The chains could have difficulties returning to an isotropic
random coil for the following reasons: (a) residual orientation of chain segments
or the persistence of metastable or precursor phases [108, 112–129], (b) aligned
chains in quasi-crystallographic registers [130–136], or (c) entanglement density
[137–145].
Figure 15a shows the capacity of L-PCL to exhibit crystalline memory. This
feature has been reported before in commercial L-PCLs by Lorenzo et al.,
employing similar SN protocols [111]. In contrast, Fig. 15b shows that the crystal-
line memory of C-PCLs is significantly smaller than that exhibited by its linear
analog. A measure of the crystalline memory is given by the width of the self-
nucleation domain (i.e., domain II) and, in particular, how far it extends beyond the
end melting point of the polymer (after the melting endotherm attains the liquid
120 R.A. Pérez-Camargo et al.

baseline). Domain II is shadowed in Fig. 15 with two shades of gray; the darker grey
shading corresponds to the crystalline memory range.
Pérez et al. [48] found that the crystalline memory of C-PCL can be erased at
lower temperatures than for L-PCLs. According to Pérez et al. [48], C-PCL chains
(partially disentangled) reach a pseudo-equilibrium random coil conformation
faster than L-PCL as a result of the lower entanglement density of C-PCL. This
experimental evidence for the difference in entanglement density is in agreement
with the argument employed by Shin et al. [43] to explain the faster overall
crystallization of cyclic polymers.

5.2 Successive Self-Nucleation and Annealing

Successive self-nucleation and annealing (SSA) is a thermal fractionation protocol

designed to deconvolute DSC melting endotherms into elementary components or
thermal fractions. This technique was developed by Müller et al. [109,
153]. Recently a review of its correct use and applications has been published [110].
SSA has been employed to study cyclic and linear PCLs by Müller et al. [48,
110], and a novel effect was found. Compared with L-PCLs, a larger annealing
capacity of C-PCL lamellar crystals was found. This is remarkable because L-PCL
chains can, in theory, be extended to twice the maximum length of C-PCL chains of
identical chain length. Hence, L-PCLs have the potential to produce thicker crystals
than C-PLCs near equilibrium conditions.
Figure 16 shows the final DSC heating scan after SSA. All samples exhibit some
thermal fractionation and, therefore, a distribution of lamellar sizes that melt at
distinct temperatures. The melting peaks have been labeled according to their origin
(see [48]). Figure 16 also shows vertical lines that indicate the values of the Ts
temperatures employed for the fractionation, and a dashed vertical line that corre-
sponds to the Ts, ideal for the 7.5 kg/mol C-PCL. This temperature represents the first
Ts employed for all samples.
To illustrate the fractionation exhibited for all the samples, consider melting
peak 1 in Fig. 16. Melting peak 1 refers to an annealed population produced mainly
during the 5 min holding time at Ts,2, although successive steps might also have
some limited influence on the size of the fraction. Melting peak 2 was produced by
Ts,3 and so on.
If we compare samples with the same molecular weight, it is interesting to note
the annealing capacity of the cyclic samples. For instance, the 7.5 kg/mol C-PCL
sample had the highest melting fraction (melting peak 1). In contrast, this fraction
was not present in other samples, because their melting range was too low.
Therefore, Ts,2 was only able to produce a thermal fraction (peak 1) for 7.5 kg/
mol C-PCL, and for all other samples it caused only self-nucleation but no
annealing. Similar behavior was found for C-PCL of 2 kg/mol, in which fraction
2 (melting peak 2) was present, as in almost all samples except for L-PCL of 2 kg/
mol. The 2 kg/mol L-PCL was the sample with the lowest melting range.
Crystallization of Cyclic Polymers 121

Fig. 16 Final heating run

l
ea
id
6
after thermal fractionation

2
s,
T

s,
T

s,
s,

s,

s,
T

T
by successive self-
nucleation and annealing. 2 1
C-PCL 7.5 k
Peaks 1–5 indicate the
melting temperatures.
3

Heat flow endo up (mW)

Vertical lines indicate the 5 4
temperatures used for 2
fractionation (Ts). Dashed
vertical line indicates the L-PCL 7.5k
3
ideal Ts for 7.5 kg/mol
C-PCL and was the first 5 4
fractionation temperature
2
used. See text for full
details. Reprinted with
permission from [48]

C-PCL 2 k
3
5 4

3
L-PCL 2k
4

5
10 mW

35 40 45 50 55 60 65 70
Temperature (°C)

The influence of the molecular weight is reflected in the results. As Mn increases,

the annealing capacity under specific SSA protocol employed also increases, and
thermal fractions with higher melting points are produced, as expected.
Figure 16 shows a higher annealing capacity of the cyclic samples, even though
the linear samples could be extended, in theory, to twice the maximum length of a
C-PCL chain of identical chain length, as represented in the model shown in
Fig. 17.
Under equilibrium conditions, linear chains have the potential to produce thicker
crystals. However, under the SSA parameters employed by Müller et al. [48, 110]
(which are dominated by kinetic factors), cyclic chains anneal and produce thicker
crystals than their linear analogs. According to Müller et al. [48, 110], the reasons
for the higher annealing capacity of cyclic chains are their lower entanglement
density and their ring topology. These characteristics facilitate annealing for cyclic
chains, compared with linear chains with a higher entanglement density and lower
diffusion coefficients. A collection of stacked lamellar crystals normally thicken at
the expense of chains located in the interlamellar regions. Such reel-in processes are
122 R.A. Pérez-Camargo et al.

(a) (b)

l l*

2L l l* L

Linear Chain
Cyclic Chain
(1)
(2) (1)
Initial (2)
Annealing (3) Initial (3)
Annealing
Equilibrium Equilibrium

Fig. 17 A possible way for (a) linear and (b) cyclic chains to thicken during annealing. (1) L-PCL
and C-PCL with initial lamellar thickness l. The square (not to scale) represents the chemical
structure of the triazole group after cyclization, which must be always in the amorphous regions
because it cannot enter the crystallographic register of PCL. (2) L-PCL and C-PCL during
annealing, in the process of thickening to a larger value, l*. (3) L-PCL with an extended chain
length of 2L and C-PCL with a once-folded configuration and extended chain length, L. Reprinted
with permission from [48]

facilitated by the lower entanglement densities and higher diffusion coefficients of

cyclic chains.

6 Reasons for Different Behavior Reported

in the Literature

The crystallization of several different cyclic polymers has been studied and in
some cases contradictory trends have been found. Experimental evidence indicates
that small amounts of linear segments in cyclic polymers have a dramatic impact on
their overall crystallization rates. Consequently, sample purity is a crucial issue to
be taken into account. On the other hand, it is possible that the conflicting trends
found in the literature could be related to the nature of the polymer.
Kapnistos et al. [33] found that the presence of only 0.07 wt% of linear chains
can dramatically alter the rheology, diffusion, chain dynamics, and relaxation times
of ring polymers. These dramatic changes are related to the bridging or threading
effect caused in ring molecules by small amounts of linear chains, which can form a
transient network that percolates throughout the material. According to Kapnistos
et al. [33], the form of the stress relaxation modulus at intermediate times is affected
by this network, producing an entanglement plateau instead of power law behavior.
Crystallization of Cyclic Polymers 123

(a) C-PCL 3k L-PCL 3k

(b) Tc=40°C Exp. Tc=40°C Ext.
C/L 95/5 3k C/L 90/10 3k Tc=44°C Exp. Tc=44°C Ext.
C/L 80/20 3k L-H fits Tc=48°C Exp. Tc=48°C Ext.
1.5

2
1/τ50% (min )

1.0

1/τ50% (min )
-1

-1
1

0.5

0.0
36 40 44 48 52 0 20 40 60 80 100
Tc (°C) wt% C-PCL
Fig. 18 (a) Overall crystallization rate (1/τ50%) as a function of isothermal crystallization
temperature (Tc) for 3 kg/mol samples of L-PCL and C-PCL and for their blends at the indicated
compositions. Solid lines represent fits to the Lauritzen–Hoffman (L-H) equation. (b) Values of
1/τ50% as a function of C-PCL content at different constant values of Tc. The plot contains
experimental data points (Exp.) and values extrapolated (Ext.) from the L-H fits to the data
shown in a. Reprinted with permission from [49]

Moreover, in blends of linear and cyclic molecules, the trend of cyclic molecules to
be threaded by linear chains has been shown by computer simulations [146–151].
The threading effects constitute strong topological interactions capable of
slowing down relaxation and chain mobility, because they represent long-lasting
entanglements. Therefore, cyclic molecules relax and diffuse more slowly than
anticipated by composition effects, as a result of the presence of small amounts of
linear chains [146–151].
The threading effects have been shown not only in computer simulations but also
experimentally. Gooβen et al. [152] performed small angle neutron scattering
(SANS) measurements for mixtures of linear and cyclic polystyrene samples, and
their results confirmed the threading of rings by linear chains. Additionally, pre-
liminary results on linear and cyclic PCL blends have been obtained by Müller
et al. [49] and are presented in Fig. 18.
Figure 18a shows the overall crystallization rate as a function of Tc for cyclic and
linear PCL samples (with Mn values of 3 kg/mol). The neat PCL samples display the
usual trend (see Fig. 10), in which C-PCL crystallizes faster than L-PCL at identical
Tc. The overall crystallization rates exhibited by the blends are between those of
neat C-PCL and L-PCL, as expected. However, a significant reduction in the overall
crystallization rate of neat C-PCL is achieved when small amounts of L-PCL (5–
10 wt%) are added. Interestingly, the addition of larger amounts (20 wt%) leads to a
lower reduction in crystallization rate.
124 R.A. Pérez-Camargo et al.

Figure 18b shows the overall crystallization rates at constant Tc values as a

function of composition. The decrease in overall crystallization rate was more
pronounced than predicted by a simple mixing law (indicated by the straight lines
in Fig. 18b) [49] when small amounts of L-PCL chains were added to C-PCL.
Figure 18b shows that the deviation of overall crystallization rate values from a
simple rule of mixing is lower as Tc increases, because the influence of diffusion is
smaller. The results shown in Fig. 18 are attributed to the threading effect of linear
chains on C-PCL [49]. It is worth noting that more experimental evidence for the
threading effect was obtained by Müller et al. [49] during non-isothermal DSC tests
and spherulitic growth rates for 3 and 12 kg/mol samples of C-PCLs and their linear
analogs (for more details see [49])
The contradictory trends observed, particularly for PCL and PTHF samples,
could be a result of the different nature of each polymer. According to Sugai
et al. [51], different crystal systems could cause opposite effects. However, more
evidence is needed in this regard.

7 Conclusions

The interesting effect of chain topology (ring versus linear) on polymer crystalli-
zation has been reviewed. It is well established that chain topology can have an
important influence on nucleation, morphology, crystallization kinetics and thermal
transitions in general. It is envisaged that such important changes can also modify
the mechanical properties, an assumption that needs to be verified in the future.
Other properties also depend on chain topology, such as glass transition tempera-
ture, visco-elastic and rheological properties, and susceptibility to degradation.
The importance of sample purity cannot be understated, especially in view of
recent rheological evidence and the results available on cyclic/linear blends. The
purity issue could be relevant when comparing results for polymers produced by
different synthetic methods or purification procedures, and it is very difficult to
assess its effect because the content of linear chains could vary from sample to
sample. Even samples produced by the same technique might have different
amounts of linear chains, depending on the molecular weight, for instance.
Although some polymeric systems have been studied extensively (namely PCL
and PTHF), the trends observed are sometimes contradictory. If we assume purity
issues are not a problem in these cases, from the evidence available in the literature
today, it seems that the effect of topology depends on the exact nature of the
polymeric structure under consideration (crystal structure and the way the chains
fold during lamellae formation). More research efforts are clearly needed to under-
stand the origin of such results.

Acknowledgements The POLYMAT/UPV/EHU team would like to acknowledge funding from

the following projects: “UPV/EHU Infrastructure: INF 14/38”; “Mineco/FEDER: SINF
130I001726XV1/Ref: UNPV13-4E-1726” and “Mineco MAT2014-53437-C2-P”. R.A.P.-C
Crystallization of Cyclic Polymers 125

gratefully acknowledges the award of a PhD fellowship by POLYMAT Basque Center for
Macromolecular Design and Engineering.

References

1. Jacob F, Wollman EL (1958) Genetic and physical determinations of chromosomal segments

in Escherichia coli. Symp Soc Exp Biol 12:75–92
2. Dulbecco R, Vogt M (1963) Evidence for a ring structure of polynoma virus DNA. Proc Natl
Acad Sci USA 50(2):236–243
3. Vologodskii A (1998) Exploiting circular DNA. Proc Natl Acad Sci USA 95(8):4092–4093
4. Harmon FG, Brockman JP, Kowalczykowski SC (2003) RecQ helicase stimulates both DNA
catenation and changes in DNA topology by topoisomerase III. J Biol Chem
278:42668–42678
5. Harmon FG, DiGate RJ, Kowalczykowski SC (1999) RecQ helicase and topoisomerase III
comprise a novel DNA strand passage function: a conserved mechanism for control DNA
recombination. Mol Cell 3(5):611–620
6. Saska I, Colgrave ML, Jones A, Anderson MA, Craik DJ (2008) Quantitative analysis of
backbone-cyclised peptides in plants. J Chromatogr B Biomed Sci Appl 872(1–2):107–114
7. Craik DJ (2009) Circling the enemy: cyclic proteins in plant defence. Trends Plant Sci 14
(6):328–335
8. Beckham HW (2011) Ring polymers: effective isolation and unique properties. In:
Hadjichristidis N, Hirao A, Tezuka Y, Du Prez F (eds) Complex macromolecular architec-
ture: synthesis and characterization, and self-assembly. Wiley, Hoboken
9. Semlyen J (2002) Cyclic polymers, 2nd edn. Kluwer, Dordrecht
10. Hoskins JN, Grayson SM (2011) Cyclic polyesters: synthetic approaches and potential
applications. Polym Chem 2:289–299
11. Schappacher M, Deffieux A (1995) Controlled synthesis of bicyclic “eight-shaped” poly
(chloroethlyl vinyl ether)s. Macromolecules 28(8):2629–2636
12. Beinat S, Schappacher M, Deffieux A (1996) Linear and semicyclic amphiphilic diblock
copolymers. 1. Synthesis and structural characterization of cyclic diblock copolymers of poly
(hydroxyethyl vinyl ether) and linear polystyrene and their linear homologues. Macromole-
cules 29(21):6737–6743
13. Kubo M, Hayashi T, Kobayashi H, Itoh T (1998) Syntheses of tadpole- and eight-shaped
polystyrenes using cyclic polystyrene as a building block. Macromolecules 31(4):1053–1057
14. Oike H, Washizuka M, Tezuka Y (2001) Designing an “a ring-with-branches” polymer
topology by electrostatic self-assembly and covalent fixation with interiorly functionalized
telechelics having cyclic ammonium groups. Macromol Rapid Commun 22(14):1128–1134
15. Oike H, Hamada M, Eguchi S, Danda Y, Tezuka Y (2001) Novel synthesis of single- and
double-cyclic polystyrenes by electrostatic self-assembly and covalent fixation with
telechelics having cyclic ammonium salt groups. Macromolecules 34(9):2776–2782
16. Tezuka Y, Komiya R, Washisuka M (2003) Designing 8-shaped polymer topology by
metathesis condensation with cyclic poly(THF) precursors having allyl groups. Macromol-
ecules 36(1):12–17
17. Jia Z, Fu Q, Huang J (2006) Synthesis of amphiphilic macrocyclic graft copolymer consisting
of poly(ethylene oxide) ring and multi-polystyrene lateral chains. Macromolecules 39
(16):5190–5193
18. Li H, Jér€ome R, Lecomte P (2006) Synthesis of tadpole-shaped copolyesters based on living
macrocyclic poly(ε-caprolactone). Polymer 47(26):8406–8413
126 R.A. Pérez-Camargo et al.

19. Li H, Riva R, Jér€

ome R, Lecomte P (2007) Combination of ring-opening polymerization and
“click” chemistry for the synthesis of an amphiphilic tadpole-shaped poly(ε-Caprolactone)
grafted by PEO. Macromolecules 40(4):824–831
20. Li H, Jér€ome R, Lecomte P (2008) Amphiphilic sun-shaped polymers by grafting macrocy-
clic copolyesters with PEO. Macromolecules 41(3):650–654
21. Pang X, Jing R, Huang J (2008) Synthesis of amphiphilic macrocyclic graft copolymer
consisting of a poly(ethylene oxide) ring and multi-poly(ε-caprolactone) lateral chains.
Polymer 49(4):893–900
22. Shi G-Y, Pan C-Y (2008) Synthesis of well-defined figure-of-eight-shaped polymers by a
combination of ATRP and click chemistry. Macromol Rapid Commun 29(20):1672–1678
23. Shi GY, Yang LP, Pan CY (2008) Synthesis and characterization of well-defined polystyrene
and poly(ε-caprolactone) hetero eight-shaped copolymers. J Polym Sci Part A Polym Chem
46(19):6496–6508
24. Dong Y-Q, Tong Y-Y, Dong B-T, Du F-S, Li Z-C (2009) Preparation of tadpole-shaped
amphiphilic cyclic PS-b-linear PEO via ATRP and click chemistry. Macromolecules 42
(8):2940–2948
25. Londsdale DE, Monteiro MJ (2010) Various polystyrene topologies built from tailored cyclic
polystyrene via CuAAC reactions. Chem Commun 46:7945–7947
26. Fan X, Wang G, Huang J (2011) Synthesis of macrocyclic molecular brushes with amphi-
philic block copolymers as side chains. J Polym Sci Part A Polym Chem 49(6):1361–1367
27. Laurent BA, Grayson SM (2011) Synthesis of cyclic dendronized polymers via divergent
“graft-from” and convergent click “graft-to” routes: preparation of modular toroidal macro-
molecules. J Am Chem Soc 133(34):13421–13429
28. Wang G, Fan X, Hu B, Zhang Y, Huang J (2011) Synthesis of eight-shaped poly(ethylene
oxide) by the combination of glaser coupling with ring-opening polymerization. Macromol
Rapid Commun 32(20):1658–1663
29. Fan X, Huang B, Wang GW, Huang J (2012) Synthesis of amphiphilic heteroeight-shaped
polymer cyclic-[poly(ethylene oxide)-b-polystyrene]2 via “click” chemistry. Macromole-
cules 45(9):3779–3786
30. Kricheldorf HR (2010) Cyclic polymers: synthetic strategies and physical properties. J Polym
Sci Part A Polym Chem 48(2):251–284
31. Bielawski CW, Benitez D, Grubbs RH (2002) An “endless” route to cyclic polymers. Science
297(5589):2041–2044
32. Mckenna GB, Hostetter BJ, Hadjichristidis N, Fetters LJ, Plazek DJ (1989) A study of the
linear viscoelastic properties of cyclic polystyrenes using creep and recovery measurements.
Macromolecules 22(4):1834–1852
33. Kapnistos M, Lang M, Vlassopoulos D, Pyckhout-Hintzen W, Richter D, Cho D, Chang T,
Rubinstein M (2008) Unexpected power-law stress relaxation of entanglement ring polymers.
Nat Mater 7:997–1002
34. Nam S, Leisen J, Breedveld V, Beckham HW (2008) Dynamics of unentangled cyclic and
linear poly(oxyethylene) melts. Polymer 49(25):5467–5473
35. Lee KS, Wegner G, Hsu SL (1987) Vibrational spectroscopic studies of linear and cyclic
alkanes: CnH2n+2, CnH2n with 24n288: chain folding, chain packing and conformations.
Polymer 28(6):889–896
36. Yang Z, Yu GE, Cooke J, Ali-Adid Z, Viras K, Matsuura H, Ryan AJ, Booth C (1996)
Preparation and crystallinity of a large unsubstituted crown ether, cyclic heptacosa
(oxyethylene)(cyclo-E27, 81-crown-27), studied by Raman spectroscopy, X-ray scattering
and differential scanning calorimetry. J Chem Soc Faraday Trans 92:3173–3182
37. Cooke J, Viras K, Yu GE, Sun T, Yonemitsu T, Ryan AJ, Price C, Booth C (1998) Large
cyclic poly(oxyethylene)s: chain folding in the crystalline state studied by Raman spectros-
copy, x-ray scattering, and differential scanning calorimetry. Macromolecules 31
(9):3030–3039
Crystallization of Cyclic Polymers 127

38. Yu GE, Sun T, Yan ZG, Price C, Booth C, Cook J, Ryan AJ, Viras K (1997) Low-molar-mass
cyclic poly(oxyethylene)s studied by Raman spectroscopy, X-ray scattering and differential
scanning calorimetry. Polymer 38(1):35–42
39. Nam S (2006) Dynamics of cyclic and linear poly(oxyethylene), and threading conformation
in their blends. Ph.D. dissertation, Georgia Institute of Technology
40. Tezuka Y, Ohtsuka T, Adachi K, Komiya R, Ohno N, Okui N (2008) A defect-free ring
polymers: size-controlled cyclic poly(tetrahydrofuran) consisting exclusively of the mono-
mer unit. Macromol Rapid Commun 29(14):1237–1241
41. C
ordova ME, Lorenzo AT, Müller AJ, Hoskins JN, Grayson SM (2011) A comparative study
on the crystallization behavior of analogous linear and cyclic poly(ε-caprolactones). Macro-
molecules 44(7):1742–1746
42. Schäler K, Ostas E, Schr€ oter K, Thurn-Albrecht T, Binder WH, Saalwächter K (2011)
Influence of chain topology on polymer dynamics and crystallization. Investigation of linear
and cyclic poly(ε-caprolactone)s by 1H solid-state NMR methods. Macromolecules 44
(8):2743–2754
43. Shin EJ, Jeong W, Brown HA, Koo BJ, Hedrick JL, Waymouth RM (2011) Crystallization of
cyclic polymers: synthesis and crystallization behavior of high molecular weight cyclic poly
(ε-caprolactone)s. Macromolecules 44(8):2773–2779
44. Kitahara T, Yamazaki S, Kimura K (2011) Effects of topological constraint and knot
entanglement on the crystal growth of polymers proved by growth rate of spherulite of cyclic
polyethylene. Kobunshi Ronbunshu 68(10):694–701
45. Takeshita H, Poovarodom M, Kiya K, Arai F, Takenaka K, Miya M, Shiomi T (2012)
Crystallization behavior and chain folding manner of cyclic, star and linear poly(tetrahydro-
furan)s. Polymer 53(23):5375–5384
46. Shin EJ, Jones AE, Waymouth RM (2012) Stereocomplexation in cyclic and linear
polylactide blends. Macromolecules 45(1):595–598
47. Su H-H, Chen H-L, Dı́az A, Casas MT, Puiggalı́ J, Hoskins JN, Grayson SM, Pérez RA,
Müller AJ (2013) New insights on the crystallization and melting of cyclic PCL chains on the
basis of a modified Thomson–Gibbs equation. Polymer 54(2):846–859
48. Pérez RA, C
ordova ME, L
opez JV, Hoskins JN, Zhang B, Grayson SM, Müller AJ (2014)
Nucleation, crystallization, self-nucleation and thermal fractionation of cyclic and linear poly
(ε-caprolactone)s. React Funct Polym 80:71–82
49. Pérez RA, L
opez JV, Hoskins JN, Zhang B, Grayson SM, Casas MT, Puiggalı́ J, Müller AJ
(2014) Nucleation and antinucleation effects of functionalized carbon nanotubes on cyclic
and linear poly(ε-caprolactones). Macromolecules 47(11):3553–3566
50. Wang J, Li Z, Pérez RA, Müller AJ, Zhang B, Grayson SM, Hu W (2015) Comparing
crystallization rates between linear and cyclic poly(ε-caprolactone) via fast-scan chip-calo-
rimeter measurements. Polymer 63:34–40
51. Sugai N, Asai S, Tezuka Y, Yamamoto T (2015) Photoinduced topological transformation of
cyclized polylactides for switching the properties of homocrystals and stereocomplexes.
Polym Chem 6:3591–3600. doi:10.1039/C5PY00158G
52. Takeshita H, Shiomi T (2013) Crystallization of cyclic and branched polymers. In: Tezuka Y
(ed) Topological polymer chemistry: progress of cyclic polymers in synthesis, properties and
functions. World Scientific, Singapore
53. Jia Z, Monteiro MJ (2012) Cyclic polymers: methods and strategies. J Polym Sci Part A
Polym Chem 50(11):2085–2097
54. Brown HA, Waymouth RM (2013) Zwitterionic ring-opening polymerization for the synthe-
sis of high molecular weight cyclic polymers. Acc Chem Res 46(11):2585–2596
55. Elupula R, Laurent BA, Grayson SM (2011) Cyclic polymers. In: Schlüter AD, Hawker CJ,
Sawamoto J (eds) Synthesis of polymers: new structures and methods. Wiley, Germany
56. Tezuka Y (2012) Topological polymer chemistry for designing multicyclic macromolecular
architectures. Polym J 44(12):1159–1169
128 R.A. Pérez-Camargo et al.

57. Dubois P, Barakat I, Jér€

ome R, Teyssie P (1993) Macromolecular engineering of polyactones
and polyactides. 12. Study of the depolymerization reactions of poly(epsilon-caprolactone)
with functional aluminum alkoxide end groups. Macromolecules 26(17):4407–4412
58. Li Y, Hoskins JN, Sreerama SG, Grayson SM (2010) MALDI–TOF mass spectral character-
ization of polymers containing an azide group: evidence of metastable ions. Macromolecules
43(14):6225–6228
59. Iwata T, Doi Y (2002) Morphology and enzymatic degradation of poly(ε-caprolactone) single
crystals: does a polymer single crystal consist of micro-crystals? Polym Int 51(10):852–858
60. Chatani Y, Okita Y, Takadoro H, Yamashita Y (1970) Structural studies of polyesters. III.
Crystal structure of poly-ε-caprolactone. Polym J 1(5):555–562
61. Hu H, Dorset DL (1990) Crystal structure of poly(ε-caprolactone). Macromolecules 23
(21):4604–4607
62. Bittiger H, Marchessault RH, Niegish WD (1970) Crystal structure of poly-ε-caprolactone.
Acta Crystallogr B 26(12):1923–1927
63. Gestı́ S, Almontassir A, Casas MT, Puiggalı́ J (2004) Molecular packing and crystalline
morphologies of biodegradable poly(alkylene dicarboxylate)s derived from 1,6-hexanediol.
Polymer 45(26):8845–8861
64. Gestı́ S, Almontassir A, Casas MT, Puiggalı́ J (2006) Crystalline structure of poly
(hexamethylene adipate). Study on the morphology and the enzymatic degradation of single
crystals. Biomacromolecules 7(3):799–808
65. Gestı́ S, Casas MT, Puiggalı́ J (2007) Crystalline structure of poly(hexamethylene succinate)
and single crystal degradation studies. Polymer 48(17):5088–5097
66. Gestı́ S, Casas MT, Puiggalı́ J (2008) Single crystal morphology and structural data of a series
of polyesters derived from 1,8-octanediol. Eur Polym J 52(7):2295–2307
67. Baratian S, Hall ES, Lin JS, Xu R, Runt J (2001) Crystallization and solid-state structure of
random polylactide copolymers: poly(L–lactide-co-D-lactide)s. Macromolecules 34
(14):4857–4864
68. Huang J, Lisowski MS, Runt J, Hall ES, Kean RT, Buehler N, Lin JS (1998) Crystallization
and microstructure of Poly(L-lactide-co-meso-lactide) copolymers. Macromolecules 31
(8):2593–2599
69. Tsuji H (2005) Poly(lactide) stereocomplexes: formation, structure, properties, degradation,
and applications. Macromol Biosci 5(7):569–597
70. Obukhov SP, Rubinstein M, Duke T (1994) Dynamics of a ring polymer in a gel. Phys Rev
Lett 73:1263–1266
71. Arrighi V, Gagliardi S, Dagger AC, Semlyen JA, Higgins JS, Shenton MJ (2004) Confor-
mation of cyclics and linear chains polymer in bulk by SANS. Macromolecules 37
(21):8057–8065
72. Hur K, Winkler RG, Yoon DY (2006) Comparison of ring and linear polyethylene from
molecular dynamics simulations. Macromolecules 39(12):3975–3977
73. Suzuki J, Takano A, Matsushita Y (2008) Topological effect in ring polymers investigated
with Monte Carlo simulation. J Chem Phys 129:034903
74. Bassett DC (2003) Polymer spherulites: a modern assessment. J Macromol Sci Part B Phys 42
(2):227–256
75. Toda A, Arita T, Hikosaka M, Hobbs JK, Miles MJ (2003) An atomic force microscopy
observation of poly(vinylidene fluoride) banded spherulites. J Macromol Sci Part B Phys 42
(3–4):753–760
76. Keith HD, Padden FJ (1984) Twisting orientation and the role of transient states in polymer
crystallization. Polymer 25(1):28–42
77. Lotz B, Cheng SZD (2005) A critical assessment of unbalanced surface stresses as the
mechanical origin of twisting and scrolling of polymer crystals. Polymer 46(3):577–610
78. Okui N, Ohno N, Unemoto S, Tezuka Y (2009) Topological effect on polymer crystallization
of linear and ring polymers. Bussei Kenkyu 92(1):51–55
Crystallization of Cyclic Polymers 129

79. Roovers J (1985) The melt properties of ring polystyrenes. Macromolecules 18

(6):1359–1361
80. Mckenna GB, Hadziioannou G, Lutz P, Hild G, Strazielle C, Straupe C, Rempp P, Kovacs AJ
(1987) Dilute solution characterization of cyclic polystyrene molecules and their zero-shear
viscosity in the melt. Macromolecules 20(3):498–512
81. Orrah DJ, Semlyen JA, Ross-Murphy SB (1988) Studies of cyclic and linear poly(dimethyl-
siloxanes): 27. Bulk viscosities above the critical molar mass for entanglement. Polymer 29
(8):1452–1454
82. Izuka A, Winter HH, Hashimoto T (1992) Molecular weight dependence of viscoelasticity of
polycaprolactone critical gels. Macromolecules 25(9):2422–2428
83. McLeish T (2002) Polymers without beginning or end. Science 297:2005–2006
84. Kawaguchi D, Masuoka K, Takano A, Tanaka K, Nagamura T, Torikai N, Dalgliesh RM,
Langridge S, Matshushita Y (2006) Comparison of interdiffusion behavior between cyclic
and linear polystyrenes with high molecular weights. Macromolecules 39(16):5180–5182
85. Lee KS, Wegner G (1985) Linear and cyclic alkanes (CnH2n+2, CnH2n) with n>100. Synthesis
and evidence for chain-folding. Maklomol Chem Rapid Commun 6(3):203–208
86. Rosen BM, Wilson CJ, Wilson DA, Peterca M, Imam MR, Percec V (2009) Dendron-
mediated self-assembly, disassembly, and self-organization of complex systems. Chem Rev
109(11):6275–6540
87. Roovers J (1988) Viscoelastic properties of polybutadiene rings. Macromolecules 21
(5):1517–1521
88. Nam S, Leisen J, Breedveld V, Beckham HW (2009) Melt dynamics of blended poly
(oxyethylene) chains and rings. Macromolecules 42(8):3121–3128
89. Porod G (1951) Die R€ ontgenkleinwinkelstreuung von dichtgepackten kolloiden Systemen
I. Teil Kolloid ZZ Polym 124(2):83–114
90. Strobl G (2007) The physics of polymers: concepts for understanding their structures and
behavior, 3rd edn. Springer, Berlin
91. Cates ME, Deutsch JM (1986) Conjectures on the statistics of ring polymers. J Phys Paris 47
(12):2121–2128
92. Iyer BVS, Lele AK, Juvekar VA (2006) Flexible ring polymers in an obstacle environment:
molecular theory of linear viscoelasticity. Phys Rev E 74(2):021805
93. Chen H-L, Li L-J, Yang W-CO, Hwang JC, Wong W-Y (1997) Spherulitic crystallization
behavior of poly(ε-caprolactone) with a wide range of molecular weight. Macromolecules 30
(6):1718–1722
94. Heck B, Hugel T, Iijima M, Sadiku E, Strobl G (1999) Steps in the transition of an entangled
polymer melt to the partially crystalline state. New J Phys 1:17.1–17.29
95. Gedde UW (1999) Polymer physics, 1st edn. Kluwer, Dordrecht
96. Phillips PJ, Rensch GJ, Taylor KD (1987) Crystallization studies of poly(ε-caprolactone).
I. Morphology and kinetics. J Polym Sci B Polym Phys 25(8):1725–1740
97. Goulet L, Prud’homme RE (1990) Crystallization kinetics and melting of caprolactone
random copolymers. J Polym Sci B Polym Phys 28(12):2329–2352
98. Guo Q, Groeninckx G (2001) Crystallization kinetics of poly(ε-caprolactone) in miscible
thermosetting polymer blends of epoxy resin and poly(ε-caprolactone). Polymer 42
(21):8647–8655
99. Kuo SW, Chan SC, Chang FC (2004) Crystallization kinetics and morphology of binary
phenolic/poly(ε-caprolactone) blends. J Polym Sci B Polym Phys 42(1):117–128
100. Müller AJ, Albuerne J, Marquez L, Raquez JM, Degée P, Dubois P, Hobbs J, Hamley IW
(2005) Self-nucleation and crystallization kinetics of double crystalline poly( p-dioxanone)-
b-poly(ε-caprolactone) diblock copolymers. Faraday Discuss 128:231–252
101. Nie K, Zheng S, Lu F, Zhu Q (2005) Inorganic–organic hybrids involving poly
(ε-caprolactone) and silica network: hydrogen-bonding interactions and isothermal crystalli-
zation kinetics. J Polym Sci B Polym Phys 43(18):2594–2603
130 R.A. Pérez-Camargo et al.

102. Liu Y, Yang X, Zhang W, Zheng S (2006) Star-shaped poly(ε-caprolactone) with polyhedral
oligomeric silsesquioxane core. Polymer 47(19):6814–6825
103. Lorenzo AT, Müller AJ, Lin MC, Chen HL, Jeng US, Prfitis D, Pitsikalis M, Hadjichristidis N
(2009) Influence of macromolecular architecture on the crystallization of (PCL2)-b-(PS2)
4-miktoarm star block copolymers in comparison to linear PCL-b-PS diblock copolymer
analogues. Macromolecules 42(21):8353–8364
104. Liu H, Huang Y, Yuan L, He P, Cai Z, Shen Y, Xu Y, Yu Y, Xiong H (2010) Isothermal
crystallization kinetics of modified bamboo cellulose/PCL composites. Carbohydr Polym 79
(3):513–519
105. Wang K, Cai L, Jesse S, Wang S (2012) Poly(ε-caprolactone)-banded spherulites and
interaction with MC3T3-E1 Cells. Langmuir 28(9):4382–4395
106. Chang H, Zhang J, Li L, Wang Z, Yang C, Takahashi I, Ozaki Y, Yan S (2010) A study on the
epitaxial ordering process of the polycaprolactone on the highly oriented polyethylene
substrate. Macromolecules 43(1):362–366
107. Blundell DJ, Keller A, Kovacs AJ (1966) A new self-nucleation phenomenon and its
application to the growing of polymer crystals from solution. J Polym Sci B Polym Lett 4
(7):481–486
108. Fillon B, Wittmann J, Lotz B, Thierry A (1993) Self-nucleation and recrystallization of
isotactic polypropylene (α phase) investigated by differential scanning calorimetry. J Polym
Sci B Polym Phys 31(10):1383–1393
109. Müller AJ, Arnal ML (2005) Thermal fractionation of polymers. Prog Polym Sci 30
(5):559–603
110. Müller AJ, Michell RM, Pérez RA, Lorenzo AT (2015) Successive self-nucleation and
annealing (SSA): correct design of thermal protocol and applications. Eur Polym J
65:132–154
111. Lorenzo AT, Arnal ML, Sánchez JJ, Müller AJ (2006) Effect of annealing time on the self-
nucleation behavior of semicrystalline polymers. J Polym Sci B Polym Phys 44
(12):1738–1759
112. Fillon B, Thierry A, Wittmann J, Lotz B (1993) Self-nucleation and recrystallization of
polymers. Isotactic polypropylene, β phase: β–α conversion and β–α growth transitions. J
Polym Sci B Polym Phys 31(10):1407–1424
113. Fillon B, Lotz B, Thierry A, Wittman J (1993) Self-nucleation and enhanced nucleation of
polymers. Definition of a convenient calorimetric “efficiency scale” and evaluation of
nucleating additives in isotactic polypropylene (α phase). J Polym Sci B Polym Phys 31
(10):1395–1405
114. Turska E, Gogolewski S (1975) Study on crystallization of nylon 6 (polycapramide). III.
Effect of “crystalline memory” on crystallization kinetics. J Appl Polym Sci 19(3):637–644
115. Gallez F, Legras R, Mercier JP (1976) Crystallization of bisphenol-A polycarbonate.
I. Influence of trimellitic acid tridecyloctyl ester on the kinetics of crystallization. J Polym
Sci Polym Phys Ed 14(8):1367–1377
116. Di Filippo G, Gonzalez ME, Gasiba MT, Müller AJ (1987) Crystalline memory on polycar-
bonate. J Appl Polym Sci 34(5):1959–1966
117. Khanna YP, Reimschuessel AC (1988) Memory effects in polymers. I. Orientational memory
in the molten state; its relationship to polymer structure and influence on recrystallization rate
and morphology. J Appl Polym Sci 35(8):2259–2268
118. Khanna YP, Kumar R, Reimschuessel AC (1988) Memory effects in polymers. III.
Processing history vs crystallization rate of nylon 6–comments on the origin of memory
effect. Polym Eng Sci 28(24):1607–1611
119. Mehl NA, Rebenfeld L (1992) Effect of melt history on the crystallization kinetics of poly
(phenylene sulfide). Polym Eng Sci 32(19):1451–1457
120. Kim SP, Kim SC (1993) Crystallization kinetics of poly(ethylene terephtalate): memory
effect of shear history. Polym Eng Sci 33(2):83–91
Crystallization of Cyclic Polymers 131

121. Khanna YP, Kuhn WP, Macur JE, Messa AF, Murthy NS, Reimshuessel AC, Schneider RL,
Sibila JP, Signorelli AJ, Taylor TJ (1995) Memory effects in polymers. V. Processing history
versus thermally induced self-orientation of unoriented poly(chlorotrifluoroethylene) films. J
Polym Sci B Polym Phys 33(7):1023–1030
122. Alfonso GC, Scardigli P (1997) Melt memory effects in polymer crystallization. Macromol
Symp 118(1):323–328
123. Mendez G, Müller AJ (1997) Evidences of the crystalline memory and recrystallization
capacity of bisphenol–A polycarbonate. J Therm Anal 50(4):593–602
124. Supaphol P, Spruiell JE (2000) Crystalline memory effects in isothermal crystallization of
syndiotactic polypropylene. J Appl Polym Sci 75(3):337–346
125. Schneider S, Drujon X, Lotz B, Wittmann JC (2001) Self-nucleation and enhanced nucleation
of polyvinylidene fluoride (α-phase). Polymer 42(21):8787–8798
126. Vasanthan N (2003) “Orientation induced memory effect” in polyamides and the relationship
to hydrogen bonding. J Appl Polym Sci 90(3):772–775
127. Supaphol P, Srimoaon P, Sirivat A (2004) Effects of crystalline memory and orientational
memory phenomena on the isothermal bulk crystallization and subsequent melting behavior
of poly(trimethylene terephthalate). Polym Int 53(8):1118–1126
128. Massa MV, Lee MSM, Dalnoki-Veress K (2005) Crystal nucleation of polymers confined to
droplets: memory effects. J Polym Sci B Polym Phys 43(23):3438–3443
129. Ziabicki A, Alfonso GC (1994) Memory effects in isothermal crystallization. I. Theory.
Colloid Polym Sci 272(9):1027–1042
130. Isayev AI, Chan TW, Shimojo K, Gmerek M (1995) Injection molding of semicrystalline
polymers. I. Material characterization. J Appl Polym Sci 55(5):807–819
131. Liedauer S, Eder G, Janeshitz-Kriegl H (1995) On the limitations of shear induced crystal-
lization in polypropylene melts. Int Polym Process 10(3):243–250
132. Somani RH, Hsiao BS, Nogales A, Srinivas S, Tsou AH, Sics I, Baltá-Calleja FJ, Ezquerra
TA (2000) Structure development during shear flow-induced crystallization of i-PP: in-situ
small-angle X-ray scattering study. Macromolecules 33(25):9385–9394
133. Somani RH, Hsiao BS, Nogales A, Fruitwala H, Srinivas S, Tsou AH (2001) Structure
development during shear flow induced crystallization of i-PP: in situ wide-angle X-ray
diffraction study. Macromolecules 34(17):5902–5909
134. Nogales A, Hsiao BS, Somani RH, Srinivas S, Tsou AH, Baltá-Calleja FJ, Ezquerra TA
(2001) Shear-induced crystallization of isotactic polypropylene with different molecular
weight distributions: in situ small- and wide-angle X-ray scattering studies. Polymer 42
(12):5247–5256
135. Janeschitz-Kriegl H, Ratajski E, Stadlbauer M (2003) Flow as an effective promotor of
nucleation in polymer melts: a quantitative evaluation. Rheol Acta 42(4):355–364
136. Azurri F, Alfonso GC (2005) Lifetime of shear-induced crystal nucleation precursors.
Macromolecules 38(5):1723–1728
137. Wunderlich B (1976) Crystal nucleation, growth, annealing, vol 2, Macromolecular physics.
Academic, New York
138. Alfonso GC, Ziabicki A (1995) Memory effects in isothermal crystallization II. Isotactic
polypropylene. Colloid Polym Sci 273(4):317–323
139. Zhu X, Li Y, Yan D, Zhu P, Lu Q (2001) Influenced of the order polymer melt on the
crystallization behavior: I. Double melting endotherms of isotactic polypropylene. Colloid
Polym Sci 279(3):292–296
140. Supaphol P, Lin JS (2001) Crystalline memory effect in isothermal crystallization of
syndiotactic polypropylenes: effect of fusion temperature on crystallization and melting
behavior. Polymer 42(23):9617–9626
141. Mamum A, Unemoto S, Okui N, Ishihara N (2007) Self-seeding effect on primary nucleation
of isotactic polystyrene. Macromolecules 40(17):6296–6303
142. Martins JA, Zhang W, Brito AM (2010) Origin of the melt memory effect in polymer
crystallization. Polymer 51(18):4185–4194
132 R.A. Pérez-Camargo et al.

143. Kawabata J, Matsuba G, Nishida K, Inoe R, Kanaya T (2011) Melt memory effects on
recrystallization of polyamide 6 revealed by depolarized light scattering and small-angle
X-ray scattering. J Appl Polym Sci 122(3):1913–1920
144. Zhang YS, Zhong LW, Yang S, Liang DH, Chen EQ (2012) Memory effect on solution
crystallization of high molecular weight poly(ethylene oxide). Polymer 53(16):3621–3628
145. Cheng SZD (2008) Phase transitions in polymers: the role of metastable states, 1st edn.
Elsevier, Amsterdam
146. Subramanian G, Shanbhag S (2008) Self-diffusion in binary blends of cyclic and linear
polymers. Macromolecules 41(19):7239–7242
147. Vasquez R, Shanbhag S (2011) Percolation of trace amounts of linear polymers in melt of
cyclic polymers. Macromol Theory Simul 20(3):205–211
148. Chapman CD, Shanbhag S, Smith DE, Robertson-Anderson RM (2012) Complex effects of
molecular topology on diffusion in entangled biopolymer blends. Soft Matter 8:9177–9182
149. Halverson JD, Grest GS, Grosberg AY, Kremer K (2012) Rheology of ring polymer melts:
from linear contaminants to ring-linear blends. Phys Rev Lett 108:038301
150. Henke SF, Shanbhag S (2014) Self-diffusion in asymmetric ring-linear blends. React Funct
Polym 80:57–60
151. Tsalikis G, Koukoulas T, Mavrantzas VG (2014) Dynamic, conformational and topological
properties of ring-linear poly(ethylene oxide) blends from molecular dynamics simulations.
React Funct Polym 80:61–70
152. Gooβen S, Brás AR, Pyckhout-Hintzen W, Wischnewski A, Richter D, Rubinstein M,
Roovers J, Lutz PJ, Jeong Y, Chang T, Vlassopoulos D (2015) Influence of the solvent
quality on ring polymer dimensions. Macromolecules 48(5):1598–1605
153. Müller AJ, Hernández ZH, Arnal ML, Sánchez JJ (1997) Successive self-nucleation/
annealing (SSA): a novel technique to study molecular segregation during crystallization.
Polym Bull 39(4):465–472
Adv Polym Sci (2017) 276: 133–182
DOI: 10.1007/12_2015_346
© Springer International Publishing Switzerland 2015
Published online: 13 December 2015

Crystallization of Precision Ethylene

Copolymers

Laura Santonja-Blasco, Xiaoshi Zhang, and Rufina G. Alamo

Abstract The crystallization and melting of polyethylene-like copolymer systems

with co-units placed at the same equal distance along the backbone are reviewed on
the basis of available thermal and structural data for a large variety of pendant and
backbone-inserted groups. Data for a series of precision halogen-containing poly-
ethylenes are used to describe the effect of size of the pendant group and the
distance between groups along the backbone on crystallization and melting behav-
iors. The effect of crystallization kinetics on polymorphism observed for systems
with co-crystallizable co-units is emphasized with data on polyethylenes containing
Cl and Br. The major characterization techniques for identifying different modes of
packing are also described. The melting behavior of alkyl-branched precision
polyethylenes is analyzed in reference to branch partitioning between crystalline
and non-crystalline regions, and also in reference to the behavior of analog poly-
ethylenes containing halogens. Furthermore, the effect of interacting functional
groups and tacticity on self-assembly and melting is analyzed using available
literature data.

Keywords Ethylene copolymers • Polyethylenes • Precision branching • Precision

copolymers • Precision substitution

Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 134
1.1 Crystallization of Ethylene Copolymers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
2 Crystallization of Polyethylenes with Precision Halogen Substitution . . . . . . . . . . . . . . . . . . . 139
2.1 Rapid Crystallization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 140
2.2 Isothermal Crystallization: Polymorphism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 150

L. Santonja-Blasco, X. Zhang, and R.G. Alamo (*)

Department of Chemical and Biomedical Engineering, FAMU-FSU College of Engineering,
2525 Pottsdamer Street, Tallahassee, FL 32310-6046, USA
e-mail: [email protected]
134 L. Santonja-Blasco et al.

3 Crystallization of Precision Alkyl-Branched Polyethylenes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 157

4 Crystallization of Precision Polyethylenes with Interacting Functional Groups . . . . . . . . . 161
4.1 Strength of Pendant Group Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 162
4.2 Functional Groups Within or Branching from the Methylene Backbone . . . . . . . . . . . 163
4.3 Effect of Tacticity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
5 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 166
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176

1 Introduction

Studies of novel, model polyolefins and polyolefin-like materials have industrial

relevance as well as strong academic interest. Polyolefins, mainly the commodities
polyethylenes and polypropylenes, now occupy a leading position among all
existing types of materials. High production and versatility of properties account
for the relevant position of these polymers over any other polyolefin material. For
example, the annual worldwide production of the major polyolefins has increased
steadily from 180 BP (billion pounds) in 2000 to over 290 BP in 2013, accounting
for >65% of all thermoplastics [1]. The production of polyolefins is expected to
increase at an even faster rate in the near future as a result of the exponential
expansion of shale gas extraction in the USA and the derived investment in ethylene
generation. Consequently, polyolefin research will continue at all levels, as small
improvements in the process or in the product can lead to a significant impact on the
world economy.
At a fundamental level of research, polyethylene contains the simplest repeating
unit, (CH2)n, yet it can be arranged in a rich variety of architectures. It is now
relatively easy and inexpensive to add functional groups, defects, or branches (short
or long) to the backbone to enable fine control of molecular structure by simple
modification of the process or by copolymerization. Structural changes in the chain
modify the level of crystallinity, crystallization rate, and the final semicrystalline
morphology, and thus alter the physical and mechanical properties of the material.
Polyethylenes have been extensively studied, yet they remain the polymers of
choice for model systems to study the fundamental laws that govern the structure–
property relations of semicrystalline polymers. The synthesis of models mimicking
polyethylenes is still a topic of major interest.
Although incorporation of large comonomer content in polyethylenes while
maintaining a uniform comonomer distribution and control of molecular weight
is now feasible with metallocene catalysts, a drawback is their inability to incorpo-
rate polar groups [2–4]. These constraints make other routes for synthesis of
ethylene copolymers with ionic polar groups or halogens attractive, especially as
a means to generate model materials for studying the effects of volume of sub-
stituent and secondary bonding on the packing behavior of polyethylenes with a
controlled content and type of defect. Strategies could then be set to design the next
generation of polyolefin materials. Special families of branched and substituted
Crystallization of Precision Ethylene Copolymers 135

polyethylene-like systems have been synthesized via acyclic diene metathesis

polymerization (ADMET), mainly by Wagener and coworkers [5]. These poly-
olefins are characterized by the placement of an alkyl branch, a halogen, or a large
variety of functional groups at equal distances along the CH2 backbone. Structur-
ally, they are a new class of polyolefins with a repeating unit,
[(CH2)x – CHR]n,
where x changes from 4 to 74, R is the branch or functional group, and the molar
mass is 30–150 K. An alternative route toward precision polyethylene-like struc-
tures is via ring-opening metathesis polymerization (ROMP) [6]. Through the
latter, the regioregularity of the addition must be maintained in order to generate
structures with equidistant control of branch placement [7–9]. Long-spaced poly-
esters, polyamides, polycarbonates, or polyacetals have also been generated via
ROMP, thus providing polyethylene-like models with functional groups directly
linked to the backbone [10–13]. The skills of organic chemists in producing these
modified model polyethylenes are truly remarkable.
The synthetic route for generation of the acyclic diene, or the cycle if the
polymerization is carried out by ROMP, is of paramount importance for further
generation of the polyethylene backbone with functional or pendant groups placed
at an exact equidistant length or very close to the equidistant length, as in most
ROMP avenues [14]. However, synthesis is not the focus of this review; hence,
although the source for the synthetic path is cited, we focus our attention on the
effect on the crystallization and melting behaviors of the punctuation placed
randomly or at a precise distance along the polyethylene backbone. We have chosen
halogens as models to establish the effect of size and content on crystallization and
to characterize polymorphism in many of these systems. Subsequently, we sum-
marize relevant crystalline behaviors of precision systems with alkyl branches and
with other functional groups.

1.1 Crystallization of Ethylene Copolymers

The crystallization of statistical ethylene-based copolymers is fairly well under-

stood. The thermodynamic and structural properties of ethylene 1-alkene copoly-
mers have been amply studied and summarized in different reviews [15–19] and the
most recent works on statistical ethylene copolymers are covered in a review by
Hu et al. in this volume [20]. The decrease in melting temperature with increasing
content of 1-alkene is understood on the basis of phase equilibria for a
two-component system made of A and B units, and the experimental data are
analyzed comparatively with predictions from equilibrium theories. When the
crystalline phase remains pure, for example only component A or the backbone
CH2 units of ethylene/1-alkene random copolymers crystallize, the experimental
melting data are analyzed on the basis of Flory’s equilibrium theory [21]. Flory’s
theory was developed for the case where the crystalline phase remains pure. Under
this premise, the equilibrium melting temperature of the copolymer (Tmocopo),
relative to that of the homopolymer (Tmo), is expressed as:
136 L. Santonja-Blasco et al.

1 1 R

o ¼ ln p ð1Þ
T mo T mcopo ΔH u

In this equation, ΔHu is the enthalpy of fusion per mole of crystalline repeating
unit, and p is the crystallizable sequence propagation probability. In statistical
copolymers, p < 1 and the equation predicts a large depression of Tmocopo. For
random copolymers, p ¼ XA (the mole fraction of crystallizable units) and Tmocopo
decreases proportionally to the content of non-crystallizable co-units. For an
ordered or block copolymer, p approaches one, thus predicting a negligible melting
point depression in the copolymer.
The concepts embodied by Eq. (1) are amply substantiated by experiments [22–
30]. Random ethylene copolymers with 1-butene or longer co-units are known to be
rejected from the crystallites. As such, their melting temperature–composition
behavior is in full agreement with theoretical predictions [15, 22]. Flory’s theory
also puts kinetic limitations on the size and content of crystallites that are formed
from a copolymer chain as a function of crystallization temperature [21, 24]. Only
crystallizable sequences longer than the critical length for a stable crystallite can
participate in the formation of crystallites. This kinetic restraint invariably leads to
a sequence length selection process in copolymer crystallization that ultimately
determines the morphology and structure of the crystalline (CR) and non-crystalline
(NCR) regions of the copolymers. Consequently, on cooling from the melt, the
copolymer crystalline morphology is complex because progressively thinner crystal-
lites are expected to develop on cooling. As the branches, or co-units, remain in the
intercrystalline regions, the concentration of co-units that surround the copolymer
crystallites becomes higher than the chain comonomer concentration during the
process of crystallization. On subsequent melting, copolymer crystallites melt over
a broad range of temperatures, because thinner crystallites melt first, followed by the
thicker crystallites, and because crystals that melt first coexist with a melt that
contains the highest content of co-units, or lowest value of XA. Even if random
copolymers with the co-unit excluded from the crystal develop a relatively narrow
distribution of crystal thicknesses, their melting behavior is expected to be broad
because of considerations of Eq. (1) and the compositional change in the inter-
crystalline regions during melting.
When the volume of the branch or pendant group is sufficiently small, some of
the co-units can participate in the crystallites. This situation has been found for the
methyl branch of random ethylene propylene copolymers [31–37], the halogen of
ethylene vinyl halides [38–42], the OH of ethylene vinyl alcohol, and other small
pendant or backbone-incorporated groups [43–45]. The thermodynamic predictions
for these cases were developed by Sanchez and Eby, considering that the crystal-
lizable co-unit adds an excess free energy ε to the crystal free energy of melting
[46, 47]. The depression of the copolymer melting point (on equilibrium basis) from
the value of the homopolymer is given by:
Crystallization of Precision Ethylene Copolymers 137

"
  #
1 1 R XC ð1
X C Þ XC

o ¼
ε þ ð1
XC Þln þ XC ln
T mo T mcopo ΔH u o
RT mcopo ð1
X B Þ XB
ð2Þ

Here, XB is the concentration of co-units B in the chain, and XC the concentration

of co-units B in the crystallites.
Although the derivations of these equations are rigorous and conceptually
viable, their application to distinguish the partitioning of the co-unit between CR
and NCR regions from experimental melting temperatures is limited for two major
reasons. First, it is impossible to observe the Tmocopo predicted by the theoretical
equations (1) and (2) under equilibrium conditions. Only crystallites formed with
the longest crystalline sequences melt at Tmocopo. The content of these sequences is
low, and the number of crystallites formed with these sequences even lower
because transport limitations reduce the effective number that can diffuse to the
crystal front. The small number of crystals from the longest sequences, even if
enabled, are not accessible by the usual experimental techniques. As a consequence,
the observed copolymer melting temperatures are much lower than theoretically
predicted. A second caveat is that both equations predict a decrease in Tmocopo with
increasing comonomer content, or decreasing XA. Hence, from experimental melt-
ing point data for random copolymers, which are usually observed to decrease with
increasing the co-unit content, one cannot conclude that the co-unit is a defect in the
crystallites. Other measurements are needed to discern the partitioning of the
copolymer co-units. For example, expansion of the unit cell is often an indication
of side branch inclusion in the crystallites, whereas solid state NMR can identify
and quantify residues inside the crystallites [48–52].
Even for random copolymers with relatively small co-units, such as those that
can be partially accommodated in the crystallites, the crystallization path is driven
mainly by selection of the most crystallizable sequence lengths. Long ethylene
sequences fold back and forth in the early stages, trapping some of the small-sized
co-units, and shorter sequences crystallize in subsequent stages with an increased
probability of pulling more co-units into the crystals, as schematically shown in
Fig. 1a. The final crystalline morphology is a complex mixture of crystallites, with a
partitioning of co-units that is difficult to control and a disordered distribution of
co-units inside and outside the crystallites.
As a result of the random distribution and the defect-like nature of the co-unit in
classical ethylene copolymers, the role of the co-unit has been directed either to
decrease the degree of crystallinity in a controlled manner [22–30] or to modify the
properties of the interlamellar regions [53–57]. The decrease in the level of crystal-
linity of random ethylene 1-alkene copolymers with increasing content of 1-alkene
is a feature amply exploited by the manufacturers of linear low-density polyethyl-
enes, whereas examples of properties affected by co-unit interactions are found in
the families of ionomers and ethylene vinyl alcohol copolymers [43, 54–57].
138 L. Santonja-Blasco et al.

Random ethylene--
based copolymers
(a)

Precision:
(b)

(i) (ii) (iii)

Fig. 1 Schematic models of crystal chain packing for polyethylenes with pendant groups placed
either randomly (a) or at an equidistant spacing along the backbone (b). Three possibilities (i–iii)
for self-assembly are featured for the latter type

Properties that depend on controlling ordered distributions of co-units in the

crystal are not accessible via classical polyolefins with a random distribution of
co-units. The present commercial strategy for modifying or enhancing the proper-
ties of ethylene-based or propylene-based copolymers is to change the comonomer
content and to change the distribution of the comonomer content across the
molecular weight distribution while maintaining the statistical intramolecular distri-
bution of the co-unit [2, 18, 58, 59]. This “tuning” of the bivariate distribution can
improve material performance, but the type of crystalline structure remains basic-
ally unchanged; thus, the basic spectrum of properties is still governed by the level
of crystallinity and the molecular mass for most statistical copolymers.
Studies of polyethylenes with co-units placed at the same distance along the
backbone serve to address whether copolymers can adopt crystalline structures
other than that adapted from the crystallization path of a random copolymer.
Depending on the partitioning of the co-unit (excluded or included in the crystalline
regions) and the distribution of the co-units in the crystallites, three viable crystal-
line structures are possible, as shown in Fig. 1b. If the co-unit is too large to be
accommodated in the crystal, the only possible packing is obtained by staggering
ethylene sequences, leaving the defect at the interface (Fig. 1b, packing scheme i).
Crystallites of this nature are thin, because their thickness is limited by the length
between co-units, and possibly have large and complex crystal–amorphous
Crystallization of Precision Ethylene Copolymers 139

interfaces. Examples of this type of crystal have been described for polyesters
[60, 61] and for precision polyethylenes with n-hexyl and n-butyl pendant groups
[62, 63].
When the co-unit finds no discrimination against entering the crystalline regions
and folding is preferential (i.e., at the same location along the backbone), the
co-units are ordered in layers inside the crystalline lamellae, resulting in unique
nanostructures at the lamellar and sub-lamellar level (Fig. 1b, packing scheme ii)
that are not feasible in classical branched polyethylenes. The third possible struc-
ture for copolymers with equidistant co-units along the backbone is obtained when
folding is random along the backbone and the distribution of the co-unit inside the
crystal is of a more random nature (Fig. 1b, packing scheme iii). Examples of
structures (ii) and (iii) enabled by precision polyethylenes with equidistant pendant
units have been found relatively recently [64–67].
Structures such as type (ii) are highly desirable, especially for long-chain macro-
molecules, because they open the window for design of unique nanostructures with
multiple length scales at the lamellar and sub-lamellar levels. Highly symmetric
layered systems enable the design of polyolefin-like materials with applications
driven by a tunable sub-lamellar crystalline structure.
We summarize here the major works on precision polyolefin-like systems that
present clear partitioning of the co-unit inside and outside the crystalline regions
and, hence, allow predictions for novel polyolefin-like systems with nanoscale
assembly at the lamellar and sub-lamellar level.

2 Crystallization of Polyethylenes with Precision Halogen

Substitution

Series of polyethylene-like systems with F, Cl, or Br atoms placed on each and

every 21st, 19th, 15th, or 9th carbon have been synthesized using ADMET, starting
from a common precursor that was further reduced to the alcohol prior to halo-
genation [68–71]. Exhaustive hydrogenation leads to repeating unit
[(CH2)x-
CHY]n
, where Y is the halogen type and x is the number of groups in the
continuous methylene segment between halogens (x varies between 8 and 20).
We term these polymers PE(x+1)Y. For example, PE21F is the polyethylene with
a fluorine atom on every 21st backbone carbon. Because the synthetic path does not
allow control of tacticity, these polymers are presumed to be atactic.
In spite of their atactic nature, these precision systems are crystalline, and
because their molar masses are usually >20,000 g/mol, they have been studied as
models to probe the effect of the size and content of halogens on the crystallization
of polyethylenes. Capitalizing on the systematic increase in van der Waals radius of
the substituent for the series, quantitative data are now available to evaluate the
degree to which the orthorhombic polyethylene lattice can tolerate atomic hydrogen
substitution. The crystalline properties and thermodynamic data of these systems
140 L. Santonja-Blasco et al.

under rapid crystallization have been studied in some detail and are reviewed in this
section. The formation of different polymorphs, observed under rapid or slow
crystallizations, are starting to be reported and are in line with the hypothetic
packing modes feasible for systems with equidistant placement of co-units, as
mentioned earlier. We review work on each of the crystallization modes in separate
sections.

2.1 Rapid Crystallization

Properties such as melting temperature and heat of fusion are often obtained via
differential scanning calorimetry (DSC) and for samples that are crystallized from
the melt at 10 C/min. Under these conditions, many precision samples crystallize
rapidly in a narrow range of temperatures. Wide-angle X-ray diffraction (WAXD)
patterns are also often obtained at room temperature for samples cooled from the
melt to infer their crystalline structure.

2.1.1 Effect of Halogen Size

The effect of the halogen’s van der Waals radius on crystalline packing and melting
behavior is extracted from data on precision systems with a F, Cl, or Br atom placed
on every 19th backbone carbon (x ¼ 18) [68]. To visualize the effect of halogen size
at the level of the unit cell, WAXD patterns are shown in Fig. 2, together with the
diffractogram of a linear polyethylene with similar molar mass (Mn ~20,000 g/mol).
In contrast to the broad WAXD patterns of ethylene vinyl halides containing a
random distribution of the halogen [39], the X-ray patterns of all precision halogen-
substituted samples are sharp, similar in nature to the pattern of the unsubstituted
linear polymer. Narrow reflections are a strong indication of a homopolymer-like
crystallization for these systems rather than a copolymer-like crystallization. In the
latter, the longest ethylene sequences are first selected for crystallization, followed
by shorter or less tactic sequences, thus resulting in a mixture of crystallites with
different thicknesses. This complexity broadens the X-ray diffraction patterns of
randomly distributed copolymers [72].
In systems with precision placement of the halogen, sharp WAXD reflections are
indicative of relatively thick crystallites that must include the halogen. In fact,
atomic force microscopy (AFM) images have demonstrated a lamellar habit with
crystal thicknesses >100 Å. As the all-trans length between halogens is just 24 Å
for all
[(CH2)18-CHY]n
systems, it is evident that the crystal stem length must
be a segment of the chain with four or five continuous repeating units. Therefore,
these precision systems fold back and forth, accommodating the halogens in the
lamellar crystallites.
Crystallization of Precision Ethylene Copolymers 141

Fig. 2 WAXD patterns of

rapidly crystallized
precision polyethylenes
with a halogen placed on
each and every 19th
backbone carbon. Reprinted
with permission from Boz
et al. [68]. Copyright 2006
American Chemical Society
Orthorhombic Linear PE

Intensity (a.u.)

Orthorhombic PE19F

Triclinic PE19Cl

Triclinic PE19Br

3 8 13 18 23 28 33 38
2q (degrees)

The fact that precision polyethylenes develop levels of crystallinity of ~70%,

which are much higher than those for random analogs (<30%), also suggests a
crystalline state built on the basis of substitutional solid solutions. In other words, in
packing backbone sequences, the substitution of hydrogen for one F, Cl, or Br atom
at an equidistant position along the backbone creates lattice distortions at levels
proportional to the solute’s van der Waals radius. This is readily apparent in the
patterns shown in Fig. 2. The small size of the F atom makes insignificant distor-
tions to the crystal lattice, and hence the orthorhombic symmetry of the unbranched
system is preserved, as seen by the presence of the (110) and (200) reflections at
21.5 and 23.5 . However, the larger sizes of Cl and Br atoms creates larger lattice
strains that promote the formation of a different crystallographic phase, a triclinic
cell with reflections at 2θ of ~19 and 22 , attributed to planes (100) and (010),
respectively. Some distortion of the lattice by accommodation of the halogen atom
is evident from the shift at lower angles of both reflections in the orthorhombic
pattern of PE19F with respect to the X-ray pattern of the unsubstituted polymer.
Shifts to lower angles are more pronounced for the reflection at ~ 22 of the triclinic
pattern of PE19Br compared with PE19Cl. Hence, the angular shifts follow expect-
ations for shifts caused by the large number of incorporated side groups in the
crystal, in spite of the atactic nature of the halogen substitution for these samples.
142 L. Santonja-Blasco et al.

The narrow melting thermograms also reflect a homopolymer-like behavior

(as shown in Fig. 3), but the melting temperatures shift at dramatically lower values
with increasing van der Waals radius of the halogen [68]. In reference to the 133 C
melting of the unsubstituted linear PE, the placement of a fluorine atom on every
19th backbone carbon lowers the melting point to 127.5 C, chlorine brings this
value to 72.7 C, and bromine to 61.5 C. Paralleling the decrease in melting temper-
ature, the degree of crystallinity decreases from ~80% to ~40%, indicating that the
halogens in the crystallites act as defects that distort the packing structure at levels
proportional to their size. This feature is also observed by the linearity of the plot of
Tm versus van der Waals radius, projecting a value of 30 C for the melting
temperature of the iodine analog, which is not yet available.

150

100
Tm (°C)

50

0
Heat Flow (W/g)

1.35 1.55 1.75 1.95

van der Waals Radius (Å)

Linear PE

PE19F

PE19Cl

PE19Br
10 W/g

10 30 50 70 90 110 130 150

Temperature (°C)
Fig. 3 Effect of halogen size on melting of rapidly crystallized precision polyethylenes. Melting
of a linear unsubstituted polyethylene is shown as reference
Crystallization of Precision Ethylene Copolymers 143

Direct evidence of the homopolymer-like crystallization, or lack or preferential

partitioning of the halogen between CR and NCR regions, is found by solid-state
NMR via direct polarization under MAS [68, 73]. For quantitative data, direct
polarization under magic angle spinning (DPMAS) is preferred over cross-
polarization (cpMAS) as the latter emphasizes the CR where cross-polarization is
most efficient. In 13C NMR under high power C–H decoupling, the methine reso-
nance is ideal for monitoring the partitioning of Cl and Br atoms between the CR
and NCR regions. The CH resonance is also unique and can be used to quantify the
content of halogen, or other pendant group, in each region, especially when both
phases contain a sufficiently large number of CHY groups and CR and NCR phases
display a resolved difference in their chemical shift (as seen in Fig. 4) [68]. CR and

Fig. 4 13C NMR spectra of (a) CR CR

(a) PE19Cl and (b) PE19Br
recorded by direct
polarization under magic NCR
angle spinning (DP MAS)
and high power 1H
decoupling. Insets show
more expanded CH 45– NCR
80 ppm regions with fits to
crystalline (CR) and
non-crystalline (NCR) 75 70 65 60
components. Reprinted with
permission from Boz PE19Cl
et al. [68]. Copyright 2006
American Chemical Society

85 75 65 55 45 35 25 15 5
δ (ppm)

(b) CR CR

NCR

NCR

75 65 55 45
PE19Br

85 75 65 55 45 35 25 15 5
δ (ppm)
144 L. Santonja-Blasco et al.

NCR resonances are resolved for CH2 at 34.1 and 31.5 ppm, respectively, as well as
for CH carbons at 67 and 64 ppm for PE19Cl, respectively. CHBr resonances are
centered at 62 ppm (CR) and 59 ppm (NCR). The ratio of intensities between CR
and NCR CHY resonances is equivalent to the ratio of CR and NCR CH2 reso-
nances in both spectra, thus indicating that CH2 and CHY are equally distributed
between both phases. In other words, there is no preference for the halogen to be in
one particular region, in agreement with the homopolymer-like crystallization
behavior extracted from the sharp WAXD and DSC thermograms. In contrast,
classical random ethylene vinyl chlorides showed no difference in resonance
between both phases in the methine region [38, 72].

2.1.2 Effect of Methylene Sequence Length Between Halogens

Melting Behavior

The effect of F, Cl, or Br substitution placed at equidistant lengths of 9, 15, 19, or

21 backbone carbons is first analyzed by the melting behavior of samples crystal-
lized under relatively rapid conditions. The endotherms are shown in Fig. 5, the
dotted lines indicating the behavior of random analogs. All melting peaks of
precision samples are sharp, which is characteristic of the fusion of homopolymers.
In contrast, the melting of random systems is broad, spreading to final melting
temperatures that are higher than those of the precision samples. This is a feature
invariably encountered when the melting of semicrystalline precision and random
analogs are compared, regardless of the branch moiety, pointing to the effect of the
sequence length in the crystallization of random copolymers [72–74]. Long ethyl-
ene sequences in random copolymers generate thicker crystallites with little halo-
gen incorporation that melt at higher values. Conversely, shorter sequences
Heat Flow (W/g)
Heat Flow (W/g)
Heat Flow (W/g)

PE21F

PE21Cl

PE15F PE19C
PE21Br
PE15Cl
PE15Br
1 W/g
5 W/g

2 W/g

PE9F PE9Cl PE9Br

50 70 90 110 130 150 0 20 40 60 80 100 120 -20 0 20 40 60 80

T (°C) T (oC) T (°C)

Fig. 5 DSC endotherms of precision halogen-containing polyethylenes (solid lines) and random
analogs (dotted lines) crystallized at 10 C/min from the melt. PE9Cl and PE9Br were quickly
quenched from the melt to
20 C prior melting. The heating rate was 10 C/min
Crystallization of Precision Ethylene Copolymers 145

generate crystallites with more defects, or generate thinner crystallites that melt at
lower temperatures, thus resulting in broad endotherms.
The effect of the halogen, as a defect in the crystal, on melting is substantiated by
the shift to lower values of the thermograms with increasing halogen content
(as seen in Fig. 5). The peak melting temperatures (Tm) are little affected by sub-
stitution of a hydrogen of the polyethylene backbone for a fluorine atom on every
9th, 15th, or 21st backbone carbon. This insignificant change in Tm confirms that the
small size of F results in only a small disturbance to the crystal lattice. Nonetheless,
the role of F as a defect is made implicit by a decrease in heat of fusion with
increasing F content in the series [71]. In contrast, increasing Cl or Br content in
precision systems leads to a large depression of the melting point; the larger the size
and the higher the content of halogen in the series, the lower the melting point.
These trends are emphasized in Fig. 6, where the melting temperature is plotted
against halogen content for crystallites of the same structure. As discussed in the
following section, these polymers can form two polymorphic structures (forms I
and II) depending on crystallization kinetics. When cooled at 10 C/min, most
develop form I; however, form I crystallites of PE9Cl and all Br-substituted
precision systems undergo fast recrystallization to form II on melting at 10 C/
min. The possibility of polymorphism and melt-recrystallization need to be taken
into account when comparing melting trends within precision polyethylene series.
Hence, to compare the thermal behaviors for the same type of structure, the low
melting peak is plotted for PE9Cl and for the bromine series in Fig. 6, instead of the
more prominent DSC peaks of Fig. 5. The melting points scale proportionally to
size and defect content. When the halogens are largely spaced, for example for the
PE21Y series, the content of halogen in the crystal is relatively low; consequently,
the depression of melting temperature by the increase in halogen size from F to Br is
~60 C. In shorter spaced precision samples, the crystallites contain larger contents
of halogens that, as defects, disturb the crystalline packing to a larger extent;
consequently, the melting temperature and heat of fusion are more dramatically
depressed by the halogen size. As seen in Fig. 6, Tm decreases by ~150 C in the
PE9Y series.

Fig. 6 Variation of melting

peak temperature with 140 F
content of halogen (moles 110
of halogen per
Tm (oC)

100 backbone carbons) for 80

rapidly crystallized 50 Cl
precision halogen-
substituted polyethylenes. 20
Br
The effect of halogen size -10
on melting temperature is
evident in the series -40
0 2 4 6 8 10 12
Mole % Halogen
146 L. Santonja-Blasco et al.

The data sets of Fig. 6 are clear examples of a melting behavior that decreases
linearly with co-unit content, but cannot and should not be related to the concepts
underlying the basis of Eq. (1), as often encountered in the literature [75]. Even
Eq. (2), derived on the basis of co-unit inclusion, might not be a good approach for
analysis of the melting behavior of precision polyethylenes. The reason is that the
derivation of Eq. (2) assumes a statistical distribution of defects in the crystallites,
which is not the case for these precision systems.

Crystal Structure

The effect of halogen size and content on crystalline packing at the level of the unit
cell can be evaluated by the set of diffractograms shown in Fig. 7 for each halo-
genated series. Clearly, the crystalline state of precision ethylene vinyl fluoride-like
copolymers is isomorphous to the crystalline state of linear polyethylene. The
orthorhombic symmetry is maintained even at high contents of fluorine, with
negligible shifts in angular reflections. In contrast, the unit cell of precision Cl-
and Br-containing polyethylenes differs from the orthorhombic packing. Rapidly
crystallized precision Cl- and Br-substituted polyethylenes pack in triclinic crystallo-
graphic cells characterized by two main reflections at ~19 and ~22 . The change
from orthorhombic to a less symmetric triclinic unit cell indicates that a reduced
order in the packing is needed to facilitate the spatial requirements for accommo-
dation of Cl and Br atoms between adjacent molecules in the crystallites. The
increasing content of Cl and Br inside the crystals expands the lattice to levels
proportional to the content of halogen in the chain, as seen for each series by the shift

PE21Br
PE21F
PE21Cl
intensity (a.u.)
Intensity (a.u.)
Intensity (a.u.)

PE19Cl
PE19Br
PE15F
PE15Cl
PE15Br
PE9Cl
PE9F

Linear PE
Linear PE
Linear PE

5 10 15 20 25 30 35 40 5 10 15 20 25 30 35 40 4 9 14 19 24 29 34 39
2θ (degrees) 2θ (degrees) 2θ (degrees)

Fig. 7 WAXD diffractograms of precision halogen-containing polyethylenes, with halogens

placed on every 21st, 19th, 15th, or 9th backbone carbon. Samples were molten and taken to
crystallize at room temperature. The dotted diffractogram for PE9Cl was collected for a sample
quenched quickly to 10 C
Crystallization of Precision Ethylene Copolymers 147

to progressively lower angles of the reflection at a 2θ of ~22.5 , which is character-

istic of the (010) plane of the triclinic lattice.
A feature of great interest is the dramatic effect of crystallization kinetics on the
packing of Cl- and Br-substituted systems, and on other precision systems with
pendant groups able to be accommodated in the crystallites. This feature is
observed in the diffractograms of PE9Cl in Fig. 7. PE9Cl crystallizes slowly at
room temperature. For this sample to crystallize faster, it needs to be quenched
quickly to temperatures below 23 C. Hence, the pattern of PE9Cl very quickly
quenched to 10 C (given by the dotted line in Fig. 7) displays the two reflections
typical of the triclinic lattice, as observed in other precision chlorine members that
crystallize quickly above room temperature. Conversely, the pattern of the speci-
men crystallized at room temperature (continuous line in Fig. 7) displays multiple
WAXD reflections, indicating a mixture of polymorphs [73]. This was the first
observation revealing that polyethylenes with precise Cl or Br substitution can pack
in at least two crystalline polymorphs that can be controlled by crystallization
kinetics [64, 65]. Mixed polymorphic behavior is also observed in the WAXD
patterns of PE21Br and PE15Br (see Fig. 7). This behavior is discussed further in
the next section.
The arrangement of halogens inside the crystallites, whether random or layered,
is a feature of extraordinary interest that can be extracted from reflections at low
angles of the WAXD patterns, which are characteristic of layered crystallites. These
reflections are especially intense for the Br series. In the WAXD patterns shown in
Fig. 7, collected under reflection, the reflections at intermediate angles (2θ of 5 –7 )
correspond closely to the distance between halogens and reflect a crystalline
halogen layered structure in polyethylenes with precise Cl and Br substitution.
Intermediate angles are absent in the systems with F substitution, as F is probably
randomly dispersed in the crystals and, hence, not layered.
Layered lamellae crystallites are often found when interchain interaction
between functional co-units drive the layered packing, as in long alkyl-based ali-
phatic polyesters [60, 76]. For polyethylenes with halogen substitution, it has been
posed that the observed layered packing is likely to be favored by preferential
folding at the Cl or Br position [73], a feature also suggested for the crystals of
model chlorinated n-alkanes [77]. It remains to be elucidated whether the drive to
form a layered structure is preferential folding, as pointed out in the Introduction,
or halogen–halogen interactions, in spite of the atactic nature of the halogen
substitution.

Degree of Crystallinity and Thermodynamic Properties

The degree of crystallinity can be extracted from the X-ray diffractograms, from the
heat of fusion if data for the heat of fusion of the pure crystal is available, from
quantitative DPMAS spectra, or from Raman spectroscopic data for orthorhombic
crystallites [68, 73]. Degrees of crystallinity obtained from WAXD patterns and the
thermodynamic properties for rapidly crystallized precision Cl-containing series
148 L. Santonja-Blasco et al.

Table 1 Thermodynamic properties of precision Cl-substituted polyethylenes

Xc ΔHm ΔHo ΔHo/bond Tmo ΔSo ΔSo/bond
Sample (WAXD)a (J/g)a (J/mol)b (J/mol) ( C) (J/mol K)b (J/mol K)
PE 0.83 238 4,014 4,014 145.5 9.6 9.6
PE21Cl 0.52 120.5 76,777 3,656 83 216 10.3
PE19Cl 0.53 107.6 61,225 3,222 75 176 9.3
PE15Cl 0.46 95.7 51,158 3,411 65 151 10.1
PE9Cl 0.34 38.6 18,397 2,044 54 56 6.2
Data taken with permission from Alamo et al. [73]. Copyright 2008 American Chemical Society
a
Data for samples cooled at 1 C/min from the melt to room temperature
b
In PE the mole is defined as a CH2 unit, whereas for the precision EVC series, the mole is the
[(CH2)x-CHCl] unit

(form I) are listed in Table 1 and plotted against Cl content in Fig. 8a. Equilibrium
melting temperatures (Tmo) were obtained by the Hoffman–Weeks extrapolation
method and the level of crystallinity was used to correct the DSC heat of fusion to
estimate the value corresponding to the 100% crystalline specimen (ΔHo). Entropy
values (ΔSo) are obtained by the ratio between ΔHo and Tmo. Values of ΔHo and
ΔSo are listed per mole of repeating unit and per bond. The level of crystallinity,
Tmo, and ΔHo per mole of repeating unit decrease substantially with increasing
halogen content, as expected for crystallites that, although layered, become more
defective with increasing content of halogen.

Crystal Thickness and Supermolecular Morphology

Crystallite thickness is an important parameter for correlation with the number of

layers in lamellar crystals and with other physical properties of precision systems.
Data on crystallite thickness and supermolecular morphology are available for the
rapidly crystallized Cl-containing series [73]. For this series, crystal thicknesses
were extracted from SAXS profiles and AFM images. Long spacing and crystal
thickness decrease with increasing halogen content, as shown in Fig. 8b, c. The data
for F-containing polyethylenes fall between the thicknesses of Cl and Br samples,
probably as a result of the low molar mass of the precision F samples analyzed
(Mw < 10,000) [71]. Focusing on the data for Cl-containing samples, a clear trend
with halogen size and content is found. At a fixed halogen content, the long spacing
and crystal thickness decrease substantially with increasing halogen size. For an all-
trans conformation, the C–C bond distance is 1.27 Å; hence, crystallites with Cl
substitution on every 21st carbon accommodate about eight repeating units in the
crystals, whereas the analog with Br substitution accommodates only about four.
Hence, although these systems crystallize as homopolymers by folding segments of
the chain including the halogen, all properties indicate that as the halogen becomes
larger it is more difficult to accommodate it inside the crystallites. As a conse-
quence, the level of crystallinity decreases and crystal thickness also decreases.
Crystallization of Precision Ethylene Copolymers 149

Fig. 8 Effect of halogen a) 100

content and type on (a) level
of crystallinity obtained
from WAXD, (b) long 80

Crystallinity (%)
spacing from SAXS, and (c)
crystal thickness from 60
analysis of the correlation
function from SAXS 40
patterns [71, 73] Linear PE
F
20
Cl
Br
0
0 2 4 6 8 10 12
Mol% Halogen
b) 300

250
Long Spacing (Ǻ)

200

150

100 Linear PE
F
50 Cl
Br
0
0 2 4 6 8 10 12
Mol% Halogen
c) 250
Linear PE
F
Crystal Thickness (Ǻ)

200 Cl
Br
150

100

50

0
0 2 4 6 8 10 12
Mol% Halogen

Within a given series, for example Cl-containing samples, the decrease in crystal
thickness from 185 Å for PE21Cl to 90 Å for PE9Cl corresponds to nine and eight
repeating units, respectively. A similar number of repeats suggests that, in spite of
the reduced level of crystallinity, when a higher content of halogen is
150 L. Santonja-Blasco et al.

accommodated as defects in the crystallites, the lamellae contains about the same
number of chain repeating units. All halogen precision systems with Mw > 10,000
display spherulitic morphology [73].

2.2 Isothermal Crystallization: Polymorphism

Although most work on the crystalline features of precision polyethylenes has been
done under uncontrolled crystallizations, often for samples cooled from the melt to
room temperature, it has been recently found that control of crystallization kinetics
allows different crystalline structures to be generated for a given precision chain.
Polyethylenes with precision halogen substitution have the ability to pack into
different polymorphs, depending on the undercooling. Schematics of two major
crystalline structures are given in Fig. 9 [64, 65]. At relatively low crystallization
temperatures, the lamellae stems pack in the all-trans conformation (form I)
whereas at relatively high crystallization temperatures, the methylene segment
between halogens zig–zags in a herringbone-like structure (form II). In this section
we review how this unique polymorphism can be identified spectroscopically or by
changes in X-ray patterns or melting behavior.
We start by reviewing X-ray patterns as a function of increasing crystallization
temperature. Figure 10 displays a composite of transmission X-ray patterns of
PE15Cl and PE21Br that were isothermally crystallized over a wide range of
temperatures (Tc). There is a clear change in crystallographic packing for Tc
below and above 53 C. At Tc < 53 C, the WAXD patterns display the two sharp
and strong reflections observed under rapid crystallization (form I). For Tc > ~53 C,
these precision polyethylenes pack in a different polymorph, form II, characterized

Form I, all-trans packing Form II, non-planar

Rapid crystallization herringbone-like packing
Slow crystallization

Fig. 9 Major polymorphs (form I and form II) of precision Cl- and Br-containing polyethylenes.
Reprinted with permission from Kaner et al. [64]. Copyright 2014 American Chemical Society
Crystallization of Precision Ethylene Copolymers 151

PE15Cl PE21Br
Tc (°C)=
Tc (ºC)=
56
65
55
Intensity (a.u.)

62
54
60

Intensity
53
58

52 54

52
45
47

23 40

2 7 12 17 22 27 32 37 42 2 7 12 17 22 27 32 37 42
2θ (deg) 2θ (deg)
Fig. 10 WAXD patterns collected at room temperature of PE15Cl and PE21Br isothermally
crystallized at the temperatures indicated. Notice the polymorphic change at Tc > 52 C

by additional reflections. The polymorphic change occurs within one degree of

undercooling, a feature that has not yet been reported for any other ADMET
precision polyethylene.
The transmission diffractograms also display strong reflections at intermediate
angles (2θ < 15 ), with periodic diffraction orders that are more relevant for form
II. These reflections correspond to spacings of ~15 Å (PE15Cl) and ~20 Å
(PE21Br), indicating that in both forms the halogens are arranged inside the crystals
in layers and that the chains are tilted at 35 (PE15Cl) or 41 (PE21Br) to the
normal of these layers [64, 65]; the larger the halogen, the greater the chain tilt.
Crystalline layered structures have been observed for precision polyethylenes with
carboxylic acids [78] and polyesters [60, 76], but the polymorphic behavior with
increasing crystallization temperature could be a feature related to intermolecular
kinks caused by small-volume pendant groups. The fact that PE15Cl and PE21Br
display a change from form I to form II at about the same crystallization temper-
ature is only fortuitous. The polymorphic transition for other precision Cl and Br
systems is expected at lower or higher crystallization temperatures, depending on
their crystallization kinetics; this is ongoing research in our laboratory. Poly-
morphism has been also observed in precision ethyl-branched polyethylenes [79]
and, although yet unreported, it is expected in methyl-branched precision systems
as well.
152 L. Santonja-Blasco et al.

A distinct characteristic distinguishing form I and form II crystallites is the

melting point. Form II melts at temperatures 10 C or 15 C higher than form I.
Hence, the polymorphic transition is easily identified by DSC experiments. Under
controlled isothermal crystallizations and subsequent melting, one notices a sharp
increase in melting temperature with increasing crystallization temperature. Exam-
ples for PE15Cl and PE21Br are given by the thermograms in Fig. 11, which were
obtained for the same isothermally crystallized samples used to record the WAXD
patterns shown in Fig. 10. Crystals formed below 52 C (form I) melt at about 61 C
whereas crystals formed at Tc > 53 C (form II) melt at about 10 C higher. The
change in crystallization temperature when crystal packing translates from form I to
form II is so sharp (within 1 C) that, in addition to large differences in free energy
of nucleation, the development of one or the other polymorph has been associated
with the thickness of the critical nucleus in the early stages of crystallization
[64]. When the estimated critical nucleus thickness is less than the length between
halogens, form I develops. If the thickness is equal to or greater than the length of
the methylene sequence between halogens, form II crystallites are formed. Indeed,
where the transition between polymorphs is experimentally observed, the esti-
mated thickness of the critical nucleus approaches the ethylene sequence length
between halogens.
The double melting at temperatures below the polymorphic transition, such as
below 52 C in the thermograms of PE21Br, are caused by melting of form I
followed by fast recrystallization in form II and further melting of the latter

Tc(°C) = Tc(°C) =
PE15Cl PE21Br
58
5W/g

65
2 W/g

57
56 62

55 60
Heat Flow (a.u.)
Heat Flow (a.u.)

54 58
53 56

52 54
52
45
50
38
47
23
40

20 30 40 50 60 70 80 90 100 20 30 40 50 60 70 80 90 100
Temperature (oC) Temperature (oC)

Fig. 11 DSC melting thermograms of PE15Cl and PE21Br isothermally crystallized at the
temperatures indicated. Heating rate was 10 C/min
Crystallization of Precision Ethylene Copolymers 153

[65]. This melting–recrystallization–melting sequence is absent in the thermograms

of precision polyethylenes substituted with the smaller halogen, and the double
melting of PE15Cl observed at Tc ¼ 53 C represents the simultaneous formation of
both forms. Hence, there is a narrow range of crystallization temperatures where
both polymorphs coexist. It has also been proven that both form I and form II are
stable [64]. For example, form I does not transform to form II when brought to a
temperature in the region of formation of form II and below the melting temperature
of form I. In other words, form I is not metastable as it needs to fully melt before
re-crystallizing into form II.
The structural differences between forms I and II can be extracted by combining
the information on layered structures that is obtained from X-ray patterns and FTIR
or Raman spectra of isothermally crystallized samples [64, 65]. Vibrational spectro-
scopy probes the conformation of backbone bonds adjacent to the halogen substi-
tution [80, 81]. This information is crucial in identifying intermolecular kinks along
crystalline halogen layers. Using model compounds, it has been shown that in the
FTIR stretching region the absorbance of the C–Cl bond depends on the conform-
ation of the adjacent C–C bonds. The absorbance at 612 cm
1 corresponds to C–Cl
stretching with vicinal C–C bonding in an all-trans conformation. The absorbance
at 665 cm
1 corresponds to C–Cl stretching when the side group is adjacent to
backbone carbons in the gauche conformation [80, 81].
As seen in the FTIR spectra of PE15Cl (Fig. 12, insets), there is a dramatic
increase in gauche bonding around the halogen substitution when crystallites
develop in form II. Indeed, it has been shown quantitatively that all carbons
adjacent to the substitution are gauche-bonded in the crystalline regions of form
II, whereas they are trans-bonded in the crystals of form I [64]. The conclusion
from this analysis is that the layer structure of form I keeps the all-trans conform-
ation, whereas the halogen layers of crystals in form II form a kink conformation
because the methine is gauche-bonded to the adjacent backbone carbons, or out of
the plane formed by the long methylene sequence if the conformation of the latter
remains all-trans (as shown in Fig. 9).
A simple direct way to prove the all-trans conformation of the methylene
segment between halogens in the crystals of form II is by analyzing the set of
progression methylene rocking bands of the FTIR spectral region between 700 and
1,100 cm
1. These bands can be analyzed in reference to progression modes of n-
alkanes of equivalent length that are known to pack in an all-trans conformation.
The n-alkane CH2 sequence is considered a linear array of m identical oscillators,
each having one degree of freedom [82, 83]. The oscillator model predicts a
frequency mode (K ) that is only a function of the difference in phase angle between
adjacent oscillators (ϕ(K )), calculated as ϕ(K ) ¼ Kπ/(m + 1), with K ¼ 1, 3, 5, . . .
Even K values are forbidden for the ideal model, and m is the number of continuous
CH2 units that adhere most closely to the ideal oscillator. The ends of the n-alkane
pendant groups and the backbone carbons adjacent to the methine of precision
polyolefins all break or affect the oscillator’s symmetry. Hence, in calculating ϕ(K )
for different K values, the length of the ideal oscillator is often shorter than the
length between pendant groups [84, 85]. For example, in the n-alkane C14H30,
154 L. Santonja-Blasco et al.

80
Tc=56°C

Absorbance
70 Absorbance Tc=53°C TT TG

TT TG
60
% of TG Conformers

500 600 700 800 900

50 Wavenumber (cm-1)
500 600 700 800 900
Wavenumber (cm-1)
40
Tc=45°C
Absorbance

30 TT

TG
20
500 600 700 800 900
10 Wavenumber (cm-1)

0
20 25 30 35 40 45 50 55 60
Temperature (°C)

Fig. 12 Percentage of gauche bonds adjacent to the CH for PE15Cl isothermally crystallized.
Notice the steep increase in gauche conformers at Tc > 52 C. Insets show FTIR spectra for
specimens crystallized at different isothermal temperatures

m ¼ 12 (14 backbone carbons minus 2 ends), and in precision PE15Cl, m ¼ 12

(14 methylene backbone carbons minus 2 adjacent to the methine).
When a progression of absorbance bands in the rocking region of the IR spectra
of precision polyethylenes follows the predicted frequency-phase angle dispersion
of the n-alkane of the same segment length, the CH2 sequences of both systems
must have the same all-trans conformation. Hence, proving that the methylene
sequences between pendant groups pack in an all-trans conformation can be easily
tested by evaluating the adherence of the frequency versus phase angle curve to the
n-alkane prediction. This simple model is extremely useful for testing periodic n-
alkane-like packing and has been successfully applied for evaluating the packing of
fatty acids [84–86] and aromatic polyesters with different methylene runs [87].
The oscillator model was also successfully applied to the spectra of PE15Cl and
PE21Br, giving evidence that the methylene segments between halogens pack in
all-trans conformation in forms I and II [64, 65]. An example of the progression
modes in the rocking-twisting region of FTIR spectra is given in Fig. 13a for forms I
and II of PE15Cl. Notice that the calculated ϕ(K ) for a sequence of 12 methylenes
closely follows the continuous line for the n-alkanes. The behaviors of form I and
form II are basically identical, thus demonstrating that the crystal packing of the
methylene run is all-trans in both (Fig. 13b). Because the bonding around the
substitution point is also trans for form I, but gauche for form II, the most plausible
structural models for crystalline packing that can be extracted from a detailed, yet
Crystallization of Precision Ethylene Copolymers 155

(a)
K=1-3

K=5 K=7
K=9 K = 11
Form II

K=5
K=7
K=6 K=9 K = 11
K=8 K = 10 K = 12
Form I

500 600 700 800 900 1000

Wavenumber (cm-1)
(b) 1100
n-alkane
PE15Cl, Tc = 51°C
PE15Cl, Tc = 56°C
Wavenumber (cm-1)

1000

900

800

700
0 20 40 60 80 100 120 140 160 180
f k (degrees)
Fig. 13 (a) FTIR spectra of PE15Cl for crystals in forms I and II. Methylene progression bands
are indicated for K ¼ 1–12. (b) Dispersion curve for n-alkanes (continuous line) and for PE15Cl
(symbols) calculated for form I (Tc ¼ 51 C) and form II (Tc ¼ 56 C, demonstrating that methylene
CH2 sequences pack in the all-trans conformation. Reprinted with permission from Kaner
et al. [64]. Copyright 2014 American Chemical Society

simple, analysis of X-ray and FTIR are those of Fig. 9. Subsequent crystallographic
analysis of fiber patterns for forms I and II of PE21Br corroborated the planar and
nonplanar structures of the two major polymorphs of these precision systems
[65]. The triclinic unit cells and lattice parameters are shown in Fig. 14. Whereas
the backbone conformation of form I is all-trans, the conformation of form II
156 L. Santonja-Blasco et al.

Form I Form II

b’

a’
a”

Crystal form a/Å b/Å c/Å α/deg β/deg γ/deg

Form I 4.90 5.75 52.7 43.2 109.8 107.9
Form II 5.00 5.65 47 77.5 112 65
Fig. 14 Crystal unit cell structures extracted from WAXD fiber patterns of PE21Br. Two Br
atoms are placed on a methine carbon at a 0.5 statistical probability to account for the atactic nature
of the chain [65]. The triclinic lattice parameters for forms I and II are listed

follows the pattern . . .TTTTTGGTTTT. . .. . .TTTTTTG0 G0 TT. . .. . ., thus perpetu-

ating lamellar symmetry.
The unique feature that emerges from the currently available studies is that
precision halogen-containing polyethylenes pack in at least two different crystalline
structures, depending on the undercooling. At large undercooling, form I develops
with molecules packing in all-trans zig–zag conformation and a layered, albeit a
little disordered, intermolecular staggering of chlorines. At relatively low under-
cooling, the staggering of chlorines or bromines is highly symmetric, leading to
intermolecular kinks or a herringbone-like conformation where the long methylene
runs zig–zag around the substitution point (Fig. 14). This polymorphism is
Crystallization of Precision Ethylene Copolymers 157

obviously related to the size of the halogen and the regularity of the substitution
along the backbone. As the halogen inside the crystal must be in layers, the larger
the size the stronger is the intermolecular halogen compression. Fluorine is too
small to cause any intermolecular effect, whereas chlorine, bromine, and larger
halogens have dramatic effects on methylene packing.
One issue that remains is the major driving force for the observed layered
crystalline structures. Three possibilities are at play: (a) preferential folding at a
given location in the backbone as earlier speculated, and perhaps directed by crystal
size; (b) intermolecular CH2 van der Waals interactions; and (c) intermolecular
halogen–halogen interactions. Interactions between functional groups have been
posed to dominate over the other two possibilities in packing aliphatic long
methylene polyesters [76], polyethylenes with equidistant sulfite groups [67], and
polyethylenes with other strongly interacting moieties [13, 66]. However, the
strength of halogen–halogen interactions over the enthalpic drive to assemble inter-
molecular methylene segments in precision halogen-containing specimens is diffi-
cult to quantify. Some works point out ineffective F–F interactions and effective
halogen–halogen interactions for Cl- and Br-containing molecules [88]. Moreover,
methylene all-trans packing may dominate in these systems as the layer symmetry
increases at higher temperatures where form II is enabled.
Controlled crystallization of precision polyethylenes with methyl branches can
help to discern whether halogen–halogen interactions are the major driving force
for formation of layered crystallites in these systems. The size of a methyl group is
similar to that of a Br atom, but the interactions between CH3 pendant groups are
significantly weaker than those between Br atoms. Unfortunately, systematic stud-
ies of the melting behavior and crystalline structure of isothermally crystallized
samples are not yet available for precision polyethylenes with methyl branches
[75, 89–92]. Although deviations from the regular all-trans geometry of the back-
bone carbons near the methyl branch have been already identified in the crystalline
regions of deuterated precision systems [92], determination of whether crystallites
of precision methyl-branched polyethylenes are also layered requires more detailed
thermal, X-ray, and spectroscopic analysis of isothermally crystallized samples. A
conclusion from the in-depth crystallization studies of PE15Cl is that understanding
all forms of crystalline packing can help clarify the multiple melting that is often
observed in many of these precision systems.

3 Crystallization of Precision Alkyl-Branched

Polyethylenes

Because of their relevance for the polyethylene industry, considerable effort has
been made by Wagener and coworkers to synthesize precision polyethylenes with
alkyl branches. Thermal data are available for series with methyl [75, 89–91, 93,
94] ethyl [79, 93–96], propyl [93, 94], butyl [93, 94, 97, 98], pentyl [93, 94], hexyl
158 L. Santonja-Blasco et al.

[63, 93, 94, 99], and up to pentadecyl [94, 100] branches, and with distances
between branches of 7–39 carbons. A polyethylene with butyl branches spaced by
75 carbons was also synthesized [101]. The synthesis of these systems and their
thermal and crystalline properties have been summarized in different reviews
[5, 102–105]; therefore, only major highlights of their crystallization behavior are
covered in this section.
The behavior of this type of precision polyethylene with respect to partitioning
of the alkyl branch between crystalline and non-crystalline regions has been
deduced from the comparative melting behavior of crystals formed under the
usual dynamic cooling from the melt. A compilation of literature data for melting
temperatures (Tm) as a function of increasing branching is given in Fig. 15. Data on
rapidly crystallized (form I) precision halogenated polyethylenes are also included
as a reference, representing systems with no discrimination for partitioning of the
pendant group. For alkyl-branched systems that display multiple melting peaks, the
lowest Tm is plotted. The only exception is for data on ethyl-branched samples, for
which low Tm for quenched and high Tm for isothermally crystallized specimens are
both included.
The linear variation of Tm with decreasing length between pendant groups can be
extrapolated to the observed melting point of the unbranched chain; the gradient of

160
Linear PE
F
120 Cl
Br
Methyl
80 Ethyl
Tm (°C)

Ethyl (low Tm)

Propyl
40 Butyl
Pentyl
Hexyl
0 Heptyl
Octyl
Nonyl
-40 Decyl
Pentadecyl
Isopropyl
-80
0 4 8 12 16 20
Mol% branches / Backbone C in repeating unit

Fig. 15 Melting temperature versus. branching composition of precision halogen-substituted

polyethylenes and precision alkyl-branched polyethylenes crystallized from the melt at 10 C/
min. Included are data of Fig. 6 for precision halogen-substituted polyethylenes and data for
precision samples with the following branches: methyl [75, 91, 94], ethyl [94, 95], propyl [93, 94]
butyl [93, 98, 101], pentyl [93, 94], hexyl [94, 99], heptyl [94], octyl [94], nonyl [94], decyl [94],
pentadecyl [94], and isopropyl [93, 94]. Data for slowly crystallized ethyl-branched samples are
also added (red triangles)
Crystallization of Precision Ethylene Copolymers 159

each line is proportional to the size of the pendant group. The trends in Fig. 15 are
clear: At a fixed branching content, systems with propyl and longer branches melt at
about the same temperature and systems with ethyl and methyl branches melt at
progressively higher values. The size and content of the pendant group nicely
correlates with the types of crystals formed.
If the pendant group is rejected to the intercrystalline regions, as is likely for
propyl and longer branches, the crystalline structure follows model (i) shown in
Fig. 1. For a fixed length of the methylene sequence, the lamellae thickness is
constant and the branch is preferentially placed at the surface of the crystallites.
Hence, the melting temperature is simply a function of the basal surface free
energy, following the Gibbs–Thomson relation. Under the assumption of lamellar
crystallites with negligible lateral surfaces and large basal surfaces, this relation
takes the form:
 
2σ e
Tm ¼ T mo 1
ð3Þ
lΔH u

Here, Tmo is the melting temperature of the infinite thick crystallite, Tm the
observed melting temperature, σe the basal surface free energy, l the lamellae
thickness, and ΔHu the heat of fusion per unit volume of pure crystal. Because
the basal surface free energy is presumably very similar for alkyl branches that
remain uncrystallized, following Eq. (3), the same Tm is expected for crystals of the
same thickness. Indeed, with Tmo ¼ 145.5 C, ΔHu ¼ 2.8  109 erg/cm3, a C–C
distance of 1.27 Å for all-trans packing, and σ e ¼ 110 erg/cm2 [24], the calculated
Tm values are very close to the observed melting temperatures (4 C) for systems
with propyl and longer branches spaced at a distance 15 methylenes. Hence, the
data for propyl and longer branches follow the predictions for model (i) in Fig. 1.
The Tm data for systems with bulkier, non-interacting groups such as cyclohexyl
and adamantyl fall below the line, suggesting a more strained basal surface or a
higher value of σ e [93].
Works on ethyl-branched precision systems are among the very few where
properties for quenched and isothermally crystallized specimens are reported
[79]. Low melting thin crystallites free of branches are formed at the lowest
crystallization temperatures or on quenching; their melting behavior is similar to
the behavior of precision systems with branches excluded from the crystallites (see
plot representing propyl and longer branches in Fig. 15). Conversely, crystallites of
ethyl-branched systems that include at least one branch are formed at higher
isothermal crystallization temperatures; consequently, they melt at higher values
(see red triangles in Fig. 15). Even under rapid cooling, ethyl-branched precision
systems spaced every 15th or every 21st carbon develop two different crystalline
structures (as seen in Fig. 16 by their dual melting behavior) [94, 95]. Therefore, on
dynamic cooling, the development of one or two populations of crystallites with
different melting peaks depends on the kinetics of the formation of each polymorph.
The ethyl-branched system spaced every 8th carbon was amorphous, an indication
160 L. Santonja-Blasco et al.

PE15Eth
Heat Flow

PE21Eth

PE39Eth

-100 -80 -60 -40 -20 0 20 40 60 80 100 120

T(°C)

Fig. 16 DSC melting endotherms for precision polyethylenes with ethyl branches placed on every
15th, 21st or 39th backbone carbons. Thermograms of PE15Eth and PE21Eth extracted from [95],
thermogram for PE39Eth from [94]

of incommensurate steric effects in accommodating the ethyl branch in the crystal-

lites at this branching level.
The work of Lieser et al. on precision methyl-branched polyethylenes with the
branch placed on every 15th or every 21st backbone carbon was a pioneer contri-
bution to an understanding of the crystallization of precision polyethylenes [90].
From X-ray powder diffraction patterns of melt-crystallized specimens and electron
diffraction of solution-crystallized samples, the authors concluded that these poly-
mers form superlattices involving two full repeating units in a triclinic symmetry.
The lamellae of the crystallites are 100–200 Å thick, in other words, at least five
times the length of the methylene sequence. Such thick crystallites provide the first
evidence that long chain segments including multiple repeating units fold back and
forth in the crystallites, hence including the branches.
The thermal properties of the series of precision methyl-branched systems indi-
cate that all these atactic models are semicrystalline, even at very high branch
content. Only polyethylenes with a methyl branch on each 5th carbon showed no
evidence of crystallization, even after annealing [91]. At this short distance, the
atactic nature of the substitution obviously prevents the formation of any stable or
metastable lattice. Crystal entities do not form because the excess free energy from
the methyl defects in the crystal overcomes the free energy of formation of any
crystallite from this highly branched structure. The plots of Tm versus composition
for the methyl-branched series (represented by diamonds in Fig. 15) fall closely on
Crystallization of Precision Ethylene Copolymers 161

the line of precision Br-substituted samples. This feature emphasizes a similar

crystallization path, and that the size of the substituent is a major drive directing
the crystallization and melting behavior of precision systems. As the van der Waals
radius of the methyl and bromine are the most similar, their distortion effects on the
crystal lattice, and thus Tm, are also very similar.
The melting temperature of the polyethylene with methyl branches on each and
every 7th carbon is
60 C, but increases by ~30 C on annealing, a feature also
consistent with the formation of different polymorphic structures [91]. It is likely
that a highly defective metastable structure that melts at
60 C develops on
dynamic cooling, whereas annealing or isothermal crystallization favors better
interchain staggering of the methylene sequences, thus increasing the symmetry
of the crystal packing and hence the melting point, similar to halogen behavior.
Other members of the methyl-branched series display single melting peaks on rapid
cooling [91]. However, it is quite likely that they would also develop polymorphism
and different levels of layered crystalline structures during controlled isothermal
crystallization. This issue has yet to be explored.
In summary, the partitioning of linear alkyl branches between crystalline and
non-crystalline regions correlates well with the melting behavior of crystals formed
under rapid cooling, and scales with the size of the pendant group. Different
polymorphic structures are predicted for systems with methyl and ethyl branches
that can be incorporated in the crystals. These predictions are based on differences
in the melting behavior of crystals formed under fast and slower crystallizations. A
compendium of crystalline unit cells that have been suggested in the literature for
precision alkyl-branched systems are listed in Table A1 of the Appendix.

4 Crystallization of Precision Polyethylenes

with Interacting Functional Groups

Interacting pendant groups in precision polyethylenes bring additional compli-

cations to the study of their crystallization behavior. The strength of the interaction
is often a function of the melt temperature. This feature translates to crystallizations
that depend on the structure or entropy of the initial melt state. However, most
available crystallization and melting data are for precision systems that were cooled
from the same melt temperature. Polymers synthesized with functional groups
directly attached to the methylene backbone (such as long spaced aliphatic or
aromatic polyesters) or with pendant functional groups placed at an exact distance
along a polyethylene-like backbone are too numerous to allow elaboration on
suggested crystalline structures and melting behavior of each type in this section.
Because of space limitations, the thermal properties of many of these precision
polyethylene-like systems are compiled in Table A2 of the Appendix. In this
section, we focus on three aspects that drastically affect the packing and melting
behavior of these systems, namely, type and strength of interactions, placement
162 L. Santonja-Blasco et al.

within or branching from the backbone, and configuration with respect to the
backbone of the functional punctuation placed at a precise distance along the
polyethylene backbone.

4.1 Strength of Pendant Group Interactions

In general, as the strength of the pendant group interaction increases, the melting
temperature also increases. This interaction can be ionic [106–112]; via hydrogen
bonding between carboxyl [61, 66, 113–115], hydroxyl [116–119], or amine
[120, 121] pendant groups; or interactions induced by bond polarity in groups
such as ketone [13, 68], ether [122, 123], acetoxy [9, 124], or acrylate [125]. Lay-
ered crystalline structures often develop in precision systems with strongly
interacting functional groups. Extensive work by Winey and coworkers on acid-
containing, and ionomer-like precision systems [106–112] have revealed that
interactions between functional or ionic groups exist, even at high melt temper-
atures, and play a role in self-assembly, which ultimately affects their uniaxial
tensile deformation behavior [115]. Moreover, crystallization from a dilute solution
of precision polyethylenes with pendant carboxylic acids leads to crystallites with
thicknesses corresponding to the length of the methylene sequence [61]. Hence,
preferential folding and structures of the type shown in Fig. 1b (model i) may
prevail in crystallizations from dilute solution, leaving the carboxylic acid groups
accommodated preferentially at the crystal surface.
Examples of increasing interactions between small-sized pendant groups and the
effect on melting behavior are given in Fig. 17 for precision systems with fluorine
[68, 71], ketone [68, 126], and hydroxyl [116] groups. As the pendant moiety is
easily accommodated in the crystal lattice, it is expected that all these systems
crystallize as homopolymers. Perturbation of the crystal lattice by the small pendant
group is relatively minor; hence, all crystalline structures developed by these long-
spaced systems are isomorphous to the orthorhombic polyethylene crystals [68, 71,
116, 118].
The F–F interaction is the weakest of the groups studied in Fig. 17 and, as such,
precision F-substituted polyethylenes display the lowest melting temperatures.
Dipole–dipole interactions increase in ketone-decorated systems, favoring
interchain interactions and resulting in ~10 C increase in the melting point of
ketone-precision systems compared with the fluorine analogs. The highest melting
points belong to the systems with hydroxyl pendant groups, which are prone to
strong hydrogen bonding [116]. The decrease in melting temperatures with increas-
ing alcohol content again makes it relevant that the –OH is primarily a defect in
crystallites that maintain the orthorhombic symmetry up to 20 mol% hydroxyls
(
OH on every 5th backbone carbon). Decreasing the space beyond five methy-
lenes leads to a change in crystallographic packing to one that approaches the
monoclinic packing of polyvinyl alcohol, and explains the observed increase in
melting temperatures for –OH contents >20 mol% (see Fig. 17). Precision random
Crystallization of Precision Ethylene Copolymers 163

Fig. 17 Effect of strength 160

of interaction between Polyethylene
pendant groups on the =O
melting temperature of F
polyethylenes with ketone 140
OH
(¼O) [68, 126], hydroxyl
(
OH) [116], or fluorine

Tm (°C)
[68, 71] groups placed at the
same precise distance along 120
the backbone

100

80
0 5 10 15 20 25 30
Mol% Group / Backbone C in repeating unit

polyamides show a similar behavior, with a minimum in melting points at much

lower content of amide groups [12]. The melting data shown in Fig. 17 for hydroxyl
pendant groups were obtained for samples synthesized via ROMP [116]. Melting
data for the same hydroxyl-branched structures synthesized via ADMET are lower
and basically invariant for systems spaced with 8–20 methylenes (Tm ~125 C)
[118]. This melting difference is probably a result of differences in tacticity
between ROMP- and ADMET-derived specimens, which may affect the strength
of hydrogen bonding.

4.2 Functional Groups Within or Branching from

the Methylene Backbone

The difference in crystalline packing and melting behavior of a methylene back-

bone precisely decorated with functional punctuations that are either directly linked
to the backbone or branching from the backbone is quite drastic, as shown by the
data compiled in Fig. 18. The melting temperatures of precise long-spaced aliphatic
polyesters [126–128], aliphatic polycarbonates [11], and aliphatic polyacetals [11]
are compared with analogs containing pendant acetoxy [124], and pendant methyl
or ethyl acrylates [125]. The melting points of the first type are displaced at much
higher values. For the first group, the melting temperatures of polyesters are the
highest, followed by polycarbonates and then polyacetals, denoting differences in
size of the functional group and in polarity. Polyethylenes with ketone groups (not
included in Fig. 18) are expected to melt at even higher temperatures than poly-
esters as they have stronger dipolar interactions, and the available data for two
precise ketone-containing polyethylenes corroborate their higher melting points
[13, 68].
164 L. Santonja-Blasco et al.

140
(CH2)x
R n
120 R: O

Polyesters, O C O
C
Polycarbonates, O O

100 Polyacetals, O O

80
Tm (°C)

60
(CH2)x
CH n

40 R' : R' O
C
O CH3
Acetoxy, O
C CH3

20 Methyl Acrylate, O
O CH3
C
Ethyl Acrylate, O

0
0 1 2 3 4 5 6 7
R Groups/ 100 CH2 Units

Fig. 18 Effect of placement, as a branch (R0 ) or inserted in backbone (R), on the melting
temperature of precision polyethylene-like polymers with increasing content of functional groups.
Data shown are for long-spaced polyesters [126, 127], polycarbonates [11], polyacetals [11], and
precision polyethylenes with acetoxy [124], methyl acrylate [125], and ethyl acrylate [125]
pendant groups

As seen in Fig. 18, the polyester spaced every 18 methylenes melts at 97 C,
whereas the analog with an acetoxy pendant group melts at 23 C. This is a drastic
depression of the melting point that cannot be accounted for simply by polar
interactions. The effect of the type of linkage on chain packing is evident. In
long-spaced polyesters, polycarbonates, or polyacetals, the functional group is
part of the polyethylene backbone; hence, it is an integral part of the crystal lattice,
with polar groups of different chains interdigitated in a close packing arrangement
[126, 127]. A crystalline layered symmetry is easily set in this precision type, as
shown recently in a precision polysulfite system [67]. Conversely, intermolecular
packing of a chain with acetoxy or acrylate pendant groups is driven to a large
extent by van der Waals interactions between the methylene runs of different
chains, with the pendant groups causing large distortions to the all-trans methylene
packing, or being rejected to the crystal–amorphous interfacial region. Crystallites
from precision chains with pendant side groups either have more defects or are
thinner than those of the first group; hence, they melt at lower temperatures.
Wagener et al. have published X-ray patterns of polyethylenes with pendant
acetoxy groups on every 23rd carbon that display low angle reflections corres-
ponding to the distance between acetoxy groups, suggesting that these groups are
also layered and are part of the crystalline symmetry [124]. Hence, layered crystal-
lites are feasible for systems with pendant-interacting functional groups. This is
inferred by the high melting points of these systems compared with the melting
Crystallization of Precision Ethylene Copolymers 165

temperatures of precision systems with large uncrystallizable pendant groups

[63, 124, 125] and by the double melting observed in some of the former type,
suggesting polymorphism [124].

4.3 Effect of Tacticity

Precision structures generated by ADMET allow unprecedented control over the

generation of models of polyethylene-like systems for understanding the effect of
defects of different types, sizes, and precise placement along the backbone on chain
folding and self-assembly of macromolecules. However, to date, it has not been
possible to control the tacticity of ADMET-generated model precision polyethyl-
enes. The question still remains as to whether the crystalline packing and properties
of the same structures with control of tacticity will remain the same or change
radically. If a higher order of symmetry is found in crystallites from isotactic or
syndiotactic precision systems, their melting temperature may also be dramatically
higher. Sporadic data for systems with control over both the sequence and stereo-
chemistry appear to point in this direction [7–9].
It is well known that tacticity and regiochemistry can have a dramatic effect
on the crystallization, thermal, and mechanical properties of polyolefins. To probe
this effect, regio- and stereoselective systems with ethylene backbones and
di-hydroxyl, alkyl, phenyl, and acetoxy pendant groups have been synthesized via
ROMP by the groups of Grubbs and Hillmyer [7, 9, 129]. Regioregular atactic
polyethylene with a methyl branch on every 8th backbone carbon is crystalline with
a melting point of 1.7 C, which is about 10 C higher than the atactic ADMET
system with a methyl branch on each 9th carbon [91]. Regioregular ROMP systems
with an ethyl, hexyl, or phenyl group on each 8th backbone carbon are amorphous
under standard DSC cooling at 10 C/min, but annealing procedures were not
attempted [7].
Both the atactic and isotactic regioregular systems with acetoxy pendant groups
on every 8th carbon are crystalline with melting temperatures of 53 C and 91 C,
respectively. These values are also higher than predicted melting points of ADMET
systems spaced at a similar distance [124]. Interestingly, the atactic but
regioselective system with acetoxy groups on the 7th carbon was amorphous.
This drastic loss of crystallinity caused by reducing the methylene length from
seven to six carbons was explained as the need for a minimum sequence length of
seven methylenes to generate crystallites [9]. Moreover, the loss of crystallinity
could also be related to a dramatic effect on melting of an odd versus even backbone
placement of the acetoxy group, similar to the odd versus even end-group effect in
n-alkane packing [130]. The atactic ADMET acetoxy system spaced by 18 methy-
lenes (5.3 mol%) melts at 23 C [124] and, following Fig. 18, a much lower melting
temperature was expected for the system with a higher content of branches. The
atactic acetoxy-branched system spaced by six methylenes (14.3 mol%) follows
this expectation, but the system spaced by seven methylenes (12.5 mol%) does not.
166 L. Santonja-Blasco et al.

Table 2 Melting temperatures and heats of fusion for precise ADMET and regioregular and
stereoselective ROMP systems
Number of CH2 groups
Pendant Synthetic between pendant ΔHm
group Tacticity route groups Tm ( C) (J/g) Reference
CH3 Atactic ADMET 8
13 39 [91]
CH3 Atactic ROMP 7 1.7 30 [7]
Acetoxy Atactic ADMET 18 23 NAa [124]
Acetoxy Atactic ROMP 6 Amorphous
Acetoxy Atactic ROMP 7 53 39 [9]
Acetoxy Isotactic ROMP 7 91 67 [9]
OH Atactic ADMET 8 120 NAa [118]
OH Atactic ROMP 7 128 67 [116]
OH Atactic ROMP 4 102 15.3 [116]
di-OH (syn-1,2 ROMP 6 111, 119 NAa [129]
diol)
di-OH (anti-1,2 ROMP 6 157 NAa [129]
diol)
a
Not available

It is feasible that, as the methylene spacer becomes shorter, the dipolar interactions
between the acetoxy groups prevail over the van der Waals CH2 interactions.
Clearly, more research on regio- and tacticity-controlled samples is needed to eluci-
date these features. Melting points and heats of fusion of similar structures synthe-
sized by ADMET and regio- and stereoselective ROMP are listed comparatively in
Table 2.
From the data in Table 2 one can see that atactic and isotactic acetoxy systems
melt at much higher temperatures than the precision analog with methyl sub-
stituents, in spite of the larger group size [9, 91, 124]. Higher melting points and
higher heats of fusion accentuate the dipolar nature of the acetoxy group. Further-
more, the precision isotactic acetoxy system shows double melting, thus suggesting
that the isotactic systems also develop different polymorphic structures [9]. A
dramatic difference in melting points was also found between the syn- and anti-
diols for stereoregular precision systems with hydroxyl groups on every 7th and 8th
backbone carbon [129], again suggesting that the relative stereochemistry has a
remarkable effect on packing of the side groups and, hence, on the properties of the
material.

5 Concluding Remarks

Precision polyethylenes with co-units placed at a periodic equal distance along the
backbone are excellent models for analyzing the effect of co-unit size, type, and
content or the distance between co-units on their crystallization and melting
Crystallization of Precision Ethylene Copolymers 167

behaviors. These are novel polyolefin-like systems with crystallization mechanisms

and crystalline properties that diverge from the most common random ethylene
copolymer systems. In this review, we focus on the crystalline properties of the
precision polyethylenes that are available via ADMET or ROMP synthetic paths.
Data on a series of precision halogen-substituted polyethylenes serve to establish
the effect of size and content of a relatively small pendant group on crystallization.
Even with the atactic configuration, if the co-unit is relatively small, all precision
systems crystallize as homopolymers. Crystalline structure and melting are affected
by the level of strain that the co-unit asserts on the crystal lattice. The crystallization
behavior under rapid and slow crystallization is contrasted for Cl- and
Br-containing systems, as the precise nature of the substitution enables unprece-
dented herringbone-like layered crystalline structures controlled by changing the
undercooling. Planar and nonplanar polymorphic structures can be identified by
X-ray diffraction, vibrational spectroscopy, and thermal analysis of isothermally
crystallized specimens.
This review also analyzes the melting behavior of alkyl-branched precision
polyethylenes, specifically branch partitioning between crystalline and
non-crystalline regions and the behavior of analog polyethylenes containing halo-
gens. Precision systems with propyl branches and longer, which are rejected from
the crystal, display the same melting temperature-branching composition relation
and their crystal thicknesses are close to the all-trans methylene sequence length.
Conversely, smaller ethyl and methyl branches are accommodated in the crystal-
lites and thus melt at higher temperatures. Different polymorphs are also enabled in
the latter by changing the undercooling.
The effects of interacting functional groups and tacticity on self-assembly and
melting are analyzed using available literature data. Strongly interacting functional
groups lead to layered crystalline structures and high melting temperatures, as seen
for polyketones, polyesters, and polysulfites. It is also demonstrated that, for the
same methylene spacer, the melting temperatures of polyethylenes with functional
groups inserted in the backbone are much higher than if the same group is added as
a side branch. Similarly, stereoregular precision systems enable formation of
crystallites with more symmetry, as they melt at higher temperatures than their
atactic counterparts.
In summary, the potential for precision polyethylene-like copolymers to gener-
ate crystalline structures not accessible by the present state-of-the-art linear
low-density polyethylenes is now becoming apparent. However, control of crystal-
lization kinetics is of paramount importance in developing multiple polymorphic
structures driven by different modes of staggering co-units in the crystallites.

Acknowledgements This material is based upon work supported by the National Science Foun-
dation under grant no. DMR1105129. Any opinions, findings, conclusions, or recommendations
expressed in this material are those of the author(s) and do not necessarily reflect the views of the
National Science Foundation. We remain grateful to Prof. Wagener and E. Boz who kindly gave us
the precision halogenated systems for study of their crystallization behavior. LSB acknowledges a
postdoctoral fellowship APOSTD/2013/036 supported by the Generalitat Valenciana and the
Universitat Politècnica de Valencia, Spain.
168 L. Santonja-Blasco et al.

Appendix

Table A1 Suggested crystallographic packing for alkyl-branched precision ADMET

polyethylenes
Number of Inclusion
CH2 groups of the
between branch in
Branch pendant the Tm ΔHm Suggested crystalline
type groups crystal ( C) (J/g) packing References
Methyl 14 Yes 39 82 Triclinic unit with hex- Lieser
agonal sub lattice et al. [90]
Methyl 20 Yes 62 103 Triclinic unit (18.75 and Lieser
21.75 )with hexagonal et al. [90]
sub lattice
Monoclinic structure Qiu
(19.1 and 22.1 ) et al. [131]
Methyl 38 Yes 92 137 Expanded orthorhombic Inci
unit cell (21.1 and 23 ) et al. [94]
Ethyl 20 Yes 15/ 58.5 Triclinic (18 and 21 ) Rojas
34 et al. [93]
Triclinic (20.1 and Hosoda
22.5 ) Sample crystal- et al. [63]
lized at 10 C for 4 days
Ethyl 38 No 76 93 Orthorhombic coexisting Inci
with a metastable mono- et al. [94]
clinic structure (~19.6 ,
21.4 and 23.4 )
Propyl 20 No 12 60 Triclinic (broader and Rojas
asymmetric peaks et al. [93]
suggesting more than
one crystalline lattice)
(~19.5 and 22.5 )
Propyl 38 No 78 71 Orthorhombic coexisting Inci
with a metastable mono- et al. [94]
clinic structure (~19.6 ,
21.4 and 23.4 )
Butyl 20 No 12 57 Triclinic (broader and Rojas
asymmetric peaks et al. [93]
suggesting more than
one crystalline lattice)
(~19.5 and 22.5 )
Butyl 38 No 75 66 Orthorhombic coexisting Zuluaga
with a metastable mono- et al. [97],
clinic structure (~19.7 , Inci
21.4 and 23.5 ) et al. [94]
(continued)
Crystallization of Precision Ethylene Copolymers 169

Table A1 (continued)
Number of Inclusion
CH2 groups of the
between branch in
Branch pendant the Tm ΔHm Suggested crystalline
type groups crystal ( C) (J/g) packing References
Pentyl 20 No 14 58 Triclinic (broader and Rojas
asymmetric peaks et al. [93]
suggesting more than
one crystalline lattice)
(~19.5 and 22.5 )
Pentyl 38 No 74 88 Orthorhombic coexisting Inci
with a metastable mono- et al. [94]
clinic structure (~19.6 ,
21.4 and 23.4 )
Hexyl 20 No 16 53 Monoclinic (~19.6 and Hosoda
23.1 ) Sample crystal- et al. [63]
lized at 0 C for 4 days
Hexyl 38 No 73 85 Orthorhombic coexisting Inci
with a metastable mono- et al. [94]
clinic structure (19.6 ,
21.4 and 23.4 )
Gem- 20 No 45 61 Reflections of hexago- Qiu
dimethyl nal, monoclinic and tri- et al. [131],
clinic packing, pointing Rojas
towards polymorphism et al. [103]
Isopropyl 38 No 77 74 Orthorhombic coexisting Inci
with a metastable mono- et al. [94]
clinic structure (~19.6 ,
21.4 and 23.4 )
Sec-butyl 20 No 9 43 Triclinic (broader and Rojas
asymmetric peaks et al. [93]
suggesting more than
one crystalline lattice)
(~19.5 and 22.5 )
Iso-butyl 38 No 73 51 Orthorhombic coexisting Inci
with a metastable mono- et al. [94]
clinic structure (~19.6 ,
21.4 and 23.4 )
Table A2 Thermal properties of precision polyethylene-like polymers with functional groups
170

CH2
sequence Tc ΔHm Tg
Group Structure (x) Tm ( C) ( C) (J/g) ( C) References
O
Ketone 18 134 127 Watson and Wagener [126]
x n
134.7 120 106 Boz et al. [68]
o
Aliphatic ester 14 95.9 78.2 125.6 Meulen et al. [127]
o
x n 15 92.4 76.6 145.5
18a 97 122 Watson and Wagener [126]
Diester O O 22 85–88 Hove et al. [60]
n
X O O
44 91–93,
100–102
Diester with propyl branches O O 22 46–48
n
X O O
44 87–92
O
Aromatic ester 9 92.8 Tasaki et al. [87]
O O
X n 10 120.3,
125.5
11 89.8, 96.5
12 119.1,
122.7
15 99.7,
101.9
16 115.2
20 117.1
NH 2
Amine 8 42b 10 Leonard et al. [120]
n 14 49
x
20 44
L. Santonja-Blasco et al.
 OH
Hydroxyl 8 120 Thompson et al. [118]
n 14 128 86
x
20 125 52 Valenti and Wagener [117]
ADMET
OH
3 113.6 36 48.5 Ramakrishnan [116]
n 4 102.6 15 45.4
x
6 125.2 59 37.8
ROMP 7 128 67 35.1
11 135 92 31.6
Acetal O O n 18 76, 82 64 147 Ortmann et al. [11]
x
19 77, 83 67 158
23 87 72 167
Carbonate O O 18 89 67 116
n
x
O 19 89 61 143
23 97 77 156
O OH
Crystallization of Precision Ethylene Copolymers

Carboxyl 8 22 Baughman et al. [78]

n
14
4
x
20 45 42
20 49 33 52 Ortmann et al. [61]
44 79, 99 85 100
Gem-dimethyl 8
47 Schwendeman et al. [132]
n
x 14 32
39 40
42
20 45 8.5 61
22
OMe
Methoxy 10
40,
30
58 35
62 Baughman et al [123]
n 14
10
22 62
x
20 40 22 78
20 40 23 80 Baughman et al. [122]
OEt
Ethoxy 10
4
35 33
65 Baughman et al [123]
n 14
33
40 36
x
20 28 10 79
171

(continued)
Table A2 (continued)
CH2
172

sequence Tc ΔHm Tg
Group Structure (x) Tm ( C) ( C) (J/g) ( C) References
O
Acetoxy 18 23 Watson and Wagener [124]
O
20 35
n
x 22 41
26 57
ADMET atactic
O
7 53 9 39 -36 Zhang et al. [9]
O
8 -46
n
x

ROMP atactic
O
7 91 59 67 -30
O

n
x

ROMP isotactic
O OMe
Methly acrylate 18 14.4 Watson and Wagener [125]
n
22 37
x
O OEt
Ethyl acrylate 18 9.8, 15.1
n
x
O
Methacrylic OH 20 13 4.5 21 Schwendeman et al. [133]
n
x
O
Methacrylate OMe 20 7
1.3 49 Schwendeman et al. [133]
n
x
O
Tert-butyl oxycabonyl (Boc) amine 8 4 Leonard et al. [120]
HN O
14 2
x n 18
8
20 2
L. Santonja-Blasco et al.
Amphiphilic Tetra(ethylene glycol) O 8
65 Berda et al. [134]
4 OH
14
3
34 19
63
x n
20 29 20 36
63
Amiphiphilic branches: Tetra(eth- R ¼ Pyrene group 20 9
12 25
21 Berda et al. [135]
ylene glycol) with different hydro-
O
4 OR
phobic segments
x n

R ¼ n-hexyl group 20 4, 11,
4, 0 28

37,
48

R ¼ n-tetradecyl 20 23 17 72

SO 3Et
Sulfonic acid ethyl ester 20 29 24 34 Opper et al. [136]
x n
SO 3Et
Aromatic sulfonic ethyl acid ester 20 24 Opper et al. [136]
Crystallization of Precision Ethylene Copolymers

x n
OH
Phosphonic acid O 8 44 Opper et al. [137]
P OH
14 45
x n
20 48, 67 23, 22
OEt
Single phosphonic ester O 8
57
P OEt
14
61
x n
20 11, 13 3 55
O O OH
Geminal phosphonic acid HO OH
14 32
OH P P
20 87 120
x n

Benzyl phosphonic acid O O 14 43

n
x
O
20 46 42
P
HO OH

Benzyl phosphonic ester O O 14
25

n
x
173

O
20 2 47
P
EtO OEt

(continued)
Table A2 (continued)
174

CH2
sequence Tc ΔHm Tg
Group Structure (x) Tm ( C) ( C) (J/g) ( C) References
Esters with perfluoroalkyl segments O O 18 98, 108 90, 99.8, Mandal et al. [138]
O O 103 12.3
n
x

O O 18 110, 128 104, 57,

O O 121 14.1
n
x

O O 18 71 66, 98.76
O O 68.4
n
x

O
Sulfite S
8 – Gaines et al. [67]
O O n
14 7 49
20 37 63
Phenyl 18
1.5,
23.6 Watson and Wagener [125]

22.5
n
x

Cyclohexyl 20 9 37 Rojas et al. [93]

n
x

Adamantyl Ad 20
8, 17 2, 8 Rojas et al. [93]

n
x Ad:
L. Santonja-Blasco et al.
Aryl ether units 20 123.3, 120.1 Song et al. [139]
O 134.9

O n
x
20 128.2, 119.9
O 133.1

O n
X
20 127.2,
O O 132.4
n
x
Crystallization of Precision Ethylene Copolymers

20 77.6, 90.4
O O
n
x
20 –
6
O O
n
x
Ph
20 – 20
O O
n
x
Ph Ph
Available data for melting temperature (Tm), crystallization temperature (Tc), heat of melting (ΔHm) and glass transition temperature (Tg) are listed
175

a
Polyesters with possible head–head, head–tail, and tail–tail orientation (1:2:1 ratio)
b
A melting peak is not apparent in the endotherm shown for this sample
176 L. Santonja-Blasco et al.

References

1. Markets and Markets (2013) Polyolefins market by types, applications & geography – Global
trends & forecasts to 2018. Markets and Markets, Dallas. www.marketsandmarkets.com
2. Brintzinger HH, Fischer D, Mulhaupt R, Rieger B, Waymouth RM (1995) Stereospecific
olefin polymerization with chiral metallocene catalysts. Angew Chem Int Ed Engl
34:1143–1170
3. Scheirs J, Kaminsky W (eds) (2000) Metallocene-based polyolefin, vol 2. Wiley, Chichester
4. Carrow BP, Nozaki K (2014) Transition-metal-catalyzed functional polyolefin synthesis:
effecting control through chelating ancillary ligand design and mechanistic insights. Macro-
molecules 47:2541–2555
5. Berda EB, Wagener KB (2011) Precision polyolefins. In: Hadjichristidis N, Hirao A,
Tezuka Y, Du Prez F (eds) Complex macromolecular architectures: synthesis, character-
ization and self-assembly, chap 10. Wiley, Singapore, p 317
6. Bielawski CW, Grubbs RH (2007) Living ring-opening metathesis polymerization.
Prog Polym Sci 32:1–29
7. Kobayashi S, Pitet LM, Hillmyer MA (2011) Regio- and stereoselective ring-opening
metathesis polymerization of 3-substituted cyclooctenes. J Am Chem Soc 133:5794–5797
8. Zhang J, Matta ME, Hillmyer MA (2012) Synthesis of sequence-specific vinyl copolymers by
regioselective ROMP of multiply substituted cyclooctenes. ACS Macro Lett 1:1383–1387
9. Zhang J, Matta ME, Martinez H, Hillmyer MA (2013) Precision vinyl acetate/ethylene
(VAE) copolymers by ROMP of acetoxy-substituted cyclic alkenes. Macromolecules 46:
2535–2543
10. Ortmann P, Mecking S (2013) Long-spaced aliphatic polyesters. Macromolecules 46:
7213–7218
11. Ortmann P, Heckler I, Mecking S (2014) Physical properties and hydrolytic degradability of
polyethylene-like polyacetals and polycarbonates. Green Chem 16:1816–1827
12. Ortmann P, Lemke TA, Mecking S (2015) Long-spaced polyamides: elucidating the gap
between polyethylene crystallinity and hydrogen bonding. Macromolecules 48:1463–1472
13. Ortmann P, Wimmer FP, Mecking S (2015) Long-spaced polyketones from ADMET co-
polymerizations as ideal models for ethylene/CO copolymers. ACS Macro Lett 4:704–707
14. Hillmyer MA, Laredo WR, Grubbs RH (1995) Ring-opening metathesis polymerization of
functionalized cyclooctenes by a ruthenium-based metathesis catalyst. Macromolecules 28:
6311–6316
15. Alamo RG, Mandelkern L (1994) The crystallization behavior of random copolymers of
ethylene. Thermochim Acta 238:155–201
16. Crist B (2003) Thermodynamics of statistical copolymer melting. Polymer 44:4563–4572
17. Mandelkern L (2002) Crystallization of polymers, vol 1. Cambridge University Press,
Cambridge, Chap 5
18. Mathot VBF, Reynaers H (2002) Crystallization, melting and morphology of homogeneous
ethylene copolymers. In: Cheng SZD (ed) Handbook of thermal analysis and calorimetry,
vol 3. Applications to polymers and plastics. Elsevier, Amsterdam, p 197
19. Li S, Register RA (2013) Crystallization in copolymers. In: Piorkowska E, Rutledge GC (eds)
Handbook of polymer crystallization, 1st edn. Wiley, Hoboken, pp 327–346
20. Hu W, Gao H, Mathot V, Chen X, Alamo RG (2015) Crystallization of Statistical Copolymers.
Adv Polym Sci (this issue)
21. Flory PJ (1955) Theory of crystallization in copolymers. Trans Faraday Soc 51:848–857
22. Alamo R, Domszy R, Mandelkern L (1984) Thermodynamic and structural properties of co-
polymers of ethylene. J Phys Chem 88:6587–6595
23. Alamo RG, Mandelkern L (1989) Thermodynamic and structural properties of ethylene co-
polymers. Macromolecules 22:1273–1277
24. Alamo RG, Mandelkern L (1991) Crystallization kinetics of ethylene copolymers. Macro-
molecules 24:6480–6493
Crystallization of Precision Ethylene Copolymers 177

25. Alamo RG, Chan EKM, Mandelkern L, Voigt-Martin IG (1992) The influence of mole-
cular weight on the melting and phase structure of random copolymers of ethylene. Macro-
molecules 25:6381–6394
26. Alamo RG, Viers BD, Mandelkern L (1993) Phase structure of random ethylene copolymers:
a study of the counit content and molecular weight as independent variables. Macromolecules
26:5740–5747
27. Isasi JR, Haigh JA, Graham JT, Mandelkern L, Alamo RG (2000) Some aspects of the
crystallization of ethylene copolymers. Polymer 41:8813–8823
28. Alizadeh A, Richardson L, Xu J, McCartney S, Marand H, Cheung YW, Chum S (1999)
Influence of structural and topological constraints on the crystallization and melting behavior
of polymers. 1. Ethylene/1-octene copolymers. Macromolecules 32:6221–6235
29. Defoor F, Groeninckx G, Reynaers H, Schouterden P, Van der Heijden B (1993) Molecular,
thermal, and morphological characterization of narrowly branched fractions of 1-octene
linear low-density polyethylene. 3. Lamellar and spherulitic morphology. Macromolecules
26:2575–2582
30. Bensason S, Minick J, Moet A, Chum S, Hiltner A, Baer E (1996) Classification of homo-
geneous ethylene-octene copolymers based on comonomer content. J Polym Sci B Polym Phys
34:1301–1315
31. Walter ER, Reding FP (1956) Variations in unit cell dimensions in polyethylene. J Polym Sci
21:561–562
32. Cole EA, Holmes DR (1960) Crystal lattice parameters and the thermal expansion of
linear paraffin hydrocarbons, including polyethylenes. J Polym Sci 46:245–256
33. Swan PR (1962) Polyethylene unit cell variations with branching. J Polym Sci 56:409–416
34. Ruiz de Ballesteros O, Auriemma F, Guerra G, Corradini P (1996) Molecular organization in
the pseudo-hexagonal crystalline phase of ethylene
propylene copolymers. Macromolecules
29:7141–7148
35. Baker CH, Mandelkern L (1965) The crystallization and melting of copolymers II—variation
in unit-cell dimensions in polymethylene copolymers. Polymer 7:71–83
36. Richardson MJ, Flory PJ, Jackson JB (1963) Crystallization and melting of copolymers of
polymethylene. Polymer 4:221–236
37. Ver Strate G, Wilchinsky ZW (1971) Ethylene-propylene copolymers: degree of crystallinity
and composition. J Polym Sci Part A-2(9):127–142
38. Stephens CH, Yang H, Islam M, Chum SP, Rowan SJ, Hiltner A, Baer E (2003) Character-
ization of polyethylene with partially random chlorine substitution. J Polym Sci Part B Polym
Phys 41:2062–2070
39. Gomez MA, Tonelli AE, Lovinger AJ, Schilling FC, Cozine MH, Davis DD (1989) Structure
and morphology of ethylene-vinyl chloride copolymers. Macromolecules 22:4441–4451
40. Stoeva S, Popov A, Rodriguez R (2004) Wide angle X-ray diffraction study of the solid-phase
chlorinated poly(ethylene). Polymer 45:6341–6348
41. Kalepky U, Fischer EW, Herchenroder P, Schelten J, Lieser G, Wegner G (1979) Character-
ization of semicrystalline random copolymers by small-angle neutron scattering. J Polym Sci
Part B Polym Phys 17:2117–2131
42. Landes BG, Harrison IR (1987) The location of comonomer units in crystallizable copolymers:
brominated polyethylene. Polymer 28:911–917
43. Bunn CW, Peiser HS (1947) Mixed crystal formation in high polymers. Nature 159:161–162
44. Chatani Y, Takizawa T, Murahashi S (1962) Crystal structures of polyketones (ethylene/
carbon monoxide copolymers). J Polym Sci 62:S27–S30
45. Wunderlich B (1976) Macromolecular physics, vol 1. Academic Press, New York
46. Sanchez IC, Eby RK (1975) Thermodynamics and crystallization of random copolymers.
Macromolecules 8:638–641
47. Helfand E, Lauritzen JI Jr (1973) Theory of copolymer crystallization. Macromolecules 6:
631–638
178 L. Santonja-Blasco et al.

48. Alamo RG, VanderHart DL, Nyden MR, Mandelkern L (2000) Morphological partitioning of
ethylene defects in random propylene-ethylene copolymers. Macromolecules 33:6094–6105
49. VanderHart DL, Alamo RG, Nyden MR, Kim MH, Mandelkern L (2000) Observation of
resonances associated with stereo and regio defects in the crystalline regions of isotactic
polypropylene: towards a determination of morphological partitioning. Macromolecules 33:
6078–6093
50. Nyden MR, VanderHart DL, Alamo RG (2001) The conformational structures of defect-
containing chains in the crystalline regions of isotactic polypropylenes. Comput Theor Polym S
11:175–189
51. De Rosa C, Auriemma F, Capitani D, Caporaso L, Talarico G (2000) Solid state 13C NMR
analysis of syndiotactic copolymers of propene with 1-butene. Polymer 41:2141–2148
52. Ruiz-Orta C, Alamo RG (2012) Morphological and kinetic comonomer partitioning in
random propylene 1-butene copolymers. Polymer 53:810–822
53. Otocka EP, Kwei TK (1968) Properties of ethylene-acrylic acid copolymers. Macromolecules
1:244–249
54. Eisenberg A, Hird B, Moore RB (1990) A new multiplet-cluster model for the morphology of
random ionomers. Macromolecules 23:4098–4107
55. Ding YS, Hubbard SR, Hodgson KO, Register RA, Cooper SL (1988) Anomalous small-
angle X-ray scattering from a sulfonated polystyrene ionomer. Macromolecules 21:
1698–1703
56. Laurer JH, Winey KI (1998) Direct imaging of ionic aggregates in Zn-neutralized poly
(ethylene-co-methacrylic acid) copolymers. Macromolecules 31:9106–9108
57. Voigt-Martin IG, Mandelkern L (1989) In: Cheremisinoff NP (ed) Handbook of polymer science
and technology, vol. 3. Marcel Dekker, New York, p 99
58. Vadlamudi M, Subramanian G, Shanbhag S, Alamo RG, Varma-Nair M, Fiscus DM, Brown GM,
Lu C, Ruff CJ (2009) Molecular weight and branching distribution of a high performance
metallocene ethylene 1-hexene copolymer film-grade resin. Macromol Symp 282:1–13
59. Mamun A, Chen X, Alamo RG (2014) Interplay between a strong memory effect of
crystallization and liquid–liquid phase separation in melts of broadly distributed ethylene-
1-alkene copolymers. Macromolecules 47:7958–7970
60. Le Fevere De Ten Hove C, Penelle J, Ivanov DA, Jonas AM (2004) Encoding crystal micro-
structure and chain folding in the chemical structure of synthetic polymers. Nat Mater 3:
33–37
61. Ortmann P, Trzaskowski J, Krumova M, Mecking S (2013) Precise microstructure self-
stabilized polymer nanocrystals. ACS MacroLett 2:125–127
62. Hosoda S, Nozue Y, Kawashima Y, Suita K, Seno S, Nagamatsu T, Wagener KB, Inci B,
Zuluaga F, Rojas G, Leonard JK (2011) Effect of the sequence length distribution on the
lamellar crystal thickness and thickness distribution of polyethylene: perfectly equisequential
ADMET polyethylene vs ethylene/α-olefin copolymer. Macromolecules 44:313–319
63. Hosoda S, Nozue Y, Kawashima Y, Utsumi S, Nagamatsu T, Wagener K, Berda E, Rojas G,
Baughman T, Leonard J (2009) Perfectly controlled lamella thickness and thickness distri-
bution: a morphological study on ADMET polyolefins. Macromol Symp 282:50–64
64. Kaner P, Ruiz-Orta C, Boz E, Wagener KB, Tasaki M, Tashiro K, Alamo RG (2014)
Kinetic control of chlorine packing in crystals of a precisely substituted polyethylene.
Toward advanced polyolefin materials. Macromolecules 47:236–245
65. Tasaki M, Yamamoto H, Hanesaka M, Tashiro K, Boz E, Wagener KB, Ruiz-Orta C,
Alamo RG (2014) Polymorphism and phase transitions of precisely halogen-substituted poly-
ethylene. (1) Crystal structures of various crystalline modifications of bromine-substituted
polyethylene on every 21st backbone carbon. Macromolecules 47:4738–4749
66. Seitz ME, Chan CD, Opper KL, Baughman TW, Wagener KB, Winey KI (2010) Nanoscale
morphology in precisely sequenced poly(ethylene-co-acrylic acid) zinc ionomers. J Am
Chem Soc 132:8165–8174
Crystallization of Precision Ethylene Copolymers 179

67. Gaines TW, Nakano T, Chujo Y, Trigg EB, Winey KI, Wagener KB (2015) Precise sulfite
functionalization of polyolefins via ADMET polymerization. ACS Macro Lett 4:624–627
68. Boz E, Wagener KB, Ghosal A, Fu R, Alamo RG (2006) Synthesis and crystallization of
precision ADMET polyolefins containing halogens. Macromolecules 39:4437–4447
69. Boz E, Nemeth AJ, Alamo RG, Wagener KB (2007) Precision ethylene vinyl bromide
polymers. Adv Synth Catal 349:137–141
70. Boz E, Nemeth AJ, Ghiviriga I, Jeon K, Alamo RG, Wagener KB (2007) Precision ethylene/
vinyl chloride polymers via condensation polymerization. Macromolecules 40:6545–6551
71. Boz E, Nemeth AJ, Wagener KB, Jeon K, Smith R, Nazirov F, Bockstaller MR, Alamo RG
(2008) Well-defined precision ethylene/vinyl fluoride polymers: synthesis and crystalline
properties. Macromolecules 4:1647–1653
72. Boz E, Ghiviriga I, Nemeth AJ, Jeon K, Alamo RG, Wagener KB (2008) Random, defect-free
ethylene/vinyl halide model copolymers via condensation polymerization. Macromolecules
41:25–30
73. Alamo RG, Jeon K, Smith RL, Boz E, Wagener KB, Bockstaller MR (2008) Crystallization
of polyethylenes containing chlorines: precise vs random placement. Macromolecules 41:
7141–7151
74. Schulz MD, Wagener KB (2014) Precision polymers through ADMET polymerization.
Macrom Chem Phys 215:1936–1945
75. Smith JA, Brzezinska KR, Valenti DJ, Wagener KB (2000) Precisely controlled methyl
branching in polyethylene via acyclic diene metathesis (ADMET) polymerization. Macro-
molecules 33:3781–3794
76. Pepels MPF, Hansen MR, Goossens H, Duchateau R (2013) From polyethylene to polyester:
influence of ester groups on the physical properties. Macromolecules 46:7668–7677
77. Gutzler F, Wegner G (1980) Synthesis and melting behavior of poly(ethylene) copolymers
obtained by polymeranalogous reaction. Colloid Polym Sci 258:776–786
78. Baughman TW, Chan CD, Winey KI, Wagener KB (2007) Synthesis and morphology of
well-defined poly(ethylene-co-acrylic acid) copolymers. Macromolecules 40:6564–6571
79. Matsui K, Seno S, Nozue Y, Shinohara Y, Amemiya Y, Berda EB, Rojas G, Wagener KB
(2013) Influence of branch incorporation into the lamella crystal on the crystallization behav-
ior of polyethylene with precisely spaced branches. Macromolecules 46:4438–4446
80. Bowner TN, Tonelli AE (1986) Infrared spectroscopy of ethylene–vinyl chloride copolymers.
J Polym Sci Part B Polym Phys 24:1631–1650
81. Tonelli AE, Bowner TN (1987) C-CL stretching vibrations in ethylene-vinyl chloride copolymers.
J Polym Sci Part B Polym Phys 25:1153–1156
82. Snyder RG, Schachtschneider JH (1963) Vibrational analysis of the n-paraffins I: assign-
ments of infrared bands in the spectra of C3H8 through n-C19H40. Spectrochim Acta 19:
85–116
83. Snyder RG (1960) Vibrational spectra of crystalline n-paraffins: part I. Methylene rocking and
wagging modes. J Mol Spectrosc 4:411–434
84. Yu GS, Li HW, Hollander F, Snyder RG, Strauss HL (1999) Comparison of the structures of
ammonium myristate, palmitate, and stearate by X-ray diffraction, infrared spectroscopy, and
infrared hole burning. J Phys Chem B 103:10461–10468
85. Yan WH, Strauss HL, Snyder RG (2000) Conformation of the acyl chains in diacylphospho-
lipid gels by IR spectroscopy. J Phys Chem B 104:4229–4238
86. Venkataraman NV, Vasudevan S (2001) Interdigitation of an intercalated surfactant bilayer.
J Phys Chem B 105:7639–7650
87. Tasaki M, Yamamoto H, Yoshioka T, Hanesaka M, Ninh TH, Tashiro K, Jeon HJ, Choi KB,
Jeong HS, Song HH, Ree MH (2014) Microscopically-viewed relationship between the
chain conformation and ultimate Young s modulus of a series of arylate polyesters with
long methylene segments. Polymer 55:1799–1808
180 L. Santonja-Blasco et al.

88. Desiraju GR, Parthasarathy R (1989) The nature of halogen-halogen interactions: are short
halogen contacts due to specific attractive forces or due to close packing of nonspherical atoms?
J Am Chem Soc 111:8725–8726
89. Wagener KB, Valenti D, Hahn SF (1997) ADMET modeling of branching in polyethylene.
The effect of a perfectly-spaced methyl group. Macromolecules 30:6688–6690
90. Lieser G, Wegner G, Smith JA, Wagener KB (2004) Morphology and packing behavior of model
ethylene/propylene copolymers with precise methyl branch placement. Colloid Polym Sci 282:
773–781
91. Baughman TW, Sworen JC, Wagener KB (2006) Sequenced ethylene-propylene copolymers:
effects of short ethylene run lengths. Macromolecules 39:5028–5036
92. Wei Y, Graf R, Sworen JC, Cheng C-Y, Bowers CR, Wagener KB, Spiess HW (2009) Local
and collective motions in precise polyolefins with alkyl branches: A combination of 2H and
13
C solid-state NMR spectroscopy. Angew Chem Int Ed 48:4617–4620
93. Rojas G, Inci B, Wei Y, Wagener KB (2009) Precision polyethylene: changes in morphology
as a function of alkyl branch size. J Am Chem Soc 131:17376–17386
94. Inci B, Lieberwirth I, Steffen W, Mezger M, Graf R, Landfester K, Wagener KB (2012)
Decreasing the alkyl branch frequency in precision polyethylene: effect of alkyl branch size
on nanoscale morphology. Macromolecules 45:3367–3376
95. Sworen JC, Smith JA, Berg JM, Wagener KB (2004) Modeling branched polyethylene:
copolymers possessing precisely placed ethyl branches. J Am Chem Soc 126:11238–11246
96. Nozue Y, Seno S, Nagamatsu T, Hosoda S, Shinohara Y, Amemiya T, Berda EB, Rojas G,
Wagener KB (2012) Cross nucleation in polyethylene with precisely spaced ethyl branches.
ACS Macro Lett 1:772–775
97. Zuluaga F, Inci B, Nozue Y, Hosoda S, Wagener KB (2009) Reducing branch frequency in
precision polyethylene. Macromolecules 42:4953–4955
98. Rojas G, Wagener KB (2009) Precisely and irregularly sequenced ethylene/1-hexene co-
polymers: a synthesis and thermal study. Macromolecules 42:1934–1947
99. Sworen JC, Wagener KB (2007) Linear low-density polyethylene containing precisely placed
hexyl branches. Macromolecules 40:4414–4423
100. Few CS, Wagener KB, Thompson DL (2014) Systematic studies of morphological changes of
precision polyethylene. Macromol Rapid Commun 35:123–132
101. Inci B, Wagener KB (2011) Decreasing the alkyl branch frequency in precision polyethylene:
pushing the limits toward longer run lengths. J Am Chem Soc 133:11872–11875
102. Berda EB, Baughman TW, Wagener KB (2006) Precision branching in ethylene copolymers:
synthesis and thermal behavior. J Polym Sci Part A Polym Chem 44:4981–4989
103. Rojas G, Berda EB, Wagener KB (2008) Precision polyolefin structure: modeling polyeth-
ylene containing alkyl branches. Polymer 49:2985–2995
104. Schulz MD, Sauty NF, Wagener KB (2015) Morphology control in precision polyolefins.
Appl Petrochem Res 5:3–8
105. Mutlu H, Montero De Espinosa L, Meier MAR (2011) Acyclic diene metathesis: a versatile tool
for the construction of defined polymer architectures. Chem Soc Rev 40:1404–1445
106. Hall LM, Stevens MJ, Frischknecht AL (2012) Dynamics of model ionomer melts of
various architectures. Macromolecules 45:8097–8108
107. Alam TM, Jenkins JE, Bolintineanu DS, Stevens MJ, Frischknecht AL, Buitrago CF,
Winey KI, Opper KL, Wagener KB (2012) Heterogeneous coordination environments in
lithium-neutralized ionomers identified using 1H and 7Li MAS NMR. Materials 5:1508–1527
108. Hall LM, Seitz ME, Winey KI, Opper KL, Wagener KB, Stevens MJ, Frischknecht AL
(2012) Ionic aggregate structure in ionomer melts: effect of molecular architecture on aggre-
gates and the ionomer peak. J Am Chem Soc 134:574–587
109. Aitken BS, Buitrago CF, Heffley JD, Lee M, Gibson HW, Winey KI, Wagener KB (2012)
Precise ionomers: synthesis and thermal/mechanical characterization. Macromolecules 45:
681–687
Crystallization of Precision Ethylene Copolymers 181

110. Buitrago CF, Alam TM, Opper KL, Aitken BS, Wagener KB, Winey KI (2013) Morpholog-
ical trends in precise acid- and ion-containing polyethylenes at elevated temperature. Mac-
romolecules 46:8995–9002
111. Buitrago CF, Jenkins JE, Opper KL, Aitken BS, Wagener KB, Alam TM, Winey KI (2013)
Room temperature morphologies of precise acid- and ion-containing polyethylenes. Macro-
molecules 46:9003–9012
112. Choi UH, Middleton LR, Soccio M, Buitrago CF, Aitken BS, Masser H, Wagener KB, Winey KI,
Runt J (2015) Dynamics of precise ethylene ionomers containing ionic liquid functionality.
Macromolecules 48:410–420
113. Buitrago CF, Opper KL, Wagener KB, Winey KI (2012) Precise acid copolymer exhibits a
face-centered cubic structure. ACS Macro Lett 1:71–74
114. Buitrago CF, Bolintineanu DS, Seitz ME, Opper KL, Wagener KB, Stevens MJ,
Frischknecht AL, Winey KI (2015) Direct comparisons of X-ray scattering and atomistic
molecular dynamics simulations for precise acid copolymers and ionomers. Macromolecules
48:1210–1220
115. Middleton LR, Szewczyk S, Azoulay J, Murtagh D, Rojas G, Wagener KB, Cordaro J, Winey
KI (2015) Hierarchical acrylic acid aggregate morphologies produce strain-hardening in
precise polyethylene-based copolymers. Macromolecules 48:3713–3724
116. Ramakrishnan S (1991) Well-defined ethylene-vinyl alcohol copolymers via hydroboration:
control of composition and distribution of the hydroxyl groups on the polymer backbone.
Macromolecules 24:3753–3759
117. Valenti DJ, Wagener KB (1998) Direct synthesis of well-defined alcohol-functionalized
polymers via acyclic diene metathesis (ADMET) polymerization. Macromolecules 31:
2764–2773
118. Thompson D, Yamakado R, Wagener KB (2014) Extending the methylene spacer length of
ADMET hydroxy-functionalized polymers. Macrom Chem Phys 215:1212–1217
119. Thompson DL, Wagener KB, Schulze U, Voit B, Jehnichen D, Malanin M (2015) Spectro-
scopic examinations of hydrogen bonding in hydroxy-functionalized ADMET chemistry.
Macrom Rapid Comm 36:60–64
120. Leonard JK, Wei Y, Wagener KB (2012) Synthesis and thermal characterization of
precision poly(ethylene-co-vinyl amine) copolymers. Macromolecules 45:671–680
121. German I, D’Agosto F, Boisson C, Tencé-Girault S, Soulié-Ziakovic C (2015) Microphase
separation and crystallization in H-bonding end-functionalized polyethylenes. Macromolecules
48:3257–3268
122. Baughman TW, van der Aa E, Lehman SE, Wagener KB (2005) Circumventing the reactivity
ratio dilemma: synthesis of ethylene-co-methyl vinyl ether copolymer. Macromolecules 38:
2550–2551
123. Baughman TW, van der Aa E, Wagener KB (2006) Linear ethylene–vinyl ether copolymers:
synthesis and thermal characterization. Macromolecules 39:7015–7021
124. Watson MD, Wagener KB (2000) Ethylene/vinyl acetate copolymers via acyclic diene meta-
thesis polymerization. Examining the effect of “long” precise ethylene run lengths. Macro-
molecules 33:5411–5417
125. Watson MD, Wagener KB (2000) Functionalized polyethylene via acyclic diene metathesis
polymerization: effect of precise placement of functional groups. Macromolecules 33:
8963–8970
126. Watson MD, Wagener KB (2000) Tandem homogeneous metathesis/heterogeneous hydro-
genation: preparing model ethylene/CO2 and ethylene/CO copolymers. Macromolecules 33:
3196–3201
127. van der Meulen I, de Geus M, Antheunis H, Deumens R, Joosten EA, Koning CE, Heise A
(2008) Polymers from functional macrolactones as potential biomaterials: enzymatic ring
opening polymerization, biodegradation, and biocompatibility. Biomacromolecules 9:
3404–3410
182 L. Santonja-Blasco et al.

128. Pepels MPF, Govaert LE, Duchateau R (2015) Influence of the main-chain configuration on
the mechanical properties of linear aliphatic polyesters. Macromolecules 48:5845–5854
129. Scherman OA, Walker R, Grubbs RH (2005) Synthesis and characterization of stereoregular
ethylene-vinyl alcohol copolymers made by ring-opening metathesis polymerization. Macro-
molecules 38:9009–9014
130. Broadhurst MG (1962) An analysis of the solid phase behavior of the normal paraffins. J Res
Nat Bur St A Phys Chem 66A:241–249
131. Qiu W, Sworen J, Pyda M, Nowak-Pyda E, Habenschuss A, Wagener KB, Wunderlich B
(2006) Effect of the precise branching of polyethylene at each 21st CH2 group on its phase
transitions, crystal structure, and morphology. Macromolecules 39:204–217
132. Schwendeman JE, Wagener KB (2005) gem-Dimethyl effects in the thermal behavior of
polyethylene. Macromol Chem Phys 206:1461–1471
133. Schwendeman JE, Wagener KB (2004) Modeling ethylene/methyl methacrylate and ethylene/
methacrylic acid copolymers using acyclic diene metathesis chemistry. Macromolecules 37:
4031–4037
134. Berda EB, Lande RE, Wagener KB (2007) Precisely defined amphiphilic graft copolymers.
Macromolecules 40:8547–8552
135. Berda EB, Wagener KB (2008) Inducing pendant group interactions in precision polyolefins:
synthesis and thermal behavior. Macromolecules 41:5116–5122
136. Opper KL, Wagener KB (2009) Precision sulfonic acid ester copolymers. Macromol Rapid
Commun 30:915–919
137. Opper KL, Markova D, Klapper M, Mullen K, Wagener KB (2010) Precision phosphonic acid
functionalized polyolefin architectures. Macromolecules 43:3690–3698
138. Mandal J, Prasad SK, Rao DSS, Ramakrishnan S (2014) Periodically clickable polyesters:
study of intrachain self-segregation induced folding, crystallization, and mesophase formation.
J Am Chem Soc 136:2538–2545
139. Song S, Miao W, Wang Z, Gong D, Chen ZR (2015) Synthesis and characterization of
precisely-defined ethylene-co-aryl ether polymers via ADMET polymerization. Polymer 64:
76–83
Adv Polym Sci (2017) 276: 183–214
DOI: 10.1007/12_2015_343
© Springer International Publishing Switzerland 2016
Published online: 13 February 2016

Supramolecular Crystals and Crystallization

with Nanosized Motifs of Giant Molecules

Xue-Hui Dong, Chih-Hao Hsu, Yiwen Li, Hao Liu, Jing Wang,
Mingjun Huang, Kan Yue, Hao-Jan Sun, Chien-Lung Wang, Xinfei Yu,
Wen-Bin Zhang, Bernard Lotz, and Stephen Z.D. Cheng

Abstract Supramolecular crystals and crystallization are general concepts used to

describe broader aspects of ordered structures and their formation in the three-
dimensional (3D) bulk and solution and in 2D thin film states at length scales ranging
from sub-nanometers to sub-micrometers. Although the fundamental crystallo-
graphic principles are still held in these structures, starting from their basic repeating
units (motifs), it is not necessary that each atomic position within their motifs
possesses translational symmetry in long range order, but could have quasi-long
range or short range order. As a result, the motif becomes the smallest unit for
constructing 3D or 2D ordered structures that maintain the long range translational
order. The formation of these supramolecular ordered structures essentially follows
the physical principle of phase transformations, involving either nucleation and

Xue-Hui Dong and Chih-Hao Hsu are equally contributed.

X.-H. Dong, C.-H. Hsu, Y. Li, H. Liu, J. Wang, M. Huang, K. Yue, H.-J. Sun, C.-L. Wang,
X. Yu, and S.Z.D. Cheng (*)
Department of Polymer Science, College of Polymer Science and Polymer Engineering,
The University of Akron, Akron, OH 44325-3909, USA
e-mail: [email protected]
W.-B. Zhang (*)
Department of Polymer Science, College of Polymer Science and Polymer Engineering,
The University of Akron, Akron, OH 44325-3909, USA
Key Laboratory of Polymer Chemistry and Physics of Ministry of Education, College of
Chemistry and Molecular Engineering, Center for Soft Matter Science and Engineering,
Peking University, Beijing 100871, China
e-mail: [email protected]
B. Lotz (*)
Institut Charles Sadron (CNRS-Université de Strasbourg), 23, Rue du Loess, 67034
Strasbourg, France
e-mail: [email protected]
184 X.-H. Dong et al.

growth or spinodal decomposition mechanisms. However, larger ordered structures

require stronger and more cooperative interactions to sustain their structures in
equilibrium or stable states. We propose utilization of collective secondary interac-
tions, similar to those found in biological and living systems, to generate sufficient
interactions and stabilize these structures. Furthermore, when the basic unit of the
structure becomes increasingly larger and heavier, thermal (density) fluctuations
during the phase transitions may not be sufficiently large to overcome transition
barriers of the basic unit. In these cases, external fields might be required to stimulate
the magnitude of thermal (density) fluctuation and/or redistribute (thus, decrease) a
single transition barrier into several stepwise transition sequences with lower bar-
riers for each transition, and thus increase the speed of phase transformations.

Keywords Crystallization • Giant molecules • Molecular nanoparticles •

Nanoatoms • Supramolecular crystal

Contents
1 Concept of Giant Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 184
2 Supramolecular Crystals and Crystallization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 187
3 Giant Shape Amphiphiles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 190
4 Giant Janus Particles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 194
5 Giant Surfactants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 198
6 Giant Tetrahedra . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 203
7 Thin Films of Giant Molecules . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 206
8 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 209
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 210

1 Concept of Giant Molecules

Rapid developments in polymer materials since Staudinger’s macromolecular

hypothesis [1] have revolutionized the whole field of science and engineering,
and generated significant impact on human society. These materials are now
popularly beneficial to our daily life. However, precise control of chemical primary
structure and composition and accurate construction of hierarchal physical struc-
tures in synthetic polymers (similar to natural polymers such as nuclei acids and
proteins) remain grand challenges. Traditional approaches in designing and syn-
thesizing polymers usually start from currently available or newly synthesized
monomer units as building blocks, within which covalent links are established
between different atoms using chemical principles. Utilizing various polymeriza-
tion reactions such as anionic, cationic, condensation, free radical, and coordina-
tion, the monomers are connected in different ways to generate polymers with
linear, branched, or other architectures [2]. This type of approach intrinsically leads
to an understanding that polymer properties are evidently associated with molecular
weight (MW) and its distribution, tacticity, and topology, but less appreciated are
the effects of controlling supramolecular structure across different length scales.
Supramolecular Crystals and Crystallization 185

Nevertheless, these supramolecular structures at different mesoscales are crucially

important for transferring and amplifying microscopic functions to macroscopic
properties.
About 60 years ago a great physicist, R. Feynman, asked a fundamental ques-
tion: “What would the properties of materials be if we could really arrange the
atoms the way we want them?” [3]. To fully address Feynman’s question, we are
required not only to know the primary chemical structures of materials, but also to
control sequential and hierarchical secondary, tertiary, and quaternary structures so
that we can completely understand how each structure at different length scales
affects the final properties of the material. To seek solutions to Feynman’s inquiry
in the context of macromolecular science, precision syntheses of primary chemical
structures and accurate control over higher level supramolecular structures are
prerequisites. In other words, simply connecting monomers in a desired sequence
is not sufficient to achieve a desired property for a specific polymer.
What we need to do is to follow the principles of chemistry and physics and
integrate them into an ever-increasing sophistication of molecular, supramolecular,
and structural design to direct development of new functional macromolecules. Yet,
the unique features of polymers (large size, large number of functional groups,
difficulties associated with purification, large degrees of conformational freedom,
and so on) have made precision synthesis of polymers yet to be demonstrated
[4]. An approach we have adopted to possibly overcome these difficulties is the
modular construction of macromolecules (a word taken in the present context to
have a broader meaning than “polymers”) using precisely defined molecular nano-
scale building blocks [5–7]. Example of precisely synthesized macromolecules are
dendrimers with a cascade topology, which are regularly branched and monodis-
perse, yet without shape persistence [8–10]. Many investigations have been carried
out on dendron–polymer conjugates [11–13] and Janus dendrimers [14–16]. When
we think of macromolecular precision syntheses, nanoscale building blocks must
possess relatively independent, modular, and well-defined three-dimensional
(3D) structure and shape so that further assembly into a variety of hierarchal
structures at different length scales is predictable and robust.
As modular building blocks, we propose the use of molecular nanoparticles
(MNPs), which are shape- and volume-persistent nano-objects with well-defined
molecular structure and specific symmetry [5]. Typical MNPs under consideration
include cage-like compounds and folded globular proteins (Fig. 1) with compact
and rigid nature as well as precisely defined symmetry and surface functionality. In
this article, we focus on the cage-like compounds. Generally speaking, the overall
molecular shape of cage-like compounds can be held either by noncovalent bonds,
as demonstrated in the “molecular flasks” reported by Fujita and coworkers [17, 18]
and the “tennis ball” reported by Rebek and colleagues [19], or by covalent bonds
as in [60]fullerene (C60) [16], polyhedral oligomeric silsesquioxane (POSS) [20],
and polyoxometalate (POM) [21–23]. To construct stable supramolecular struc-
tures, “collective physical secondary interactions” are necessary, which are gener-
ated by the functional groups on the periphery of MNPs. These interactions
constitute the enthalpic driving force for further assembly of MNPs into hierarchal
structures under the packing constraints imposed by the overall molecular shape
186 X.-H. Dong et al.

Fig. 1 Typical examples of molecular nanoparticles

(entropic effect), leading to a variety of unconventional structures and phase

behaviors, as predicted by computer simulation [24–27].
A large variety of MNPs with different sizes, symmetries, and surface functional
groups can be used as versatile nanoscale building blocks. In view of their incom-
pressible and impenetrable features, we describe them as “nanoatoms.” This term is
evocative of the term “artificial atoms” that has been used to describe quantum dots
[28] and even metal nanoparticles [29, 30]. It is also similar to “nanoscale atoms,”
as used recently by Nukolls and coworkers to refer to pseudo-spherical molecular
clusters employed as “atoms” in constructing binary crystalline solids with unique
electronic and magnetic properties [31]. We use “giant molecules” to describe the
macromolecules built from MNP nanoatoms or their conjugates with other macro-
molecules such as polymers and dendrimers [5]. Although the term “giant mole-
cules” has been used exchangeably with “macromolecules” or more generally to
describe structures with a large number of atoms, we attempt to distinguish it from
traditional macromolecules. Giant molecules can be viewed as large-sized analogs
of small molecules modularly constructed from MNP nanoatom building blocks,
and emphasize that giant molecules are monodisperse and precisely defined
macromolecules [5].
We have designed, synthesized, and investigated the supramolecular lattices of
several categories of giant molecules, including giant surfactants, giant shape
amphiphiles, giant Janus particles, and giant polyhedra [5, 6]. It is certain that
many more types of giant molecules will appear in the future. However, even in
these few giant molecule categories, highly diverse, thermodynamically stable or
metastable phase structures and assembly processes in the bulk, thin film, and
solution have been recognized [6, 32–40]. Giant surfactants, simulating small
molecular surfactants, are polymer tail-tethered nanoatoms in which the head and
tail components are chemically very different, generating amphiphilicity [6, 32, 33,
35–37, 39–45]. Giant shape amphiphiles and giant Janus particles consist of
covalently bonded MNPs with distinct shapes, where self-assembly is not only
driven by chemical interactions but is also largely influenced by the packing
constraints of each individual shape [46, 47]. Giant polyhedra are synthesized
either by large MNPs with polyhedral shape or by deliberately placing nanoatoms
at the vertices of a polyhedron [38]. These giant molecules capture the fundamental
Supramolecular Crystals and Crystallization 187

structural and functional features of their small-molecule counterparts but are about
a thousand times bigger in volume; they are recognized to be size-amplified
versions of their smaller counterparts [5].
In this article, we focus on how these giant molecules are modularly assembled
together via collective secondary interactions to form various ordered supramolec-
ular structures with identified lattices. We call this type of ordered structure
“supramolecular crystals,” and their formation process “supramolecular
crystallization.”

2 Supramolecular Crystals and Crystallization

Before we describe supramolecular crystals and crystallization in detail, we need to

briefly review what crystals are and how they form. A crystal is a homogenous solid
whose constituents (such as atoms, molecules, or ions) are arranged in a highly
ordered microscopic structure, forming a crystal lattice that periodically extends in
exactly the same arrangement over three directions in a real space [48]. The
smallest building block that repeats itself in the crystal lattice is called the motif,
which can be constructed of atoms, ions, parts of a molecule, a whole molecule, or
even several molecules. A crystal must possess long-range positional, bond-
orientational, and molecular-orientational orders. These orders are based on the
motif, in which every atom must be located at a specific space position that can be
orderly repeated. There exist 7 Lattice system and 14 Bravais lattices. To determine
a crystal structure, X-ray and/or electron diffraction experiments are required to
obtain diffraction patterns based on the Bragg equation in its reciprocal lattice. Note
that a reciprocal lattice is a Fourier transformation of its corresponding lattice in
real space. In most cases, the length scale associated with the determination is in the
range of a fraction of nanometer to one nanometer [49].
Compared with crystallization of simple, small molecules, polymer crystalliza-
tion takes place even in supercooled liquids far from thermodynamic equilibrium
(crystallization temperatures are far below the equilibrium melting temperature at a
constant pressure). In the case of a crystallizable polymer, the Gibbs free energy of
the isotropic liquid in undercooling is always higher than that of the crystal solid.
The traditional crystallization concept is that the crystal nucleus is small in volume
and large in surface. The volume free energy term is negative towards the overall
Gibbs free energy and stabilizes the nucleus, whereas the surface free energy term is
positive in Gibbs free energy and destabilizes the nucleus. In the initial stage of
crystallization, the surface free energy wins the competition between these two free
energy terms, leading to a free energy barrier. The driving force to overcome the
free energy barrier and trigger the nucleation process (either primary or surface
nucleation) originates from the thermal (density) fluctuations of the liquid [50].
Beyond the general thermodynamics of a crystallization process, detailed
descriptions of polymer crystallization are based on two different approaches, as
188 X.-H. Dong et al.

deduced from experimental observation of structural and morphological data or

scattering data. These two viewpoints have led to dissimilar models for detailed
microscopic descriptions of polymer crystallization, although all of these theoret-
ical models are kinetic in nature. In the past half century, tremendous development
in this research area has been achieved. If we look back, those exciting events
include the discovery of polymer single crystals, lamellae, and chain folding.
Explosive growth followed, based on detailed studies of crystallization, annealing,
and melting; determination and control of molecular conformations, crystal struc-
tures, and morphologies; heated debates over crystal-growth models and the valid-
ity and extent of chain folding; elucidation and exploitation of mechanical, optical,
electrical, and other physical properties; crystallization under various confined
environments having low dimensions in space; and many others. In addition, new
syntheses provided a number of macromolecules with different chemical and
topological structures. Their crystallization behaviors have also stimulated many
new observations and discoveries such as molecularly double-twisted helical single
crystals [51–53].
However, major issues and questions still remain in polymer crystallization and
are yet to be resolved: (1) Understanding of the microscopic structure and dynamics
of supercooled liquids is still not complete (e.g., are supercooled liquids homoge-
neous?). (2) Current theoretical models in analytical forms all adopt mean-field
approaches and, thus, lose the details of molecular dynamic pathways during
crystallization; the specific trajectory of the pathway of one individual chain
molecule during crystallization can be very different from others. Although the
models can provide reasonable explanations for major, but not all, experimental
results observed and reported so far, they cannot predict the crystallization behav-
ior. (3) What is the nucleation barrier? This free energy barrier must have both
enthalpic and entropic origins, which are crucially associated with the molecular
structure and chain dynamics at microscopic length scales; namely, how does a
chain molecule that starts out in a 3D random coil conformation convert into
predominately chain-folded lamellar crystals? [50, 54, 55]
We now introduce the concept of “supramolecular crystals and crystallization.”
First and by far the most important notion is that they are crystals and must be in a
highly ordered microscopic, supramolecular lattice that periodically extends over
three directions in real space. The building block (motif), which repeats itself in the
lattice, contains many molecules and/or clusters within the motif, and those mol-
ecules and/or clusters often do not have specific space positions at the atomic and/or
molecular scales to generate a repetition scheme in the motif. In a supramolecular
crystal, although the positional, bond-orientational, and molecular-orientational
long-range orders are retained at the length scale of motif repetition (we have to
replace the bond-orientational order to be motif-orientational order at this length
scale), these long-range orders are not necessary for the structures within the motifs
at smaller length scales. In essence, molecules and/or clusters within the motif can
either have long-range, quasi-long-range, or short-range order. In such a way,
supramolecular lattices are constructed on the basis of the repeatability of motifs
to describe the ordered structures and are called “supramolecular crystals.” This
Supramolecular Crystals and Crystallization 189

concept is different from the protein crystallography often used to deal with large
crystal structures. In protein crystals, each ordered atomic position in the crystal is
perfectly retained. A simple example of supramolecular lattice formation is the
nanophase separation process as a result of immiscibility between components.
Flexible diblock copolymers with two immiscible blocks in the strong phase
separation region can form various supramolecular crystals (e.g., double gyroid
and body-centered cubic lattices) based on the composition of the blocks. The
length scale of supramolecular crystals ranges from nanophase-separated, ordered
structures at a length scale of below 10 nm to “colloidal crystals” that are usually
micrometers in length.
In general, secondary interactions play an important role in ordered phase
structural formation processes. However, these secondary interactions often do
not possess sufficiently strong interaction to hold supramolecular lattices with a
size of up to hundreds of nanometers [6]. To stabilize these lattice structures,
stronger interactions are necessary. It is possible to introduce collective secondary
interactions, as found in biological and living systems, to strengthen the stability of
the lattice. For example, a single hydrogen bond energy is relatively weak (~20–
40 kJ/mol) and much lower than that of a chemical bond (a single carbon–carbon
bond energy is ~145 kJ/mol). However, combination of 14 hydrogen bonds at a
single location simultaneously generates an interaction energy of 280–560 kJ/mol,
which is strong enough to stabilize the supramolecular lattice. In principle, for the
formation of supramolecular crystals, a balance of both enthalpic and entropic
terms must be achieved to establish a thermodynamically stable or metastable
system with a minimum free energy.
Supramolecular crystallization mechanisms are usually identical to, but broader
in scope than, crystallization of small molecules and polymers. Classic phase
transition processes involve either nucleation and growth, which illustrate a phase
transition from a metastable state to an equilibrium state with a free energy barrier,
or spinodal decomposition, which describes a phase transition from an unstable
state to an equilibrium state without overcoming a free energy barrier [49–51]. In
addition, supramolecular crystallization processes also take place in hierarchal
structures at different length scales. In many cases, the overall supramolecular
lattice formation requires several stages of sequential ordering structure formation
to complete. The overall barrier of the phase transformation could thus be separated
into several steps to speed up the transition kinetics (Ostwald’s stage rule)
[50]. When the size of building blocks becomes increasingly large with heavier
masses, the dynamics of these building blocks is more sluggish. Overcoming a free
energy barrier or simply diffusing/moving to an equilibrium state through thermal
(density) fluctuations becomes more difficult and time consuming. External force
fields (such as mechanical, electric, or magnetic) are helpful in providing assistance
for the building blocks to move and/or overcome the barriers.
More complicated pathways for structural formation may exist in giant mole-
cules. In many cases, the supramolecular crystallization ends at a state that is not
thermodynamically the most stable but, rather, the molecules are trapped in a
metastable state. This probably occurs for kinetic reasons. Generally speaking, a
190 X.-H. Dong et al.

metastable phase possesses a local free energy minimum, but not a global minimum
as in thermodynamic equilibrium states [50]. Whether this metastable state can be
experimentally observed during its phase transition process relies on the height of
transition free energy barrier leading to this metastable state. Once a metastable
state exists, its lifetime is determined by the height of the transition free energy
barrier that prevents it from transferring to a more stable state. One example is the
polymorphism observed in supramolecular crystals, which is very similar to that in
common crystals. Another example is the observation of various kinetically
trapped, metastable morphologies in micelles. To judge whether a supramolecular
crystal is in a metastable or a thermodynamically equilibrium state is, however,
difficult in most cases for giant molecules. This is because, unless the metastable
phase can be transferred to a more stable or a final equilibrium state in the
experimentally accessible time period, obtaining the thermodynamic properties of
these metastable and stable phases is extremely challenging.
In the following sections, we focus on this interesting topic and describe various
aspects of supramolecular crystals and crystallizations based on our recently
obtained experimental results with giant molecules. Understanding of this topic is
certainly just beginning. It requires many further scientific efforts to make this
concept fruitful.

3 Giant Shape Amphiphiles

When the motif increases in size, its shape starts to play an increasingly important
role in determining the self-assembly behavior and characteristics of the supramo-
lecular structure. This is even more significant when the motifs have rigid confor-
mations that pose specific packing constraints. Recently, Glotzer and coworkers
outlined a roadmap for predicting the phase formations from well-defined rigid
Archimedian polyhedra [24]. These polyhedra could not only form crystals, liquid
crystals, and amorphous glasses, but they also self-assemble into plastic crystals
and quasicrystals which are relatively more restricted regarding the shape of
components for building the motifs. There is good reason to expect that, when
components of distinct shapes are coupled together, the packing constraints of each
shaped component imposes an additional dimension for controlling their self-
assembled structure besides the differences in their compositions and chemical
interactions. These compounds are termed “shape amphiphiles,” with emphasis on
the role of shape in assembly [26]. Shape could be the single driving force for self-
assembly. If we put a plane–cube shape amphiphile to assemble, we anticipate that
in the condensed state, planes would preferentially stack together as long as the
cube arrangement allows. In combination with designed chemical interactions,
shape amphiphiles exhibit rich and versatile phase structures. It is also noteworthy
that in order for shape to take an important role in assembly, the component should
have a relatively rigid conformation and large size. Therefore, most shape amphi-
philes form supramolecular crystals whose supramolecular lattice is well defined
Supramolecular Crystals and Crystallization 191

even though detailed atomic order may be absent. Over the past decade, researchers
have developed a sizable family of shape amphiphiles [26, 27, 56, 57]. We now
look briefly at their structures and self-assembly principles.
The first example of shape amphiphiles was reported in 2003 when a disc-like
columnar mesogen was linked with a rod-like mesogen [56]. The former prefers to
stack one on top of another as a result of π–π interactions, whereas the latter align
with each other as a result of dipole–dipole interactions. Consequently, the disc–rod
amphiphiles promote the mixing of disc liquid crystal and rod liquid crystals.
However, the structure of such an amphiphile has not yet been revealed. Years
later, Glotzer and colleagues used computer simulations to predict the self-
assembly of various shape amphiphiles, including polymer-tethered nanoparticles
and nanorods. [26, 27]. Many novel and intriguing hierarchical structures were
predicted, which stimulated further research in this direction. Our group was among
the first of several groups to look into the crystallization of such amphiphiles. For
accurate control of the chemical structure, we first applied the concept of “click”
chemistry [58–60] to the synthesis of this class of shape amphiphiles. It is believed
that precise chemical structure is the basis for study of the physical behavior of
these novel materials. C60 and POSS were identified as the prototype molecular
nanoparticles because of their well-defined structure and the ease of their functiona-
lization [5]. Various shape amphiphiles were then designed and synthesized.
Fullerene is one type of carbon allotrope with wide-ranging applications
[61, 62]. It is known to be very hydrophobic and tends to aggregate randomly.
One often needs to chemically modify it to improve its compatibility with other
materials, but the formation of ordered structures is difficult [63]. To drive the
formation of ordered structures, the following aspects must be recognized: (1) C60
is a conformationally rigid, incompressible sphere; (2) when closed, it is impene-
trable to most atoms and functional groups and thus the structure is relatively
independent; and (3) it is likely to form aggregates as a result of strong π–π
interactions. Therefore, we realized that we could either take advantage of the
strong aggregation of C60 to interplay with other interactions to guide the formation
of ordered structures, or use even stronger interactions of other components to
override that of C60 and serve as template for the arrangement of C60. Alternatively,
we could use a conformationally rigid, incompressible, impenetrable counterpart of
C60 to assist the formation of an ordered structure.
We first investigated a class of plane–sphere shape amphiphiles [64–67]. Por-
phyrin is a known discotic mesogen when functionalized with multiple alkyl chains
of varying lengths. By linking donor porphyrin with acceptor C60, we hoped that
columnar liquid crystals would still be preferred. There are a variety of ways to
connect the two components with varying side chain lengths, linkages, and number
of C60 molecules. The molecules shown in Fig. 2a, d differ by the presence of two
long alkyl chains modification and were studied in detail. The results indicate that
the π–π stacking of porphyrin overrides the aggregation of C60 molecules, and the
formation of columnar liquid crystals persists in both cases [64–67]. The former
gives a hexagonal columnar phase with an unusual 12944 helical structure for each
column [67]. The C60s were found to interact within each column to form three
192 X.-H. Dong et al.

Fig. 2 Self assembly of plane–sphere shape amphiphiles: (A, D) chemical structures and
(B, C and E, F) corresponding supramolecular structures [64]

pendant, continuous helical channels along the column (Fig. 2). This is a typical
supramolecular double cable structure that contains parallel arrays of hole- and
electron-transporting channels. Upon removal of the alkyl groups on C60,
intercolumnar C60–C60 interactions take over and, hence, a rectangular columnar
liquid crystal phase in an orthorhombic unit cell forms (Fig. 2d–f) [64]. Neverthe-
less, in this case, the C60s form separate continuous domains parallel to the column
of porphyrin, which is still one type of supramolecular double cable structure.
Although their performances have yet to be optimized, preliminary device tests
have shown higher photovoltaic conversion efficiency than their phenyl-C61-
butyric acid methyl ester (PCBM) counterpart. Therefore, even for the same type
of shape amphiphiles, the self-assembled structures can be very different,
depending on the detailed primary structure. It should also be recognized that,
although precise prediction of physical structure is not possible, the key structural
features apparently do not change, as reflected in the supramolecular double cable
structure in both cases. It is anticipated that cycles of iteration are required to
understand the structure–property relationship and optimize the properties of the
material [68].
The strategy was also extended to other shape amphiphiles. For example, a series
of rod–sphere shape amphiphiles, C60-oligofluorene (OF) conjugates, were pre-
pared [69]. This time, the framework was held constant for the rod–sphere shape
amphiphile owing to a rigid spiro linkage. The only thing that changed was the
length of the alkyl side chains. Interestingly, all three crystal structures exhibit very
similar molecular packing schemes with alternating layers of the C60 and OF
Supramolecular Crystals and Crystallization 193

Fig. 3 Crystal structure of a rod–sphere shape amphiphile, OFn-C60: molecular packing model
projection on ab plane (A), bc plane (B), and ac plane (C) [69]

components (Fig. 3), although their symmetry is very different. It seems that,
although the interaction are both π–π interactions, their geometry is very different,
with spherical C60 being isotropic and rod-like OF being anisotropic. This imposes
different packing constraints, which leads to maximum contact between C60 mol-
ecules, but not OF conjugates. The alkyl chains fill in the gaps between C60 spheres,
thereby leaving more space for close packing of the rod-like planar OF components.
In the above examples, C60 acts as a 3D sphere and interacts with the rest of the
molecule as a relatively independent unit. We anticipate that this is also true for
other molecular nanoparticles.
Polyhedral oligomeric silsesquioxanes are perhaps the smallest of silicon
nanoparticles. Of these, T8 POSS possesses the highest Oh symmetry and is cube-
like. With a shape-persistent cage (diameter ~ 1.0 nm), it is an ideal inorganic
counterpart of C60 [70–74]. A cube–sphere shape amphiphile, POSS-C60, can be
easily designed and prepared by covalently linking POSS and C60 together via a
simple ester bond [34, 46, 47]. Upon crystallization, it exhibits polymorphism,
forming hexagonal and orthorhombic crystals [46]. In both cases, the double-
layered structure can be observed, because POSS and C60 are chemically incom-
patible but geometrically complementary to each other. Phase separation between
POSS and C60 is the very first step, after which the molecular nanoparticles in each
layer start to further organize into intriguing hierarchical structures. Interestingly,
because POSS is generally considered as insulating and C60 is regarded as
194 X.-H. Dong et al.

conducting or even superconducting upon doping, this unique structure can be

proposed as an as-assembled “nanocapacitor” [68]. Moreover, changing the molar
ratio of the components to two POSS to one C60 creates an asymmetric shape
amphiphile, which packs into a three-layered structure with interdigitated C60 as the
center sandwiched layer [47].
Replacing C60 with another POSS molecule and linking both POSS molecules
by a rigid ribbon creates interesting packing, as shown in a novel cube–plane–cube
shape amphiphile POSS-PDI-POSS [75]. There are two competing interactions, the
π-π stacking interaction and the close-packing of POSS. Because the POSS has a
larger volume and the steric hindrance is considerable, the stacking is frustrated
beyond dimer formation. As a balance, the two perylene diimide (PDI) molecules
are packed face-to-face to maximize the contact and the POSS cages bend away
from the plane and block the opposite side of the PDI. Finally, the dimer serves as
the basic building block for crystal construction. The principles hold true no matter
whether the linker between POSS and PDI is a rigid phenylene linkage or a
relatively flexible propylene linkage. It should be noted that in all of these crystals,
although the basic structural features are evident, the atomic order may not be as
precise as in small-molecule crystals. This is reflected in the relatively diffuse
diffraction spots of these crystals and, sometimes, by streaks caused by crystal
defects. Within these supramolecular crystals, the molecular nanoparticles are
relatively independent of the rest of the molecule. They interact with the rest
basically through the collective physical interactions from the functional groups
on the surface. Their inner structures and compositions are kept identical, which
allows us to at least qualitatively predict the molecular packing using the coarse-
grained model. This, in our opinion, is a valuable direction of research on supra-
molecular crystals.

4 Giant Janus Particles

Molecular nanoparticle (MNP)-based Janus molecules represent the simplest

two-component model systems in giant molecules. The word “Janus” comes from
the double-faced Roman god and reflects the asymmetric nature of particles with
two regions of distinct surface chemistry, polarity, and/or interactions [76–78]; the
name was first proposed by de Gennes in his 1991 Nobel lecture [79]. As a new
class of nano-Janus grains that bridge the gap between conventional inorganic/
polymeric colloidal Janus particles [80–82] and molecular Janus entities (i.e.,
amphiphilic dendrimers [16] or polymer brushes [83]), MNP-based Janus particles
(or so-called molecular Janus nanoparticles) usually possess 3D volume and shape-
persistent nanostructure with molecular precision and high uniformity. Fine tuning
of the molecular symmetry and surface chemistry of each MNP could efficiently
manipulate the hierarchical structure formation of the whole molecular Janus
particle in the solid state.
Supramolecular Crystals and Crystallization 195

The Janus feature requires breaks in both geometrical and chemical symmetry,
which serve as the driving force to mediate the self-assembly and hierarchical
structure formation of the molecule. The geometric sense describes the decrease
in overall molecular symmetry during the mono- and multifunctionalization of
MNPs [84]. The chemical sense refers to the incorporation of surface functionalities
with different interactions into the other part of the molecule to generate
amphiphilicity. In the family of molecular Janus nanoparticles, one of the simplest
ways to achieve symmetry breaking is to connect two or more immiscible MNPs
together via short covalent linkage(s).
Following this premise, a dumbbell-shaped molecular Janus particle was
designed and synthesized by closely coupling together two MNP units with similar
structure but distinct surface functional groups: carboxylic acid-functionalized
POSS (APOSS) and isobutyl-functionalized POSS (BPOSS) [85]. Because a short
linkage is used to connect these two immiscible POSS molecules together (hydro-
philic versus hydrophobic), the Oh symmetry of the T8 cage is reduced to C3v
symmetry along the long axis of the molecule. The amphiphilic feature of the whole
particle (APOSS-BPOSS) is thus created, allowing the formation of a solid-state
bilayered structure with a head-to-head, tail-to-tail type of packing arrangement
(Fig. 4). Moreover, the supramolecular crystalline packing of BPOSS particles
within each layer is also formed at low temperatures, further generating a hierar-
chical structure with an orthorhombic unit cell and a symmetry group of P21212.
This can be confirmed by selected area electron diffraction (SAED) patterns from
stacked single lamellar crystals with a flat-on arrangement (Fig. 4) and further
illustrated by computer simulation. Each atomic position of the motif does not
possess 3D translational symmetry, yet the unit cell does. This structure is thus a
typical supramolecular crystal. Furthermore, the ordered crystal structure in the
layers disappears upon heating as a typical first-order transition, whereas the
bilayered structure persists throughout as a supramolecular liquid crystal phase.
Notably, the formation of such a bilayered structure with possible supramolec-
ular ordering at the meso-length scale is highly favored for the 3D packing of Janus
particles, provided the volume/shape and the coverage of surface functionalities are
commensurate. Introducing incommensurate factors into the Janus particle system
can offer opportunities for construction of new frustrated supramolecular structures
beyond the bulk supramolecular crystals [86]. For example, introduction of a strong
electrostatic repulsive force between the amphiphilic layers, which prevents crystal
growth along the lamellar normal direction in a polar solvent (i.e., dimethyl-
formamide, methanol, and acetonitrile), can result in the final formation of ultrathin
2D supramolecular crystals with extra-large specific area ratio (Fig. 4). Specifically,
the collective hydrogen bonding within the APOSS layers is partially suppressed by
titrating with tetraalkylammonium hydroxide (TBAOH) to a certain molar ratio.
Well-defined 2D supramolecular crystals with definite and uniform thickness can
be formed by slow evaporation of the resulting solution. Atomic force microscopy
(AFM) shows an average thickness that is nearly equal to the c-axes of the unit cell
of BPOSS-APOSS bulk crystals without counterions. The structure of 2D supra-
molecular crystals of BPOSS-APOSS has been revealed by SAED. They consist of
196 X.-H. Dong et al.

Fig. 4 Self-assembly of APOSS-BPOSS molecular Janus particles into 3D stacked lamellae

crystal or 2D nanosheets under different conditions: (a) in the solid state or in melt [85] and (b)
in polar solvent or mixing with tetraalkylammonium hydroxide [86]

two inner BPOSS crystalline layers that are sandwiched by two outer APOSS layers
covered by counterions. In polar solvents, solvation of the counterions creates
partially charged 2D supramolecular crystals and, thus, generates electrostatic
repulsive interactions that block supramolecular crystal growth along the layer
normal direction. The role of counterions has been illustrated by tuning both their
number and size. Only a certain degree of neutralization of carboxylic groups can
result in well-defined 2D supramolecular crystals.
BPOSS-C60 conjugates, another typical molecular Janus nanoparticle, self-
assembles into a double-layered lamellar supramolecular crystal as a result of the
immiscible nature of BPOSS and C60 with similar sizes [46]. Interestingly,
depending on growth conditions, polymorphism with two different crystal struc-
tures was observed as a result of the distinct packing ordering sequence of those two
MNPs at the initial stage. For example, an orthorhombic unit cell is formed when
BPOSS cages initiate the crystallization process, whereas a hexagonal unit cell is
generated when C60 packs into the first layer and dominates the supramolecular
structure formation (Fig. 5A). The self-assembled alternating C60 and POSS layers
along the c-axis of the crystals with d-spacing of several nanometers finally induce
the bilayered structure in both crystal forms. The hexagonal lattice is the thermo-
dynamically more stable phase, based on the supramolecular crystal packed struc-
ture and density difference.
Supramolecular Crystals and Crystallization 197

Fig. 5 (A) Self-assembled structures of BPOSS-C60 [46]: (a) chemical structure of BPOSS-C60;
(b) molecular packing in orthorhombic crystal and (c) the corresponding morphology; (d ) molec-
ular packing in crystal hexagonal crystal and (e) the corresponding morphology. (B) Self-
assembled structures of diBPOSS-C60 [47]: (a) chemical structure of diBPOSS-C60 and (b)
molecular packing in orthorhombic crystal

We further extended the dumbell-like Janus grains into “Mickey Mouse”-like

molecular Janus particles via the development of a diBPOSS-C60 shape amphiphile
that consists of one C60 covalently linked with two BPOSS MNPs [47]. Considering
the incommensurate volume of two immiscible domains (two BPOSS versus one
C60), the formation of a conventional bilayered structure is not possible. Alterna-
tively, the unbalanced volume of two BPOSS domains and one C60 domain could
induce the formation of a “one-and-half-layered” supramolecular packing scheme,
whereby one single layer of interdigitated C60 molecules is sandwiched between
two layers of BPOSS cages (Fig. 5B). Such a packing scheme along the c-axis could
generate a 3D orthorhombic supramolecular crystal lattice with Pnnm symmetry.
Another group of MNPs, POMs, are also utilized to build up molecular Janus
particles as a result of their intrinsic electrostatic interactions and size comparable
with other preceding MNPs. POMs are defined as discrete metal-oxygen anionic
clusters and are composed of transition and/or actinide metals in their high oxida-
tion states with oxygen atoms to be shared. A simple example is BPOSS-Lindqvist
POM (see Fig. 6a), where Lindqvist POM is a molybdenum-oxygen cluster with
two negative charges and covalently bonded to BPOSS via the Sonogashira cou-
pling reaction [86]. When slowly evaporated from low polarity solvent (e.g.,
acetone), the molecular dyad crystallizes into a 3D triclinic supramolecular lattice
198 X.-H. Dong et al.

Fig. 6 (A) Chemical structure of BPOSS-Lindqvist POM. (B) Stacked 3D crystals grown from
acetone solution. (C) Well-defined 2D nanocrystals grown from acetonitrile solution [86]

that is dictated by the BPOSS cages. However, when crystallized from polar
solvents (e.g., acetonitrile, DMF, and methanol), the counterions of Lindqvist
POMs are solvated and prevent 3D crystal growth (see Fig. 6b). As a result, 2D
supramolecular crystals with definite and uniform size of several nanometers are
generated (Fig. 6c).

5 Giant Surfactants

Giant surfactants are a subcategory of giant molecules and consist of MNPs as head
(s) and polymer chains as tail(s). Thus, they contain the essential geometrical
features of small-molecule surfactants yet with amplified size [5, 6]. In contrast
to giant shape amphiphiles and Janus particles, the conformational entropy origi-
nating from the flexible nature of polymeric tails plays an important role in the
system. In this sense, giant surfactants possess the duality of both small molecule
surfactants and block polymers [5]. Although the general principles of phase
separation and crystallization of (block) polymers are beyond the scope of this
review, giant surfactants possess their own unique characteristics during the for-
mation of supramolecular crystals. First, clustering of functional groups at the
periphery of the MNPs generates collective secondary interactions that act cumu-
latively and cooperatively and are indispensable for the construction and stabiliza-
tion of the supramolecular crystals. Second, the presence of shape-persistent MNPs
imposes external constrains on the polymer tails during supramolecular
Supramolecular Crystals and Crystallization 199

Fig. 7 (A) SAXS pattern and (B) bright field TEM images of ordered phases (from left to right:
Lam, DG, Hex, and BCC) from DPOSS-PS. The phase diagram shown at the bottom [36]

crystallization, which alters the pathway of Gibbs free energy minimization and
leads to unconventional structures with varied metastabilities. Third, the flexible
polymer tails provide more packing freedom than available in giant shape amphi-
philes and Janus particles, resulting in highly diverse, thermodynamically stable/
metastable ordered structures that behave similarly to block polymers.
These unique characteristics are illustrated by a representative giant surfactant
with a hydroxyl-functionalized POSS (DPOSS) head and a polystyrene (PS) tail
(DPOSS-PS) [36]. In the bulk, diverse supramolecular ordered structures, including
lamellae (Lam), bicontinuous double gyroids (DG), hexagonally packed cylinders
(Hex), and body-centered cubic packed spheres (BCC), are observed with increas-
ing volume fraction of the PS tail, as evidenced by small-angle X-ray scattering
(SAXS) patterns and transmission electron microscopy (TEM) images (Fig. 7). The
structures are analogous to the classic phases of block copolymers, yet with much
smaller domain spacings. The collective hydrogen bonding between DPOSS mol-
ecules enhances the interaction parameters, which provides a versatile platform for
engineering technologically relevant nanostructures with feature sizes of less than
10 nm. It is worth noting that, although the molecular structure of DPOSS-PS
prohibits access to the phase diagram for a volume fraction of PS less than 0.64 as a
result of the fixed size of one block, new molecular designs with multiple heads
could be applied to explore the other half of the phase diagram [36].
Another remarkable feature is that the supramolecular structures constructed by
giant surfactants are very sensitive to the molecular architecture, which is signifi-
cantly weakened in the case of block copolymers as a result of the elasticity and
compressibility of random coils [36, 40]. We compared the phase behaviors of
200 X.-H. Dong et al.

topological isomer pairs with identical chemical composition but different archi-
tectures to highlight this sensitivity. For example, a bicontinuous DG supramolec-
ular structure was observed in DPOSS-PS35 with a single PS tail, but a Hex phase
was seen for the corresponding topological isomer with the same PS volume
fraction but two tails (DPOSS-2PS17) [36]. It is thus not sufficient to use a single
order parameter (volume fraction) to describe the phase behavior of giant surfac-
tants. Additional parameters associated with the molecular architecture should also
be considered.
Interestingly, the sensitivity of molecular architecture is also observed in solu-
tion self-assembly. Unusual nanostructured colloidal particles have been observed
in a topological isomer pair consisting of a carboxylic acid-functionalized C60
(AC60) head and PS tail(s) (AC60-PS44 and AC60-2PS23) by slow addition of a
selective solvent into their solution in 1,4-dioxane [36]. Spherical colloidal particles
were observed for AC60-PS44, but double truncated conical particles for AC60-
2PS23 (Fig. 8). Both showed unimodal narrow size distribution, as confirmed by
dynamic light scattering (DLS). A zoom-in characterization by TEM showed that
the spherical particles have an onion-like inner structure, whereas the double
truncated conical particles exhibit a hexagonal inner structure (Fig. 8c, f), which
is in good agreement with their corresponding bulk structures (Lam for AC60-PS44
and Hex for AC60-2PS23). The surface of these particles is covered mainly by
anionic AC60, and the inner part undergoes further self-organization into different
finer nanostructures through phase separation between AC60 and PS tails. The
formation of these colloidal particles is a result of strong collective interactions
and conformational rigidity of MNPs, which is not observed in traditional block
copolymer self-assembly.

Fig. 8 SEM and TEM images of colloidal nanoparticles of (A–C) AC60-PS44 and (D–F) AC60-
2PS23 under similar conditions [36]
Supramolecular Crystals and Crystallization 201

Fig. 9 (A) Illustration and chemical structure of the FPOSS-based giant surfactant. (B–D) The
resultant nanostructures in solution: 2D hexagonally patterned nanosheets (B, C) and laterally
structured vesicles (D) [35]

Large 2D hexagonally patterned colloidal nanosheets with a size of tens of

micrometers were observed in solution assembly of a giant surfactant with a
fluorinated polyhedral oligomeric silsesquioxane (FPOSS) head and a polysty-
rene-block-poly(ethylene oxide) (PS-b-PEO) diblock copolymer tail (FPOSS-PS-
b-PEO) (Fig. 9) [35]. With vigorous stirring, the strong shear force interrupts the
formation of large sheets, resulting in small debris with faceted or hexagonal shapes
(Fig. 9c). According to the relative hydrophobicity of the three components, the
nanosheets are 2D layers with hexagonally connected FPOSS domains embedded
inside a continuous PS matrix that is further covered by the PEO corona. The
nanosheets have a thicker boundary than the interior, which partially releases the
excess rim-cap energy and stabilizes this unusual structure. Further increasing the
water content induces the formation of vesicles with hexagonally patterned walls
(Fig. 9d).
The rigid conformation of MNPs imposes packing constraints on the polymer
tails, leads to intriguing unconventional supramolecular structures that are usually
inaccessible for free polymers. A typical example is the observation of exactly
defined half-stemmed polymer lamellar supramolecular crystals with polymer
chain-ends remaining trapped in the middle of the lamellar crystal as defects.
These structures have been obtained using giant surfactants having a crystalline
PEO tail (PEO-POSS and PEO-C60) (Fig. 10) [32]. Upon crystallization from the
melt, PEO is sandwiched between two layers of MNPs. To balance the cross-
sectional areas of the MNPs and the PEO stems, half-stemmed crystals with
precisely controlled defects were formed, as evidenced by lamellar thickness and
melting temperature (Fig. 10c, d). Note that the half-stemmed crystals are
202 X.-H. Dong et al.

Fig. 10 (A) Chemical structure of PEO-C60 and PEO-POSS. (B) Illustration of integral folding
and half-stemmed crystals. (C) Relationship between lamellar thickness, L, and crystallization
temperature Tx; insets show the corresponding SAXS pattern. (D) Relationship between melting
temperature, Tm, and crystallization temperature Tx; insets show ultrafast heating chip DSC
thermograms [32]

thermodynamically unstable and further reorganize into integral folded crystals.

With the geometric restrictions imposed by MNPs, these half-stemmed PEO crys-
tals are settled in free-energy minima and experimentally accessible.
Furthermore, the geometric constraints and selective affinities of MNPs could
promote phase separation between two tail blocks in giant surfactants with block
copolymer tails. Giant surfactants with a low molecular weight PS-b-PEO diblock
copolymer tail (FPOSS-PS-b-PEO and AC60-PS-b-PEO) could self-assemble into a
library of diverse ordered structures in the bulk [87, 88]. The FPOSS-based giant
surfactants form lamellae with alternating MNP domain and block polymer domain
[87]. We found that the fixed volume and shape of FPOSS generates a spatially
confined environment for tethered PS-b-PEO tails. The incommensurate cross-
sectional area between the MNP head and polymer tail results in chain
overcrowding and stretching, which provides additional entropic driving force to
facilitate the phase separation of block copolymer tails. The chain stretching leads
to unexpected phase separation between low molecular weight PS and PEO blocks,
and subsequent formation of hierarchical lamellar structures among the three
immiscible components. On the other hand, the hydrogen bonding between the
AC60 and the PEO blocks also facilitates nanophase separation of originally
disordered low molecular weight PS-b-PEO block copolymers, affording the
Supramolecular Crystals and Crystallization 203

formation of various ordered nanostructures, including Lam, DG, and Hex phases
[88]. In these ordered structures, the AC60 and PEO blocks are associated with each
other in one domain and the PS blocks segregate into another. These examples
provide an efficient and practical strategy for the design and preparation of giant
surfactants for the construction of ordered nanostructures for technologically rele-
vant applications.

6 Giant Tetrahedra

For the concept of supramolecular crystals, there is no example more convincing

than the case of giant tetrahedral molecules. The assembly/packing of building
blocks with specific shape and symmetry in 3D space is a long-lasting topic in
scientific research. Typical polyhedra are among the most intriguing 3D structures.
Figure 11 shows three different types of tetrahedra as examples.
The first type includes classic tetrahedra (type I in Fig. 11), which have closed
structures with flat faces, straight edges, and sharp corners or vertices. Glotzer and
coworkers simulated hundreds of convex polyhedra whose assembly arises solely
from their anisotropic shape, and developed simple criteria to predict how polyhe-
dra pack into ordered structures such as crystals, liquid crystals, and quasicrystals,
or remain amorphous [24, 89]. Meanwhile, the synthetic advances in generation of
inorganic polyhedral nanocrystals (e.g., tetrahedra, cubes, truncated cubes,
cuboctahedra, octahedra, and rhombic dodecahedra) with good control of size,
shape, and surface chemistry have provided some simple illustrations of the
entropy-driven packing of polyhedra [90–94].
The second type of polyhedra includes closed framework structures, which have
straight edges and sharp corners or vertices but are without flat faces (type II in
Fig. 11). With the aid of coordination interactions using transition metals as
“nodes” and organic ligands as “spaces,” many complex 3D topologies possessing
polyhedral geometries or symmetries have been revealed. The development of
elaborate design principles during the last few decades has led to diverse polyhedral
structures, ranging from simple Platonic solids including tetrahedron, octahedron,

Fig. 11 Three different types of tetrahedra building blocks

204 X.-H. Dong et al.

Fig. 12 Chemical structures of giant tetrahedra 1–4. The cartoons in the top left are the
corresponding simplified schemes [38]

cube, dodecahedron, and icosahedron, to more complex Archimedean solids [95–

97].
The third type of polyhedral structure possesses polyhedral symmetry with only
sharp corners or vertices, all connected to one center (type III in Fig. 11). The best
example of this type of polyhedra is the methane, CH4, structure with a tetrahedral
shape. The colloidal nanotetrapods or nano-octapods also belong to this type
[98, 99]. This type of polydedra can be easily constructed starting from nanoatoms.
If nanoatoms are placed on the apexes of a rigid polyhedron linker to form a larger
faceted giant molecule, the molecules amplify the symmetry of the linkers and
result in giant polyhedra, reminiscent of the classic small-molecule valence shell
electron pair repulsion structures. Of all the polyhedra, the tetrahedron is the
simplest to start with. Recently, we presented an experimental approximation of
giant tetrahedra by coupling four POSS MNPs with different surface functionalities
to a rigid tetrahedral core [38]. These giant tetrahedra have precise and uniform
molecular weights. By introducing different numbers of hydrophilic and hydropho-
bic POSS NMPs, it was possible to achieve symmetry breaking of the positional
interactions in giant tetrahedra generated by the collective (multiple) hydrogen
Supramolecular Crystals and Crystallization 205

Fig. 13 SAXS patterns of structures 2a (A), 3c (B), 3a (C), and 4b (D) after corresponding
thermal treatments. The corresponding BF TEM images of thin-sectioned samples are shown
below; insets show the corresponding fast Fourier transform patterns [38]

bonding between the hydrophilic POSS MNPs and crystallization of the hydropho-
bic POSS MNPs. Detailed chemical structures of these giant tetrahedra are shown
in Fig. 12 (structures 1–4).
In the giant tetrahedron 1 with four hydrophobic POSS MNPs, tetrahedral cores
adopt an interpenetrated stacking manner to form geometrically locked columns
along the c-axis, surrounded by a shell of crystalline BPOSS cages. When one
hydrophobic BPOSS is replaced by a hydrophilic POSS in giant tetrahedra 2, there
exist the two competing interactions of collective hydrogen bonding between the
hydrophilic POSS MNPs and crystallization of the hydrophobic BPOSS MNPs. At
room temperature, the latter dominates and forces the giant tetrahedra into a
frustrated three-layer packing mode, as crystallization of BPOSS MNPs always
prefers to create flat interfaces. After melting the crystals, the hydrophilic DPOSS
MNPs aggregate via collective hydrogen bonding to form the spherical core. The
hydrophobic BPOSS MNPs undergo a 2D scrolling and construct shells of the
spheres located on the surfaces to fit into an A15 supramolecular crystal lattice, as
demonstrated by both SAXS and bright field (BF) TEM observations (see
Fig. 13A). Figure 14a illustrates the assembly mechanism. It is surprising that
two types of supramolecular spheres with different coordination environments
can be resolved clearly under TEM (Fig. 13A). It is speculated the Frank–Kasper
A15 supramolecular lattice formation must be associated with the deformability of
the selectively assembled spheres and shape polydispersity.
The giant tetrahedra of series 3 are more symmetric in terms of both volume
fractions and interactions. Above the melting temperature of crystalline BPOSS,
double layered Lam (see Fig. 13B for the SAXS and TEM analyses, and Fig. 14b
for a schematic illustration of the mechanism) or DG (Fig. 13C and Fig. 14b)
supramolecular structures could be identified. However, in giant tetrahedra 4,
206 X.-H. Dong et al.

Fig. 14 Selective assembly mechanisms and packing models for giant tetrahedra 2 (A), 3 (B),
and 4 (C) [38]

BPOSS MNPs fail to crystallize at room temperature as a result of the low BPOSS
volume fraction. After annealing, an inverse Hex structure (Fig. 13D and Fig. 14c)
having hydrophilic POSS MNPs in the continuous matrix and hydrophobic BPOSS
MNPs in the core domain appears, rather than a spherical supramolecular structure
as in the giant tetrahedra 2. This use of rigid giant polyhedra as building blocks to
construct diverse supramolecular crystal lattices including the “metal alloy”-like
Frank–Kasper structure opens up a wide field of supramolecular crystals with
unexpected structures and properties.

7 Thin Films of Giant Molecules

The diverse self-assembly and crystallization behaviors of giant molecules in the

bulk state have been described extensively in the previous sections. Similar to the
concept that adding nanoatoms to a macromolecular system imposes an additional
dimension for controlling supramolecular structure, factors such as surface tension
and substrate interactions could further enrich the self-assembled structures of giant
molecules in thin films.
Self-assembly in thin films with only a few repeating motifs of supramolecular
lattices along the film normal is greatly affected by surface conditions [100–
102]. To minimize the free energy, the components with low surface energy
preferentially emerge to the free surface, whereas components with affinity for
substrates wet the substrate [103]. The subtle balance of enthalpy and entropy that
determines the lowest free energy state can be manipulated in thin films by tuning
Supramolecular Crystals and Crystallization 207

Fig. 15 Molecular structures of giant surfactants (A) AC60-PS-PEO, (B) FPOSS-PS-PEO, (C)
PS-AC60-PEO, and (D) PS-FPOSS-PEO. Hydrophilic components are labeled in blue, hydropho-
bic components in red, and omniphobic components in green (C.-H. Hsu et al., unpublished work)

the secondary interactions between molecules and substrate and the molecular
confinement close to surfaces.
Supramolecular crystallization of giant surfactants with a PS-b-PEO block
copolymer have been studied in thin films (Fig. 15) (C.-H. Hsu et al., unpublished
work). Using rational molecular design, four categories of giant surfactants with
distinct molecular topologies and different MNP surface functionalities were
constructed as a comprehensive model system. Two MNPs with different affinities
to the tail blocks (i.e., hydrophilic AC60 and omniphobic FPOSS) are utilized as
heads of the giant surfactants. The MNP is tethered to a PS-b-PEO block copolymer
at two specific positions, the end of the PS block (MNP-PS-PEO) and the junction
point [PS-(MNP)-PEO], resulting in topological isomer pairs with almost
identical chemical compositions but different architectures. The molecular weights
of PS-b-PEO were controlled within the weak segregation region to highlight the
enhanced immiscibility between polymer blocks in the presence of MNPs.
The self-assembly behaviors of specifically designed giant surfactants are sys-
tematically studied in thin films using grazing incidence X-ray scattering and TEM,
focusing on the effects of head surface functionalities and molecular architectures
on nanostructure formation. With fixed length of the PEO block, changing the
molecular weight of the PS block leads to phase formation and transition. As a
result of the distinct affinity, the AC60-based giant surfactants form two-component
208 X.-H. Dong et al.

Fig. 16 Ternary phase diagram of three-component giant surfactant self-assembly in thin films.
Self-assemblies of AC60-based and FPOSS-based giant surfactants are depicted by blue and green
lines, respectively. Each self-assembled phase is represented by the colored model and supported
by the corresponding grazing incidence X-ray diffraction pattern. Solid lines represent the linear
giant surfactants and dashed lines represent star-like surfactants. The black dashed line indicates
compositions with equal volume fraction of MNP and PEO (C.-H. Hsu et al., unpublished work)

morphologies, whereas three-component morphologies are found in the FPOSS-

based surfactants. A library of interesting morphologies is observed. For linear
giant surfactants, the AC60-based giant surfactants show a transition sequence of
LAM, modulated lamellae (ML), HEX, and BCC structures, whereas three-
component undulated lamellae (UL3), three-component lamellae (LAM3), perfo-
rated lamellae (PL3), and cylinder-within-perforated lamellae (CPL3) structures are
sequentially observed in the FPOSS-based samples, with increasing the molecular
weight of the PS block (Fig. 16) (C.-H. Hsu et al., unpublished work). Within these
structures, AC60s are distributed in the PEO domain without long range positional
order, whereas FPOSS forms separated domains with mesomorphic packing. The
topological isomers of AC60-based giant surfactants exhibit different dimensions
yet similar or identical morphologies; however, those FPOSS-based giant surfac-
tants show strong topological dependence on the stable morphologies. A stretching
parameter of the PS block has been utilized to interpret and characterize the phase
transitions (C.-H. Hsu et al., unpublished work). The versatile self-assembled
morphologies suggest that giant surfactants are an excellent platform for producing
well-controlled supramolecular structures, and that topology can serve as an addi-
tional crucial factor for finely tuning the size and the geometry of structures.
Supramolecular Crystals and Crystallization 209

8 Concluding Remarks

In summary, we have collected as many experimental observations of ordered

supramolecular structures via self-assembly processes as possible and introduced
the concept of supramolecular crystals and crystallization. This concept extends
conventional crystals and crystallization to a broader research field and includes
ordered structures with crystallographic building blocks within which atoms in
motifs might not possess classic positional, bond-orientational, and molecular-
orientational long-range order. Various examples are given of different types of
giant molecules, including giant shape amphiphiles, Janus particles, giant surfac-
tants, and giant tetrahedra self-assembled (“crystallized”) in bulk, solution, and film
states. Although they undergo different assembly processes in forming ordered
structures (supramolecular crystallization), the classic phase transformation theory
and concepts are still the guiding principles for these formation mechanisms. The
central issue is the following: What are the features of the packing scheme in these
supramolecular crystals of giant molecules with nanosized motifs? First, within
their motifs, long-range ordered atomic positions with translational symmetry are
not required. The motifs are the smallest repeating building blocks in the crystals.
Second, topologies of the giant molecules are crucially important in the formation
of these supramolecular crystals as a result of their relatively rigid molecular shape
persistency as individual packing units. Third, the design and introduction of
collective secondary interactions can stabilize crystal structures with nanosized
motifs and ordered structures. Fourth, the selective assembly process leads to
stepwise structure formation and makes modular self-organization possible at
different length scales. It is expected that the investigation of ordered supramolec-
ular structures can also be extended from supramolecular crystals to supramolecular
liquid crystals, plastic crystals, and quasicrystals. For a broader perspective, in
parallel with finding new phase structures, there must be further developments in
many research fields in chemistry and physics. For example, in chemistry, we need
to improve synthetic ability from “well-defined” to “precisely defined” chemistry.
Traditionally, one starts to construct a material on the basis of an atomic platform.
How do we think of a modular approach using building blocks (nanoatoms) to
construct our materials at the nanoscale or even higher length scales, in addition to
the classic random-coil polymers? How do we design structural and functional
synthons to transfer and amplify molecular functions to macroscopic properties? In
physics, how do we describe a non-equilibrium state during the phase transitions?
How do we deal with an entropy-dominated transformation? When we describe a
phase transition, we always think that it is in an infinitely large space. If the space
becomes limited and even smaller than the lower limit for application of statistical
mechanics, what could we do? Finally, we are at present studying the structure and
properties of materials at one length and time scale. How could we expand our
understanding to multiple lengths, space and time scales? All of these questions
need to be answered through the joint efforts of chemists, physicists, bioscientists,
life scientists, and engineers.
210 X.-H. Dong et al.

Acknowledgments This work was supported by the National Science Foundation

(DMR-1408872). Use of the Advanced Photon Source at Argonne National Laboratory was
supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences,
under contract DE-AC02-06CH11357.

References

1. Staudinger H (1920) Ber Dtsch Chem Ges 53:1073

2. Odian G (2004) Principles of polymerization, 4th edn. Wiley-Interscience, New York
3. Feynman RP (1960) Eng Sci 23:22
4. Ober CK, Cheng SZD, Hammond PT, Muthukumar M, Reichmanis E, Wooley KL, Lodge TP
(2009) Macromolecules 42:465
5. Zhang W-B, Yu X, Wang C-L, Sun H-J, Hsieh I-F, Li Y, Dong X-H, Yue K, van Horn RM,
Cheng SZD (2014) Macromolecules 47:1221
6. Yu X, Li Y, Dong X-H, Yue K, Lin Z, Feng X, Huang M, Zhang W-B, Cheng SZD (2014) J
Polym Sci B Polym Phys 52:1309
7. Zhang W-B, Cheng SZD (2015) Chin J Polym Sci 33:797
8. Tomalia DA, Christensen JB, Boas U (2012) Dendrimers, dendrons, and dendritic polymers:
discovery, applications, and the future. Cambridge University Press, Cambridge
9. Newkome GR, Moorefield CN, Vogtle F (2001) Dendrimers and dendrons: concepts, syn-
theses, applications. Wiley, Weinheim
10. Fréchet JMJ, Tomalia DA (2001) Dendrimer and other dendritic polymers. Wiley, West
Sussex
11. Roovers J, Comanita B (1999) Dendrimers and dendrimer-polymer hybrids. In: Roovers J
(ed) Branched polymers I, vol 142, Adv. Polym. Sci. Springer, Berlin, pp 179–228
12. del Barrio J, Oriol L, Sanchez C, Serrano JL, Di Cicco A, Keller P, Li MH (2010) J Am Chem
Soc 132:3762
13. Dong X-H, Lu X, Ni B, Chen Z, Yue K, Li Y, Rong L, Koga T, Hsiao BS, Newkome GR, Shi
A-C, Zhang W-B, Cheng SZD (2014) Soft Matter 10:3200
14. Percec V, Imam MR, Peterca M, Leowanawat P (2012) J Am Chem Soc 134:4408
15. Rosen BM, Wilson CJ, Wilson DA, Peterca M, Imam MR, Percec V (2009) Chem Rev
109:6275
16. Percec V, Wilson DA, Leowanawat P, Wilson CJ, Hughes AD, Kaucher MS, Hammer DA,
Levine DH, Kim AJ, Bates FS, Davis KP, Lodge TP, Klein ML, DeVane RH, Aqad E, Rosen
BM, Argintaru AO, Sienkowska MJ, Rissanen K, Nummelin S, Ropponen J (2010) Science
328:1009
17. Fujita M, Oguro D, Miyazawa M, Oka H, Yamaguchi K, Ogura K (1995) Nature 378:469
18. Tominaga M, Suzuki K, Kawano M, Kusukawa T, Ozeki T, Sakamoto S, Yamaguchi K,
Fujita M (2004) Angew Chem Int Ed 43:5621
19. Branda N, Wyler R, Rebek J (1994) Science 263:1267
20. Hartmann-Thompson C (2011) Applications of polyhedral oligomeric silsesquioxanes.
Springer, Dordrecht
21. Coronado E, Gomez-Garcia CJ (1998) Chem Rev 98:273
22. Katsoulis DE (1998) Chem Rev 98:359
23. Dolbecq A, Dumas E, Mayer CR, Mialane P (2010) Chem Rev 110:6009
24. Damasceno PF, Engel M, Glotzer SC (2012) Science 337:453
25. Horsch M, Zhang Z, Glotzer S (2005) Phys Rev Lett 95:056105
26. Glotzer SC, Horsch MA, Iacovella CR, Zhang Z, Chan ER, Zhang X (2005) Curr Opin
Colloid Interface Sci 10:287
27. Glotzer SC, Solomon MJ (2007) Nat Mater 6:557
Supramolecular Crystals and Crystallization 211

28. Bratschitsch R, Leitenstorfer A (2006) Nat Mater 5:855

29. Kastner MA (1993) Phys Today 46:24
30. Dmitrii F, Rosei F (2007) Angew Chem Int Ed 46:6006
31. Roy X, Lee C-H, Crowther AC, Schenck CL, Besara T, Lalancette RA, Siegrist T, Stephens
PW, Brus LE, Kim P, Steigerwald ML, Nuckolls C (2013) Science 341:157
32. Dong X-H, van Horn R, Chen Z, Ni B, Yu X, Wurm A, Schick C, Lotz B, Zhang W-B, Cheng
SZD (2013) J Phys Chem Lett 4:2356
33. Su H, Zheng J, Wang Z, Lin F, Feng X, Dong X-H, Becker ML, Cheng SZD, Zhang W-B, Li
Y (2013) ACS Macro Lett 2:645
34. Lin Z, Lu P, Hsu CH, Yue K, Dong XH, Liu H, Guo K, Wesdemiotis C, Zhang WB, Yu X,
Cheng SZD (2014) Chem Eur J 10:11630
35. Ni B, Huang M, Chen Z, Chen Y, Hsu CH, Li Y, Pochan D, Zhang WB, Cheng SZD, Dong
XH (2015) J Am Chem Soc 137:1392
36. Yu X, Yue K, Hsieh IF, Li Y, Dong XH, Liu C, Xin Y, Wang HF, Shi AC, Newkome GR, Ho
RM, Chen EQ, Zhang WB, Cheng SZD (2013) Proc Natl Acad Sci USA 110:10078
37. Wang Z, Li Y, Dong X-H, Yu X, Guo K, Su H, Yue K, Wesdemiotis C, Cheng SZD, Zhang
W-B (2013) Chem Sci 4:1345
38. Huang M, Hsu C-H, Wang J, Mei S, Dong X, Li Y, Li M, Liu H, Zhang W, Aida T, Zhang
W-B, Yue K, Cheng SZD (2015) Science 348:424
39. Yu X, Zhong S, Li X, Tu Y, Yang S, van Horn RM, Ni C, Pochan DJ, Quirk RP,
Wesdemiotis C, Zhang W-B, Cheng SZD (2010) J Am Chem Soc 132:16741
40. Yu X, Zhang WB, Yue K, Li X, Liu H, Xin Y, Wang CL, Wesdemiotis C, Cheng SZD (2012)
J Am Chem Soc 134:7780
41. Dong X-H, Zhang W-B, Li Y, Huang M, Zhang S, Quirk RP, Cheng SZD (2012) Polym
Chem 3:124
42. Li Y, Dong X-H, Guo K, Wang Z, Chen Z, Wesdemiotis C, Quirk RP, Zhang W-B, Cheng
SZD (2012) ACS Macro Lett 1:834
43. Zhang W-B, Li Y, Li X, Dong X, Yu X, Wang C-L, Wesdemiotis C, Quirk RP, Cheng SZD
(2011) Macromolecules 44:2589
44. Ni B, Dong X-H, Chen Z, Lin Z, Li Y, Huang M, Fu Q, Cheng SZD, Zhang W-B (2014)
Polym Chem 5:3588
45. Yue K, Liu C, Guo K, Wu K, Dong X-H, Liu H, Huang MJ, Wesdemiotis C, Cheng SZD,
Zhang W-B (2013) Polym Chem 4:1056
46. Sun H-J, Tu Y, Wang C-L, van Horn RM, Tsai C-C, Graham MJ, Sun B, Lotz B, Zhang W-B,
Cheng SZD (2011) J Mater Chem 21:14240
47. Lin M-C, Hsu C-H, Sun H-J, Wang C-L, Zhang W-B, Li Y, Chen H-L, Cheng SZD (2014)
Polymer 55:4514
48. Benz K-W, Neumann W (2014) Introduction to crystal growth and characterization. Wiley,
Weinheim
49. De Graef M, Mchenry ME (2012) Structure of materials. Cambridge University Press,
Cambridge
50. Cheng SZD (2008) Phase transitions in polymers the role of metastable states. Elsevier,
Amsterdam
51. Li CY, Cheng SZD, Ge JJ, Bai F, Zhang JZ, Mann IK, Harris FW, Chien L-C, Yan D, He T,
Lotz B (1999) Phys Rev Lett 83:4558
52. Li CY, Cheng SZD, Ge JJ, Bai F, Zhang JZ, Mann IK, Chien L-C, Harris FW, Lotz B (2000) J
Am Chem Soc 122:72
53. Lotz B, Cheng SZD (2005) Polymer 46:577
54. Cheng SZD, Lotz B (2003) Philos Transact A Math Phys Eng Sci 361:517
55. Cheng SZD, Lotz B (2005) Polymer 46:8662
56. Date RW, Bruce DW (2003) J Am Chem Soc 125:9012
57. Yang Y, Zimmt MB (2015) J Phys Chem B 119:7740
58. Hoyle CE, Bowman CN (2010) Angew Chem Int Ed 49:1540
212 X.-H. Dong et al.

59. Kolb HC, Finn MG, Sharpless KB (2001) Angew Chem Int Ed 40:2004
60. Meldal M, Tornøe CW (2008) Chem Rev 108:2953
61. Kadish KM, Ruoff RS (2000) Fullerenes: chemistry, physics, and technology. Wiley-
Interscience, New York
62. Martin N, Giacalone F (2009) Fullerene polymers: synthesis, properties and applications.
Wiley-VCH, Weinheim
63. Hirsch A, Brettreich M (2005) Fullerenes: chemistry and reactions. Wiley-VCH, Weinheim
64. Wang CL, Zhang WB, van Horn RM, Tu Y, Gong X, Cheng SZD, Sun Y, Tong M, Seo J, Hsu
BB, Heeger AJ (2011) Adv Mater 23:2951
65. Wang CL, Zhang WB, Yu X, Yue K, Sun HJ, Hsu CH, Hsu CS, Joseph J, Modarelli DA,
Cheng SZD (2013) Chem Asian J 8:947
66. Wang C-L, Zhang W-B, Hsu C-H, Sun H-J, van Horn RM, Tu Y, Anokhin DV, Ivanov DA,
Cheng SZD (2011) Soft Matter 7:6135
67. Wang C-L, Zhang W-B, Sun H-J, van Horn RM, Kulkarni RR, Tsai C-C, Hsu C-S, Lotz B,
Gong X, Cheng SZD (2012) Adv Energy Mater 2:1375
68. Zhang W-B, Wang X-M, Wang X-W, Liu D, Han S-Y, Cheng SZD (2015) Prog in Chem
27:1333
69. Teng FA, Cao Y, Qi YJ, Huang M, Han ZW, Cheng SZD, Zhang WB, Li H (2013) Chem
Asian J 8:1223
70. Cordes DB, Lickiss PD, Rataboul F (2010) Chem Rev 110:2081
71. Pielichowski K, Njuguna J, Janowski B, Pielichowski J (2006) Adv Polym Sci 201:225
72. Roll MF, Asuncion MZ, Kampf J, Laine RM (2008) ACS Nano 2:320
73. Tanaka K, Chujo Y (2012) J Mater Chem 22:1733
74. Kuo S-W, Chang F-C (2011) Prog Polym Sci 36:1649
75. Ren X, Sun B, Tsai C-C, Tu Y, Leng S, Li K, Kang Z, van Horn RM, Li X, Zhu M,
Wesdemiotis C, Zhang W-B, Cheng SZD (2010) J Phys Chem B 114:4802
76. Jiang S, Chen Q, Tripathy M, Luijten E, Schweizer KS, Granick S (2010) Adv Mater 22:1060
77. Cleland AN (2009) Phys Today 62:68
78. Walther A, Müller AHE (2008) Soft Matter 4:663
79. de Gennes P (1992) Angew Chem Int Ed 31:842
80. Dendukuri D, Pregibon DC, Collins J, Hatton TA, Doyle PS (2006) Nat Mater 5:365
81. Nie Z, Li W, Seo M, Xu S, Kumacheva E (2006) J Am Chem Soc 128:9408
82. Nisisako T, Torii T, Takahashi T, Takizawa Y (2006) Adv Mater 18:1152
83. Walther A, Andre X, Drechsler M, Abetz V, Muller AHE (2007) J Am Chem Soc 129:6187
84. Li Y, Guo K, Su H, Li X, Feng X, Wang Z, Zhang W, Zhu S, Wesdemiotis C, Cheng SZD,
Zhang W-B (2014) Chem Sci 5:1046
85. Li Y, Zhang WB, Hsieh IF, Zhang G, Cao Y, Li X, Wesdemiotis C, Lotz B, Xiong H, Cheng
SZD (2011) J Am Chem Soc 133:10712
86. Liu H, Hsu CH, Lin Z, Shan W, Wang J, Jiang J, Huang M, Lotz B, Yu X, Zhang WB, Yue K,
Cheng SZD (2014) J Am Chem Soc 136:10691
87. Dong X-H, Ni B, Huang M, Hsu C-H, Chen Z, Lin Z, Zhang W-B, Shi A-C, Cheng SZD
(2015) Macromolecules 48:7171. doi:10.1021/acs.macromol.5b01661
88. Lin Z, Lu P, Hsu C-H, Sun J, Zhou Y, Huang M, Yue K, Ni B, Dong X-H, Li X, Zhang W-B,
Yu X, Cheng SZD (2015) Macromolecules 48:5496
89. Haji-Akbari A, Engel M, Keys AS, Zheng X, Petschek RG, Palffy-Muhoray P, Glotzer SC
(2009) Nature 462:773
90. Henzie J, Grünwald M, Widmer-Cooper A, Geissler PL, Yang P (2012) Nat Mater 11:131
91. Greyson EC, Barton JE, Odom TW (2006) Small 2:368
92. Boles MA, Talapin DV (2014) J Am Chem Soc 136:5868
93. Quan Z, Fang J (2010) Nano Today 5:390
94. Liao C-W, Lin Y-S, Chanda K, Song Y-F, Huang MH (2013) J Am Chem Soc 135:2684
95. Chakrabarty R, Mukherjee PS, Stang PJ (2011) Chem Rev 111:6810
96. Furukawa H, Cordova KE, O’Keeffe M, Yaghi OM (2013) Science 341:1230444
Supramolecular Crystals and Crystallization 213

97. Guillerm V, Kim D, Eubank JF, Luebke R, Liu X, Adil K, Lah MS, Eddaoudi M (2014) Chem
Soc Rev 43:6141
98. Miszta K, de Graaf J, Bertoni G, Dorfs D, Brescia R, Marras S, Ceseracciu L, Cingolani R,
van Roij R, Dijkstra M, Manna L (2011) Nat Mater 10:872
99. Manna L, Milliron DJ, Meisel A, Scher EC, Alivisatos AP (2003) Nat Mater 2:382
100. Park I, Park S, Park HW, Chang T, Yang HC, Ryu CY (2006) Macromolecules 39:315
101. Ludwigs S, Boker A, Voronov A, Rehse N, Magerle R, Krausch G (2003) Nat Mater 2:744
102. Knoll A, Horvat A, Lyakhova KS, Krausch G, Sevink GJA, Zvelindovsky AV, Magerle R
(2002) Phys Rev Lett 89:035501
103. Yin Y, Sun P, Jiang R, Li B, Chen T, Jin Q, Ding D, Shi A-C (2006) J Chem Phys 124:184708
Adv Polym Sci (2017) 276: 215–256
DOI: 10.1007/12_2015_327
© Springer International Publishing Switzerland 2015
Published online: 11 August 2015

Self-Nucleation of Crystalline Phases Within

Homopolymers, Polymer Blends,
Copolymers, and Nanocomposites

R.M. Michell, A. Mugica, M. Zubitur, and A.J. Müller

Abstract Self-nucleation (SN) is a special nucleation process triggered by self-

seeds or self-nuclei that are generated in a given polymeric material by specific
thermal protocols or by inducing chain orientation in the molten or partially molten
state. SN increases the nucleation density of polymers by several orders of magni-
tude, producing significant modifications to their morphology and overall crystal-
lization kinetics. In fact, SN can be used as a tool for investigating the overall
isothermal crystallization kinetics of slow-crystallizing materials by accelerating
the primary nucleation stage in a previous SN step. Additionally, SN can facilitate
the formation of one particular crystalline phase in polymorphic materials. The SN
behavior of a given polymer is influenced by its molecular weight, molecular
topology, and chemical structure, among other intrinsic and extrinsic characteris-
tics. This review paper focuses on the applications of DSC-based SN techniques to
study the nucleation, crystallization, and morphology of different types of poly-
mers, blends, copolymers, and nanocomposites.

R.M. Michell
Grupo de Polı́meros USB, Departamento de Ciencia de los Materiales, Universidad Sim

on
Bolı́var, Apartado 89000, Caracas 1080-A, Venezuela
A. Mugica and M. Zubitur
Faculty of Chemistry, POLYMAT and Polymer Science and Technology Department,
University of the Basque Country UPV/EHU, Paseo Manuel de Lardizabal 3, 20018 Donostia/
San Sebastián, Spain
A.J. Müller (*)
Grupo de Polı́meros USB, Departamento de Ciencia de los Materiales, Universidad Sim

on
Bolı́var, Apartado 89000, Caracas 1080-A, Venezuela
Faculty of Chemistry, POLYMAT and Polymer Science and Technology Department,
University of the Basque Country UPV/EHU, Paseo Manuel de Lardizabal 3, 20018 Donostia/
San Sebastián, Spain
IKERBASQUE, Basque Foundation for Science, Bilbao, Spain
e-mail: [email protected]
216 R.M. Michell et al.

Keywords Crystallization rate
Melt memory
Self-nucleation

Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 216
2 Brief Description of the DSC Experimental Protocol Required for Study of Self-
Nucleation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 217
3 Crystalline Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222
4 Self-Nucleation as a Tool for Ascertaining the Origin of Fractionated Crystallization 227
5 Influence of Confinement on Self-Nucleation Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
6 Effect of Chain Topology on Self-Nucleation Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
7 Self-Nucleation and Preferential Polymorphism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239
8 Self-Nucleation Before Isothermal Crystallization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 240
9 Molecular Weight . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 244
10 Self-Nucleation and Nucleating Agents Efficiency Determination . . . . . . . . . . . . . . . . . . . . . . 245
11 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247

1 Introduction

The understanding and control of polymer crystallization is an important goal for

both academy and industry. The solidification of a semicrystalline polymer from the
melt usually involves overall crystallization, a process that includes primary nucle-
ation and crystal growth. Primary nucleation can have a dominating influence on
overall crystallization and in the resulting morphology. Self-nucleation
(SN) provides a way to enhance primary crystal nucleation in polymers. SN in
polymeric materials has attracted much attention since 1966, when Keller et al. [1]
employed it for the first time to produce single crystals crystallized from solution
with similar sizes. Fillon et al. [2] in 1993 studied SN of melt-crystallized isotactic
polypropylene (PP), establishing a simple experimental protocol employing differ-
ential scanning calorimetry (DSC).
SN can be defined as a thermal protocol for the production of self-nuclei within a
polymer melt. As a consequence, nucleation density increases significantly as self-
nuclei are produced. For any given polymer, self-nuclei should in theory be its best
nucleating agent. If self-nuclei are composed of crystal fragments remaining from
partial melting experiments, such crystal fragments have the capacity to produce
epitaxial nucleation because they have the same crystal structure as the polymer
being crystallized. On the other hand, self-nuclei can also be constituted by regions
of the melt where segmental orientation has not been erased. Such regions produce
crystalline memory effects that can induce nucleation upon cooling from non-
isotropic melts [1–4]
In this paper, the application of the DSC SN technique is reviewed since its
inception in 1993. Results obtained from study of fractionated crystallization,
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 217

confinement, nucleation efficiency, or crystalline memory in homopolymers,

blends, copolymers, and nanocomposites are presented.

2 Brief Description of the DSC Experimental Protocol

Required for Study of Self-Nucleation

The typical SN experimental procedure can be described as follows [2, 3] (see

Fig. 1):
1. Erasure of crystalline thermal history. The sample is heated to a temperature of
approximately 25–30 C above the DSC melting temperature peak and kept in
the melt for 3 min. The exact temperature can vary from polymer to polymer
because crystalline memory is a function of molecular weight, thermal history,
and intermolecular interactions. The objective is to erase all crystalline memory
and obtain an isotropic melt. All thermally sensitive nuclei are destroyed,
leaving only the temperature-resistant heterogeneous nuclei of unknown nature,
such as catalyst residues or impurities.
2. Creation of a standard semicrystalline state. The sample is cooled from the
homogeneous melt at a constant rate (usually 10 or 20 C/min) down to a
temperature low enough to allow the sample to crystallize until saturation. The
crystallization temperature peak recorded during this cooling scan is defined as
the standard crystallization temperature (or standard Tc), because it is only a
function of the density of thermally stable nuclei in the sample. The sample is
held at the minimum temperature for 3 min.

Fig. 1 Scheme of a self- 3'

nucleation (SN) experiment 200
[2]
180
5' T s

160
Temperature (ºC)

10 °C/min

in
10 °C/m

140

120

100

80

60

40

Time (min)
218 R.M. Michell et al.

3. Thermal conditioning at a temperature denoted Ts. At this temperature, the

sample could melt, self-nucleate. or self-nucleate and anneal. The sample is
heated at a constant rate (the same rate employed in step 2) from the chosen
minimum temperature up to a selected self-seeding or self-nucleation tempera-
ture (frequently denoted Ts), and then held at this Ts temperature for 5 min.
4. Cooling at a constant rate from Ts down to the minimum temperature chosen in
step 2. During cooling, the sample crystallizes at a temperature that depends on
the previous treatment (step 3) at Ts.
5. Final melting. Subsequent heating at a constant rate (the same rate as in steps 2–
4) from the minimum temperature chosen in step 2 up to the maximum melting
temperature established in step 1.
For this experimental protocol, the most important parameters are [2, 4]:
(a) Heating and cooling rates. Typically the heating and cooling scans
employed are 10 or 20 C/min; however, the SN protocol can also be
performed employing higher rates [5–7]
(b) Ts temperature
(c) Time spent at Ts
An example of the experimental data obtained during SN is shown in Fig. 2. The
cooling scans after the isothermal step at Ts are presented in Fig. 2a and the
subsequent heating scans are shown in Fig. 2b. Finally, the standard heating scan,
with the domains indicated, is shown in Fig. 2c. The SN domains were defined by
Fillon et al. [2] for isotatic PP. The three SN domains are described below (see
Fig. 2):
Domain I or melting domain. The polymer is under domain I when complete
melting occurs and the crystalline history of the material is erased. Therefore,
all crystalline memory is erased and the melt is isotropic. For the PP studied in
Fig. 2, complete melting is found at 169 C. At this Ts and above, no change is
detected in peak crystallization temperature (Tc), as compared with the standard
crystallization temperature obtained with Ts ¼ 200 C. The PP in Fig. 2 is under
domain I at any Ts temperature larger or equal to 169 C. In this domain, the
nucleation density for the PP sample employed in Fig. 2 is constant and its value
is approximately 106 nuclei/cm3. Both crystallization and melting DSC scans are
identical within domain I (see red curves in Fig. 2).
Domain II or self-nucleation domain. In domain II, the Ts range employed is low
enough to produce self-nuclei, but high enough to avoid annealing. Therefore,
domain II is easily identified when after 5 min at a given Ts, the peak crystal-
lization temperature of the sample increases in comparison with the standard
value. According to Fig. 2a, the start of domain II for the PP sample employed
occurs at a Ts ¼ 168 C. Additionally, the subsequent melting trace in Fig. 2b
does not reveal any sign of annealing. This is the distinctive behavior of domain
II or exclusive SN domain.
The nature of the self-nuclei generated in domain II has been the subject of
debate. Fillon et al. [2] considered that small crystal fragments remain in domain
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 219

(a) (b)

Heat Flow Endo Up (mW)

Heat Flow Endo Up (mW)
Ts=200ºC
Ts=200ºC, Dom I
Ts=169ºC
Ts=169ºC, Dom I
Ts=168ºC
Ts=168ºC, Dom II
Ts=167ºC
Ts=167ºC, Dom II
Ts=166ºC
Ts=166ºC, Dom II
Ts=165ºC

Ts=164ºC Ts=165ºC, Dom III

20 mW

15 mW
Ts=164ºC, Dom III

50 75 100 125 150 175 200 50 75 100 125 150 175 200
Temperature (ºC) Temperature (°C)
(c)
140
Flow Endo Up (mW)

DII DI
DIII 135

130

Tc (°C)
125

120
2 mW Heat

115

150 155 160 165 170 175 180

Temperature (°C)
Fig. 2 (a) DSC cooling scans (at 10 C/min) for PP after 5 min at the indicated Ts. (b) Subsequent
heating scans (at 10 C/min) after the cooling runs shown in (a). The arrow points to a close-up of a
small high temperature peak. (c) Representation of the self-nucleation domains for PP homopol-
ymer superimposed on top of the standard DSC melting trace. Data points represent crystallization
temperature peaks (plotted using the right-hand y-axis) as a function of Ts values (on the x-axis).
Reprinted from Müller et al. [6], Copyright (2015), with permission from Elsevier

II, or in other words, that Ts is high enough to melt almost all of the polymer
crystals, but low enough to leave “small” crystal fragments that can act as self-
nuclei. This explanation is probably true for the lowest temperatures of domain
II. Lorenzo et al. [4] considered that the presence of residual segmentation
orientation in the melt is enough to produce self-nuclei by inducing the so-called
memory effect, especially in the highest temperature range within domain
II. Further discussion of the nature of the self-nuclei and the importance of the
crystalline memory is addressed in the next sections.
220 R.M. Michell et al.

Ts DI

High Ts DII

Low Ts DII

Ts DIII

Fig. 3 Molecular representation of the three different self-nucleation domains (DI, DII, and DIII).
Reprinted from Müller et al. [6], Copyright (2015), with permission from Elsevier

The generation of self-nuclei increases the number of nuclei in PP from about

106 nuclei/cm3 (in domain I) to 1012–1013 nuclei/cm3, with a concomitant
reduction in spherulitic size (see Fig. 3).
The SN effect within domain II can slightly alter the subsequent melting trace
after crystallization. Small increases in peak melting temperature can be
observed, as well as the disappearance of partial melting and recrystallization
effects that may have been present in the standard melting scan within domain
I. In Fig. 2, the polypropylene standard melting behavior in domain I is charac-
terized by a bimodal melting peak, or a main melting peak with a right-hand side
high temperature shoulder, easily appreciated in Fig. 2c. This bimodality is a
result of reorganization processes experienced by PP during the heating scan
(i.e., by a process of partial melting of thinner crystals, which re-crystallize into
thicker crystals). The first melting peak is a result of fusion of the crystals formed
during the standard cooling scan, whereas the second (or high temperature
shoulder) is a result of fusion of re-crystallized PP during the scan. Figure 2b
clearly shows that, after SN, the melting peak bimodality within domain II
disappears. This behavior is caused by the enhanced nucleation density, which
makes possible the crystallization of PP at higher temperatures with thicker
crystals that do not need to reorganize during the heating scan.
Domain III or self-nucleation and annealing domain. When Ts is too low, partial
melting is produced and the unmolten crystals anneal during the 5 min at Ts.
Figure 2b shows that, at Ts ¼ 165 C, the melting endotherm exhibits a small high
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 221

temperature sharp peak (shown in a close up and indicated with an arrow) that is
a result of the melting of annealed crystals. Careful inspection reveals that, at
Ts ¼ 164 C, the crystallization exotherm shows a high temperature tail (indicat-
ing immediate recrystallization when the material is cooled from Ts), which is
also a sign that the sample is in domain III.
Figure 2c shows the transition temperatures between SN domains on top of the
standard PP melting endotherm. The same figure also shows the variation in the
peak crystallization temperature (Tc) with Ts (Tc data is associated with the right-
hand y-axis, and the x-axis with Ts temperature). This plot represents the typical
Tc variations during SN behavior [2]. Tc values remain constant in domain I,
whereas in domain II they increase sharply, as peak Tc values are proportional to
nucleation density.
Figure 3 illustrates the different SN domains at a molecular level. In domain I
the polymer chains are completely molten, forming isotropic random coils or a
homogeneous melt. When Ts decreases to domain II, the melt is no longer
uniform and two possible representations are considered in Fig. 3. At higher
temperatures within domain II, the melt retains some residual segmental orien-
tation, or crystalline memory, which produces SN. On the other hand, at lower
temperatures within domain II, small fragments of crystals are found, as postu-
lated by Fillon et al. [2]. In domain III, the temperature is low enough to partially
melt the polymer, while the unmolten crystals anneal into thicker crystals.
Figure 4 illustrates the SN behavior of poly(butylene succinate) (PBS)
[8]. Two polarized optical micrographs show the typical morphology after
cooling the sample from the melt in domain I (after the melt memory is erased)
or from domain II (after SN). The dramatic increase in nucleation density causes
an exponential increase in the number of spherulites being nucleated and,
therefore, their size is reduced, because they impinge on one another before
they can grow. This illustrates how the morphology of a semicrystalline sample
can be significantly altered by SN.
Figure 4 additionally shows how PBS possesses a remarkable crystalline
memory. Note that the melting process is completely over by 120 C. Neverthe-
less, the material experiences significant SN until 133 C, as judged by the
increase in Tc values (Tc data is associated with the right-hand y-axis). In the
case of PBS, Fig. 3 can be used to illustrate a strong crystalline memory effect. It
is clear that crystalline memory is the only explanation for the behavior of the
material in the high Ts region within domain II, because full melting had been
attained, as judged by the return of the DSC trace to the base line after melting at
temperatures above 120 C. Wide-angle X-ray scattering (WAXS) experiments
also indicate complete melting at temperatures above 120 C [8].
222 R.M. Michell et al.

DIII DII DI
Ts= 116ºC Ts= 145ºC
Heat Flow Endo Up (mW)

105

50mm 50mm

Tc (º)
90

75
3 mW

116ºC 134ºC

100 110 120 130 140 150

Temperature (ºC)

Fig. 4 Self-nucleation domains for PBS homopolymer, superimposed on top of its standard DSC
melting trace. Insets show polarized light optical micrographs taken during cooling from
Ts ¼ 145 C (domain I) and Ts ¼ 116 C (domain II). The data points represent peak crystallization
temperatures (plotted using the right handy-axis) as a function of Ts values (on the x-axis).
Reprinted with permission from Arandia et al. [8]. Copyright (2015) American Chemical Society

3 Crystalline Memory

The controversial idea that polymer crystallization from the melt could start from
some kind of initial transient state, precursor, or mesophase has been discussed in
several works [4, 9–63]. Such precursors could be correlated with the nucleation
density and also with melt memory.
“Melt memory” is a term employed to describe the phenomenon that occurs
when a semicrystalline polymer melt retains a partial memory of its previous
crystalline structure [2, 4, 27, 45, 47, 56, 64–76] and it could be related to the
existence of precursors, according to some authors [4, 77].
Self-nuclei could also originate from crystalline memory. Lorenzo et al. [4] have
argued that crystalline memory is observed in semicrystalline polymers in the
quiescent state as a consequence of their high molecular weight and the availability
of multiple chain conformations (from random coil in the melt to linear segments
within crystals). The transition from a semicrystalline morphology to homogeneous
melt is not instantaneous, and it requires high temperatures and finite times.
The melt must be heated to high temperatures (typically, 25–30 C above the
peak melting temperature; however, the exact temperature depends on the polymer
under study) for a short time (typically ~3–5 min) or to lower temperatures (always
above the melting temperature) for longer times. This process erases the memory of
previous processing and thermal treatments. The memory effect could originate
from some embryos (molecular clusters) that survive the thermal treatment and then
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 223

become active nuclei once the temperature is lowered [26, 44]. However, it is not
clear why some embryos could survive the high temperatures applied. Another
explanation given in the literature for the memory effect is the presence of precursor
structures, which are analogous to embryos because they could eventually trans-
form into nuclei. The origin of precursors differs for any given polymer. According
to some authors, some precursors are formed by shear flow in the melt [44, 78–
86]. Shear flow is the origin of polymer chain orientation and their partial align-
ment. These precursors can be destroyed by thermal treatment if the temperature is
high enough and the time sufficient. Crystalline precursors can also be produced if
the temperature and/or time are not sufficient to erase the crystalline memory of the
previous chain conformational state in the crystal [4].
According to Somani et al. [79], precursors do not have crystallographic order,
can survive for a long time after the shear stops (more than 2 h), and are layer-like
superstructures [79].
Strobl and coworkers [18, 34, 35] explained memory effects by considering that
a precursor is a conceptual object that contains crystallites. During crystallization,
the object volume is filled with crystals and grows with crystallization time. During
melting, the object is progressively emptied. At low melting temperatures, some
crystals remain inside the object (causing melt memory effects upon cooling).
However, at high temperatures, when the memory effect is erased, the object is
completely empty. Additionally, the authors postulate that the melt memory effect
is related to the creation rate of the crystals within the precursor. These objects or
precursors have never been observed without the presence of crystals, but Häfele
et al. [34, 35] consider that indirect experimental evidence points towards their
presence before crystallites are formed.
Lorenzo et al. [4] studied the origin of isotactic polypropylene (PP) self-nuclei
produced in domain II and the influence of SN time at specific Ts values. Figure 5
shows cooling scans after partial melting at three different Ts values and for
different annealing times. The standard cooling scan is also shown for comparison.
Figure 5 demonstrates that the existence of self-nuclei not only depends on Ts,
but also depends on the time spent at Ts (ts). For all Ts values employed, when ts
increases the nucleating effect decreases or completely disappears, and the exo-
thermic peak is progressively shifted to lower values until it reaches the equilibrium
at the same Tc displayed by the material during the standard cooling run. Another
interesting effect found by Lorenzo et al. [4] is that the time needed to reach the
standard Tc is lower for the higher Ts temperatures (see Fig. 5a). At Ts ¼ 168 C,
only 60 min are needed to decrease nucleation density to values similar to those of
the standard sample. However, when lower Ts values (within domain II) are
employed (Fig. 5a), an isotropic melt is not achieved at the times employed
(120 min).
According to Fillon et al. [2], domain II originates from the presence of residual
crystalline fragments; however, Lorenzo et al. [4] considered the possibility of the
existence of noncrystalline structures that could act as self-nuclei, especially in the
high Ts range within domain II.
224 R.M. Michell et al.

(a) Standard (b) Standard (c) Standard

Heat Flow, Endo Up

Flow, Endo Up
Heat Flow, Endo Up

0 min
0 min
0 min
1 min 1 min
1 min 5 min
5 min
5 min 30 min
30 min 60 min

24 mW Heat
30 min 60 min 120 min
120 min 120 min
24 mW

24 mW
Ts = 166°C Ts = 168°C
Ts = 167°C
50 75 100 125 150 175 200 50 75 100 125 150 175 200 50 75 100 125 150 175 200
Temperature (ºC) Temperature (ºC) Temperature (ºC)

Fig. 5 (a–c) DSC cooling scans (at 10 C/min) after the indicated annealing times for three
different Ts temperatures (within domain II) for PP. The standard cooling scan is also included
in each figure. Reprinted with permission from Lorenzo et al. [4]. Copyright (2006) John Wiley
and Sons

(a) 139 (b)

Ts values: Small crystal
135 fragments
168ºC
Heat Flow, Endo Up

131 167ºC 0 min

166ºC
1 min
Tc (°C)

127
5 min
123 30 min
119 120 min

115
0.5 mW

Tc Standard
Ts = 166°C
111
0 30 60 90 120
165 170 175 180 185 190
ts (min) Temperature (ºC)

Fig. 6 (a) Variation in peak crystallization temperature (Tc) for PP as a function of the annealing
time (ts) at the given Ts temperatures. (b) Heating DSC scans (at 10 C/min) after the indicated
annealing times at Ts ¼ 166 C. Reprinted with permission from Lorenzo et al. [4]. Copyright
(2006) John Wiley and Sons

Lorenzo et al. [4] performed a special thermal protocol to explore the nature of
self-nuclei. They ran the conventional protocol for SN, but instead of cooling the
sample after the time spend at Ts, they immediately heated the sample until melting.
Any crystals remaining in the sample should melt in the subsequent heating run.
Figure 6b shows the result for the lowest Ts temperature within domain II.
Lorenzo et al. [4] performed the experiments reported in Fig. 6b at Ts ¼ 166 C,
because it gives the largest probability of survival of crystalline fragments. The
presence of a crystalline phase in the sample partially molten at 166 C was
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 225

detectable only at ts ¼ 0 min. When the annealing time was increased, the signal
disappeared. As a consequence, no crystalline fragments were present, or at least
not in the necessary quantities to produce a signal in the DSC.
The previous results show that the SN experimental protocol is very sensitive to
annealing time and to cooling and heating rates, therefore it is a nonequilibrium
procedure. Also, self-nuclei are more likely to arise from a pre-order phase in the
melt than from residual crystalline fragments. Figure 5 shows that SN persists at
Ts ¼ 166 C, even after 120 min of annealing. However, in Fig. 6b it seems evident
that after only 1 min there are no traces of crystalline fragments.
According to Lorenzo et al. [4], application of the SN thermal protocol involves
the presence of some pre-order state (or precursor) in domain II. Such precursors
(or residual orientation in melt regions of the sample) are produced by incomplete
erasing of the crystalline order. Lorenzo et al. [4] also studied the influence of SN in
shear modulus relaxation experiments.
Figure 7 shows shear modulus (G(t)) values extracted at constant times plotted as
a function of the Ts employed. The vertical lines are drawn to illustrate the limits
between domains, according to parallel DSC experiments. Figure 7 shows differ-
ences in the relaxation behavior of the shear modulus in the melt depending on its
crystalline memory. The curves show three ranges at low, intermediate, and high
values of G(t). The transition between the different values corresponds to the three
different SN domains. At higher temperatures within domain I (200–169 C), the
values of G(t) are very similar and exhibit small increments as the temperature
decreases. On the other hand, at temperatures belonging to domain II, small
increases in temperature produce a large increase in G(t). This fact could be related
to the existence of residual segmental orientation in the melt. Finally, when the
temperature decreases to 165 C (within domain III), the values of G(t) increase
remarkably as a result of unmolten crystals present in the melt.
Gurarslan et al. [33] developed a method to produce highly oriented polymer
melts, employing cyclodextrins (CDs). These water-soluble polysaccharides are
nanometric in size and have internal hydrophobic cavities and external hydrophilic
groups. The cavities can be filled with polymer chains and produce crystalline
compounds, originating an arrangement of extended polymer chains (see Fig. 8).
The authors employed poly(L-lactic acid), poly(ε-caprolactone), and nylon-6 in
conjunction with CDs in order to form crystalline compounds. Water was used to
remove CDs and obtain coalesced polymeric samples of each material.
The coalesced samples were amorphous, but they had a strong pre-orientation.
The influence of the previous state in the subsequent crystallization was studied by
DSC. Figure 9 shows the first cooling scan for two PCL samples, one as-received
and the other as a coalesced sample. The difference in the crystallization temper-
ature was more than 25 C, resembling the changes produced by the SN process. An
interesting fact of this system is that chain orientation remained in the sample, even
after days of exposure to temperatures above the melting temperature. It is possible
that the orientation given during this particular treatment is even larger than that
produced by the crystallization process. The process for production of a homoge-
neous random melt for these structures takes much longer times than for a conven-
tional crystalline sample. According to Gurarslan et al. [33], the proximity of the
226 R.M. Michell et al.

60 D III D II DI

40

0.5 s
6 1.0 s
G (t) (kPa)
2.0 s
5 3.0 s

0
164 166 168 170 200
Ts (°C)

Fig. 7 Shear modulus (G(t)) values at constant time intervals plotted versus Ts (see text for
details). Vertical lines indicate the limits between domains. Reprinted with permission from
Lorenzo et al. [4]. Copyright (2006) John Wiley and Sons

Fig. 8 Extended unentangled polymer chains in a coalesced sample (right) and the randomly
coiling entangled chains in polymer solution or melt (left). Reprinted with permission from
Gurarslan et al. [33]. Copyright (2012) John Wiley and Sons

crystals to the amorphous zone, in an as-received sample, facilitates chain mobility

and allows the formation of a homogeneous melt. However, the coalesced sample
exhibits uniform order throughout, making the creation of random coiled chains in
the melt more difficult.
Gurarslan et al. [33] employed 2 wt% of the coalesced sample as self-nucleating
agent in the as-received PCL sample and found that the crystallization temperature
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 227

Fig. 9 DSC cooling curves

of as-received and Coalesced PCL
coalesced PCL samples
determined at 20 C/min.
Adapted with permission
from Gurarslan

Heat Flow, Endo up

et al. [33]. Copyright (2012) As-received PCL
John Wiley and Sons

0 20 40 60
Temperature (ºC)

of the self-nucleated sample was the same as that of the coalesced sample. This
experiment demonstrates that a small quantity of pre-order structure is able to self-
nucleate a polymeric sample and that there is no need for crystalline fragments.
Melt memory effects are also viewed by Luo and Sommer as responsible for SN
[43]. The authors performed molecular dynamic simulations and primitive path
analysis to study the process of crystallization from the melt state. According to
their predictions, the disentanglement process necessary for crystallization occurs
from locally unentangled chains and their posterior folding. On the other hand, they
also found that re-entanglement is a slow process and could originate memory
effects during the heating–recooling process (i.e., self-nucleation).

4 Self-Nucleation as a Tool for Ascertaining the Origin

of Fractionated Crystallization

Self-nucleation has many useful applications in polymer characterization. Coinci-

dent and fractionated crystallization phenomena are frequently observed in blends
and copolymers, and SN can be employed to investigate the origin of the crystal-
lization process [3, 8, 87–100]. The work performed by Morales et al. [93] has been
chosen as an example of the application of SN to a polymer blend.
Morales et al. [93] studied the crystallization of PP and LLDPE blends
employing different melt mixing techniques. The minor component was PP, and
it was dispersed as small isolated particles in the LLDPE matrix. The crystallization
process of the PP microdomains (MDs) occurs at lower temperatures (see Fig. 10)
than those typical for bulk PP.
Once PP is divided into many small and isolated droplets, the heterogeneities
responsible for PP nucleation are also divided among the different droplets,
228 R.M. Michell et al.

Fig. 10 DSC cooling scans

at 10 C/min for 80/20
LLDPE/PP blends after iPP
complete melting at
Ts ¼ 200 C and partial

Heat Flow, Endo up

melting at Ts ¼ 162 C. Data
taken from Morales 80/20, TS= 200 ºC
et al. [93]

80/20, TS= 162 ºC

15 mW
30 60 90 120 150
Temperature (ºC)

whereas some heterogeneities can migrate to the LLDPE matrix component during
processing. When the number of droplets is of the same order of magnitude
(or larger) than the number of heterogeneities originally present in bulk PP, many
droplets are free of highly active heterogeneities and nucleate at lower temperatures
(higher supercooling). The average number of PP particles in the blends is about
1011 particles/cm3 [93], which is much greater than typically present in commercial
PP (approximately 106 heterogeneities/cm3 [2]). Such a large difference normally
leads to statistically clean droplets and a large depression in their crystallization
temperature. For more details on the process of fractionated crystallization, the
reader is referred to the literature [3, 101–105].
The PP/LLDPE blends studied by Morales et al. [93] experienced coincident
crystallization for both PP droplets and LLDPE matrix when the blend was cooled
from the melt. Therefore, a single main crystallization peak is observed in the DSC
cooling run, corresponding to the 80/20 blend where crystalline memory has been
erased or is in domain I (see the DSC cooling curve corresponding to Ts ¼ 200 C in
Fig. 10). On the other hand, the subsequent heating scan shows two separate
melting peaks (not shown in Fig. 10) associated with the fusion of PP and
LLDPE, indicating that both components were able to crystallize during the previ-
ous cooling from the melt. The difference between the melting temperatures of
LLDPE and PP is large enough to apply SN to the PP phase while the LLDPE phase
is in the melt.
Morales et al. [93] applied SN to the PP component of the blend with the purpose
of injecting self-nuclei into every PP droplet. The blend sample was molten at
200 C for 3 min to erase any previous thermal history, then it was cooled at 10 C/
min to room temperature to create a standard crystalline state. Subsequently, the
sample was heated at 10 C/min up to the SN temperature (Ts ¼ 162 C), which
belongs to domain II, for 5 min. As PP is self-nucleated, the increase in the
crystallization temperature produces total separation from the crystallizing peak
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 229

of the LLDPE phase, as shown in the cooling curve corresponding to Ts ¼ 162 C in

Fig. 10. Therefore, the coincident crystallization process is a result of the lack of
highly active heterogeneities in most PP droplets. The self-nucleated blend (where
only the PP phase was self-nucleated) shows two crystallization peaks, the self-
nucleated PP at Tc ¼ 135 C and the LLDPE matrix at Tc ¼ 108–109 C. It should be
noted that the crystallization temperature of the LLDPE matrix also increased in
comparison with the original blend. This increment is associated with the nucleat-
ing effect of the self-nucleated PP on LLDPE [93].
Another typical case observed for immiscible blends is the occurrence of
fractionated crystallization. In these blends, the crystallization process of the
dispersed phase can be fractionated into several exothermic processes upon cooling
from the melt. Arnal et al. [87] studied the crystallization of an 80/20 PS/PP blend,
and found that the PP phase within the blend exhibits fractionated crystallization
(see Fig. 11). The blend has the typical sea-island morphology, with PP droplets
within a PS matrix (see Fig. 11a). The number of droplets for this blend is reported
by Arnal et al. [87] to be in the order of 1011 particles/cm3 and the PP employed
contains approximately 107 heterogeneities/cm3. According to Fig. 11, the single
exotherm (109–111 C) present in bulk PP is fractionated into four peaks (labeled A,
B, C and D) after blending.
Fractionated crystallization is a result of the presence of different types of
heterogeneities in the PP phase (dispersed as droplets) within the 80/20 PS/PP
blend, each one with a different activation energy. The most active heterogeneity
originates crystallization peak A in Fig. 11 (at the highest Tc) and it is the same type
of heterogeneity as that responsible for the nucleation of bulk PP. When PP is in the
bulk state, the heterogeneity with the lowest specific interfacial energy difference is
activated at lower supercooling (high Tc) and dominates the crystallization of the
polymer via secondary nucleation.
The crystallization peaks labeled B and C for the 80/20 PS/PP blend in Fig. 11
were attributed by Arnal et al. to crystallization processes originating from different
types of less active heterogeneities, which need larger undercoolings to be acti-
vated. In the case of peak D, the largest undercooling is required for a small fraction
of the material to crystallize. It is likely that it corresponds to the crystallization of a
fraction of heterogeneity-free PP droplets. Its exact origin is debatable because it
could correspond to the crystallization of a group of clean droplets whose nucle-
ation either starts from the surface of the droplets (or the interface between PP and
PS) or by homogeneous nucleation inside the PP droplets (see [3, 101–105]).
Fractionated crystallization in immiscible blends with sea-island morphologies
has been explained by the lack of highly active heterogeneities in every dispersed
phase droplet. The different peaks at lower crystallization temperatures than for the
bulk polymer can be attributed to nucleation triggered by less active heterogeneities
and, eventually, to surface or homogeneous nucleation. One way to demonstrate the
validity of this explanation is to inject nuclei in every droplet by adding a nucleating
agent [106] or by SN [87]. Then, all the low temperature crystallization peaks
should disappear and the dispersed phase should crystallize in a single peak at low
undercoolings [87, 93, 106, 107].
230 R.M. Michell et al.

Fig. 11 (a) SEM image of

(a)
the 80/20 PS/PP blend. (b)
DSC cooling scans at 10 C/
min for the samples
indicated at Ts ¼ 162 C
(blue) and Ts ¼ 161 C (red).
The single exotherm (A)
present in bulk iPP is
fractionated into peaks A, B,
C, and D after blending.
Adapted with permission
from Arnal
et al. [87]. Copyright (1998)
John Wiley and Sons

2 mm

(b)
iPP A
Heat Flow, Endo Up

iPP SN 162°C

D C
B
A
x4
80/20

80/20 SN 161°C 2mW

10mW

40 60 80 100 120 140

Temperature (°C)

Figure 11b shows the DSC cooling scan of bulk PP after SN at 162 C. The Ts
employed belongs to domain II and it is the lowest temperature in this domain or the
ideal SN temperature (because it produces maximum SN without annealing). This
Ts temperature was employed to generate the maximum quantity of self-nuclei and,
hence, produce the largest shift in Tc. The introduction of self-nuclei (the most
active nuclei theoretically possible) in PP produces an increase of 28 C in its peak
crystallization temperature. A similar SN procedure was employed in the blend
(with Ts ¼ 161 C, as the ideal SN temperature is 1 C lower in the blend). In Fig. 11,
the cooling scan after SN shows that the four exothermic peaks present in the
untreated blend disappear and that there is only one crystallization peak at 134.5 C
(2 C lower than in self-nucleated bulk PP). This result confirms that SN generates at
least one self-nuclei in each MD.
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 231

5 Influence of Confinement on Self-Nucleation Behavior

The SN technique has been employed to study the nucleation and crystallization of
block copolymer components. In block copolymers, factors such as the volumetric
fraction and the degree of segregation affect the type of confinement and, as a
consequence, influence the SN behavior. Several works dealing with the crystalli-
zation of semicrystalline block copolymers have reported the SN of crystallizable
block(s) [3, 94–98, 108–119]. In these studies, three kinds of behavior have been
reported:
1. Presence of three self-nucleation domains. The classical behavior described by
Fillon et al. [2], with three SN domains, has been observed for strongly segre-
gated block copolymers, when the studied blocks are continuous phases or
percolated MDs and the nucleation is induced by heterogeneous nucleation.
The classical behavior has also been observed in miscible or weakly segregated
block copolymers that crystallize from a homogeneous or weakly segregated
melt [3]. Figure 12a shows the SN domains for the PE block within a strongly
segregated PS-b-PE-b-PCL triblock terpolymer with weight percent ratio of
27/37/36 (denoted S27E37C36). In this case, the PE phase is percolated and
crystallization can easily spread by secondary nucleation. The material exhibits
three SN domains.
2. Absence of domain II. Several block copolymer systems only exhibit domains I
and III. The absence of the SN domain (domain II) occurs within isolated MDs
and its disappearance has been linked to confinement [3]. Figure 12b shows the
SN domains for the PE block within S57E27C16, a strongly segregated PS-b-PE-
b-PCL triblock terpolymer. PE forms isolated MDs in this case. The material
exhibits only domains I and III.
3. Absence of domain II and partition of domain III into two subdomains. In cases
where the injection of self-nuclei into every MD is difficult, in view of the very
large number of MDs, domain III is split into two domains [3]:
(a) Domain IIIA, where annealing without SN occurs
(b) Domain IIISA, where SN and annealing are simultaneously observed;
domain IIISA is the exact equivalent of the standard domain III established
by Fillon et al. [2]
Figure 12c shows the SN domains for the PE block within S50E15C35, a
strongly segregated PS-b-PE-b-PCL triblock terpolymer with only 15%
PE. The PE block forms isolated MDs that are strongly confined by the other
two blocks. The material exhibits only domains I and III; additionally, domain
III is split into two domains, IIIA and IIISA [98, 109].
By changing the composition of a block copolymer, it is possible to switch
from one behavior to another. Müller et al. [98] studied a series of different block
copolymers to establish the influence of the composition, molecular weight, and
chemical structure on homogeneous nucleation, fractionated crystallization, and
self nucleation behavior. One of the systems studied was PS-b-PEO diblock
232 R.M. Michell et al.

Fig. 12 Self-nucleation
domains for the PE block
(a)
within the following IIISA II I

Heat Flow, Endo Up

strongly segregated triblock
terpolymers: (a) S27E37C36,
(b) S57E27C16, and (c)
S35E15C50. IIIA domain
where annealing without
self-nucleation occurs, IIISA
domain where annealing
and self-nucleation occur
simultaneously. Adapted

1 mW
with permission from
Balsamo
et al. [109]. Copyright 60 65 70 75 80 85 90 95
(2000) John Wiley and Sons
Temperature (ºC)

(b)
IIISA I
Heat Flow, Endo Up
1 mW

60 65 70 75 80 85 90 95
Temperature (ºC)

(c) IIISA I
Heat Flow, Endo Up

IIIA
1 mW

70 75 80 85 90 95
Temperature (ºC)
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 233

Fig. 13 Cooling and

heating DSC scans (10 C/

Heat Flow, Endo Up

46
min) for two PS-b-PEO S 39 EO 61
diblock copolymers.
Adapted with permission
from Müller
46
et al. [98]. Copyright (2002) S 39 EO 61
American Chemical Society

19
S 81 EO 19

Cold
Tg Tg

10 mW
cryst.

2 mW
19
S 81 EO 19

-90 -60 -30 0 30 60 90 120

Temperature (°C)

copolymers with two different compositions: S39EO6145 and S81EO1919 (the

subscripts indicate the weight composition and the superscripts the molecular
weight of the entire copolymer in kilograms per mole). In the first copolymer, the
PEO block constitutes the continuous phase (i.e., PS forms cylinders in a PEO
matrix). In the second copolymer, PEO is dispersed as nanodroplets in a PS
matrix. Figure 13 shows the DSC cooling and heating scans for these
copolymers.
As expected, the DSC cooling scan of the S39EO6146 diblock copolymer
exhibits a single crystallization peak at 37 C (see Fig. 13), as PEO constitutes
the continuous phase and the crystallization starts from heterogeneous nucle-
ation. On the other hand, for the S81EO1919 diblock copolymer, PEO is dispersed
as nanodroplets [98, 108, 120, 121] and crystallization develops inside isolated
MDs. This copolymer is strongly segregated, and the crystallization of the PEO
nanodroplets occurs under rigid confinement (no break-out is possible), as PS
vitrifies at around 75 C (because of the low molecular weight of the PS block).
Additionally, the typical number of heterogeneities for the PEO block is several
orders of magnitude (i.e., approximately seven orders of magnitude) lower than
the number of PEO nanodroplets. Therefore, the majority of these MDs are free
of heterogeneities, and the crystallization can start by surface (or interphase)
nucleation or by homogeneous nucleation. The extremely low crystallization
temperature (from 30 to 50 C) suggests that the nucleation is more likely to
be homogeneous for this copolymer, because Tc is very close to the glass
transition temperature (Tg) of the PEO block. The SN behavior of these copol-
ymers is shown in Fig. 14.
The S39EO6146 diblock copolymer exhibits classical SN behavior, as shown in
Fig. 14a, where the three domains can be detected. This is the expected behavior
234 R.M. Michell et al.

(a) (b)
46 19
Heat Flow, Endo Up S 39 EO 61 S 81EO 19

Heat Flow, Endo Up

III I
"Standard"
"Confined"

3 Domains

II

1 mW
2mW

IIISA IIIA I
20 30 40 50 60 70 80 90 0 10 20 30 40 50 60
Temperature (ºC) Temperature (°C)

Fig. 14 Self-nucleation domains corresponding to the PEO block for (a) S39EO6146 diblock
copolymer and (b) S81EO198.5 diblock copolymer (see text for details)

because PEO constitutes the matrix phase. However, for the S81EO198.5 diblock
copolymer, PEO is dispersed as weakly confined nanodroplets surrounded by
vitrified PS phase at the temperatures where PEO can crystallize (as discussed
above). In the case of highly confined polymeric phases, domain II disappears
and domain III splits into two. In addition, annealing occurs before SN, making
it impossible to detect domain II [98].
The triblock terpolymer S62B27C1162 (PS-b-PB-b-PCL) has a peculiar mor-
phology consisting of a PS matrix in which core–shell cylinders (where PCL is
the core and PB the shell) are dispersed (see cartoon in Fig. 15b). In this case, the
isolated PCL cylindrical cores are surrounded by PB soft shells. The copolymer
is strongly segregated, but PCL is only weakly confined. Even so, the segrega-
tion strength is large enough to prevent any break-out and the PCL has to
crystallize within the confinements of the cylindrical cores. Figure 15 shows
how the crystallization temperature changes during the process of SN at different
Ts temperatures. The crystallization temperature of the PCL block is extremely
low (40 to 50 C), indicating that homogenous nucleation probably triggers
crystallization, as Tc is very close to the vitrification temperature of PCL. The
transition between domain I and domain IIIA can be clearly seen in Fig. 15a, as
well as the subsequent transition to domain IIISA. Domain II is absent as a
consequence of the confinement.
Müller et al. [98] also studied the crystallization of PE and PEO blocks within
E24EP57EO1969 triblock terpolymer. It should be noted that the poly(ethylene-
alt-propylene) (PEP) block cannot crystallize. This copolymer was synthesized
via hydrogenation of the B24EP56EO2067 precursor employing the Wilkinson
catalyst. Wilkinson catalysis provides active heterogeneities to the PEO block.
Figure 16 shows the cooling scan of the unpurified E24EP57EO1969 sample. Upon
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 235

a)
b)
Heat Flow, Endo Up 140 °C

Heat Flow, Endo Up

63 °C
62 °C IIISA I
61 °C S
60 °C
59 °C B
58 °C
IIIA
C
57 °C
56 °C
55 °C
54 °C
5 mW

53 °C

10 mW
51 °C
57 59
49 °C
-80 -60 -40 -20 0 20 40 60 35 40 45 50 55 60 65
Temperature (°C) Temperature (°C)
Fig. 15 (a) Cooling scans at 10 C/min after partial melting at the indicated Ts. (b) Self-nucleation
domains for S62B27C1162 triblock terpolymer. Inset shows core–shell structure of the terpolymer;
C PCL core, B PB shell, S PS matrix. Adapted with permission from Müller et al. [98]. Copyright
(2002) American Chemical Society

Unpurified
Tc1(PEO) Tc2(PEO) Tc(PE)
Heat Flow, Endo Up

Tm(PEO) Tm(PE)

Purified

Tc(PE)
Tc(PEO)

Tm(PEO) Tm(PE)
4 mW

-60 -30 0 30 60 90 120

Temperature (°C)
Fig. 16 Cooling and heating DSC scans of original and purified E24EP57EO1969 triblock copol-
ymer. Adapted with permission from Müller et al. [98]. Copyright (2002) American Chemical
Society

cooling from the melt, the PE block crystallizes first (at high temperatures) and
then, at lower temperatures, crystallization of the PEO block takes place in two
peaks, exhibiting fractionated crystallization. A large part of the PEO block
crystallizes close to 20 C, whereas a smaller fraction crystallizes at lower
236 R.M. Michell et al.

temperatures (20 C and below). To investigate the origin of the fractionated

crystallization and whether or not the Wilkinson catalyst provided impurities for
the heterogeneous nucleation of a fraction of PEO MDs, the sample was purified
to remove as much of the Wilkinson catalyst as possible. The sample was
refluxed in a toluene solution with concentrated HCl and all the catalyst was
successfully removed.
Crystallization of the PEO block in purified E24EP57EO1969 triblock copoly-
mer is different from crystallization in the unpurified sample. In the purified
sample, the exotherm that was present at around room temperature completely
disappears. Therefore, this exotherm can be assigned to the nucleating effect of
the Wilkinson catalyst. The purified PEO block crystallizes in two lower tem-
perature peaks, the largest exotherm being located around 27 C and a very
small exotherm at 47 C. The first exotherm could originate from crystalliza-
tion of the PEO block after heterogeneous nucleation from a weakly nucleating
heterogeneity or surface nucleation. The lowest crystallization exotherm clearly
originates from crystallization after homogeneous nucleation of the PEO block.
The purification process also affects the self nucleation behavior. Figure 17
shows the SN domains of the PEO block within the original and purified
E24EP57EO1969 triblock terpolymer. The PEO block in the unpurified version
of E24EP57EO1969 exhibits classical behavior, with three domains. This is inter-
esting because PEO in this block copolymer forms isolated MDs (i.e., a mixture
of spheres and cylinders). Therefore, the presence of the Wilkinson catalyst is
probably responsible for the observation of domain II. On the other hand, the
PEO block within the purified E24EP57EO1969 exhibits the expected behavior of
a largely confined phase and domain II disappears. This is shown in Fig. 17,
where a direct transition from domain I to III can be observed [98].

(a) (b)
Heat Flow, Endo Up

Heat Flow, Endo Up

IIIIIISA II
SA
II II IIIIIISA
SA
II
3 mW

3 mW

20 25 30 35 40 45 50 55 60 65 70 20 25 30 35 40 45 50 55 60 65 70
Temperature (°C) Temperature (°C)

Fig. 17 (a) Self-nucleation domains for unpurified E24EP57EO1969. (b) Self-nucleation domains
for purified E24EP57EO1969
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 237

6 Effect of Chain Topology on Self-Nucleation Behavior

Pérez et al. [119] studied the crystallization of linear (L-PCL) and cyclic (C-PCL)
PCL chains of almost identical molecular weights. They found that the transition
temperatures between domains are affected by chain topology (see Table 1). In fact,
domain II is much wider in L-PCLs than in C-PCLs. In addition, the transition
temperature between domain I and domain II (TsI–II) occurs in L-PCLs at higher
values than the final melting temperature of the sample, as compared with C-PCLs.
In other words, L-PCL samples have a much higher crystalline memory than C-
PCLs of similar molecular lengths. Crystalline memory effects are a function of
entanglement density [26, 40, 44, 63, 122–126], and the entanglement density in L-
PCLs is significantly higher than in C-PCL as a result of its free chain ends.
Therefore, partially disentangled PCL cyclic chains can reach a random coil
conformation, or a uniform melt, faster than more entangled linear chains. As a
consequence, the crystalline memory of C-PCLs can be erased at lower tempera-
tures than for L-PCLs.
Another parameter that influences SN behavior is the branch content in PE [50,
127, 128]. Reid et al. [50] studied the crystallization of a series of PE-co-1-butene
copolymers with different amounts of 1-butene (ethylene branches). They found
that when the branch content increases, the transition temperature between domain
I and domain II decreases. They attributed this tendency to the reduced crystallinity
and thinner lamellae formed by highly branched copolymers. The process of
producing a homogeneous random melt is facilitated if the quantities of crystals
are reduced and if they are thinner. As a consequence, domain I can occur at lower
temperatures [50].
Reid et al. [50] calculated the equilibrium melting temperatures (T0m ) for the PE-
co-1-butene copolymers, and found that the value of TsI–II was higher than T0m for
the copolymers with low branch content. However, when the branch content
reached 4.53 mol% the TsI–II was similar or lower than T0m (see Fig. 18). This
unique behavior was present only in the copolymers; the linear PE studied by Reid
et al. showed the expected behavior; that is, for any temperature above T0m the melt
was always homogeneous and no evidence of self nucleation was reported (see
Fig. 19) [4, 20, 31, 69, 124, 125].
Reid et al. attributed the presence of crystalline memory above T0m to residual
sequence segregations in the melt, originating from diffusion restriction of the
crystalline sequences in reaching the homogeneous random melt. The restricted

Table 1 Transition temperatures for the self-nucleation domains, for the samples indicated [119]
Material DI to DII ( C) DII to DIII ( C)
C-PCL (2,000 g/mol) 64 60
C-PCL (7,500 g/mol) 69 61
L-PCL-OH (2,000 g/mol) 60 51
L-PCL-OH (7,500 g/mol) 67 56
238 R.M. Michell et al.

Fig. 18 Plot of the

difference between the 25
temperature of transition
from domain I to domain II

TSI-II - Tmocopo (°C)

(TsI–II) and the final melting 15
temperature of the
copolymer (T0m ) versus ethyl 5
branching content. Adapted
with permission from Reid
et al. [50]. Copyright (2013) -5
American Chemical Society
-15
0 1 2 3 4 5 6
Branching Content (mol %)

Fig. 19 Ts versus peak (a)

crystallization temperature 210
after partial melting for (a)
linear polyethylene 190
fractions with different
molecular weights and (b) 170
Ts (°C)

polyethylene with fixed

molecular weight and 150
varying branching content.
The horizontal lines 130
Tm0 = 145.5ºC
indicate T0m for each 110 Molecular Weight
sample. Adapted with
permission from Reid 90
et al. [50]. Copyright (2013) 109 111 113 115 117
American Chemical Society Crystallization Temperature (°C)
(b)
210 5.68% 4.14% 3.6% 2.2% 0%

190
Ts (°C)

170

150

130

110

90
40 60 80 100 120
Crystallization Temperature (°C)

diffusion of ethylene linear sequences is caused by the branches, since there are no
other differences between the homopolymers and copolymers. Ethyl or longer
branches do not participate in the crystalline regions and during the crystallization
process they are segregated to the amorphous phase surrounding the crystals. The
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 239

authors argue that when the melting process begins, the diffusion of linear
sequences is hindered by chain branches. As a consequence, the melt needs a longer
time and/or higher temperatures to allow diffusion and subsequent homogenization
of the melt.

7 Self-Nucleation and Preferential Polymorphism

Self-nucleation can be employed to trigger specific crystallization of a polymorphic

polymer sample. Increasing the number of crystals using a specific polymorphic
modification is a practical option for improving the performance of a polymeric
material [129–138].
Dai et al. [132] studied the effect of SN treatment on the thermal properties and
structure of PP. The PP employed was synthesized using a metallocene catalyst.
The standard cooling scan shows the formation of equal amounts of α- and γ-
polymorphism. The authors studied the SN behavior from 122 to 160 C. The
relative amount of α- and γ-crystals was determined by WAXS. In the DSC heating
scans, the lower temperature melting peak corresponds to the fusion of γ-crystals,
whereas the high temperature peak corresponds to the fusion of α-crystals. Dai
et al. established the following series of regions with specific combinations of
polymorphs (see Fig. 20) [132]:

Fig. 20 DSC heating scans at 5 C/min of m-PP after self-nucleation at the temperatures indicated
followed by cooling at (a) low Ts (122–140 C) and (b) high Ts (142–152 C). Dai
et al. [132]. Copyright (2002) John Wiley and Sons
240 R.M. Michell et al.

Region I: Ts ¼ 122–130 C, and Ts values are lower than the melting temperature of
the γ-phase. At these temperatures, the γ-crystals are annealed and the α-phase
remains unaffected. The amount of α-crystals increases in the subsequent
cooling scans as a result of the epitaxial recrystallization within existing α-
lamellae as substrates.
Region II: Ts ¼ 122–130 C, and Ts values are higher than the melting temperature
of the γ-phase, but lower than the melting temperature of α-crystals. In this
region, α-crystals anneal and the γ-phase is melted. As in region I, the amount of
α-crystals increases in the subsequent cooling scans.
Region III: Ts ¼ 140–148 C, and Ts values are higher than the melting temperature
of the α-phase, but not high enough for complete melting. Therefore, in this
region, complete melting of γ-crystals takes place but α-crystals are partially
melted. When Ts increase, the nucleation of α-crystals decreases, originating a
depression in the crystallization and melting temperatures on subsequent DSC
scans. The amount of γ-phase increases and exceeds the amount of α-crystals as
a result of the effective nucleation of γ-crystals by the α-phase.
Region IV: Ts ¼ 150–160 C, and Ts values are higher than those employed in region
III. Only a small amount of self-nuclei remain. Subsequent cooling originates
recrystallization at lower temperatures. As in region III, the relative amount of γ-
crystals is higher than that of α-crystals.
Kang et al. [133, 134] studied α- and β-PP crystals. They identified three regions
where the relative amount of each polymorphic form changes according to the Ts
employed during partial melting.
On the other hand, Cavallo et al. [129, 130] studied in detail the self-nucleation
and cross-crystallization between the polymorphic phases of poly(1-butene). Other
studies based on the SN and control of polymorphism were performed for polyam-
ide 11 [131], poly(vinylidenefluoride) [135, 136], poly(butylene adipate) [138], and
chocolate [137].

8 Self-Nucleation Before Isothermal Crystallization

It is well known that some polymers or crystallizable blocks within block copoly-
mers have slow crystallization kinetics. In some cases, the overall crystallization
kinetics is so slow that isothermal DSC experiments cannot be performed. Accel-
eration of the nucleation step can be accomplished by previous SN of the polymer.
Under these conditions, the crystallization rate of self-nucleated material increases
such that isothermal DSC can be performed [4, 69, 76, 96, 123, 139–152]. It is well
known that overall crystallization kinetics obtained by isothermal DSC experiments
contains both nucleation and growth components. If the isothermal crystallization
starts from an ideally self-nucleated material, the experimental data may corre-
spond exclusively to crystal growth [142].
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 241

(a) (b)
42
D77C23 42
D77C23
41
D65C35
Heat Flow, Endo Up

Heat Flow, Endo Up

41
D65C35

11
D40C60 PPDX
PCL
PCL + PPDX 12
D40C60
45
D23C77

35
D23C77
5 mW

5 mW
-20 0 20 40 60 80 100 120 -20 0 20 40 60 80 100 120
Temperature (ºC) Temperature (ºC)
Fig. 21 (a) DSC cooling scans (10 C/min) for PPDX-b-PCL diblock copolymers. (b) Subsequent
heating scans (10 C/min). Reproduced from Müller et al. [96] with permission of The Royal
Society of Chemistry

Müller et al. [96] studied the isothermal crystallization and SN of PPDX-b-PCL

diblock copolymers. The cooling and subsequent heating scans for the PPDX-b-
PCL samples are shown in Fig. 21. The nomenclature employed is the same as that
employed above, where subscripts indicate the composition (in wt%) and the
superscripts the molecular weight of the block copolymer in kilograms per mole.
Figure 21a shows that only one crystallization exotherm is observed for all
copolymers. This single crystallization peak corresponds to the coincident crystal-
lization of both PCL and PPDX blocks [153]. Parallel real-time WAXS experi-
ments demonstrated that both phases crystallize in the same temperature range. It
was shown that the PPDX block crystallization occurs at lower temperatures than
neat PPDX because of the topological restrictions induced by covalent bonding
with the highly flexible PCL block. The depression in the PPDX temperature is
large enough to overlap with crystallization of the PCL block, which is triggered
almost immediately when the PPDX block starts to crystallize. The subsequent
heating scans demonstrate the separate melting of the crystals of each block.
Figure 22 shows how the SN of the PPDX block within the D77C2342 block
copolymer can separate the crystallization of each component. In Fig. 22a, evolu-
tion of the separation of the two blocks is evident. When the number of self-nuclei
increases (as Ts decreases), the coincident crystallization peak splits into two
(Ts ¼ 110 C) and, finally, into three exotherms (Ts ¼ 108 C). For Ts ¼ 108 C, the
PPDX block crystallization is self-nucleated and crystallizes at higher temperatures
in a complex bimodal exotherm. The annealing of PPDX crystals begins when
Ts ¼ 104 C, and in the subsequent heating scan the presence of a second higher
temperature peak is evident. On the other hand, the exothermic peak around 30 C is
242 R.M. Michell et al.

(a) (b)
115 °C DI PPDX
PCL + PPDX PCL
115°C
110°C DII
1 mW Heat Flow, Endo Up

Heat Flow, Endo Up

x10 110°C
108°C χ PPDX
= 60.3 DII
c 108°C

107ºC χ PPDX = 81.1 DII 107ºC

c
Annealing
104ºC DIII

2 mW
PCL PPDX 104ºC

0 20 40 60 80 100 0 25 50 75 100 125

Temperature (°C) Temperature (°C)

Fig. 22 Self-nucleation of the PPDX block within D77C2342. (a) DSC cooling scans from the
indicated Ts temperatures; χcPPDX represents the amount of PPDX that has been self-nucleated. (b)
Subsequent heating scans. The arrow points to the melting endotherm of annealed crystals during
treatment at Ts ¼ 104  C. Reproduced from Müller et al. [96] with permission of The Royal
Society of Chemistry

associated with crystallization of the PCL block, whereas all the PPDX crystallizes
at much higher temperatures. Figure 22a also shows that the self-nucleated PPDX
block crystals can nucleate the PCL block crystals (as the crystallization tempera-
ture of the PCL block increases).
Employing the conventional method to determine the overall isothermal crys-
tallization kinetics by DSC, the crystallization rate of the PPDX block was
extremely low so no significant exothermic signal was possible. Müller et al. [96]
developed a different methodology to study the crystallization kinetics of the PPDX
block. The samples were first self-nucleated (steps 1–3 in Fig. 23a) and then
quenched (at 80 C/min) to their isothermal crystallization temperature (Tc, indi-
cated in Fig. 23b). Self-nucleation accelerates the overall kinetics as a result of an
increase in the number of active nuclei in the sample. The effect is equivalent to
starting the isothermal DSC measurements shown in Fig. 23b with samples whose
nucleation process had already finished. Self-nucleation should provide all the
necessary nuclei for crystallization; as a consequence, only secondary nucleation
or growth is measured during the isothermal DSC runs shown in Fig. 23b.
For comparison, Müller et al. [96] also applied the same procedure indicated in
Fig. 23a for a PPDX homopolymer. The experimental data was analyzed employing
Lauritzen–Hoffman theory (LH). The value obtained for Kτg (a value proportional
to the energy barrier for primary and secondary nucleation) was lower than that
obtained for the same sample employing conventional isothermal crystallization
from the melt without SN. The value of Kτg shifts from 31.0  104 K2 for the
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 243

(b)

Heat Flow, Endo Up

60 ºC

(a) 62 ºC
Heat Flow (mW), Endo Up

1 64 ºC

2 68 ºC
3
70 ºC

0.5 mW
Tc

-20 0 20 40 60 80 100 120 3 6 9 12 15 18

Temperature (ºC) Time (min)

Fig. 23 Examples of DSC isothermal crystallization scans for the PPDX block within D77C2342
diblock copolymer. The measurements were performed after the PPDX block had been previously
self-nucleated at Ts ¼ 110 C and then quenched (80 C/min) to the indicated Tc temperatures. (a)
Thermal protocol applied to the sample: the sample is first heated to a completely molten state
(at 130 C), indicated by point 1. Then, the sample is cooled at 10 C/min to obtain a “standard”
crystalline state (point 2). The sample is then ideally self-nucleated by heating to point 3. Finally,
the sample is quenched (at 80 C/min) to the isothermal crystallization temperature desired (Tc).
(b) Isothermal DSC curves recorded after applying the procedure indicated in (a). Reproduced
from Müller et al. [96] with permission of The Royal Society of Chemistry

conventional method to 17.0  104 K2 after SN. Kτg is proportional to the energy
barrier for overall crystallization. This result indicates that the energy barrier for
nucleation and growth is reduced by performing the nucleation in the previous
SN step.
On the other hand, the same LH analysis was performed on spherulitic growth
data for neat PPDX homopolymer. The value obtained for the secondary nucleation
constant for PPDX spherulitic growth was K gG ¼ 17:2  104 K2 , a value in remark-
able agreement with the Kτg value obtained by DSC after SN. This result shows that
SN saturated the polymer with active nuclei such that the DSC measurements
correspond to growth only. The same analysis was performed for PPDX-b-PCL
block copolymers. The obtained values of Kτg were higher than for neat PPDX,
indicating that presence of the PCL block hinders secondary nucleation. Such
hindrance is responsible for the coincident crystallization of both blocks during
non-isothermal crystallization.
This technique of performing SN just before isothermal DSC measurements has
been applied to segmented copolymers of poly(ether ester), based on poly(ethylene
glycol) and poly(ethylene terephtalate) [139], PP [63, 133, 141, 151], PCL[142],
PPDX [142], poly(ε-caprolactone)-block-poly(propyleneadipate) [143], poly(eth-
ylene naphthalate) [144], poly(propylene suberate) [145], poly(propylene
244 R.M. Michell et al.

terephthalate)/SiO2nanocomposites [146], poly(propylene azelate) [147], poly(pro-

pylene sebacate) [147], poly(ethylene azelate) [148], sindiotactic poly(propylene)
(sPP) [69, 123], poly(propylene terephthalate) [149], poly(trimethylene terephthal-
ate) [76], and poly(lactic acid) [150].
Additionally, non-isothermal crystallization kinetics has also been performed
after SN. In all cases, the crystallization rate was accelerated for the SN system
compared with crystallization from the isotropic melt [154, 155].
Self-nucleation can be used as a tool for tailoring the morphology of block
copolymers in thin films. The nanostructure of thin films has an influence on their
surface properties, which could be of great interest for the fabrication of photonic
crystals and bioanalytical devices. Shultze et al. [156] studied the SN behavior and
its influence on the surface morphology of polybutadiene-b-poly(ethylene oxide).
According to their results, it is possible to modify the periodicity of the block
copolymer phase separation using SN treatment [156].

9 Molecular Weight

Another molecular parameter that influences SN behavior is molecular weight.

Melt memory depends on entanglement density, which increases with molecular
weight. Self-nucleation behavior is closely related to melt memory effects, as
discussed above. As a consequence, it is expected that the temperature of transition
between domains is dependent on molecular weight [50, 134, 157, 158].
Reid et al. [50] reported the effect of molecular weight on the crystallization of
random ethylene-co-1-butene copolymers and linear PE samples. They performed a
series of crystallization experiments after partial melting and found that, for low
molecular weight samples (M < 4.500 g/mol), the crystallization temperature upon
subsequent cooling does not depend on the Ts employed, at least in the range
studied. However, when the molecular weight was higher, a shift in Tc to higher
temperatures was observed, indicating SN effects.
The process of erasing previous crystalline memory needs high temperatures.
Reid et al. [50] have shown (see Fig. 24) that the critical melting temperature (TsI–II)
needed to reach homogeneous copolymer melts depends on the molecular weight
(for an ethylene-co-1-butene copolymer with 2.2 mol% of ethyl branches). They
found that the transition between domains I and II strongly depends on molar mass.
They concluded that the origin of the SN process is associated with the entangled
melt dynamics. The partial melting at Ts involves the movement of chains, which
are constrained by a series of obstacles such as entanglements, loops, ties, and
knots. To accomplish a homogeneous melt, the original crystalline structure must
be completely destroyed, and this process needs chain movements to overcome
obstacles that prevent chain relaxation. When the molecular weight increases, the
number of obstacles also increases, making the chain movements needed to achieve
a fully homogeneous random melt more difficult.
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 245

Fig. 24 Critical melting 180

temperature (TsIII ) plotted
as a function of molar mass 170
of random ethylene 160
1-butene copolymers with
2.2 mol% ethyl branches.

TsI-II(°C)
150
The dashed line represents
the equilibrium melting 140
temperature. Adapted with
130
permission from Reid
et al. [50]. Copyright (2013) 120
American Chemical Society
110
100 1000 10000 100000 1000000
Mw (g/mol)

For the ethylene-co-1-butene copolymer with 2.2 mol% of ethyl branches, TsI–II
increases linearly with the logarithm of molar mass (Fig. 24). One particular finding
of Reid et al. is that the value of TsI–II is always higher than the calculated
equilibrium melting temperature (T0m ) of the copolymer as long as the molecular
weight is higher than 1,300 g/mol. For molecular weights below this value, the
copolymer exhibits a TsI–II that is equal to or lower than T0m . Additionally, 1,300 g/
mol also corresponds to the critical molecular weight for the development of
entanglements in polyethylene. Hence, ethylene/α-olefin copolymer samples that
are free of entanglements do not have memory effects above T0m .

10 Self-Nucleation and Nucleating Agents Efficiency

Determination

Fillon et al. [159] developed a method for evaluating the efficiency of a nucleating
agent employing SN. Self-nuclei are commonly considered to be the best nuclei for
the polymer under study because they have the ideal crystallography for epitaxial
nucleation. Also, the number of self-nuclei increases with a decrease of Ts within
domain II. Therefore, the minimum temperature within domain II is the ideal SN
temperature, because it causes maximum production of self-nuclei without any
annealing of unmolten crystal fragments (if present). The peak Tc value reached
after SN at the ideal Ts is therefore the maximum value of the crystallization
temperature induced by SN without the influence of annealing.
The above facts are the basis of Eq. (1), where calculation of the nucleation
efficiency (NE) is related to the shift in crystallization temperature produced by the
nucleating agent under consideration, in comparison with the shift produced by the
maximum quantity of self-nuclei generated in the process of ideal SN [159]. NE can
be calculated by the following simple expression:
246 R.M. Michell et al.

T CNA  T CP
NE ¼ ; ð1Þ
T CMP  T CP

where TCNA is the peak crystallization temperature of the polymer with the nucle-
ating agent, TCP is the peak crystallization temperature of the neat polymer (without
any nucleating agent and with a standard thermal history), and TCMP is the maxi-
mum crystallization temperature of the ideally self-nucleated neat polymer sample
(i.e., SN is performed with the minimum Ts value within domain II).
NE is a convenient parameter for comparing the efficiencies of different nucle-
ating agents in a quantitative way and with respect to a unique property of every
polymer (i.e., its SN ability).
Equation (1) has been recently used to determine nucleation efficiency in several
publications that deal with the use of nucleating agents to increase factors such as
the crystallinity degree, mechanical properties, optical properties, and processing
window [75, 160–183].
A typical nucleating agent never reaches an NE value higher than 100%, because
a value higher than 100% in Eq. (1) implies that the nucleating agent is better than
self-nuclei. However, Müller et al. [168, 175, 184] found that, in some cases, the NE
of carbon nanotubes (CNTs) yields values above 100%, referring to this phenom-
enon as “supernucleation.”
Figure 25 shows CNT nucleation efficiency for PE, PCL, and PEO. Values well
above 100% were found, demonstrating that CNTs are better nucleating agents than
the corresponding homopolymer self-nuclei in these specific cases. The reasons for
supernucleation are still being studied. It is not clear what are the exact conditions
needed to achieve such high levels of nucleating efficiency. Two factors are
believed to be important: the interactions between the polymer and the nucleating
agent and excellent dispersion of the nanofiller [168, 175, 184].

Fig. 25 Efficiency of 500

CNTs as nucleating agent
for the indicated samples.
Nucleation Efficiency (%)

Reprinted from Müller 400

et al. [168], Copyright
(2011), with permission
300
from Elsevier

200

100
Maximun self-nuclei efficiency

0 5 10 15 20 25 30 35 40
CNT Content (%)
MWNT/PE MWNT- g-PCL MWNT- g-POE
SWNT/PE DWNT/PE
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 247

11 Conclusions

The self-nucleation technique (SN), applied using a simple thermal protocol

performed with a standard differential scanning calorimeter, is a powerful and
convenient tool for studying polymer nucleation and crystallization. SN increases
nucleation density by orders of magnitude, thus affecting the morphology, crystal-
linity degree, and (potentially) the mechanical properties of the polymer under
study. SN can accelerate the overall crystallization of polymeric materials
(by exponentially enhancing the primary nucleation step) and can be used as a
tool for investigating crystallization in slow-crystallizing materials.
SN is affected by the architecture (e.g., homopolymer versus copolymer) and
topology (e.g., ring versus linear, or branched versus linear) of polymer chains,
molecular weight, chemical structure, and molecular orientation, among other
factors. Additionally, SN provides a mean to quantify the nucleation efficiency
(NE) of additives such as nucleating agents and nanofillers, establishing a relative
scale that allows meaningful comparisons between the NE of different additives.
The SN technique is now well established and, as shown in this review, has been
applied to the study of homopolymers, random and graft copolymers, polymer
blends, polymorphic polymers, and nanocomposites.

Acknowledgments The POLYMAT/UPV/EHU team would like to acknowledge funding from

the following projects: “UPV/EHU Infrastructure: INF 14/38”; “MINECO/FEDER:
SINF130I001726XV1/Ref.: UNPV13-4E-1726” and “MINECO MAT2014-53437-C2-2-P”.

References

1. Blundell DJ, Keller A, Kovacs AJ (1966) A new self-nucleation phenomenon and its
application to the growing of polymer crystals from solution. J Polym Sci B Polym Lett
4:481–486
2. Fillon B, Wittmann JC, Lotz B, Thierry A (1993) Self-nucleation and recrystallization of
isotatic polypropylene (α phase) investigated by differential scanning calorimetry. J Polym
Sci B 31:1383–1393
3. Müller AJ, Balsamo V, Arnal ML (2005) Nucleation and crystallization in diblock and
triblock copolymers. Adv Polym Sci 190:1–63
4. Lorenzo AT, Arnal ML, Sánchez JJ, Müller AJ (2006) Effect of annealing time on the self-
nucleation behavior of semicrystalline polymers. J Polym Sci Polym Phys 44:1738–1750
5. Vanden Poel G, Mathot VBF (2007) High performance differential scanning calorimetry
(HPer DSC): a powerful analytical tool for the study of the metastability of polymers.
Thermochim Acta 461:107–121
6. Müller AJ, Michell RM, Pérez RA, Lorenzo AT (2015) Successive self-nucleation and
annealing (SSA): correct design of thermal protocol and applications. Eur Polym J 65:132–
154
7. Müller AJ, Arnal ML (2005) Thermal fractionation of polymers. Prog Polym Sci 30:559–603
8. Arandia I, Mugica A, Zubitur M et al (2015) How composition determines the properties of
isodimorphic poly(butylene succinate-ran-butylene azelate) random biobased copolymers:
from single to double crystalline random copolymers. Macromolecules 48:43–57
248 R.M. Michell et al.

9. Allegra G (ed) (2005) Advances in polymer science series: interphases and mesophases in
polymer crystallization, vol I–III. Springer, Berlin
10. Bassett DC, Turner B (1974) On the phenomenology of chain-extended crystallization in
PE. Philos Mag 29:925–955
11. Rastogi S, Hikosaka M, Kawabata H, Keller A (1991) Role of mobile phases in the
crystallization of polyethylene. Part 1. Metastability and lateral growth. Macromolecules
24:6384–6391
12. Okada T, Saito H, Inoue T (1992) Time-resolved light scattering studies on the early stage of
crystallization in isotactic polypropylene. Macromolecules 25:1908–1911
13. Imai M, Mori K, Mizukami T, Kaji K, Kanaya Y (1992) Structural formation of poly
(ethylene terephthalate) during the induction period of crystallization: 2. Kinetic analysis
based on the theories of phase separation. Polymer 33:4457–4462
14. Ezquerra TA, L
opez-Cabarcos E, Hsiao BS, Balta-Calleja FJ (1996) Precursors of crystalli-
zation via density fluctuations in stiff-chain polymers. Phys Rev E Stat Phys Plasmas Fluids
Relat Interdiscip Topics 54:989–992
15. Terrill NJ, Fairclough PA, Towns-Andrews E, Komanschek BU, Young RJ, Ryan AJ (1998)
Density fluctuations: the nucleation event in isotactic polypropylene crystallization. Polymer
39:2381–2385
16. Olmsted PD, Poon WCK, McLeish TCB, Terrill NJ, Ryan AJ (1998) Spinodal-assisted
crystallization in polymer melts. Phys Rev Lett 81:373–376
17. Matsuba G, Kaji K, Nishida K, Kanaya T, Imai M (1999) Conformational change and
orientation fluctuations of isotactic polystyrene prior to crystallization. Polym J 31:722–727
18. Strobl G (2000) From the melt via mesomorphic and granular crystalline layers to lamellar
crystallites: a major route followed in polymer crystallization. Eur Phys J E 3:165–183
19. Lotz B (2000) What can polymer crystal structure tell about polymer crystallization pro-
cesses. Eur Phys J E 3:185–194
20. Cho K, Saheb DN, Choi J, Yang H (2002) Real time in situ X-ray diffraction studies on the
melting memory effect in the crystallization of β-isotactic polypropylene. Polymer 43:1407–
1416
21. Lotz B (2005) Analysis and observation of polymer crystal structures at the individual stem
level. Adv Polym Sci 180:17–44
22. Li L, De Jeu WH (2005) Flow-induced mesophases in crystallizable polymers. Adv Polym
Sci 181:75–120
23. Allegra G, Meille SV (2005) Pre-crystalline, high-entropy aggregates: a role in polymer
crystallization. Adv Polym Sci 191:87–135
24. Kaji K, Nishida K, Kanaya T, Matsuba G, Konishi T, Imai M (2005) Spinodal crystallization
of polymers: crystallization from the unstable melt. Adv Polym Sci 191:187–240
25. Muthukumar M (2005) Modeling polymer crystallization. Adv Polym Sci 191:241–274
26. Alfonso GC, Ziabicki A (1995) Memory effects in isothermal crystallization II. Isotactic
polypropylene. Colloid Polym Sci 273:317–323
27. Alfonso GC, Scardigli P (1997) Melt memory effects in polymer crystallization. Macromol
Symp 118:323–328
28. Balzano L, Rastogi S, Peters G (2011) Self-nucleation of polymers with flow: the case of
bimodal polyethylene. Macromolecules 44:2926–2933
29. Bastiaansen CWM, Meyer HEH, Lemstra PJ (1990) Memory effects in polyethylenes:
influence of processing and crystallization history. Polymer 31:1435–1440
30. Cavallo D, Zhang L, Portale G, Alfonso GC, Janani H, Alamo RG (2014) Unusual crystal-
lization behavior of isotactic polypropylene and propene/1-alkene copolymers at large
undercoolings. Polymer 55:3234–3241
31. Cho K, Saheb DN, Yanga H, Kanga BI, Kim J, Lee SS (2003) Memory effect of locally
ordered α-phase in the melting and phase transformation behavior of β-isotactic polypropyl-
ene. Polymer 44:4053–4059
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 249

32. Di Lorenzo ML, Righet MC (2004) Morphological analysis of poly(butylene terephthalate)

spherulites during fusion. Polym Bull 53:53–62
33. Gurarslan A, Joijode AS, Tonelli AE (2012) Polymers coalesced from their cyclodextrin
inclusion complexes: what can they tell us about the morphology of melt-crystallized poly-
mers. J Polym Sci B Polym Phys 50:813–823
34. Häfele A, Heck B, Kawai T, Kohn P, Strobl G (2005) Crystallization of a poly(ethylene-co-
octene): I. A precursor structure and two competing mechanisms. Eur Phys J E 16:207–216
35. Häfele A, Heck B, Hippler T, Kawai T, Kohn P, Strobl G (2005) Crystallization of poly
(ethylene-co-octene): II. Melt memory effects on first order kinetics. Eur Phys J E 16:217–
224
36. Horst RH, Winter HH (2000) Stable critical gels of a copolymer of ethene and 1-butene
achieved by partial melting and recrystallization. Macromolecules 33:7538–7543
37. Gao H, Vadlamudi M, Alamo RG, Hu W (2013) Monte Carlo simulations of strong memory
effect of crystallization in random copolymers. Macromolecules 46:6498–6506
38. He Y, Xu Y, Wei J, Fan Z, Li S (2008) Unique crystallization behavior of poly(l-lactide)/poly
(d-lactide) stereocomplex depending on initial melt states. Polymer 49:5670–5675
39. Jorda R, Wilkes GL (1988) Rapid recrystallization of freshly melted spherulites. Polym Bull
19:409–415
40. Kawabata J, Matsuba G, Nishida K, Inoue R, Kanaya T (2011) Melt memory effects on
recrystallization of polyamide 6 revealed by depolarized light scattering and small-angle X-
ray scattering. J Appl Polym Sci 122:1913–1920
41. Li X, Su F, Ji Y et al (2013) Influence of the memory effect of a mesomorphic isotactic
polypropylene melt on crystallization behavior. Soft Matter 9:8579–8588
42. Li X, Ma Z, Su F et al (2014) New understanding on the memory effect of crystallized iPP.
Chin J Polym Sci 32:1224–1233
43. Luo C, Sommer JU (2013) Disentanglement of linear polymer chains toward unentangled
crystals. ACS Macro Lett 2:31–34
44. Martins JA, Zhang W, Brito AM (2010) Origin of the melt memory effect in polymer
crystallization. Polymer 51:4185–4194
45. Massa MV, Lee MSM, Dalnoki-Veress K (2005) Crystal nucleation of polymers confined to
droplets: memory effects. J Polym Sci B Polym Phys 43:3438–3443
46. Maus A, Hempel E, Thurn-Albrecht T, Saalwächter K (2007) Memory effect in isothermal
crystallization of syndiotactic polypropylene - role of melt structure and dynamics. Eur Phys J
E 23:91–101
47. Mendez G, Müller AJ (1997) Evidences of the crystalline memory and recrystallisation
capacity of bisphenol-A polycarbonate. J Therm Anal 50:593–602
48. Prox M, Pomnimit B, Yarga J, Ehrenstein GW (1990) Thermoanalytical investigations of
self-reinforced polyethylene. J Therm Anal 36:1675–1684
49. Rault J, Robelin E (1980) Memory effect in liquid polyolefine. Polym Bull 2:373–381
50. Reid BO, Vadlamudi M, Mamun A et al (2013) Strong memory effect of crystallization above
the equilibrium melting point of random copolymers. Macromolecules 46:6485–6497
51. Stribeck N, B€osecke P, Bayer R, Camarillo AA (2005) Structure transfer between a polymer
melt and the solid state. Investigation of the nanostructure evolution in oriented polyethylene
by means of continuous X-ray scattering. Progr Colloid Polym Sci 130:127–139
52. Varga J, Menczel J, Solti A (1976) Memory effect of low-density polyethylene crystallized in
a stepwise manner. J Therm Anal 10:433–440
53. Varga J, Menczel J, Solti A (1979) The melting of high-pressure polyethylene subjected to
stepwise heat treatment. J Therm Anal 17:333–342
54. Varga J, Schulek-T
oth F, Ille A (1991) Effect of fusion conditions of β-polypropylene on the
new crystallization. Colloid Polym Sci 269:655–666
55. Varga J, Schulek-T
oth E (1996) Crystallization, melting and spherulitic structure of Β-
nucleated random propylene copolymers. J Therm Anal 47:941–955
250 R.M. Michell et al.

56. Vasanthan N (2003) “Orientation induced memory effect” in polyamides and the relationship
to hydrogen bonding. J Appl Polym Sci 90:772–775
57. Wang M, Hu W, Ma Y, Ma Y (2005) Orientational relaxation together with polydispersity
decides precursor formation in polymer melt crystallization. Macromolecules 38:2806–2812
58. Xu J, Ma Y, Hu W, Rehahn M, Reiter G (2009) Cloning polymer single crystals through self-
seeding. Nat Mater 8:348–353
59. Yamato M, Kimura T (2006) Relationship between magnetic alignment and the crystalliza-
tion condition of isotactic polystyrene. Sci Technol Adv Mater 7:337–341
60. Zheng C, Zhang X, Dong X et al (2006) Variations of regular conformation structures in melt
of syndiotactic polypropylene. Polymer 47:7813–7820
61. Zhou T, Yang H, Ning N, Xiang Y, Du R, Fu Q (2010) Partial melting and recrystallization of
isotactic polypropylene. Chin J Polym Sci 28:77–83
62. Zhu H, Yang H, Zhao Y, Wang D (2011) The dynamic crystallization and multiple melting
behavior of polypropylene in the in-reactor alloy: a differential scanning calorimetry study. J
Appl Polym Sci 121:1372–1383
63. Zhu X, Li Y, Yan D, Zhu P, Lu Q (2001) Influence of the order of polymer melt on the
crystallization behavior: I. Double melting endotherms of isotactic polypropylene. Colloid
Polym Sci 279:292–296
64. Turska E, Gogolewski S (1975) Study on crystallization of nylon 6 (polycapramide). III.
Effect of “crystalline memory” on crystallization kinetics. J Appl Polym Sci 19:637–644
65. Di Filippo G, Gonzalez ME, Gasiba MT, Müller AJ (1987) Crystalline memory on polycar-
bonate. J Appl Polym Sci 34:1959–1966
66. Mehl NA, Rebenfeld L (1992) Effect of melt history on the crystallization kinetics of poly
(phenylene sulfide). Polym Eng Sci 32:1451–1457
67. Fillon B, Lotz B, Thierry A, Wittman JC (1993) Self-nucleation and enhanced nucleation of
polymers. Definition of a convenient calorimetric “efficiency scale” and evaluation of
nucleating additives in isotactic polypropylene (α phase). J Polym Sci B Polym Phys
31:1395–1405
68. Fillon B, Thierry A, Wittman JC, Lotz B (1993) Self-nucleation and recrystallization of
polymers. Isotactic polypropylene, β phase: β-α conversion and β-α growth transitions. J
Polym Sci B Polym Phys 31:1407–1424
69. Supaphol P, Spruiell JE (2000) Crystalline memory effects in isothermal crystallization of
syndiotactic polypropylene. J Appl Polym Sci 75:337–346
70. Gallez F, Legras R, Mercier JP (1976) Crystallization of bisphenol-A polycarbonate.
I. Influence of trimellitic acid tridecyloctyl ester on the kinetics of crystallization. J Polym
Sci Polym Phys Ed 14:1367–1377
71. Khanna YP, Reimschuessel AC (1988) Memory effects in polymers. I. Orientational memory
in the molten state; its relationship to polymer structure and influence on recrystallization rate
and morphology. J Appl Polym Sci 35:2259–2268
72. Khanna YP, Kumar R, Reimschuessel AC (1988) Memory effects in polymers. III.
Processing history vs. crystallization rate of nylon 6-comments on the origin of memory
effect. Polym Eng Sci 28:1607–1611
73. Kim SP, Kim SC (1993) Crystallization kinetics of poly(ethylene terephthalate): memory
effect of shear history. Polym Eng Sci 33:83–91
74. Khanna YP, Kuhn WP, Macur JE et al (1995) Memory effects in polymers. V. Processing
history versus thermally induced self-orientation of unoriented poly(chlorotrifluoroethylene)
films. J Polym Sci B Polym Phys 33:1023–1030
75. Schneider S, Drujon X, Lotz B, Wittmann JC (2001) Self-nucleation and enhanced nucleation
of polyvinylidene fluoride (α-phase). Polymer 42:8787–8798
76. Supaphol P, Srimoaon P, Sirivat A (2004) Effects of crystalline and orientational memory
phenomena on the isothermal bulk crystallization and subsequent melting behavior of poly
(trimethylene terephthalate). Polym Int 53:1118–1126
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 251

77. Rieger J (2003) Polymer crystallization viewed in the general context of particle formation
and crystallization. In: Sommer JU, Reiter G (eds) Polymer crystallization: observations,
concepts and interpretations. Springer, Berlin, pp 7–16
78. Azzurri F, Alfonso GC (2008) Insights into formation and relaxation of shear-induced
nucleation precursors in isotactic polystyrene. Macromolecules 41:1377–1383
79. Somani RH, Yang L, Hsiao BS (2002) Precursors of primary nucleation induced by flow in
isotactic polypropylene. Physica A 304:145–157
80. Azzurri F, Alfonso GC (2005) Lifetime of shear-induced crystal nucleation precursors.
Macromolecules 38:1723–1728
81. Cavallo D, Azzurri F, Balzano L, Funari SS, Alfonso GC (2010) Flow memory and stability
of shear-induced nucleation precursors in isotactic polypropylene. Macromolecules 43:9394–
9400
82. Gai JG, Cao Y (2013) Structure memory effects and rheological behaviors of polyethylenes
in processing temperature window. J Appl Polym Sci 129:354–361
83. Janeschitz-Kriegla H, Ratajski E (2005) Kinetics of polymer crystallization under processing
conditions: transformation of dormant nuclei by the action of flow. Polymer 46:3856–3870
84. Somani RH, Sics I, Hsiao BS (2006) Thermal stability of shear-induced precursor structures
in isotactic polypropylene by rheo-X-ray techniques with couette flow geometry. J Polym Sci
B Polym Phys 44:3553–3570
85. Tao F, Bonnaud L, Auhl D, Struth B, Dubois P, Bailly C (2012) Influence of shear-induced
crystallization on the electrical conductivity of high density polyethylene carbon nanotube
nanocomposites. Polymer 53:5909–5916
86. Zhang W, Martins JA (2006) Evaluation of the effect of melt memory on shear-induced
crystallization of low-density polyethylene. Macromol Rapid Commun 27:1067–1072
87. Arnal ML, Matos ME, Morales RA, Santana O, Müller AJ (1998) Evaluation of the fraction-
ated crystallization of dispersed polyolefins in a polystyrene matrix. Macromol Chem Phys
199:2275–2288
88. Arnal ML, Müller AJ, Maiti P, Hikosaka M (2000) Nucleation and crystallization of isotactic
poly(propylene) droplets in an immiscible polystyrene matrix. Macromol Chem Phys
201:2493–2504
89. C
ordova ME, Lorenzo AT, Müller AJ, Gani L, Tence-Girault S, Leibler L (2011) The
influence of blend morphology (co-continuous or sub-micrometer droplets dispersions) on
the nucleation and crystallization kinetics of double cerystalline polyethylene/polyamide
blends prepared by reactive extrusion. Macromol Chem Phys 212:1335–1350
90. Gao Y, Liu H (2007) Crystallization behavior of dry-brush PEO-PS block copolymer and
PEO homopolymer blend. J Appl Polym Sci 106:2718–2723
91. Ibarretxe J, Groeninckx G, Bremer L, Mathot VBF (2009) Quantitative evaluation of
fractionated and homogeneous nucleation of polydisperse distributions of water-dispersed
maleic anhydride-grafted grafted polypropylene. Polymer 50:4584–4595
92. Luo C, Han X, Gao Y, Liu H, Hu Y (2009) Crystallization behavior of “wet brush” and “dry
brush” blends of PS-b-PEO-b-PS/h-PEO. J Appl Polym Sci 113:907–915
93. Morales RA, Arnal ML, Müller AJ (1995) The evaluation of the state of dispersion in
immiscible blends where the minor phase exhibits fractionated crystallization. Polym Bull
35:379–386
94. Müller AJ, Arnal ML, L
opez-Carrasquero F (2002) Nucleation and crystallization of PS-b-
PEO-b-PCL triblock copolymers. Macromol Symp 183:199–204
95. Müller AJ, Albuerne J, Esteves LM et al (2004) Confinement effects on the crystallization
kinetics and self-nucleation of double crystalline poly(p-dioxanone)-b-poly(ε-caprolactone)
diblock copolymers. Macromol Symp 215:369–382
96. Müller AJ, Albuerne JML, Raquez JM et al (2005) Self-nucleation and crystallization
kinetics of double crystalline poly(p-dioxanone)-b-poly(e-caprolactone) diblock copolymers.
Faraday Discuss 128:231–252
252 R.M. Michell et al.

97. Arnal ML, L
opez-Carrasquero F, Laredo E, Müller AJ (2004) Coincident or sequential

crystallization of PCL and PEO blocks within polystyrene-b-poly(ethylene oxide)-b-poly(ε-
caprolactone) linear triblock copolymers. Eur Polym J 40:1461–1476
98. Müller AJ, Balsamo V, Arnal ML, Jakob T, Schmalz H, Abetz V (2002) Homogeneous
nucleation and fractionated crystallization in block copolymers. Macromolecules 35:3048–
3058
99. Manaure AC, Müller AJ (2000) Nucleation and crystallization of blends of poly(propylene)
and ethylene/α-olefin copolymers. Macromol Chem Phys 201:958–972
100. Molinuevo CH, Mendez GA, Müller AJ (1998) Nucleation and crystallization of PET
droplets dispersed in an amorphous PC matrix. J Appl Polym Sci 70:1725–1735
101. Castillo RV, Müller AJ (2009) Crystallization and morphology of biodegradable or biostable
single and double crystalline block copolymers. Prog Polym Sci 34:519–560
102. Müller AJ, Balsamo V, Arnal ML (2007) Crystallization in block copolymers with more than
one crystallizable block. In: Reiter G, Strobl G (eds) Lecture notes in physics: progress in
understanding of polymer crystallization, vol 714. Springer, Berlin, pp 229–259
103. Müller AJ, Arnal ML, Lorenzo AT (2013) Crystallization in nano-confined polymeric
systems. In: Piorkowska E, Rutledge G (eds) Handbook of polymer crystallization. Wiley,
New York, pp 347–378
104. Michell RM, Blaszczy-Lezak I, Mijangos C, Müller AJ (2013) Confinement effects on
polymer crystallization: from droplets to alumina nanopores. Polymer 54:4059–4077
105. Michell RM, Blaszczyk-Lezak I, Mijangos C, Müller AJ (2014) Confined crystallization of
polymers within anodic aluminum oxide templates. J Polym Sci B Polym Phys 52:1179–1194
106. Santana OO, Müller AJ (1994) Homogeneous nucleation of the dispersed crystallisable
component of immiscible polymer blends. Polym Bull 32:471–477
107. Manaure AC, Morales RA, Sánchez JJ, Müller AJ (1997) Rheological and calorimetric
evidences of the fractionated crystallization of iPP dispersed in ethylene/α-olefin copolymers.
J Appl Polym Sci 66:2481–2493
108. Chen HL, Wu JC, Lin TL, Lin JS (2001) Crystallization kinetics in microphase-separated
poly(ethylene oxide)-block-poly(1,4-butadiene). Macromolecules 34:6936–6944
109. Balsamo V, Paolini Y, Ronca G, Müller AJ (2000) Crystallization of the polyethylene block
in polystyrene-b-polyethylene-b-polycaprolactone triblock copolymers, 1. Self-nucleation
behavior. Macromol Chem Phys 201:2711–2720
110. Schmalz H, Muller AJ, Abetz V (2003) Crystallization in ABC triblock copolymers with two
different crystalline end blocks: influence of confinement on self-nucleation behavior.
Macromol Chem Phys 204:111–124
111. Boschetti-de-Fierro A, Lorenzo AT, Müller AJ, Schmalz H, Abetz V (2008) Crystallization
kinetics of PEO and PE in different triblock terpolymers: effect of microdomain geometry
and confinement. Macromol Chem Phys 209:476–487
112. Huang CL, Jiao L, Zeng JB, Zhang JJ, Yang KK, Wang YZ (2013) Fractional crystallization
and homogeneous nucleation of confined PEG microdomains in PBS-PEG multiblock copol-
ymers. J Phys Chem B 117:10665–10676
113. Castillo RV, Müller AJ, Lin MC, Chen HL, Jeng US, Hillmyer MA (2008) Confined
crystallization and morphology of melt segregated PLLA-b-PE and PLDA-b-PE diblock
copolymers macromolecules. Macromolecules 41:6154–6164
114. Lin MC, Chen HL, Lin WF, Huang PS, Tsai JC (2012) Crystallization of isotactic polypro-
pylene under the spatial confinement templated by block copolymer microdomains. J Phys
Chem B 116(40):12357–12371
115. Müller AJ, Lorenzo AT, Arnal ML, Boschetti de Fierro A, Abetz V (2006) Self-nucleation
behavior of the polyethylene block as function of the confinement degree in polyethylene-
block-polystyrene diblock copolymers. Macromol Symp 240:114–122
116. Lorenzo AT, Arnal ML, Müller AJ, Boschetti-de-Fierro A, Abetz V (2007) Nucleation and
isothermal crystallization of the polyethylene block within diblock copolymers containing
polystyrene and poly(ethylene-alt-propylene). Macromolecules 40:5023–5037
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 253

117. Lorenzo AT, Müller AJ, Lin MC et al (2009) Influence of macromolecular architecture on the
crystallization of (PCL2)-b-(PCL2) 4-miktoarm star block copolymers in comparison to
linear PCL-b-PS diblock copolymer analogues. Macromolecules 42:8353–8364
118. Pan Y, Yu X, Shi T, An L (2010) Nucleation and crystallization of H-shaped (PS)2PEG(PS)2
block copolymers. Chin J Polym Sci 28:347–355
119. Pérez RA, C
ordova ME, L
opez JV et al (2014) Nucleation, crystallization, self-nucleation
and thermal fractionation of cyclic and linear poly(ε-caprolactone)s. React Funct Polym
80:71–82
120. Chen H, Hsiao S, Lin T, Yamauchi K, Hasegawa H, Hashimoto T (2001) Microdomain-
tailored crystallization kinetics of block copolymers. Macromolecules 34:671–674
121. Hamley IW (1998) The physics of block copolymers. Oxford University Press, London
122. Wunderlich B (1976) Macromolecular physics. Crystal nucleation, growth, annealing, vol
2. Academic, New York
123. Supaphol P, Lin JS (2001) Crystalline memory effect in isothermal crystallization of
syndiotactic polypropylenes: effect of fusion temperature on crystallization and melting
behavior. Polymer 42:9617–9626
124. Mamun A, Unemoto S, Okui N, Ishihara N (2007) Microdomain-tailored crystallization
kinetics of block copolymers. Macromolecules 40:6296–6303
125. Zhang YS, Zhong LW, Yang S, Liang DH, Chen EQ (2012) Memory effect on solution
crystallization of high molecular weight poly(ethylene oxide). Polymer 53:3621–3628
126. Cheng SZD (2008) Phase transitions in polymers: the role of metastable states, 1st edn.
Elsevier Science, Amsterdam
127. Mamun A, Chen X, Alamo RG (2014) Interplay between a strong memory effect of
crystallization and liquid-liquid phase separation in melts of broadly distributed ethylene-1-
alkene copolymers. Macromolecules 47:7958–7970
128. Wang Y, Lu Y, Zhao J, Jiang Z, Men Y (2014) Direct formation of different crystalline forms
in butene-1/ethylene copolymer via manipulating melt temperature. Macromolecules
47:8653–8662
129. Cavallo D, Gardella L, Portale G, Müller AJ, Alfonso GC (2013) On cross- and self-
nucleation in seeded crystallization of isotactic poly(1-butene). Polymer 54:4637–4644
130. Cavallo D, Gardella L, Portale G, Müller AJ, Alfonso GC (2014) Self-nucleation of isotactic
poly(1-butene) in the trigonal modification. Polymer 55:137–142
131. Chocinski-Arnault L, Gaudefroy V, Gacougnolle JL, Rivière A (2002) Memory effect and
crystalline structure in polyamide 11. J Macromol Sci B 41:777–785
132. Dai PS, Cebe P, Capel M (2002) Thermal analysis and X-ray scattering study of metallocene
isotactic polypropylene prepared by partial melting. J Polym Sci B Polym Phys 40:1644–
1660
133. Kang J, Zhang J, Chen Z et al (2014) Isothermal crystallization behavior of β-nucleated
isotactic polypropylene with different melt structures. J Polym Res 21:506
134. Kang J, Chen Z, Chen J et al (2015) Crystallization and melting behaviors of the β-nucleated
isotactic polypropylene with different melt structures - the role of molecular weight.
Thermochim Acta 599:42–51
135. Na B, Pan H, Lv R, Zhu J, Li C (2012) A facile route to ordered γ phase in poly(vinylidene
fluoride) via melt annealing and re-crystallization. Mater Lett 85:37–39
136. Pan H, Na B, Lv R, Li C, Zhu J, Yu Z (2012) Polar phase formation in poly(vinylidene
fluoride) induced by melt annealing. J Polym Sci B Polym Phys 50:1433–1437
137. Schenk H, Peschar R (2004) Understanding the structure of chocolate. Radiat Phys Chem
71:829–835
138. Wu MC, Woo EM (2005) Effects of α-form or β-form nuclei on polymorphic crystalline
morphology of poly(butylene adipate). Polym Int 54:1681–1688
139. Gu Q, Wu L, Wu D, Shen D (2001) Effect of self-nucleation on the crystallization of
segmented copolymer poly(ether ester). J Appl Polym Sci 81:498–504
254 R.M. Michell et al.

140. Kang J, Li J, Chen S et al (2013) Investigation of the crystallization behavior of isotactic

polypropylene polymerized with different Ziegler-Natta catalysts. J Appl Polym Sci
129:2663–2670
141. Kang J, Li J, Chen S et al (2013) Hydrogenated petroleum resin effect on the crystallization of
isotactic polypropylene. J Appl Polym Sci 130:25–38
142. Lorenzo AT, Müller AJ (2008) Estimation of the nucleation and crystal growth contributions
to overall crystallization energy barrier. J Polym Sci B Polym Phys 46:1478–1487
143. Nanaki SG, Papageorgiou GZ, Bikiaris DN (2012) Crystallization of novel poly(ε-caprolactone)-
block-poly(propylene adipate) copolymers. J Therm Anal Calorim 108:633–645
144. Papageorgiou GZ, Achilias DS, Karayannidis GP (2010) Estimation of thermal transitions in
poly(ethylene naphthalate): experiments and modeling using isoconversional methods. Poly-
mer 51:2565–2575
145. Papageorgiou GZ, Panayiotou C (2011) Crystallization and melting of biodegradable poly
(propylene suberate). Thermochim Acta 523:187–199
146. Papageorgiou ZG, Bikiaris DN, Achilias DS (2012) Spherulite growth rates of in situ
prepared poly(propylene terephthalate)/SiO2 nanocomposites. J Therm Anal Calorim
114:431–440
147. Papageorgiou GZ, Achilias DS, Bikiaris DN (2009) Crystallization kinetics and melting
behaviour of the novel biodegradable polyesters poly(propylene azelate) and poly(propylene
sebacate). Macromol Chem Phys 210:90–107
148. Papageorgiou GZ, Bikiaris DN, Achilias DS, Karagiannidis N (2010) Synthesis, crystalliza-
tion, and enzymatic degradation of the biodegradable polyester poly(ethylene azelate).
Macromol Chem Phys 211:2585–2595
149. Sisti L, Finelli L, Lotti N, Berti C, Munari A (2003) Memory effect in melting behaviour,
crystallization kinetics and morphology of poly(propylene terephthalate). ePolymers 54:1–19
150. Xu Y, Wang Y, Xu T, Jingjing Z, Liu C, Shen C (2014) Crystallization kinetics and
morphology of partially melted poly(lactic acid). Polym Test 37:179–185
151. Zhu X, Li Y, Yan D, Fang Y (2001) Crystallization behavior of partially melting isotactic
polypropylene. Polymer 42:9217–9222
152. Zhao Y, Vaughan AS, Sutton SJ, Swingler SG (2001) On nucleation and the evolution of
morphology in a propylene/ethylene copolymer. Polymer 42:6599–6608
153. Albuerne J, Marquez L, Müller AJ et al (2003) Nucleation and crystallization in double
crystalline poly(p-dioxanone)-b-poly(ε-caprolactone) diblock copolymers. Macromolecules
36:1633–1644
154. Cai J, Li T, Han Y, Zhuang Y, Zhang X (2006) Nonisothermal crystallization kinetics and
morphology of self-seeded syndiotactic 1,2-polybutadiene. J Appl Polym Sci 100:1479–1491
155. Zheng H, Wang B, Zheng G, Wang Z, Dai K, Liu C (2014) Study on crystallization kinetics
of partially melting polyethylene aiming to improve mechanical properties. Ind Eng Chem
Res 53:6211–6220
156. Schulze R, Arras MML, Helbing C et al (2014) How the calorimetric properties of a
crystalline copolymer correlate to its surface nanostructures. Macromolecules 47:1705–1714
157. Kanga J, Penga H, Wanga B et al (2015) Investigation on the self-nucleation behavior of
controlled-rheology polypropylene. J Macromol Sci B 54:127–142
158. Tidick P, Fakirov S, Avramova N, Zachmann HG (1984) Effect of the melt annealing time on
the crystallization of nylon-6 with various molecular weights. Colloid Polym Sci 262:445–
449
159. Fillon B, Lotz B, Thierry A, Wittmann JC (1993) Self-nucleation and enhanced nucleation of
polymers. Definition of a convenient calorimetric “efficiency scale” and evaluation of
nucleating additives in isotactic polypropylene (α phase). J Polym Sci B 31:1395–1405
160. Abraham F, Schmidt HW (2010) 1,3,5-Benzenetrisamide based nucleating agents for poly
(vinylidene fluoride). Polymer 51:913–921
161. Anderson KS, Hillmyer MA (2006) Melt preparation and nucleation efficiency of polylactide
stereocomplex crystallites. Polymer 47:2030–2035
Self-Nucleation of Crystalline Phases Within Homopolymers, Polymer Blends. . . 255

162. Bouza R, Marco C, Naffakh M, Barral L, Ellis G (2011) Effect of particle size and a
processing aid on the crystallization and melting behavior of iPP/red pine wood flour
composites. Compos Part A 42:935–949
163. Dai J, Shen Y, Yang J, Huang T, Zhang N, Wang Y (2014) Crystallization and melting
behaviors of polypropylene admixed by graphene and β-phase nucleating agent. Colloid
Polym Sci 292:923–933
164. Fanegas N, G
omez MA, Marco C, Jiménez I, Ellis G (2007) Influence of a nucleating agent
on the crystallization behaviour of isotactic polypropylene and elastomer blends. Polymer
48:5324–5331
165. Gahleitner M, Kheirandish GS, Wolfschwenger J (2011) Nucleation of polypropylene homo
and copolymers. Int Polym Proc 26:2–20
166. Laoutid F, Estrada E, Michell RM, Bonnaud L, Müller AJ, Dubois P (2013) The influence of
nanosilica on the nucleation, crystallization and tensile properties of PP-PC and PP-PA
blends. Polymer 54:3982–3993
167. Libster D, Aserin A, Garti N (2007) Advanced nucleating agents for polypropylene. Polym
Adv Technol 18:685–695
168. Müller AJ, Arnal ML, Trujillo M, Lorenzo AT (2011) Super-nucleation in nanocomposites
and confinement effects on the crystallizable components within block copolymers,
Miktoarm star copolymers and nanocomposites. Eur Polym J 47:614–629
169. Patil N, Invigorito C, Gahleitner M, Rastogi S (2013) Influence of a particulate nucleating
agent on the quiescent and flow-induced crystallization of isotactic polypropylene. Polymer
54:5883–5891
170. Pérez RA, L
opez JV, Hoskins JN et al (2014) Macromolecules nucleation and antinucleation
effects of functionalized carbon nanotubes on cyclic and linear poly(ε-caprolactones). Mac-
romolecules 47:3553–3566
171. Priftis D, Sakellariou G, Hadjichristidis N, Penott E, Lorenzo AT, Müller AJ (2009) Surface
modification of multiwalled carbon nanotubes with biocompatible polymers via ring opening
and living anionic surface initiated polymerization. Kinetics and crystallization behavior. J
Polym Sci A Polym Chem 47:4379–4390
172. Pucciariello R, Villani V, Giammarino G (2011) Thermal behaviour of nanocomposites based
on linear-low-density poly(ethylene) and carbon nanotubes prepared by high energy ball
milling. J Polym Res 18:949–956
173. Song P, Wei Z, Liang J, Chen G, Zhang W (2012) Crystallization behavior and nucleation
analysis of poly(l-lactic acid) with a multiamide nucleating agent. Polym Eng Sci 52:1058–
1068
174. Trujillo M, Arnal ML, Müller AJ et al (2007) Thermal and morphological characterization of
nanocomposites prepared by in-situ polymerization of high-density polyethylene on carbon
nanotubes. Macromolecules 40:6268–6276
175. Trujillo M, Arnal ML, Müller AJ et al (2012) Supernucleation and crystallization regime
change provoked by MWNT addition to poly(ε-caprolactone). Polymer 53:832–841
176. Wu Y, Ling HS (2012) The role of surface charge of nucleation agents on the crystallization
behavior of poly(vinylidene fluoride). J Phys Chem B 116:7379–7388
177. Xu S, Zhao X, Ye L (2013) Mechanical and crystalline properties of monomer casting Nylon-
6/SiO2 composites prepared via in situ polymerization. Polym Eng Sci 53:1809–1822
178. Xu Y, Wu L (2013) Synthesis of organic bisurea compounds and their roles as crystallization
nucleating agents of poly(l-lactic acid). Eur Polym J 49:865–872
179. Yang Z, Zhang Z, Tao Y, Mai K (2008) Effects of polyamide 6 on the crystallization and
melting behavior of β-nucleated polypropylene. Eur Polym J 44:3754–3763
180. Zhang Q, Chen Z, Wang B, Chen J, Yang F, Kang J (2015) Effects of melt structure on
crystallization behavior of isotactic polypropylene nucleated with α/β compounded nucleat-
ing agents. J Appl Polym Sci 132:41355
256 R.M. Michell et al.

181. Zhuravlev E, Wurma A, P€ otschke P, Androsch R, Schmelzer JWP, Schick C (2014) Kinetics
of nucleation and crystallization of poly(ε-caprolactone) - multiwalled carbon nanotube
composites. Eur Polym J 52:1–11
182. Sabino MA, Ronca G, Müller AJ (2000) Heterogeneous nucleation and self-nucleation of
poly(p-dioxanone). J Mater Sci 35:5071–5084
183. Wang K, Mai K, Han Z, Ze H (2001) Interaction of self-nucleation and the addition of a
nucleating agent on the crystallization behavior of isotactic polypropylene. J Appl Polym Sci
81:78–84
184. Trujillo M, Arnal ML, Müller AJ et al (2008) Thermal fractionation and isothermal crystal-
lization of polyethylene nanocomposites prepared by in situ polymerization. Macromolecules
41:2087–2095
Adv Polym Sci (2017) 276: 257–288
DOI: 10.1007/12_2015_325
© Springer International Publishing Switzerland 2015
Published online: 2 December 2015

Crystal Nucleation of Polymers at High

Supercooling of the Melt

René Androsch and Christoph Schick

Abstract Analysis of the crystallization kinetics of numerous polymers has

revealed a bimodal dependence of the gross crystallization rate on temperature,
often leading to the occurrence of two crystallization-rate maxima at widely
different temperatures. This review discusses possible reasons for this observation,
including temperature-controlled changes in the mechanism of primary crystal
nucleation, activation of growth at different crystal faces, and formation of different
crystal polymorphs as a result of variation of the supercooling. It is suggested that
crystallization proceeds via homogeneous crystal nucleation at high supercooling
of the melt, which is supported by estimation of the nucleation density from
morphological analyses, crystallization experiments performed on heterogeneity-
free droplets, and a link between the time scales of molecular relaxations in the
glassy state and primary crystal nucleation. The final part of this review presents an
example of the application of Tammann’s nuclei development method to obtain
nucleation rates in polymer glasses.

Keywords Crystal morphology
Homogenous crystal nucleation
Nucleation

density
Nucleation rate
Tammann’s nuclei development method

Contents
1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 258
2 Gross Crystallization Rate of Polymers at Widely Different Supercooling . . . . . . . . . . . . . . 261

R. Androsch (*)
Martin-Luther-University Halle-Wittenberg, Center of Engineering Sciences, 06099 Halle/
Saale, Germany
e-mail: [email protected]
C. Schick
University of Rostock, Institute of Physics, Wismarsche Strasse 43-45, 18051 Rostock,
Germany
258 R. Androsch and C. Schick

3 Nuclei Density by Analysis of the Semicrystalline Morphology . . . . . . . . . . . . . . . . . . . . . . . . . 263

4 Homogeneous Nucleation at High Supercooling: Evidence from Droplet Experiments . 266
5 Sequence of Enthalpy Relaxation, Homogeneous Crystal Nucleation, and Crystal
Growth in the Glassy Amorphous Phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 268
6 Mesophase Formation During Crystallization at High Supercooling . . . . . . . . . . . . . . . . . . . . . 272
7 Crystallization of Poly(L-lactic acid) at High Supercooling: Application of Tammann’s
Nuclei Development Method to Obtain Nucleation Rates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 274
8 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 279
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 281

1 Introduction

Melt-crystallization of macromolecules is restricted to the temperature range

between the equilibrium melting temperature (Tm,0) and a temperature T1, which
is assumed to be 30–50 K below the glass transition temperature (Tg). Whereas the
high-temperature limit of the temperature range of crystallization is thermodynam-
ically controlled, the low-temperature limit is related to the decreasing mobility of
molecular segments, with T1 interpreted as the temperature at which no transport
across the liquid–crystal phase boundary can be expected. Superposition of the
effects of increasing thermodynamic driving force for the liquid–crystal phase
transition and decreasing mobility of molecular segments with increasing
supercooling (where supercooling is defined as the difference between Tm,0 and
the crystallization temperature) typically leads, first, to an increase in the crystal-
lization rate in response to a lower crystallization temperature and, subsequently, to
a decrease as the temperature approaches Tg [1, 2].
Crystallization proceeds via the stages of primary crystal nucleation and crystal
growth. Primary nucleation involves the formation of a nucleus of supercritical size
(i.e., a small crystal that does not disappear with time). Nucleation is controlled by
the interplay between the gain and loss of free enthalpy as a result of the liquid–
crystal phase transformation and the formation of a liquid–crystal interface, respec-
tively. In general, with increasing size of the nucleus at a given temperature, the
total free-enthalpy change goes through a maximum, which defines the critical size
of the nucleus and the free-enthalpy barrier required to allow growth of the nucleus,
connected with a lowering of the free enthalpy. The formation of such a nucleus can
occur in the bulk phase (homogeneous crystal nucleation) or on pre-existing
surfaces/heterogeneities (heterogeneous crystal nucleation) [3–6]. In the latter
case (i.e., in the presence of active sites at surfaces), the total interfacial stress at
the nucleus surface is considered lower than in the case of homogeneous nucleation,
and for this reason heterogeneous nucleation is often thermodynamically favored,
particularly in the temperature range of low supercooling. The critical size of the
nucleus and the free-enthalpy barrier both decrease with increased supercooling
(i.e., with increased thermodynamic driving force for the phase transformation).
From thermodynamic considerations, it is expected that the nucleation frequency
Crystal Nucleation of Polymers at High Supercooling of the Melt 259

for a given nucleation mechanism increases exponentially with the supercooling of

the melt; however, this is opposed by kinetic restraints (i.e., increasing character-
istic time of transport of motifs) [1–6]. As a function of temperature, the dominating
mechanism of nucleation can change because it is controlled by the critical size of
the nucleus and the work required for creation of the surfaces of the crystal nuclei,
both of which depend on the supercooling.
Formation of primary nuclei is followed by their growth to crystals at faces that
are often parallel or nearly parallel to the chain axis. The kinetics of lateral crystal
growth is often modeled by the Hoffman–Lauritzen (HL) approach, in which
different growth regimes are based on the relationship between the crystal surface
nucleation rate and the rate of lateral spreading of such secondary crystal nuclei
across the growth face [7–9]. To account for the large entropic barrier for attaching
a molecular stem at the growth face, several alternative models have been intro-
duced. However, all these models assume that crystals grow directly into the melt;
examples include the model of molecular nucleation by Wunderlich [10] and, more
recently introduced, the model of Strobl proposing that molecular segments at the
crystal growth front pass a transient mesomorphic stage [11–13]. This notwith-
standing, the temperature dependence of the crystal growth rate qualitatively obyes
similar rules as in case of the temperature dependence of the nucleation rate. The
crystal growth rate increases with decreasing temperature as a result of the increas-
ing thermodynamic driving force for the phase transformation, passes through a
maximum, and then decreases as a result of the reduced mobility of amorphous
molecular segments at the growth faces.
If the rates of both primary crystal nucleation and crystal growth are not too high,
they can be measured separately by optical microscopy [14]. The number of
spherulites forming at a given crystallization temperature provides information
about the nucleation density, whereas the crystal growth rate is evaluated by the
time evolution of the spherulite radius. Such analyses, however, can only be
performed if the nucleation density is relatively low, that is, at low supercooling
when the distance between nuclei is of the order of several micrometers. Alterna-
tively, in such a case, the nucleation density can be evaluated using high-resolution
imaging techniques for analysis of the submicron spherulite or crystal density after
complete crystallization.
A different, although frequently applied, technique for gaining information
about the crystallization rate is calorimetry, which is based on analysis of the
time-dependence of the release of exothermic heat flow during the crystallization
process [14–16]. Although it is impossible to identify the absolute number of
simultaneously growing entities contributing to the increase in crystallinity, it has
been shown that application of the Avrami model can give information about
nucleation and growth mechanisms [17–20]. Also, analysis of the temperature
dependence of the gross crystallization rate allows detection of changes in both
growth regimes [21] and nucleation mechanisms with temperature, because the
activation energies for the different processes change. This has been shown with the
recent introduction of fast scanning chip calorimetry (FSC) for analysis of polymer
crystallization at any supercooling of the melt [22–24]. FSC allows not only
260 R. Androsch and C. Schick

adjustment of well-defined states of amorphous structure prior to crystal nucleation

and growth, but also analysis of the kinetics of crystallization even at the temper-
atures of maximum crystallization rate, related to both the short FSC instrumental
time constant and high cooling capacity. It is worth noting that the maximum
cooling rate in conventional differential scanning calorimeters (DSC) is of the
order of only a few hundred degrees Kelvin per minute [25–27], similar to that
for commercial hot-stage microscopy systems, although efforts have been made to
widen the cooling-rate range [28–30]. However, such a low cooling rate does not
often allow large supercooling of the melt because crystallization may begin before
the analysis temperature has been reached [31]. This is particularly true for poly-
mers containing heterogeneous nucleators, regardless of whether they are present
by chance or added purposely. FSC, in contrast, depending on the particular device,
allows cooling of the samples at rates up to 103–106 K s1 [32, 33], which for most
polymers is sufficiently fast to completely suppress crystallization of the melt when
cooling to below T1. This gives the possibility of studying crystallization not only
at low supercooling but in the entire temperature range between Tm, 0 and T1, so
that supercooling-controlled changes in nucleation and growth mechanisms can be
easily recognized. Although not the focus of this review, variation of the cooling
scheme permits (a) controlled generation and freezing of nuclei populations, which
then can be grown to crystals on re-heating or at elevated temperature in order to
obtain information about their density, or (b) to obtain glasses of different instabil-
ities, allowing sub-Tg crystal nucleation and even growth. In addition to the fast-
cooling capacity, FSC also allows analysis of the kinetics of crystallization at
temperatures of maximum crystallization rate. In many cases, the characteristic
crystallization time is less than a second, which does not permit its quantification by
conventional DSC, because of the high instrument time constant and low cooling
capacity. Note that the FSC time constant, depending on the particular sensor and
device employed, is of the order of milliseconds or faster [32, 33].
FSC analysis of the crystallization kinetics of numerous polymers has revealed a
bimodal dependence of the gross crystallization rate on temperature, often leading
to the occurrence of two crystallization-rate maxima at different temperatures, or to
a distinctly broadened crystallization-rate maximum as a result of superposition of
contributions from two different crystallization mechanisms [34–45]. This experi-
mental observation has been explained by a change in the mechanism of primary
crystal nucleation, although without further knowledge and structural analyses such
interpretation is unjustified. The occurrence of two crystallization-rate maxima at
different temperatures can be caused by many reasons, including a change in the
mechanism of growth, activation of growth at different crystal faces, or formation
of different equilibrium or nonequilibrium crystal polymorphs as a function of the
supercooling. Thus, the initial assumption of a qualitative change in the nucleation
mechanism requires further research.This paper is structured to first provide exam-
ples of the bimodal temperature dependence of gross crystallization rate for differ-
ent polymers. This is followed by presentation of research supporting the
assumption that crystallization at different supercooling of the melt is governed
by different mechanisms of nucleation. It includes (a) using imaging techniques for
Crystal Nucleation of Polymers at High Supercooling of the Melt 261

analysis of the semicrystalline morphology of various polymers, each crystallized

at different supercooling, to yield conclusions about the nucleation density;
(b) discussion of the crystallization behavior of heterogeneity-free polymer phases;
and (c) linking the time scale of the molecular relaxations in the glassy state to the
time scale of primary nucleation. In the two final parts of this review, mesophase
formation during crystallization at high supercooling is discussed, and an example
of the application of Tammann’s nuclei development method to obtain nucleation
rates in polymer glasses is presented.

2 Gross Crystallization Rate of Polymers at Widely

Different Supercooling

FSC has been applied to study the gross crystallization rate of several polymers in a
wide range of different supercooling of the melt [34–47]. Samples were heated to
above the melting temperature and after equilibration of the melt cooled to the
crystallization temperature at a rate high enough to avoid crystallization during
cooling, typically with a rate of the order of magnitude of 103 K s1. Isothermal
crystallization leads to observation of an exothermic peak, which often is analyzed
to obtain the so-called peak-time of crystallization, that is, the time of maximum
heat-flow rate during crystallization or the half-transition time, both being measures
of the crystallization rate. Figure 1 shows typical examples of the temperature-
dependence of the characteristic crystallization time obtained for isotactic polypro-
pylene (iPP) [34, 35], polyamide 66 (PA 66) [37], polyamide 11 (PA 11) [38], and
poly(butylene terephthalate) (PBT) [39–41]. Note that in all of the presented
experiments, data sets are shown so as to emphasize the occurrence of two distinct
crystallization-rate maxima or crystallization-time minima. In particular, for all of
the examples, there were numerous data collected from DSC experiments at low
supercooling showing characteristic crystallization times exceeding 102 s. More-
over, the observation of a bimodal dependence of the crystallization rate on
supercooling is not restricted to the example polymers shown in Fig. 1 as similar
data were reported in the literature for polyamide 6 (PA 6) [36], poly(ε-caprolactone)
(PCL) [42, 43], a variety of random copolymers of propylene with 1-alkenes [44, 45],
and syndiotactic polypropylene [46].
The data shown in Fig. 1 demonstrate that, with increasing supercooling, the
characteristic time of crystallization decreases as a result of increasing thermody-
namic driving force for the crystallization, presumably leading to a continuous
increase in both the rate of crystal nucleation and growth. It is important to note
that, as a result of the recording of an integrated exothermic heat-flow rate signal,
distinction between these contributions a priori and without modeling is impossible.
The characteristic time of crystallization passes through a first minimum and begins
to increase, probably because the decreasing mobility of chain segments hinders
growth. However, at even larger supercooling there is further increase in the
262 R. Androsch and C. Schick

Fig. 1 Characteristic time

of crystallization of iPP [34,
35], PA 66 [37], PA 11 [38],
and PBT [41] (from top to
bottom) as a function of the
crystallization temperature.
The approximate position of
Tg is indicated in each plot.
The squares and circles
indicate crystallization
processes of different
kinetics. The arrow in the
PBT plot indicates
increasing crystallization
rate. Furthermore, the plots
contain information about
crystal polymorphs forming
at different supercooling.
The lines connecting data
points are drawn to guide
the eye
Crystal Nucleation of Polymers at High Supercooling of the Melt 263

crystallization rate, as indicated by the arrow in the PBT data plot shown in Fig. 1.
Possible reasons have been outlined above and include a sudden increase in the
nucleation density, as a result of changes in the nucleation mechanism, and an
increase in the crystal growth rate. The latter could be the result of a change in the
mechanism of secondary nucleation, activation of crystal growth at faces of lower
surface free energy, or growth of different crystal polymorphs.
In Fig. 1, in addition to the crystallization-rate data, information is provided about
the structure of the ordered phase formed at different supercooling. For iPP, PA
66, PA 11, and PA 6 (not shown), it is proven that during crystallization at low and
high supercooling there is formation of crystals and mesophases, respectively. For
PBT and poly(ethylene terephthalate) (PET; not shown) such supercooling-controlled
crystal/mesophase polymorphism was not detected. As such, at least for some of the
examples shown, it cannot be excluded that the low-and high-temperature
crystallization-rate maxima are the result of formation of different crystal polymorphs.
Detailed information and further discussion on this subject are given in Sect. 6.

3 Nuclei Density by Analysis of the Semicrystalline

Morphology

To obtain at least semiquantitative information about nucleation densities at tem-

peratures associated with high- and low-temperature crystallization processes for
all the examples shown in Fig. 1, detailed morphological analyses were performed
at both the nanometer and micrometer length scales. Figure 2 shows, as a typical
example, polarizing optical microscopy (POM) micrographs of PBT crystallized at
130 C and 70 C, that is, at the temperatures of the high- and low-temperature
crystallization rate maxima, respectively. Crystallization at the higher temperature
reveals a space-filling spherulitic superstructure, with the size of spherulites being
5–20 μm. Observation of spherulites formed during crystallization at low
supercooling, in general, indicates the presence of lamellae, which began to grow
from a former primary crystal nucleus to yield a birefringence pattern with radial
symmetry [48–51]. In the case of the sample crystallized at 70 C, a spherulitic
superstructure was not observed, because the sample appeared black and featureless
between crossed polarizers. It was concluded that in this case crystals are irregu-
larly arranged in the amorphous phase when forming at high supercooling of the
melt. The inset in the right-hand image of Fig. 2 is a soft zoom of a spherulite,
illustrating rotation of the Maltese cross to 45 off the polarizer axes (parallel to the
image borders), indicating formation of so-called abnormal spherulites, which are a
specific feature of PBT crystallization [52, 53]. Qualitatively similar POM images
showing spherulites formed during crystallization at low supercooling and a fea-
tureless birefringence pattern after crystallization at high supercooling or cold-
crystallization (implying nucleation at high supercooling) have also been obtained
for iPP and its random copolymers with 1-alkenes [54–58], for PA 6 [59], and for
PA 11 [38].
264 R. Androsch and C. Schick

Fig. 2 POM images of PBT crystallized at 70 C (left) and 130 C (right), that is, at the temper-
atures of maximum rate of the low- and high-temperature crystallization processes, respectively
(see Fig. 1). The inset is a soft zoom of a spherulite, illustrating rotation of the Maltese cross 45
off the polarizer axes (parallel to the image borders) and indicating formation of so-called
abnormal spherulites, which are a specific feature of PBT crystallization. Reprinted from [41],
Copyright (2015), with permission from Elsevier

Further details of the semicrystalline morphology and nucleation density were

collected by atomic force microscopy (AFM) and transmission electron microscopy
(TEM). Figure 3 shows AFM images of the nanometer length scale surface struc-
ture of PBT [41], PA 11 [38], and PA 6 [60] of different crystallization histories.
The images in the left- and right-hand columns of Fig. 3 were obtained from
samples crystallized at high and low supercooling, respectively (i.e., at tempera-
tures related to the low- and high-temperature crystallization-rate maximum in
Fig. 1). In all the examples presented in Fig. 3, there was formation of laterally
extended lamellae during crystallization at low supercooling of the melt, as con-
cluded from the observation of spherulitic birefringence patterns (see Fig. 2, right).
In contrast, during crystallization at temperatures related to the low-temperature
crystallization rate maxima shown in Fig. 1, formation of lamellae was not detected.
Instead, the AFM images reveal the presence of small and apparently isometric
nodular domains with a size in the order of 10 nm. It is furthermore visible that
these domains do not form a higher-order superstructure at the micrometer length
scale, which is in accord with the featureless POM structure.
Although the observation of lamellae and spherulites after crystallization at low
supercooling is well described in the literature [61], the nanometer-scale morphol-
ogy of polymers crystallized at high supercooling has been less investigated.
However, earlier reports in this field of research are consistent with the images
shown in Fig. 3 (left-hand column), such that there is formation of nodular domains
with a size of less than about 10 nm. Pioneering work in this area has been
performed by Geil [62–64] using an ultraquenching technique and electron micros-
copy to identify nanoscale granular structures in a large number of polymers after
quenching. These observations of a nodular morphology in quenched samples were
initially doubted as being possibly related to instrumental artifacts of electron
microscopy [65, 66]. However, the observations were then confirmed in many
Crystal Nucleation of Polymers at High Supercooling of the Melt 265

Fig. 3 AFM images of PBT [41], PA 11 [38], and PA 6 [60] (from top to bottom). Images in the
left and right columns are of samples crystallized at high and low supercooling, respectively, that
is, at temperatures related to the low- and high-temperature crystallization-rate maximum in Fig. 1.
Scale bars: 50 nm. Images in the top and center rows were adapted from [41], Copyright (2015),
with permission from Elsevier, and from [38], Copyright (2013) American Chemical Society,
respectively. The bottom left image was adapted from [60], Copyright (2012), with permission
from Elsevier

independent studies using a wide variety of different preparation techniques and

instrumentation [67–74].
Most important in the context of this review is the observation that, on crystal-
lization at comparatively high temperature, the distance between former crystal
nuclei is of the order of several micrometers. It cannot be shorter than the lateral
size of lamellae and can be safely quantified by the diameter of spherulites formed
during radial growth of lamellae. In the case of crystallization at high supercooling,
the nucleation density is several orders of magnitude higher, as the distance
between neighboring crystals, and therefore former nuclei, is only 10–20 nm. A
rough estimation of the number of independently grown crystals, that is, formed via
primary nucleation (as shown in the top left image in Fig. 3, representing the
structure of PBT crystallized at 70 C), yields a value of about 1015 nuclei mm3.
266 R. Androsch and C. Schick

In the case of crystallization at 130 C, spherulites with a diameter of about 10 μm

were detected (see right image in Fig. 2), indicating a nucleation density of only
about 106 nuclei mm3 [41]. In other words, the nucleation density for crystalliza-
tion at temperatures close to the low-temperature crystallization-rate maximum is at
least nine orders of magnitude higher than for crystallization at temperatures close
to the high-temperature crystallization-rate maximum.

4 Homogeneous Nucleation at High Supercooling:

Evidence from Droplet Experiments

Both the increase in the gross crystallization rate and the change in semicrystalline
morphology during crystallization of the melt at high supercooling consistently
indicate a distinct increase in the nucleation density. Assuming that the high-
temperature crystallization process occurs via heterogeneous nucleation, the
increase in nucleation rate at high supercooling could be the result of activation
of a different kind of heterogeneous nuclei or formation of homogeneous nuclei. In
order to prove or disprove the hypothesis of homogeneous nucleation experimen-
tally, nucleation/crystallization experiments in droplets of sufficiently small size
were performed, as suggested long ago by Turnbull [75–79]. In more recent
investigations, submicron-sized heterogeneity-free droplets of iPP [80–82] and
PA 6 [83] were slowly cooled from the melt, revealing crystallization only at
high supercooling in the temperature range of the low-temperature maximum of
the crystallization rate. As an example, Figure 4 shows the temperatures of crys-
tallization of bulk PA 6 and dispersed PA 6 droplets, both as a function of the
cooling rate. Crystallization in PA 6 droplets, regardless of the rate of cooling of the
melt in the investigated range of 5–300 K min1, occurred only at temperatures of
110–120 C (i.e., at high supercooling) and at distinctly lower temperatures than in
bulk PA 6. Because of the absence of heterogeneous nuclei in droplets, it was

Fig. 4 Temperature of
crystallization of bulk PA
6 (squares) and dispersed
PA 6 droplets (diamonds) as
a function of the cooling
rate [83]. Adapted with
permission from [83],
Copyright (2006) Wiley
Crystal Nucleation of Polymers at High Supercooling of the Melt 267

Fig. 5 FSC (top) and DSC (bottom) cooling scans of iPP, demonstrating the occurrence of high-
and low-temperature crystallization processes, labeled H and L, respectively, measured during
cooling at different rates [80, 84]. The top set of curves was obtained for bulk iPP containing
heterogeneous nuclei, whereas the bottom curve was obtained for heterogeneity-free droplets.
Adapted from [84], Copyright (2012), with permission from Elsevier (top part), and from [80],
Copyright (2011) Wiley (bottom part)

concluded that crystallization at low temperatures proceeds via homogeneous

nucleation.
Similar observation of the drastically reduced temperature of crystallization
during slow cooling has been reported for iPP droplets. Whereas crystallization
of the bulk phase of iPP (containing impurities) typically occurs at 110–120 C,
when cooling at rates usually applied in DSC, crystallization in droplets was only
observed well below 50 C (i.e., at temperatures corresponding to the
low-temperature crystallization-rate maximum).
Figure 5 shows FSC cooling scans for bulk and heterogeneity-free droplet iPP
[84]. The top curve in Fig. 5 was measured on cooling at 2 K s1 and reveals
crystallization at about 90 C. This crystallization event (labeled ‘H’) is associated
with the high-temperature crystallization process initiated by heterogeneous nucle-
ation. With increasing cooling rate, the high-temperature crystallization process
becomes less intense, and is almost completely suppressed during cooling at about
200 K s1. The decrease in the enthalpy of crystallization of the high-temperature
crystallization process (indicated by the decreasing peak area in Fig. 5) is paralleled
by the evolution of a low-temperature crystallization event (labeled ‘L’ in Fig. 5)
below 50 C. It was concluded that if the high-temperature crystallization process is
suppressed as a result of fast cooling, then crystallization continues by a different
nucleation mechanism at lower temperatures. Fast cooling, in order to avoid
crystallization via heterogeneous nucleation at high temperature, is not required if
heterogeneous nuclei are absent, as shown by droplet-crystallization experiments.
268 R. Androsch and C. Schick

Fig. 6 AFM images of iPP. The left image was obtained from a rapidly cooled sample annealed at
ambient temperature containing heterogeneous nuclei/impurities [85]. The right image shows the
nanometer scale structure of a heterogeneity-free droplet [80]. Scale bars: 100 nm. Adapted with
permission from [85], Copyright (2009) IOP publishing (left image) and from [80], Copyright
(2011) Wiley (right image)

Analysis of the crystallization kinetics of heterogeneity-free droplets of iPP

revealed a crystallization peak at 44 C, although cooling was performed at a low
rate of 10 K min1 (Fig. 5, bottom curve) [80]. Comparison of the temperature of
crystallization of iPP droplets with the temperature of the low-temperature crystal-
lization event in bulk iPP indicates an identical mechanism of crystallization (i.e.,
prevalence of homogenous nucleation).
Further striking evidence for identical mechanisms of nucleation in
heterogeneity-free droplets and the bulk phase of polymers when crystallizing at
high supercooling is provided by analysis of the nucleation density using micros-
copy. Figure 6 shows the AFM surface structures of two different samples of iPP.
The left image was obtained on an iPP that contained heterogeneous nuclei/impu-
rities and which was rapidly cooled from the melt to avoid the high-temperature
crystallization, before annealing at ambient temperature [85]. The right image in
Fig. 6 shows the nanometer-scale structure of a heterogeneity-free droplet with a size
of about 500 nm [80, 81]. In both cases, the same nodular morphology of the ordered
phase and nucleation density were detected, as discussed for the left-hand images in
Fig. 3, obtained for PA 6, PA 11, and PBT after crystallization at temperatures
related to the low-temperature crystallization rate maximum shown in Fig. 1.

5 Sequence of Enthalpy Relaxation, Homogeneous Crystal

Nucleation, and Crystal Growth in the Glassy
Amorphous Phase

Formation of crystal nuclei and crystal growth of polymers is predicted to cease at a

temperature 30–50 K below Tg (i.e., at a temperature where no further transport
across the liquid–crystal phase boundary is expected) [1, 2]. In the context of
Crystal Nucleation of Polymers at High Supercooling of the Melt 269

investigating the nucleation and crystallization behaviors of polymers at high

supercooling of the melt, nucleation rates have also been measured in the glassy
state for PCL [42, 43], iPP [86], iPB-1 [47], poly(L-lactic acid) (PLLA) [87, 88],
PET [89], and PA 6 [90] by both calorimetry and microscopy. It was found
experimentally that the rate of nuclei formation of the low-temperature crystalliza-
tion process is fastest in the temperature range between Tg and (Tg + 50) K and that
the nucleation rate decreases monotonically with decreasing temperature. These
results point to predominance of a single nucleation mechanism in the analyzed
temperature range, regardless of whether nucleation occurs in the glassy or
devitrified amorphous phase.
Specific nucleation experiments have been performed to quantitatively describe
homogeneous nucleation of polymers in the glassy state, aiming to identify a link
between the rate of nuclei formation and the structure of the glass, with the latter
being controlled by the conditions of its formation and by the temperature and time
of annealing. Figure 7 shows the change in enthalpy of a polymer during fast
cooling followed by isothermal annealing at the temperature Ta. At high tempera-
ture, the enthalpy of the polymer melt decreases during cooling, as expected from
the short relaxation time of the system. With reference to the enthalpy of the liquid
phase (Fig. 7, left), the experimentally observed enthalpy remains at a higher level
after vitrification of the melt at Tg. It is known that, for a given polymer, the
difference between the enthalpy of the glass and the liquid depends on both the
cooling rate (which controls Tg) and the annealing temperature (which controls the
enthalpy values of the liquid and glass) [91–93]. The enthalpy decreases during

Fig. 7 Left: Plot of the enthalpy of an initially fully amorphous and liquid polymer as a function of
temperature during cooling to the glassy state, and during subsequent isothermal annealing at the
annealing temperature Ta. Vertical arrows indicate the enthalpy decrease during annealing and
then again during crystal growth. Arrows on the experimental curve indicate cooling. Right: FSC
heating scans showing enthalpy relaxation (1), crystal nucleation (2), and crystal growth (3) during
annealing, with these processes identified in their exact time sequence (see text for detailed
description)
270 R. Androsch and C. Schick

annealing such that, at first, the enthalpy relaxes to the value of the liquid state. This
is followed by a further decrease towards the enthalpy of the crystal phase as a
result of crystal growth (vertical arrows in Fig. 7, left). Prior to crystal growth,
formation of nuclei is required which, however, does not lead to a measurable
change in enthalpy. The exact time sequence of the various processes (enthalpy
relaxation, crystal nucleation, and crystal growth) has been evaluated by analysis of
FSC heating scans, recorded after annealing for different times (Fig. 7, right). The
top set of curves in Fig. 7 demonstrate that only enthalpy relaxation occurs on
annealing of an amorphous sample in the glassy state for short periods of time, as it
causes an enthalpy-recovery peak at Tg on subsequent heating. Note that the heating
rate selected for recording the heating scan and the previous cooling rate were
sufficiently high to suppress nuclei formation. Otherwise, there would have been
detected an exothermic cold-crystallization peak followed by endothermic melting,
which was not the case. If annealing is continued after complete relaxation of the
glass, then cold-crystallization and melting are observed (Fig. 7, curve 2). The areas
of both peaks increase with annealing time; however, they are always identical in
area but of opposite sign. In other words, the total change in enthalpy during this
annealing-time period is close to zero, indicating absence of crystallization. Even-
tually, on extended annealing, the area of the endothermic melting peak exceeds the
area of the exothermic cold-crystallization peak (Fig. 7, curve 3), which proves that
the enthalpy of the sample decreases further as a result of the formation of crystals
at the annealing temperature. Note that the endothermic low-temperature peak in
curve 3, close to the glass transition, is a result of both enthalpy recovery and
melting of the small crystals formed at Ta. Melting, however, overlaps with
exothermic reorganization and cold-crystallization, leading to observation of the
high-temperature melting peak. Simultaneously, as a result of the formation of
small crystals, there is a significant shift of Tg towards a higher temperature,
indicating immobilization of the amorphous phase, probably as a result of the
formation of a rigid amorphous fraction (RAF). Note that, as a result of the covalent
linkage of crystals and the amorphous structure in partially crystallized polymers,
the mobility of amorphous chain segments in regions near the crystals is reduced.
This often leads to an increase in Tg and the formation of fractions of amorphous
structure with different properties [94–100].
Analysis of FSC curves (Fig. 7, right) allows determination of the time required
to complete the process of enthalpy relaxation, and of characteristic times for
formation of crystal nuclei and crystals. These times are shown in Fig. 8 as a
function of the annealing temperature for the specific case of PA 6, focusing on the
temperature range close to Tg. The temperature range of the glass transition on
cooling at 103 K s1, applied in this particular investigation of PA 6, is indicated in
Fig. 8 by the shaded area. It was concluded from visual inspection of the various
data sets that formation of crystal nuclei in glassy PA 6 only occurred if the process
of enthalpy relaxation of the glass is complete. The experimentally assessable
decrease in enthalpy of the glass to the value of the liquid state at identical
temperature corresponds to a densification of the glass, involving cooperative
rearrangement of molecular segments on a length scale of a few nanometers.
Crystal Nucleation of Polymers at High Supercooling of the Melt 271

Fig. 8 Time to complete enthalpy relaxation (diamonds), onset time of formation of crystal nuclei
(squares), and onset time of crystallization (circles) as a function of the temperature of annealing
initially fully amorphous PA 6 [90]. The shaded area indicates the temperature range of the glass
transition on cooling at a rate of 103 K s1. Adapted with permission from [90], Copyright (2014)
Elsevier

The data in Fig. 8 suggest that densification of the glass and disappearance of the
relatively large scale motions connected to enthalpy relaxation are complete prior
to formation of crystal nuclei. In other words, completion of the enthalpy relaxation
of the glass is considered a pre-requisite for the formation of crystal nuclei in the
glass. The experimentally evidenced link between the processes of enthalpy relax-
ation and crystal nucleation suggests that the latter occurs homogeneously in the
bulk by rearrangement of molecular segments, but is not initiated on heterogene-
ities for which such a delay in nucleation is not expected.
The interpretation of a bimodal temperature dependence of the gross crystalli-
zation rate and crystallization in heterogeneity-free droplets in terms of bulk
homogeneous crystal nucleation is still controversial and not yet unequivocally
agreed upon [101, 102]. In a recent review of confinement effects on polymer
crystallization it was suggested that the ordering process in droplets is initiated by
surface nucleation, because it requires a lower supercooling than homogeneous
nucleation in the bulk of the droplets [102]. Surface nucleation, however, is
excluded as an explanation for the experimental results presented above because
it has been proven that the surface morphology of samples crystallized at high
supercooling is indifferent to the structure in the bulk. TEM and small-angle X-ray
scattering (SAXS) analyses, in particular for the specific case of iPP, provide
similar information as the AFM surface analysis with respect to the higher nucle-
ation density than that obtained on crystallization at low supercooling. In the case of
surface nucleation, different nucleation densities at the surface and in the bulk are
expected, but have not been proven. A link between enthalpy relaxation of the glass
(occurring in the bulk) and crystal nucleation has been confirmed for PCL [42, 43],
iPB-1 [47], and PA 6 [90], and favors a bulk nucleation process. However, surface
nucleation, regardless of whether at the surface of droplets or in bulk samples,
cannot always be excluded as contributing to the crystallization process of
272 R. Androsch and C. Schick

polymers at high supercooling as there is little evidence described in the literature

[102, 103]. For example, there are several reports that crystallization of PET at high
supercooling is faster in the regions near the surface than in the bulk [104–107],
with the results explained in terms of higher mobility of chain segments located
near the surface. In a more recent study, the morphology/habit of crystals of cold-
crystallized PET was evaluated using electron microscopy and AFM [108]. It was
found that crystals formed on heating the amorphous glass to a temperature higher
than Tg are of lamellar shape in the bulk and almost isometric in habit at the surface.
This is consistent with the notion that the nucleation density in the bulk is lower
than at the surface, probably related to the relatively low bulk nucleation rate. PET,
like PLLA, belongs to a small group of polymers that reveal formation of spheru-
lites during cold-crystallization (i.e., after nuclei formation at high supercooling or
in the glassy state). For the polymers discussed above, including iPP, PA 6, PA
66, PA 11, and PBT, such a result has not yet been reported, nor different surface
and bulk crystal morphologies. In other words, there is a lack of evidence for
surface nucleation as a dominant nucleation mechanism during crystallization at
high supercooling.

6 Mesophase Formation During Crystallization at High

Supercooling

In addition to the change in the nucleation mechanism/density, the low-temperature

crystallization-rate maximum could be caused by an increase in the crystal growth
rate. Such an increase could be a result of activation of growth at different crystal
faces when lowering the crystallization temperature, or a result of the formation of a
different crystal polymorph. In fact, for iPP, PA 6, PA 66, PA 11, and PLLA, crystal
formation at high temperature is replaced by formation of conformationally disor-
dered crystals or mesophases at low temperature, with the latter being metastable at
the condition of their formation. In all cases, the conformationally disordered
crystals or mesophases convert irreversibly into the stable crystal polymorph with
heating. The exact mechanism of the transformation, that is, whether it occurs via
melting and melt-recrystallization or within a solid–solid phase transition, has not
yet been fully investigated. Important in the context of the discussion of the
crystallization kinetics, the temperature ranges of formation of crystals and
mesophases coincide with the high- and low-temperature crystallization-rate max-
ima in Fig. 1, respectively.
Detailed information on the crystal/mesophase polymorphism is available for
iPP, with major findings summarized in several reviews [109–114]. It was found
[112–114] that crystallization at low supercooling results in formation of mono-
clinic α-crystals, in which left- and right-handed helices are arranged in layers to
achieve high lateral packing density. Quenching in ice water followed by annealing
at ambient temperature led to formation of the mesophase with a pseudohexagonal
Crystal Nucleation of Polymers at High Supercooling of the Melt 273

symmetry and helix reversals as the main conformational defects. Quantitative

thermal analysis of the specific heat capacity of semi-mesomorphic iPP revealed
almost no increase beyond the level of the vibrational heat capacity of solid iPP,
which suggests that the mesophase co-exists with rigid amorphous structure and
only contains immobilized chain segments; therefore, it is termed conformationally
disordered glass [115]. The mesophase is metastable and can only be annealed at
temperatures lower than the stability limit of about 80 C [115, 116]. At higher
temperatures, connected with the onset of helix mobility [117], the mesophase
becomes unstable and converts to monoclinic α-crystals with a latent heat of 7%
of the heat of fusion of monoclinic crystals. It is worth noting that the mesophase–
crystal phase transition at temperatures higher than 80 C can be suppressed by
heating at rates faster than 30,000 K s1 to a temperature higher than the melting
temperature of the α-phase. Although it has been suggested that the mesophase–
crystal phase transition occurs in the solid state, such a transition relies on slow
heating, which allows fast local melting and immediate but slower recrystallization.
Fewer details are available regarding the mesophase/crystal polymorphism of
polyamides, which is controlled by the crystallization temperature, as in the case of
iPP. Although the structure (and thermodynamic properties) of crystals forming
from the quiescent melt at low supercooling are known for PA 6, PA 66, and PA
11 [118–123], little information has been reported on the mesophases forming at
low temperature. Melt-crystallization at low supercooling typically leads to forma-
tion of α-crystals, in which a planar/sheet-like orientation of hydrogen bonding
between the amide groups of neighboring chain segments is achieved after cooling
to room temperature. In contrast, crystallization at high supercooling of the melt
leads to formation of a pseudohexagonal mesophase in which the molecular stems
exhibit rotational symmetry, leading to an arbitrary orientation of hydrogen bonds
between neighboring chains [38, 124–127]. It is important to note that the
pseudohexagonal mesophase forming at high supercooling is different from the
high-temperature modification of the α-form. Crystals formed from the melt at low
supercooling show a Brill transition [128, 129]. At temperatures lower than the Brill
transition temperature, depending on the particular polyamide, crystals exhibit
monoclinic or triclinic symmetry. At higher temperatures, the anisotropy of lateral
chain packing is lost, leading to the observation of a pseudohexagonal unit cell.
Regarding the stability of the mesophase that forms at high supercooling of the
melt, it is known that on heating it converts irreversibly into stable crystals [38, 130,
131], but the exact mechanism has not yet been investigated. In other words, it is
unknown whether and under what conditions, with respect to heating rate or
transformation temperature, the mesophase converts to crystals via melting and
melt-recrystallization, or within a solid–solid phase transformation.
For iPP, PA 6, PA 66, and PA 11 it has been shown that the low-temperature
crystallization-rate maximum in the curves of Fig. 1 is connected with both a
tremendous increase in the nucleation rate/density and the growth of a different
crystal polymorph. Unfortunately, for all these polymers, data on the growth rate of
the low-temperature polymorph are not yet available, and therefore it cannot be
excluded that the low-temperature crystallization-rate maximum in Fig. 1 is a result
274 R. Androsch and C. Schick

of both an increase in the nucleation density and an increase in growth rate. This
notwithstanding, it is also true for all these polymers that the mesophase is at best
metastable at the temperature of its formation with respect to the crystal phase that
forms from the melt at low supercooling. It must be assumed that formation of the
mesophase at high supercooling is kinetically controlled, but the exact thermody-
namics are unknown. The mesophase can be considered to represent an intermedi-
ate phase, with a local minimum of the free enthalpy along the path from the
unstable melt to the stable crystal phase in accord with Ostwald’s rule of stages
[132], or even as a frozen-in. It can be speculated that, as a result of the extremely
high nucleation density of perhaps one nucleus in a cube with a side length of 5–
10 nm, the mobility of internuclei amorphous chain segments is reduced by
mesophase formation at high supercooling. Thus, conformational defects are
entrapped in the ordered phase without the possibility of their removal due to
vitrification of the whole system, with the latter proven (e.g., for PCL) by obser-
vation of a distinct increase in Tg proportional to the crystalline content [42, 43].
Of the example polymers shown in Fig. 1 (employed for demonstration of the
bimodal temperature dependence of the crystallization rate), PBT does not exhibit
supercooling-controlled crystal/mesophase polymorphism as shown by iPP and
polyamides. It has recently been demonstrated that the structures of crystals
forming at temperatures associated with the low- and high-temperature crystalliza-
tion-rate maxima in Fig. 1 are identical [41]. PBT can form two different crystal
structures [133] and a mesophase [134], with the latter only being observed on
stretching the amorphous phase. Crystallization of the quiescent melt leads to
formation of triclinic α-crystals, which are characterized by a non-extended chain
conformation of the butylene unit between the planar phenylene rings. The
α-structure can reversibly transform into the β-structure under tension, with the
butylene units then adopting an extended all-trans conformation, causing a slight
change in the unit cell parameters with respect to those of the α-phase. However,
analysis of the X-ray structure of samples of different cooling/crystallization
history led to the conclusion that observation of the two crystallization-rate maxima
at different temperatures (as shown in Fig. 1) is not related to the formation of
different crystal polymorphs of different growth rate [41, 135] but related to the
large increase in nucleation density.

7 Crystallization of Poly(L-lactic acid) at High

Supercooling: Application of Tammann’s Nuclei
Development Method to Obtain Nucleation Rates

A further example of a polymer showing a distinct crystal/mesophase polymor-

phism controlled by the crystallization temperature is poly(L-lactic acid) (PLLA).
The crystallization behavior of PLLA is discussed separately because there has not
yet been proven a change in the nucleation mechanism with variation in
Crystal Nucleation of Polymers at High Supercooling of the Melt 275

temperature. Instead, there is evidence that the nucleation kinetics is decoupled

from the kinetics of crystal growth and the formation of different crystal poly-
morphs. PLLA forms different ordered structures as a function of the conditions of
crystallization [136, 137]. Crystallization of the quiescent melt at temperatures
higher than about 110 C leads to formation of α-crystals with two antiparallel
aligned helical chain segments packed in an orthorhombic unit cell [138–140]. For-
mation of α-crystals is replaced by formation of pseudohexagonal α0 -crystals at
temperatures lower than about 110 C. The α0 -form is considered to be a conforma-
tionally disordered α-crystal with slightly increased lattice spacings. It irreversibly
transforms upon heating into the stable α-form [141–145], either within a solid–
solid state transformation [143–145] or via melting followed by melt-
recrystallization [146, 147].
Typically, the density of primary crystal nuclei forming in a supercooled melt is
assessed by hot-stage microscopy. This allows counting the number of spherulites
per unit volume at a given analysis temperature, either after complete crystalliza-
tion or during the isothermal crystallization process. In the latter case, information
about the kinetics of nuclei formation can also be obtained. As a general rule, for a
given nucleation scheme (i.e., both homogeneous nucleation or nucleation on a
substrate/impurity), the nucleation density increases with decreasing temperature as
a result of the decreasing work required to produce a critical-sized nucleus from the
melt. Using optical microscopy, it can be observed that the number of spherulites
increases and their size decreases. An example of such analysis is presented in
Fig. 9 (right), which shows POM micrographs obtained for PLLA, isothermally
crystallized at different temperatures between 126 C and 81 C. As expected, with
decreasing crystallization temperature, the number of spherulites (i.e., the number

Fig. 9 Right: POM images of PLLA isothermally crystallized at different temperatures between
81 C and 126 C [148]. Left: Spherulite density of PLLA as a function of crystallization temper-
ature [136]. Scale bars: 50 μm. The POM images were adapted from [148], Copyright (2013), with
permission from Elsevier. The graph on the left was adapted from [136], Copyright (2013), with
permission from Elsevier
276 R. Androsch and C. Schick

of nucleation sites) increases, leading to a reduction in spherulite size. Quantitative

information about the temperature-dependence of the nucleation density is given in
the graph shown in Fig. 9 (left), which indicates the spherulite number of PLLA per
square millimeter as a function of the crystallization temperature [136, 148–
151]. The data suggest increasing nucleation density with decreasing temperature
in the analyzed temperature range between 95 C and 140 C, reaching a plateau
value at around 90 C. At temperatures lower than 95 C, which is 30–40 K above Tg,
such analysis of the nucleation density via measurement of the spherulite density
fails because their number then is too high to be reliably analyzed.
To obtain data on the nucleation kinetics/density at higher supercooling of the
melt, specifically designed nucleation experiments can be performed, as described
by Tammann over a century ago [152]. Tammann’s two-stage crystal nuclei
development method implies formation of nuclei at large supercooling and follows
their isothermal growth at higher temperatures, utilizing the often widely different
temperatures of maximum rate of primary crystal nucleation and crystal growth. In
other words, this method allows detection of nuclei that were formed at low
temperature by growing them at a higher temperature to experimentally detectable
sizes. In the specific case of PLLA, such experiments have been performed using
both POM and FSC [87, 88].
For demonstration, Fig. 10 shows POM images of initially amorphous samples
of PLLA, annealed for different times between 2 and 1,000 min at temperatures
between 50 C and 70 C (i.e., slightly below Tg) and then cold-crystallized at an
elevated temperature of 120 C for 10 min. The temperature–time profile is also
shown in Fig. 10. Annealing at 50 C for less than 100 min is not connected with
nuclei formation, as concluded from the constant low number of spherulites grow-
ing at 120 C. However, if the annealing time exceeds 100 min then there is an
increase in spherulite density as a result of formation of additional nuclei at 50 C.
With increasing annealing temperature, an increased spherulite number is observed
on annealing for a shorter time, ultimately providing information about the tem-
perature dependence of the nucleation rate.
It is worth noting that application of Tammann’s two-stage crystal nuclei
development method is not restricted to analysis of the nucleation rate at temper-
atures below Tg. Rather, it is required that the growth of nuclei in the nucleation
stage is either completely absent or at least distinctly slower than the nucleation
rate. This is particularly true at temperatures below Tg, although it has been shown
that crystal growth can also be observed under such conditions [42, 43, 86,
90]. However, for the specific case of PLLA, Tammann’s two-stage crystal nuclei
development method has also been applied for the temperature range between Tg
and 110 C (using FSC) in order to complete the analysis of the nucleation behavior
presented in Fig. 9. Figure 11 (left) shows FSC heating curves recorded using a rate
of 200 K s1, which is sufficiently fast to suppress cold-crystallization of PLLA on
heating. The set of curves was obtained for initially fully amorphous samples,
which were annealed at 70 C for different times between 0 and 10,000 s, to allow
nuclei formation, and then heated to 120 C to allow growth of nuclei for 300 s.
Afterwards, the samples were rapidly cooled to 60 C and heated to 200 C to gain
Crystal Nucleation of Polymers at High Supercooling of the Melt 277

Fig. 10 POM images of initially amorphous PLLA, annealed in the glassy state at different
temperatures for different times (as indicated), and then cold-crystallized at 120 C for 10 min
[87]. Scale bar: 100 μm. Also shown is the time–temperature profile for Tammann’s nuclei
development method. Adapted with permission from [87], Copyright (2013) American Chemical
Society

information about the crystal fraction formed in the growth stage. It can be seen
that, with increasing annealing time in the nucleation stage, the area of the endo-
thermic peak increases, indicating an increase in the crystal fraction during the
300 s growth step. Analysis of the peak area or enthalpy of melting as a function of
the time of annealing in the nucleation stage provides information about the
nucleation kinetics, quantified, for example, as the onset-time of nuclei formation
(shown in Fig. 11, right). The data reveal that the rate of nuclei formation is fastest
at about 95 C, which is in agreement with studies of the temperature dependence of
nucleation density (shown in Fig. 9). In extension of earlier work on the nucleation
density of PLLA (performed by microscopy), it was not only shown that on
lowering the temperature to values below 95 C the nucleation rate decreases
progressively in the investigated temperature range, but also that the temperature
dependence of the nucleation rate in the high-supercooling temperature range does
278 R. Androsch and C. Schick

Fig. 11 Left: Apparent heat capacity of PLLA as a function of temperature, obtained on heating at
200 K s1 after prior annealing in the nucleation stage at 70 C for different times between 0 and
10,000 s, followed by cold-crystallization in the growth stage at 120 C for 300 s. Right: The
diamonds and triangles indicate the onset time of crystal nucleation, obtained by plotting the
enthalpy of melting determined from the FSC curves in the left plot as a function of the annealing
time in the nucleation stage. For comparison, the squares show half-times of crystallization of
PLLA as a function of temperature, as obtained from the literature [153]. Adapted from [88],
Copyright (2013), with permission from Elsevier

not seem to be affected by the glass transition. The data shown in Fig. 11 also
reveal, by extrapolation, that annealing glassy PLLA at ambient temperature, (i.e.,
about 30 K below Tg) leads to formation of crystal nuclei after 107–108 s. Note that
this observation might be of importance from an application point of view because
such ordering is perhaps connected with changes in the mechanical behavior, at
least if nucleation is followed by growth.
Figure 11 (right) shows that the gross crystallization rate is about three orders of
magnitude slower than the nucleation rate at the same temperature. The crystalli-
zation half-time data were adapted from the literature [153], and are scaled with the
right axis to emphasize the occurrence of two qualitatively different crystallization
processes, with the crossover observed at around 120 C. Crystallization at temper-
atures higher than about 120 C is connected with formation of α-crystals, whereas
at lower temperatures conformationally disordered α0 -crystals form. Inspection of
the gross crystallization rate (Fig. 11, right) indicates that the maximum crystalli-
zation rate of α0 -crystals is almost one order of magnitude higher than that of
α-crystals. However, the data in Fig. 9 prove that the α/α0 polymorphism is not
connected with a change in the mechanism of crystal nucleation. Rather, it has been
shown in numerous studies that the rates of growth of α0 - and α-crystals are very
different [153–155]. The growth rate of the different crystal polymorphs has been
measured by analysis of spherulite growth rates using hot-stage microscopy, and a
typical finding reported in the literature is presented in Fig. 12. The data suggest
that the temperature of maximum growth rate of α-crystals/spherulites is about
Crystal Nucleation of Polymers at High Supercooling of the Melt 279

Fig. 12 Spherulite growth

rate for PLLA as a function
of the crystallization
temperature. Adapted from
[155], Copyright (2005),
with permission from
Elsevier

130 C, whereas that of α0 -crystals/spherulites is at 110–115 C, with the maximum

growth rate of the latter being almost twice than that of α-crystals/spherulites.
Regarding nucleation of the crystallization process of PLLA at high
supercooling of the melt, there exist a few further independent studies [156–
159]. In particular, it has been shown that cold-crystallization is faster than melt-
crystallization at identical temperatures, with the observed results being explained
by formation of additional crystal nuclei on the thermal pathway to the crystalliza-
tion temperature. Parameters in such nucleation experiments were the rate of
cooling to, and the residence time at, various minimum temperatures before heating
of the sample to the cold-crystallization temperature. In a further work [159], melt-
quenched PLLA was aged below Tg, and changes in the structure of the initially
fully amorphous specimen were monitored in situ by time-resolved infrared spec-
troscopy. The data provided evidence for local ordering and the formation of a
mesophase. It has been suggested that the formed mesophase becomes disordered in
conjunction with the physical aging peak at the glass transition temperature, while
simultaneously enhancing cold-crystallization on continuation of heating. The
discussed reports [156–159] all prove that nuclei/locally ordered structures develop
at low temperatures and that these structures increase the crystallization kinetics at
elevated temperatures.

8 Summary

Analysis of the temperature-dependence of the gross crystallization rate of a large

number of polymers, including PCL, iPP, PA 6, PA 66, PA 11, and PBT, has
revealed qualitatively different crystallization behaviors at temperatures below and
above about (Tg + 30) K. In several cases, two distinct crystallization-rate maxima
are observed at low and high supercooling. Possible reasons for this observation
could be a change in the mechanism of primary crystal nucleation, in the forming
crystal structure/polymorph, or of the crystal morphology/growth face in response
to variation of the supercooling, all of which are discussed in this review.
280 R. Androsch and C. Schick

Recent research supports the notion that the low-temperature crystallization-rate

maximum is caused by a drastic increase in the number of primary crystal nuclei.
For iPP, PA 6, PA 11, and PBT it has been shown that crystallization at low
supercooling proceeds via spherulitic growth of lamellae, whereas at high
supercooling the formation of lamellae and spherulites is often not observed.
Instead, there is formation of small ordered domains or crystals, which do not
form a higher-order superstructure, indicating their independent growth. A rough
estimation of the nucleation density on crystallization at low and high supercooling
gave values of 106 and 1015 nuclei mm3, respectively, with the increase occurring
in a relatively narrow temperature interval. Although the exact mechanism of
nucleation is not known for either of these temperature ranges, it is speculated
that crystallization at low and high supercooling is related to heterogeneous and
homogeneous nucleation, respectively. Heterogeneous nucleation is assumed for
the low supercooling temperature range, because purposely added nucleation
agents are active in just this temperature range [160, 161], increasing the crystal-
lization rate only at those temperatures. In contrast, the low-temperature crystalli-
zation-rate maximum is not affected by the addition of heterogeneous nuclei,
suggesting that it is related to formation of homogeneous nuclei.
A homogeneous nucleation mechanism for crystallization at high supercooling
is supported by analysis of the crystallization behavior of heterogeneity-free drop-
lets. In such cases, crystallization at low supercooling is not detected but only
occurs in exactly the same temperature range for which the low-temperature
crystallization-rate maximum is observed in samples containing heterogeneous
nucleators. Moreover, the density of primary crystal nuclei in such heterogeneity-
free droplets and in bulk samples containing impurities when crystallized at high
supercooling is seemingly identical. Further evidence favoring homogeneous
nucleation at high supercooling is provided by establishment of a link between
the time scales of the densification/relaxation of the glass and the formation of
nuclei. It has been shown that nuclei are able to form only after completion of the
relaxation of the glass by cooperative rearrangement of molecular segments on a
length scale of a few nanometers. Nuclei formation is then assumed to occur
without cooperative displacement of segments, but only with local changes of
conformations.
Final conclusions cannot be drawn about the effect of crystal/mesophase poly-
morphism on the gross crystallization rate and its bimodal temperature-dependence
in the cases of iPP, PA 6, PA 66, and PA 11, because the growth rates for mesophase
formation at high supercooling are unknown. It might be possible that the
low-temperature crystallization-rate maximum is also caused by an increased
growth rate, which, however, does not contradict the above conclusions about the
effect of a change in the nucleation mechanism.
In contrast to iPP, PA 6, PA 66, and PA 11, which show a distinct supercooling-
controlled crystal/mesophase polymorphism that is paralleled by observation of a
large difference in nucleation densities, the α/α0 -crystal polymorphism of PLLA
seems completely decoupled from the primary crystal-nucleation step, including its
kinetics. For PLLA, crystal growth rates have also been measured at high
Crystal Nucleation of Polymers at High Supercooling of the Melt 281

supercooling, suggesting that formation of the low-temperature α0 -phase has a

higher growth rate than the high-temperature α-crystal polymorph. The transition
of formation of α-crystals at low supercooling to formation of α0 -crystals at high
supercooling, occurring around 110 C, however, is not connected with a drastic
increase in the nucleation density.
For PBT, in contrast, it has been proven that crystallization at low and high
supercooling leads to formation of the same crystal polymorph. For PCL, informa-
tion on the formation of different crystal polymorphs during crystallization at
different supercooling is not yet available.

References

1. Hoffmann JD, Davis GT, Lauritzen JI Jr (1976) The rate of crystallization of linear polymers
with chain folding. In: Hannay HB (ed) Treatise on solid state chemistry, crystalline and
noncrystalline solids, vol 3. Plenum, New York
2. Wunderlich B (1976) Crystal nucleation, growth, annealing, vol 2, Macromolecular physics.
Academic, New York
3. Becker R (1938) Die Keimbildung bei der Ausscheidung von metallischen Mischkristallen.
Ann Phys 32:128–140
4. Turnbull D, Fisher JC (1949) Rate of nucleation in condensed systems. J Chem Phys
17:71–73
5. Turnbull D (1950) Kinetics of heterogeneous nucleation. J Chem Phys 18:198–203
6. Binsbergen FL (1977) Natural and artificial heterogeneous nucleation in polymer crystalli-
zation. J Polym Sci Polym Symp 59:11–27
7. Hoffman JD, Lauritzen JI Jr (1961) Crystallization of bulk polymers with chain folding:
theory of growth of lamellar spherulites. J Res Natl Bur Stand 65A:297–336
8. Hoffman JD (1964) Theoretical aspects of polymer crystallization with chain folds: bulk
polymers. SPE Trans 4:315–362
9. Lauritzen JI Jr, Hoffman JD (1973) Extension of theory of growth of chain-folded polymer
crystals to large undercoolings. J Appl Phys 44:297–336
10. Wunderlich B, Mehta A (1974) Macromolecular nucleation. J Polym Sci Polym Phys
12:255–263
11. Strobl G (2000) From the melt via mesomorphic and granular crystalline layers to lamellar
crystallites: a major route followed in polymer crystallization? Eur Phys J E3:165–183
12. Strobl G (2005) A thermodynamic multiphase scheme treating polymer crystallization and
melting. Eur Phys J E18:295–309
13. Strobl G (2006) Crystallization and melting of bulk polymers: new observations, conclusions
and a thermodynamic scheme. Prog Polym Sci 31:398–442
14. Long Y, Shanks RA, Stachurski ZH (1995) Kinetics of polymer crystallization. Prog Polym
Sci 20:651–701
15. Booth A, Hay JN (1969) The use of differential scanning calorimetry to study polymer
crystallization kinetics. Polymer 10:95–104
16. Chan TW, Isayev AI (1994) Quiescent polymer crystallization: modeling and measurements.
Polym Eng Sci 34:461–471
17. Lorenzo AT, Arnal ML, Albuerne J, Müller AJ (2007) DSC isothermal polymer crystalliza-
tion kinetics measurements and the use of the Avrami equation to fit the data: guidelines to
avoid common problems. Polym Test 26:222–231
18. Avrami M (1939) Kinetics of phase change. I General theory. J Chem Phys 7:1103–1112
282 R. Androsch and C. Schick

19. Avrami M (1940) Kinetics of phase change. II Transformation-time relations for random
distribution of nuclei. J Chem Phys 8:212–224
20. Avrami M (1941) Granulation, phase change, and microstructure kinetics of phase change III.
J Chem Phys 9:177–183
21. Vyazovkin S, Sbirrazzuoli N (2004) Isoconversional approach to evaluating the Hoffman–
Lauritzen parameters (U* and Kg) from the overall rates of nonisothermal crystallization.
Macromol Rapid Commun 25:733–738
22. Adamovsky SA, Minakov AA, Schick C (2003) Scanning microcalorimetry at high cooling
rate. Thermochim Acta 403:55–63
23. Adamovsky S, Schick C (2004) Ultra-fast isothermal calorimetry using thin film sensors.
Thermochim Acta 415:1–7
24. Minakov AA, Adamovsky SA, Schick C (2005) Non adiabatic thin-film (chip)
nanocalorimetry. Thermochim Acta 432:177–185
25. Wunderlich B (2005) Thermal analysis of polymeric materials. Springer, Berlin
26. Pijpers TFJ, Mathot VBF, Goderis B, Scherrenberg RL, van der Vegte EW (2002) High-
speed calorimetry for the study of the kinetics of (de)vitrification, crystallization, and melting
of macromolecules. Macromolecules 35:3601–3613
27. Kolesov IS, Androsch R, Radusch HJ (2004) Non-isothermal crystallization of polyethylenes
as function of cooling rate and concentration of short chain branching. J Therm Anal Calorim
78:885–895
28. Ding Z, Spruiell JE (1996) An experimental method for studying nonisothermal crystalliza-
tion of polymers at very high cooling rates. J Polym Sci Polym Phys 34:2783–2804
29. Boyer SAE, Haudin JM (2010) Crystallization of polymers at constant and high cooling rates:
a new hot-stage microscopy set-up. Polym Test 29:445–452
30. Märtson T, Ots A, Krumme A, Lohmus A (2010) Development of a faster hot-stage for
microscopy studies of polymer crystallization. Polym Test 29:127–131
31. Kamal MR, Chu E (1983) Isothermal and non-isothermal crystallization of polyethylene.
Polym Eng Sci 23:27–31
32. van Herwaarden S, Iervolino E, van Herwaarden F, Wijffels T, Leenaers A, Mathot V (2011)
Design, performance and analysis of thermal lag of the UFS1 twin-calorimeter chip for fast
scanning calorimetry using the Mettler-Toledo Flash DSC 1. Thermochim Acta 522:46–52
33. Minakov AA, Schick C (2007) Ultrafast thermal processing and nanocalorimetry at heating
and cooling rates up to 1 MK/s. Rev Sci Instrum 78:073902
34. De Santis S, Adamovsky S, Titomanlio G, Schick C (2007) Isothermal nanocalorimetry of
isotactic polypropylene. Macromolecules 40:9026–9031
35. Silvestre C, Cimmino S, Duraccio D, Schick C (2007) Isothermal crystallization of isotactic
poly(propylene) studied by superfast calorimetry. Macromol Rapid Commun 28:875–881
36. van Drongelen M, Meijer-Vissers T, Cavallo D, Portale G, Vanden Poel G, Androsch R
(2013) Microfocus wide-angle X-ray scattering of polymers crystallized in a fast scanning
chip calorimeter. Thermochim Acta 563:33–37
37. Rhoades AM, Williams JL, Androsch R (2015) Crystallization of polyamide 66 at
processing-relevant cooling conditions and at high supercooling. Thermochim Acta
603:103–109
38. Mollova A, Androsch R, Mileva D, Schick C, Benhamida A (2013) Effect of supercooling on
crystallization of polyamide 11. Macromolecules 46:828–835
39. Pyda M, Nowak-Pyda E, Heeg J, Huth H, Minakov AA, Di Lorenzo ML et al (2006) Melting
and crystallization of poly(butylene terephthalate) by temperature-modulated and superfast
calorimetry. J Polym Sci Polym Phys 44:1364–1377
40. Schawe JEK (2014) Influence of processing conditions on polymer crystallization measured
by fast scanning DSC. J Therm Anal Calorim 116:1165–1173
41. Androsch R, Rhoades AM, Stolte I, Schick C (2015) Density of heterogeneous and homo-
geneous crystal nuclei in poly(butylene terephthalate). Eur Polym J 66:180–189
Crystal Nucleation of Polymers at High Supercooling of the Melt 283

42. Zhuravlev E, Schmelzer JWP, Wunderlich B, Schick C (2011) Kinetics of nucleation and
crystallization in poly(ε-caprolactone). Polymer 52:1983–1997
43. Wurm A, Zhuravlev E, Eckstein K, Jehnichen D, Pospiech D, Androsch R, Wunderlich B,
Schick C (2012) Crystallization and homogeneous nucleation kinetics of poly(ε-caprolactone)
(PCL) with different molar masses. Macromolecules 45:3816–3828
44. Mileva D, Androsch R (2012) Effect of co-unit type in random propylene copolymers on the
kinetics of mesophase formation and crystallization. Colloid Polym Sci 290:465–471
45. Cavallo D, Gardella L, Alfonso GC, Mileva D, Androsch R (2012) Effect of comonomer
partitioning on the kinetics of mesophase formation in random copolymers of propene and
higher α-olefins. Polymer 53:4429–4437
46. Supaphol P, Spruiell JE (2001) Isothermal melt- and cold-crystallization kinetics and subse-
quent melting behavior in syndiotactic polypropylene: a differential scanning calorimetry
study. Polymer 42:699–712
47. Stolte I, Androsch R, Di Lorenzo ML, Schick C (2013) Effect of aging the glass of isotactic
polybutene-1 on Form II nucleation and cold-crystallization. J Phys Chem
B117:15196–15203
48. Keller A (1955) The spherulitic structure of crystalline polymers. Part I. Investigations with
the polarizing microscope. J Polym Sci 17:291–308
49. Keller A (1955) The spherulitic structure of crystalline polymers. Part II. The problem of
molecular orientation in polymer spherulites. J Polym Sci 17:351–364
50. Keller A, Waring JRS (1955) The spherulitic structure of crystalline polymers. Part III.
Geometrical factors in spherulitic growth and the fine‐structure. J Polym Sci 17:447–472
51. Magill JH (2001) Review spherulites: a personal perspective. J Mater Sci 36:3143–3164
52. Roche EJ, Stein RS, Thomas EL (1980) Electron microscopy study of the structure of normal
and abnormal poly(butylene terephthalate) spherulites. J Polym Sci Polym Phys
18:1145–1158
53. Stein RS, Misra A (1980) Morphological studies on poly(butylene terephthalate). J Polym Sci
Polym Phys 18:327–342
54. Piccarolo S (1992) Morphological changes in isotactic polypropylene as a function of cooling
rate. J Macromol Sci Phys B31:501–511
55. Zia Q, Androsch R, Radusch HJ, Piccarolo S (2006) Morphology, reorganization, and
stability of mesomorphic nanocrystals in isotactic polypropylene. Polymer 47:8163–8172
56. Zia Q, Androsch R, Radusch HJ (2010) Effect of structure at the micrometer and nanometer
length scales on the light transmission of isotactic polypropylene. J Appl Polym Sci
117:1013–1020
57. Mileva D, Zia Q, Androsch R, Radusch HJ, Piccarolo S (2009) Mesophase formation in poly
(propylene-ran-1-butene) by rapid cooling. Polymer 50:5482–5489
58. Mileva D, Androsch R, Radusch HJ (2009) Effect of structure on light transmission in
isotactic polypropylene and random propylene-1-butene copolymers. Polym Bull
62:561–571
59. Mileva D, Kolesov I, Androsch R (2012) Morphology of cold-ordered polyamide 6. Colloid
Polym Sci 290:971–978
60. Mileva D, Androsch R, Zhuravlev E, Schick C (2012) Morphology of mesophase and crystals
of polyamide 6 prepared in a fast scanning chip calorimeter. Polymer 53:3994–4001
61. Wunderlich B (1973) Crystal structure, morphology defects, vol 1, Macromolecular physics.
Academic, New York
62. Gezovich DM, Geil PH (1968) Morphology of quenched polypropylene. Polym Eng Sci
8:202–209
63. Hsu CC, Geil PH, Miyaji H, Asai K (1983) Structure and properties of polybutylene
crystallized from the glassy state. II. Electron microscopy. J Macromol Sci Phys
B22:489–496
64. Yeh GSY, Geil PH (1967) Crystallization of polyethylene terephthalate from the glassy state.
J Macromol Sci Phys B1:235–249
284 R. Androsch and C. Schick

65. Meyer M, Van der Sande J, Uhlmann DR (1978) On the structure of glassy polymers.
VI. Electron microscopy of polycarbonate, poly(ethylene terephthalate), poly(vinyl chloride),
and polystyrene. J Polym Sci Polym Phys 16:2005–2014
66. Kanig G (1983) Application of the short-time staining for the electron microscopic investi-
gation of the crystallization of polyethylene. Colloid Polym Sci 261:373–374
67. Caldas V, Brown GR, Nohr RS, MacDonald JG, Raboin LE (1994) The structure of the
mesomorphic phase of quenched isotactic polypropylene. Polymer 35:899–907
68. Ogawa T, Miyaji M, Asai K (1985) Nodular structure of polypropylene. J Phys Soc Jpn
54:3668–3670
69. Wang ZG, Hsiao BS, Srinivas S, Brown GM, Tsou AH, Cheng SZD et al (2001) Phase
transformation in quenched mesomorphic isotactic polypropylene. Polymer 42:7561–7566
70. Grubb DT, Yoon DY (1986) Morphology of quenched and annealed isotactic polypropylene.
Polym Commun 27:84–88
71. Zia Q, Androsch R, Radusch HJ, Ingolič E (2008) Crystal morphology of rapidly cooled
isotactic polypropylene: a comparative study by TEM and AFM. Polym Bull 60:791–798
72. Miyamoto Y, Fukao K, Yoshida T, Tsurutani N, Miyaji H (2000) Structure formation of
isotactic polypropylene from the glass. J Phys Soc Jpn 69:1735–1740
73. Manabe N, Yokota Y, Minami H, Uegomori Y, Komoto T (2002) A TEM study on melt-
crystallized poly(butylene terephthalate). J Electron Microsc 51:11–19
74. Schaper A, Hirte R, Ruscher C, Hillebrand R, Walenta E (1986) The electron microscope
characterization of the fine structure of nylon 6: I. The supermolecular structure in melt-cast,
isotropic bulk material. Colloid Polym Sci 264:649–658
75. Cormia RL, Price FP, Turnbull D (1962) Kinetics of crystal nucleation in polyethylene.
J Chem Phys 37:1333–1340
76. Cormia RL, Turnbull D (1961) Kinetics of crystal nucleation in some normal alkane liquids.
J Chem Phys 34:820–827
77. Burns JR, Turnbull D (1966) Kinetics of crystal nucleation in molten isotactic polypropylene.
J Appl Phys 37:4021–4026
78. Koutsky JA, Walton AG, Baer E (1967) Nucleation of polymer droplets. J Appl Phys
38:1832–1839
79. Gornick F, Ross GS, Frolen LJ (1967) Crystal nucleation in polyethylene: the droplet
experiment. J Polym Sci Polym Symp 18:79–91
80. Langhe DS, Hiltner A, Baer E (2011) Transformation of isotactic polypropylene droplets
from the mesophase into the α-Phase. J Polym Sci Polym Phys 49:1672–1682
81. Jin Y, Hiltner A, Baer E, Masirek R, Piorkowska E, Galeski A (2006) Formation and
transformation of smectic polypropylene nanodroplets. J Polym Sci Polym Phys
44:1795–1803
82. Ibarretxe J, Groeninckx G, Bremer L, Mathot VBF (2009) Quantitative evaluation of
fractionated and homogeneous nucleation of polydisperse distributions of water-dispersed
maleic anhydride-grafted-polypropylene micro- and nano-sized droplets. Polymer
50:4584–4595
83. Salmer
on Sánchez M, Mathot V, Vanden Poel G, Groeninckx G, Bruls W (2006) Crystalli-
zation of polyamide confined in sub-micrometer droplets dispersed in a molten polyethylene
matrix. J Polym Sci Polym Phys 44:815–825
84. Mileva D, Androsch R, Cavallo D, Alfonso GC (2012) Structure formation of random
isotactic copolymers of propylene and 1-hexene or 1-octene at rapid cooling. Eur Polym J
48:1082–1092
85. Zia Q, Androsch R (2009) Effect of atomic force microscope tip geometry on the evaluation
of the crystal size of semicrystalline polymers. Meas Sci Technol 20:097003 (4pp)
86. Mileva D, Androsch R, Zhuravlev E, Schick C, Wunderlich B (2012) Homogeneous nucle-
ation and mesophase formation in glassy isotactic polypropylene. Polymer 53:277–282
87. Androsch R, Di Lorenzo ML (2013) Crystal nucleation in glassy poly(L-lactic acid). Macro-
molecules 46:6048–6056
Crystal Nucleation of Polymers at High Supercooling of the Melt 285

88. Androsch R, Di Lorenzo ML (2013) Kinetics of crystal nucleation of poly(L-lactic acid).

Polymer 54:6882–6885
89. Illers KH (1971) Geordnete Strukturen in “amorphem” Polyäthylenterephthalat. Kolloid Z Z
Polym 245:393–398
90. Androsch R, Schick C, Schmelzer JWP (2014) Sequence of enthalpy relaxation, homoge-
neous crystal nucleation and crystal growth in glassy polyamide 6. Eur Polym J 53:100–108
91. Hodge IM (1994) Enthalpy relaxation and recovery in amorphous materials. J Non-Cryst
Solids 169:211–266
92. Moynihan CT, Easteal AJ, De Bolt MA, Tucker J (1976) Dependence of the fictive temper-
ature of glass on cooling rate. J Am Ceram Soc 59:12–16
93. Moynihan CT, Easteal AJ, Wilder J, Tucker J (1974) Dependence of the glass transition
temperature on heating and cooling rate. J Phys Chem 78:2673–2677
94. Wunderlich B (2003) Reversible crystallization and the rigid–amorphous phase in semicrys-
talline macromolecules. Prog Polym Sci 28:383–450
95. Schick C, Krämer L, Mischok W (1985) Der Einfluß struktureller Veränderungen auf den
Glasübergang in teilkristallinem Polyethylenterephthalat I. Isotherme Kristallisation. Acta
Polym 36:47–53
96. Schick C, Fabry F, Schnell U, Stoll G, Deutschbein L, Mischok W (1988) Der Einfluß
struktureller Veränderungen auf den Glasübergang in teilkristallinem Poly(ethylenterephthalat)
2. Charakterisierung der übermolekularen Struktur. Acta Polym 39:705–710
97. Schick C, Wigger J, Mischok W (1990) Der Einfluß struktureller Veränderungen auf den
Glasübergang in teilkristallinem Poly(ethylenterephthalat) 3. Der Glasübergang in den
zwischenlamellaren Bereichen. Acta Polym 41:137–142
98. Androsch R, Wunderlich B (2005) The link between rigid amorphous fraction and crystal
perfection in cold-crystallized poly(ethylene terephthalate). Polymer 46:12556–12566
99. Zia Q, Mileva D, Androsch R (2008) The rigid amorphous fraction in isotactic polypropyl-
ene. Macromolecules 41:8095–8102
100. Kolesov I, Androsch R (2012) The rigid amorphous fraction of cold-crystallized polyamide
6. Polymer 53:4070–4077
101. 12th Laehnwitz Seminar on Calorimetry: Interplay between nucleation, crystallization, and
the glass transition, 10–15 June 2012, Rostock, Germany.
102. Michell RM, Blaszczyk-Lezak I, Mijangos C, Müller AJ (2013) Confinement effects on
polymer crystallization: from droplets to alumina nanopores. Polymer 54:4059–4077
103. Carvalho JL, Dalnoki-Veress K (2010) Homogeneous bulk, surface, and edge nucleation in
crystalline nanodroplets. Phys Rev Lett 105:237801
104. Hayes NW, Beamson G, Clark DT, Law DSL, Raval R (1996) Crystallization of PET from
the amorphous state: observation of different rates for surface and bulk using XPS and FTIR.
Surf Interface Anal 24:723–728
105. De Cupere VM, Rouxhet PG (2002) Surface crystallization of poly(ethylene terephthalate)
studied by atomic force microscopy. Polymer 43:5571–5576
106. Jukes PC, Das A, Durell M, Trolley D, Higgins AM, Geoghegan M, MacDonald JE, Jones
RAL, Brown S, Thompson P (2005) Kinetics of surface crystallization in thin films of poly
(ethylene terephthalate). Macromolecules 38:2315–2320
107. Durell M, MacDonald JE, Trolley D, Wehrum A, Jukes PC, Jones RAL, Walker CJ, Brown S
(2002) The role of surface-induced ordering in the crystallization of PET films. Europhys Lett
58:844–850
108. Zia Q, Ingolič E, Androsch R (2010) Surface and bulk morphology of cold-crystallized poly
(ethylene terephthalate). Colloid Polym Sci 288:819–825
109. Wunderlich B, Grebowicz J (1984) Thermotropic mesophases and mesophase transitions of
linear, flexible macromolecules. Adv Polym Sci 60(61):1–59
110. Auriemma F, De Rosa C, Corradini P (2005) Solid mesophases in semicrystalline polymers:
structural analysis by diffraction techniques. Adv Polym Sci 181:1–74
286 R. Androsch and C. Schick

111. Androsch R, Di Lorenzo ML, Schick C, Wunderlich B (2010) Mesophases in polyethylene,

polypropylene, and poly(1-butene). Polymer 51:4639–4662
112. Natta G, Peraldo M, Corradini P (1959) Smectic mesomorphic form of isotactic polypropyl-
ene. Rend Accad Naz Lincei 26:14–17
113. Natta G, Corradini P (1960) Structure and properties of isotactic polypropylene. Nuovo
Cimento 15(Suppl):40–51
114. Natta G (1960) Progress in the stereospecific polymerization. Makromol Chem 35:94–131
115. Grebowicz J, Lau SF, Wunderlich B (1984) The thermal properties of polypropylene.
J Polym Sci Polym Symp 71:19–37
116. Mileva D, Androsch R, Zhuravlev E, Schick C (2009) The temperature of melting of the
mesophase of isotactic polypropylene. Macromolecules 42:7275–7278
117. Schaefer D, Spiess HW, Suter UW, Fleming WW (1990) Two-dimensional solid-state NMR
studies of ultraslow chain motion: glass transition in atactic poly(propylene) versus helical
jumps in isotactic poly(propylene). Macromolecules 23:3431–3439
118. Bunn CW, Garner EV (1947) The crystal structures of two polyamides (‘nylons’). Proc R Soc
Lond A Math Phys Sci 18:39–68
119. Aelion R (1948) Preparation and structure of some new types of polyamides. Ann Chim Appl
3:5–61
120. Slichter WP (1959) Crystal structures in polyamides made from ω-amino acids. J Polym Sci
36:259–266
121. Little K (1959) Investigation of Nylon “texture” by X-ray diffraction. Br J Appl Phys
10:225–230
122. Advanced Thermal Analysis System (ATHAS) Data Base. Implemented into Springer-
Materials, available at materials.springer.com
123. Wunderlich B (2008) Thermal properties of aliphatic Nylons and their link to crystal structure
and molecular motion. J Therm Anal Calorim 93:7–17
124. Schmidt GF, Stuart HA (1958) Gitterstrukturen mit räumlichen Wasserstoffbrückensystemen
und Gitterumwandlungen bei Polyamiden. Z Naturforsch A 13:222–225
125. Ziabicki A (1959) Über die mesomorphe β-Form von Polycapronamid und ihre Umwandlung
in die kristalline Form α. Kolloid Z 167:132–141
126. Cavallo D, Gardella L, Alfonso GC, Portale G, Balzano L, Androsch R (2011) Effect of
cooling rate on the crystal/mesophase polymorphism of polyamide 6. Colloid Polym Sci
289:1073–1079
127. Haberkorn H, Illers KH, Simak P (1979) Molekülordnung und Kristallinität in Polyhexa-
methylenadipamid. Colloid Polym Sci 257:820–840
128. Brill R (1942) Über das Verhalten von Polyamiden beim Erhitzen. J Prakt Chem 161:49–64
129. Brill R (1956) Beziehungen zwischen Wasserstoffbindung und einigen Eigenschaften von
Polyamiden. Makromol Chem 18:294–309
130. Androsch R, Stolp M, Radusch HJ (1996) Crystallization of amorphous polyamides from the
glassy state. Acta Polym 47:99–104
131. Fichera A, Malta V, Marega C, Zannetti R (1988) Temperature dependence of the polymor-
phous phases of nylon 6. Makromol Chem 189:1561–1567
132. Threlfall T (2003) Structural and thermodynamic explanations of Ostwald’s rule. Org Process
Res Dev 7:1017–1027
133. Yokouchi M, Sakakibara Y, Chatani Y, Tadokori H, Tanaka T, Yoda K (1976) Structures of
two crystalline forms of poly(butylene terephthalate) and reversible transition between them
by mechanical deformation. Macromolecules 9:266–273
134. Son K (2000) Formation of polymorphic structure and its influences on properties in
uniaxially stretched polybutylene terephthalate films. J Appl Polym Sci 78:412–423
135. Bornschlegl E, Bonart R (1980) Small angle X-ray scattering studies of poly(ethylene
terephthalate) and poly(butylene terephthalate). Colloid Polym Sci 258:319–331
136. Saeidlou S, Huneault MA, Li H, Park CB (2012) Poly(lactic acid) crystallization. Prog Polym
Sci 37:1657–1677
Crystal Nucleation of Polymers at High Supercooling of the Melt 287

137. Cocca M, Androsch R, Righetti MC, Malinconico M, Di Lorenzo ML (2014) Conforma-

tionally disordered crystals and their influence on material properties: the cases of isotactic
polypropylene, isotactic poly(1-butene), and poly(L-lactic acid). J Mol Struct 1078:114–132
138. De Santis P, Kovacs AJ (1968) Molecular conformation of poly(S-lactic acid). Biopolymers
6:299–306
139. Kalb B, Pennings AJ (1980) General crystallization behaviour of poly(L-lactic acid). Polymer
21:607–612
140. Hoogsteen W, Postema AR, Pennings AJ, Ten Brinke G, Zugenmaier P (1990) Crystal
structure, conformation, and morphology of solution-spun poly(L-lactide) fibers. Macromol-
ecules 23:634–642
141. Pan P, Zhu B, Kai W, Dong T, Inoue Y (2008) Effect of crystallization temperature on crystal
modifications and crystallization kinetics of poly(L-lactide). J Appl Polym Sci 107:54–62
142. Pan P, Kai W, Zhu B, Dong T, Inoue Y (2007) Polymorphous crystallization and multiple
melting behavior of poly(L-lactide): molecular weight dependence. Macromolecules
40:6898–6905
143. Zhang J, Tashiro K, Domb AJ, Tsuji H (2006) Confirmation of disorder α form of poly(L-
lactic acid) by the X-ray fiber pattern and polarized IR/Raman spectra measured for
uniaxially-oriented samples. Macromol Symp 242:274–278
144. Zhang J, Duan Y, Sato H, Tsuji H, Noda I, Yan S, Ozaki Y (2005) Crystal modifications and
thermal behavior of poly(L-lactic acid) revealed by infrared spectroscopy. Macromolecules
38:8012–8021
145. Kawai T, Rahman N, Matsuba G, Nishida K, Kanaya T, Nakano M, Okamoto H, Kawada J,
Usuki A, Honma N, Nakajima K, Matsuda M (2007) Crystallization and melting behavior of
poly(L-lactic acid). Macromolecules 40:9463–9469
146. Androsch R, Schick C, Di Lorenzo ML (2014) Melting of conformationally disordered
crystals (α0 -phase) of poly(L-lactic acid). Macromol Chem Phys 215:1134–1139
147. Androsch R, Zhuravlev E, Schick C (2014) Solid-state reorganization, melting and melt-
recrystallization of conformationally disordered crystals (α0 -phase) of poly(L-lactic acid).
Polymer 55:4932–4941
148. Yasuniwa M, Tsubakihara S, Iura K, Ono Y, Dan Y, Takahashi K (2006) Crystallization
behavior of poly(L-lactic acid). Polymer 47:7554–7563
149. Tsuji H, Takai H, Saha SK (2006) Isothermal and non-isothermal crystallization behavior of
poly(L-lactic acid): effects of stereocomplex as nucleating agents. Polymer 47:3826–3837
150. Tsuji H, Ikada Y (1996) Crystallization from the melt of poly(lactide)s with different optical
purities and their blends. Macromol Chem Phys 197:3483–3499
151. Li X, Li Z, Zhong G, Li L (2008) Steady – shear-induced isothermal crystallization of poly(L-
lactide) (PLLA). J Macromol Sci Phys 47:511–522
152. Tammann G (1898) Number of nuclei in supercooled liquids. Z Phys Chem 25:41–479
153. Di Lorenzo ML (2006) The crystallization and melting processes of poly(L-lactic acid).
Macromol Symp 234:176–183
154. Di Lorenzo ML (2001) Determination of spherulite growth rates of poly(L-lactic acid) using
combined isothermal and non-isothermal procedures. Polymer 42:9441–9446
155. Di Lorenzo ML (2005) Crystallization behavior of poly(L-lactic acid). Eur Polym J
41:569–575
156. De Santis F, Pantani R, Titomanlio G (2011) Nucleation and crystallization kinetics of poly
(lactic acid). Thermochim Acta 522:128–134
157. Sánchez MS, Mathot VBF, Vanden Poel G, Ribelles JLG (2007) Effect of cooling rate on the
nucleation kinetics of poly(L-lactic acid) and its influence on morphology. Macromolecules
40:7989–7997
158. Hernández Sánchez F, Molina Mateo J, Romero Colomer FJ, Salmer
on Sánchez M, G
omez
Ribelles JL, Mano JF (2005) Influence of low-temperature nucleation on the crystallization
process of poly(L-lactide). Biomacromolecules 6:3283–3290
288 R. Androsch and C. Schick

159. Zhang T, Hu J, Duan Y, Pi F, Zhang J (2011) Physical aging enhanced mesomorphic structure
in melt-quenched poly(L-lactic acid). J Phys Chem B 115:13835–13841
160. Androsch R, Monami A, Kucera J (2014) Effect of an alpha-phase nucleating agent on the
crystallization kinetics of a propylene/ethylene random copolymer at largely different
supercooling. J Cryst Growth 408:91–96
161. Zhuravlev E, Wurm A, P€ otschke P, Androsch R, Schmelzer JWP, Schick C (2014) Kinetics
of nucleation and crystallization of poly(ε-caprolactone) – multiwalled carbon nanotube
composites. Eur Polym J 52:1–11
Index

A strain-induced, 167, 177

Anisotropy, strain-induced, 173 Crystal modification, control, 78
Anthracene, 59–62 Cyclic voltammetry (CV), 83
Arrhenius shift factor, 158
Atomic force microscopy (AFM), 69, 75, 83,
97, 99, 110, 114, 120, 131 D
Autoepitaxy, 62 Dicyclohexylterephthalamide, 57
Differential radial distribution function
(DRDF), 170
B Dilatometry, 218, 243
Ballistic cooling, 129, 145, 148 extended, 249
Banded–nonbanded transition (BNB), 114 Discrete elastic-viscoelastic stress split
Banded polymer spherulite, 95 (DEVSS), 253
Bifurcation of growth, 26
Biomorphs, 98
Borax, 96 E
Entanglements, 128, 197
Epitaxy, 23, 32, 46, 55, 58, 87
C Ewald sphere, 102, 106, 107, 145
Chain tilt, 95 Extended-chain crystals (shishs), 62
Chirality, 102, 121 Extended pom-pom (XPP) model, 248
Clapeyron equation, 154, 249, 261, 275
Compressibility, 243, 257, 265
Concomitant crystallization, 1 F
Continuous-cooling-transformation (CCT) Fast scanning calorimetry (FSC), 216
diagrams, 148 Flow-induced crystallization (FIC), 244
Cooling, fast, 212 Flow instabilities, 271
undercooling, 4, 16, 24, 31, 40, 49, 147, Folded-chain lamellar crystals (kebabs), 62
152, 209, 275, 282 Fourier transform infrared spectroscopy
Couette cell, 249 (FTIR), 67–69, 129
Cross-nucleation, 1
Crystallization, 55
flow-induced, 243 G
kinetics, 167, 207, 261 Gibbs–Thomson relation, 6
shear-induced, 127 Growth, bifurcation, 26

289
290 Index

Growth (cont.) Millisecond time resolution, 127

transformation, 26 Modelling, 207
transition, 26, 48, 79 Molecular vibration spectroscopy, 68
Guinier approximation, 142 Momentum balance, 254
Monte Carlo simulations, 68

H
Halos, amorphous, 170 N
Heat balance, 221, 231, 233, 257 Nanocalorimetry, 95
Hencky strain rate, 159 Nanofocus X-ray scattering, 95
Hermans’ orientation factor, 251, 270, 274 Natural rubber, 167, 168
Heteroepitaxy, 64 Nucleation, 1, 4
Hexamethyleneadipate (HA), 18 cross-nucleation, 1
High-density polyethylene (HDPE), 95, 97, density, 208
102, 107 point-like, 264, 282
High shear rates, 252 shish-kebab, 279
Hoffman–Lauritzen theory, 28, 100, 275
Homoepitaxy, 62
O
Organic field-effect transistors (OFETs), 56, 76
I Orientation factor, 251, 270
Impingement, 281 Ostwald’s rule, 1, 3–6, 24, 42

K P
Kebabs, 62, 142, 160, 245, 270, 279 Parameter sensitivity analysis, 267
Kolmogoroff–Avrami expression, 210, 247 Photoswitchers, 79, 81
KP-model, 100 Pirouette dilatometer, 218, 249
Polanyi sphere, 107
Polarized optical microscopy (POM), 11, 43,
L 110, 130, 151, 271
Lamellae, 44, 62, 95 Poly(1,3-dioxolan) (PDOL), 19, 24
branching, 243 Poly(3-hexylthiophene) (P3HT), 69
chirality, 95, 106–108 Poly(3-hydroxybutyrate) (PHB), 74, 95, 120
daughters, 245, 263, 270, 287 Poly(3-hydroxypropionate) (PHP), 17
edge-on, 74 Poly(3-hydroxyvalerate) (PHV), 120
handedness, 107 Poly(3-methylthiophene) (P3MT), 83
nanoscale, 31 Poly(butylene adipate) (PBA), 5, 17
thickness, 5, 6, 71, 118, 154, 226 Poly(butylene terephthalate) (PBT), 113
twisted, 99–106, 119 Poly(butylene-2,6-naphthalate) (PBN), 18
Lauritzen–Hoffman theory, 28, 100, 275 Poly(caprolactone) (PCL), 66, 77, 102
Lennard–Jones potential, 24 Poly(di-n-hexylsilane), 74
Linear low density polyethylene (LLDPE), Poly(ethylene adipate) (PEA), 72
130, 152 Poly(ethylene oxide) (PEO), 19
Linear viscoelastic envelope (LVE), 158 Poly(ethylene sebacate), 77
Log-conformation representation (LCR), 253 Poly(ethylene terephthalate) (PET), 113, 208
Poly(heptamethylene terephthalate)
(PHepT), 18
M Poly(hexamethylene terephthalate) (PHexT),
Mannitol, 23 11, 18, 32
Mechanical cycling, 189 Poly(lactic acid) (PLA), 97, 246
5-methyl-2-[(2-nitrophenyl)amino]-3- Poly(L-lactic acid) (PLLA), 21, 33, 102
thiophenecarbonitrile, 23 Poly(methylmethacrylate) (PMMA), 74
Index 291

Poly(pivalolactone) (PPVL), 3, 30, 35 Self-poisoning, polymorphic, 20

Poly(propylene adipate) (PPA), 95, 97, 110 Semicrystalline polymers, 2, 7, 55, 57, 68, 74,
Poly(tetrafluoroethylene) (PTFE), 76, 77, 86 79, 111, 128–138, 158, 170, 244
Poly(trimethylene terephthalate) (PTT), 95, 97, cross-nucleation, 24
102, 113 non-isothermal crystallization, 207
Poly(vinylidene fluoride) (PVDF), 2, 15, 57, polymorphs, 2
95, 97, 99 Porod behavior, 141
Poly(vinylpyrrolidone), 23 Sentmanat extension rheometer (SER), 156
Polyethylene, 25, 59, 95, 207 Shear layer thickness, 271
high-density (HDPE), 97, 102, 107 Shish, 62, 243, 248, 273
linear low density (LLDPE), 130, 152, Shish-kebab, 62, 131–159, 243, 245, 280
209, 224 Slit flow, 149–152, 217, 250
shish-kebab, 131 Small-angle light scattering (SALS), 228
Poly(1-butene), isotactic (i-PBu), 3, 13, 42, Small-angle oscillatory shear (SAOS), 255
102, 246 Small-angle X-ray scattering (SAXS), 104,
Poly(4-methyl-1-pentene), isotactic 107, 127, 135, 264
(i-P4M1P), 2, 14 Spherulites, 11, 16, 18, 62, 96, 130, 260
Polymorphic self-poisoning, 20 banding, 97, 116
Polymorphism, 1, 72, 78, 208, 243 bandwidth, 115
Polymorphs, 1, 98, 157, 209, 275, 282 branches, 38
concomitant crystallization, 7 crystallization kinetics, 247, 278
cross-nucleation, 22 flow-induced, 287
ROY, 23 isotropic, 262
Polyoctenamer, 77 nucleated, 30, 36, 243
Polyoxymethylene, 77 Strain-induced anisotropy, 173
Poly( p-phenylene ethynylene)s, 79 Strain-induced crystallization, 167, 177
Polypropylene, 2, 207, 243 Strain rate, Hencky, 159
isotactic (iPP), 2, 11, 57, 148, 208, 246 Strain-regulation process, 196
syndiotactic (sPP), 2, 71 Strain relaxation, 167
Polystyrene, isotactic, 75, 76, 102 Streamers, 259
syndiotactic, 11, 14, 21 Streamline-upwind Petrov–Galerkin (SUPG),
Pressure, 217 253
Processing, 127, 243 Stretching, 57, 167, 195, 253
Propene/ethylene, ballistic cooling, 145 uniaxial, 127, 142
Structure regulation, 74
Synchrotron radiation, 95
Q
Quiescent state, 169, 176, 180
Quinacridone, 57 T
Tazofelone (TZF), 33
Temperature-induced crystallization (TIC),
R 169
Real-time fast structuring, 127 Terphenyl, 59, 61, 72
Recovery, 113, 191–194, 199, 201 Thompson scattering, 135
Rheo-SAXS/WAXD, 156
ROY polymorphs, 23
Rubber elasticity, 178 U
Undercooling, 4, 16, 24, 31, 40, 49, 147, 152,
209, 275, 282
S
Schneider rate equations, 210, 230, 247, 262,
275, 278, 288 V
Screw dislocation, 100, 109, 116 Viscoelastic fluid model, 255
Selected-area electron diffraction (SAED), 113 Viscoelasticity, nonlinear, 243, 253, 260, 269
292 Index

W X
Wide-angle X-ray diffraction (WAXD), 12, 18, X-ray diffraction, 12, 15, 25, 43, 69, 85,
34, 43, 67, 129, 131, 213, 250 167–175, 202, 213, 247, 250
Wide-angle X-ray scattering (WAXS), 104, X-ray photoelectron spectrum (XPS), 84
127, 129, 161 X-ray scattering, nanofocus, 95, 102