NanoScience and Technology
NanoScience and Technology
Series Editors:
P. Avouris B. Bhushan D. Bimberg K. von Klitzing H. Sakaki R. Wiesendanger
The series NanoScience and Technology is focused on the fascinating nano-world, mesoscopic physics, analysis with atomic resolution, nano and quantum-effect devices, nanomechanics and atomic-scale processes. All the basic aspects and technology-oriented developments in this emerging discipline are covered by comprehensive and timely books.
The series constitutes a survey of the relevant special topics, which are presented by leading experts in the field. These books will appeal to researchers, engineers, and advanced
students.
Applied Scanning Probe Methods II
Scanning Probe Microscopy
Techniques
Editors: B. Bhushan and H. Fuchs
Applied Scanning Probe Methods VI
Characterization
Editors: B. Bhushan and S. Kawata
Applied Scanning Probe Methods III
Characterization
Editors: B. Bhushan and H. Fuchs
Applied Scanning Probe Methods VII
Biomimetics
and Industrial Applications
Editors: B. Bhushan and H. Fuchs
Applied Scanning Probe Methods IV
Industrial Application
Editors: B. Bhushan and H. Fuchs
Roadmap
of Scanning Probe Microscopy
Editors: S. Morita
Scanning Probe Microscopy
Atomic Scale Engineering
by Forces and Currents
Editors: A. Foster and W. Hofer
Nanocatalysis
Editors: U. Heiz and U. Landman
Single Molecule Chemistry
and Physics
An Introduction
By C. Wang and C. Bai
Atomic Force Microscopy, Scanning
Nearfield Optical Microscopy
and Nanoscratching
Application
to Rough and Natural Surfaces
By G. Kaupp
Applied Scanning Probe Methods V
Scanning Probe Microscopy
Techniques
Editors: B. Bhushan, H. Fuchs,
and S. Kawata
Nanostructures
Fabrication and Analysis
Editor: H. Nejo
Fundamentals of Friction and Wear
on the Nanoscale
Editors: E. Gnecco and E. Meyer
Lateral Alignment
of Epitaxial Quantum Dots
Editor: O. Schmidt
Nanostructured Soft Matter
Experiment, Theory, Simulation
and Perspectives
Editor: A.V. Zvelindovsky
A.V. Zvelindovsky
(Ed.)
Nanostructured
Soft Matter
Experiment, Theory, Simulation
and Perspectives
With 261 Figures
Dr. A.V. Zvelindovsky (Ed.)
Centre for Materials Science
Department of Physics, Astronomy and Mathematics
University of Central Lancashire
Preston
Lancashire PR1 2HE
United Kingdom
Series Editors:
Professor Dr. Phaedon Avouris
Professor Dr., Dres. h.c. Klaus von Klitzing
IBM Research Division
Nanometer Scale Science & Technology
Thomas J. Watson Research Center
P.O. Box 218
Yorktown Heights, NY 10598, USA
Max-Planck-Institut
für Festkörperforschung
Heisenbergstr. 1
70569 Stuttgart, Germany
Professor Dr. Bharat Bhushan
University of Tokyo
Institute of Industrial Science
4-6-1 Komaba, Meguro-ku
Tokyo 153-8505, Japan
Ohio State University
Nanotribology Laboratory
for Information Storage
and MEMS/NEMS (NLIM)
Suite 255, Ackerman Road 650
Columbus, Ohio 43210, USA
Professor Dr. Dieter Bimberg
TU Berlin, Fakutät Mathematik/
Naturwissenschaften
Institut für Festkörperphyisk
Hardenbergstr. 36
10623 Berlin, Germany
Professor Hiroyuki Sakaki
Professor Dr. Roland Wiesendanger
Institut für Angewandte Physik
Universität Hamburg
Jungiusstr. 11
20355 Hamburg, Germany
A C.I.P. Catalogue record for this book is available from the Library of Congress
ISSN 1434-4904
ISBN 978-1-4020-6329-9 (HB)
ISBN 978-1-4020-6330-5 (e-book)
Published by Springer,
P.O. Box 17, 3300 AA Dordrecht, The Netherlands
In association with
Canopus Publishing Limited,
27 Queen Square, Bristol BS1 4ND, UK
www.springer.com and www.canopusbooks.com
All Rights Reserved
© Canopus Publishing Limited 2007
No part of this work may be reproduced, stored in a retrieval system, or transmitted in any form or by
any means,electronic,mechanical,photocopying,microfilming,recording or otherwise,without written
permission from the Publisher, with the exception of any material supplied specifically for the purpose
of being entered and executed on a computer system, for exclusive use by the purchaser of the work.
Preface
“The scientist does not study nature because it is useful; he studies
it because he delights in it, and he delights in it because it is beautiful.
If nature were not beautiful, it would not be worth knowing, and if
nature were not worth knowing, life would not be worth living.”
Henri Poincaré (1854 - 1912)
The ancient Greeks, quite ingeniously, realised that all materials and their
(now known as macroscopic) properties, including life itself, are due to a
limited number of tiny, constantly moving building blocks and the connections (now called interactions) between these blocks. Receiving both scientific
and non-scientific opposition, the idea faded and, despite some renaissance of
atomistic ideas in the 17-19th centuries, it still took more than two thousand
years, until the time of Einstein, for the idea of microscopic building blocks to
be fully accepted. These ideas, begun during the golden age of physics in the
20th century, have led to a comprehensive understanding of such states of matter as gases and solids, which in turn have completely revolutionised everyday
life in the developed world by introducing technological wonders such as modern cars, air traffic, semiconductor chips for computers and nuclear power.
Another state of matter, fluids, appeared to be much more difficult to tackle,
even in the case of simple liquids like liquid argon, a research favourite in the
field. Legend tells that Lev D. Landau, Physics Nobel Laureate, was said to
have commented that there could be no theoretical physics of liquids, as they
have no small parameters. Nonetheless, as the 20th century advanced, it also
became possible to treat even this most slippery of subjects due, in part, to
the introduction of computers and the development of computer simulation
methods like molecular dynamics. The 20th century brought yet another revolution: the industrial production of novel classes of materials, which simply
did not exist before. For instance, almost every aspect of our everyday life
would change immeasurably if plastics should disappear and life would turn
“blind”, “deaf” and rather miserable without liquid crystals for computer
screens or mobile phones. Such new materials were given the name complex
fluids, and their building blocks are not simply atoms or small molecules, but
include block copolymers, surfactants, amphiphiles, colloids, liquid crystals,
biomacromolecules, such as proteins and DNA, and various composites of the
above. Complex fluids possess features of both fluids (for instance, they can
flow) and solids (they can have an internal structure often with various well
VI
Preface
resolved symmetry groups). These structures have a characteristic scale for
their building blocks which is in the range of nanometers to microns, but the
building blocks can be made (synthesised) with various degrees of complexity,
so more than one size scale can be involved. Some structures can be formed
spontaneously from a homogeneous mixture of the building blocks, a process
referred to as self-assembly, which can be hierarchical and occur on various
time scales depending on the complexity of the building blocks. Self-assembly
is related to self-organization, which makes complex fluids similar to living
matter, so they can serve as model systems for biological systems and bioinspired materials. In the last decades of the 20th century the term complex
fluids started to be substituted by a more general one that is better suited
to the overall concept of condensed matter: soft matter. The transition between millennia was marked by a burst of soft matter research, due, in part,
to the fact that computers had then reached a level of power allowing the
simulation of experimental size systems, thus enabling the very first “virtual
experiments” of such complex systems to be performed. This development
made the links between theory and experiment truly symbiotic.
Nanostructured soft materials, even apart from future technological perspectives beyond our imagination, are fascinating and beautiful. This research
field is growing so fast that there has been no single book that provided an
overview of the many different perspectives on both fundamental concepts
and recent advances in the field. A group of very enthusiastic contributors has
now filled this gap; and the present book is the first comprehensive monograph on nanostructured soft matter. It covers materials ranging in size from
short amphiphilic molecules to block copolymers to proteins and also discusses
colloids, hybrids, microemulsions and bio-inspired materials such as vesicles.
Each chapter is written by active world-class researchers in the field who offer
the reader an interdisciplinary view from differing perspectives. They combine the experimental approaches of Chemistry and Physics, e.g. scattering
techniques, electron and Atomic Force microscopy, with various Theoretical
Physics, Mathematics and advanced computer modelling methods. We hope
the book will be useful for both active and starting researchers as well as for
undergraduate students; or, citing one of the anonymous referees of the original proposal for this book: “There is something for everyone in this book and
it would represent a very useful text for those both operating at the forefront
of nano-science and those entering the field . . . ”
I wish to thank the publishers at Canopus for assistance in the production
of this book. I also thank Drs. R. McCabe, S. V. Kuzmin and N. Kiriushcheva.
My editorial effort is dedicated to Prof. A. V. Zatovsky (1942-2006), who first
introduced me to the wonders of Soft Matter.
Preston, Lancashire, January 2007
AVZ
Contents
Preface
A. V. Zvelindovsky (ed.) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . V
Part I Experimental Advances
Microemulsion Templating
W. F. C. Sager . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3
Nanofabrication of Block Copolymer Bulk and Thin Films:
Microdomain Structures as Templates
Takeji Hashimoto and Kenji Fukunaga . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
Characterization of Surfactant Water Systems by X-Ray
Scattering and 2 H NMR
Michael C. Holmes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
Polyelectrolyte Diblock Copolymer Micelles:
Small Angle Scattering Estimates of the Charge Ordering in
the Coronal Layer
Johan R. C. van der Maarel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
Structure and Shear-Induced Order in Blends of a Diblock
Copolymer with the Corresponding Homopolymers
I. W. Hamley, V. Castelletto and Z. Yang . . . . . . . . . . . . . . . . . . . . . . . . . . . 159
Electric Field Alignment of Diblock Copolymer Thin Films
T. Xu, J. Wang and T. P. Russell . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 171
Control of Block Copolymer Microdomain Orientation from
Solution using Electric Fields:
Governing Parameters and Mechanisms
Alexander Böker . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
VIII
Contents
Structure and Dynamics of Cylinder Forming Block
Copolymers in Thin Films
Larisa Tsarkova . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
Part II Mathematical and Theoretical Approaches
Mathematical Description of Nanostructures with Minkowski
Functionals
G.J.A. Sevink . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 269
Scaling Theory of Polyelectrolyte and Polyampholyte Micelles
Nadezhda P. Shusharina and Michael Rubinstein . . . . . . . . . . . . . . . . . . . . . 301
The Latest Development of the Weak Segregation Theory of
Microphase Separation In Block Copolymers
I. Ya. Erukhimovich . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 327
Coarse-Grained Modeling of Mesophase Dynamics in Block
Copolymers
Zhi-Feng Huang and Jorge Viñals . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 371
Effective Interactions in Soft Materials
Alan R. Denton . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 395
Part III Computer Simulations
Ab-initio Coarse-Graining of Entangled Polymer Systems
J.T. Padding and W.J. Briels . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 437
Computer Simulations of Nano-Scale Phenomena Based on
the Dynamic Density Functional Theories:
Applications of SUSHI in the OCTA System
Takashi Honda and Toshihiro Kawakatsu . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461
Monte Carlo Simulations of Nano-Confined Block Copolymers
Qiang Wang . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 495
Understanding Vesicles and Bio-Inspired Systems with
Dissipative Particle Dynamics
Julian C. Shillcock . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 529
Theoretical Study of Nanostructured Biopolymers Using
Molecular Dynamics Simulations: A Practical Introduction
Danilo Roccatano . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 555
Contents
IX
Understanding Liquid/Colloids Composites with Mesoscopic
Simulations
Ignacio Pagonabarraga . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 587
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 617
List of Contributors
Alexander Böker (Germany)
Physikalische Chemie II, Universität
Bayreuth.
Wim J Briels (The Netherlands) Computational Biophysics,
Faculty of Science and Technology,
University of Twente, Enschede.
Valeria Castelletto (United
Kingdom) Department of Chemistry, University of Reading.
Alan R Denton (USA) Department of Physics, North Dakota State
University, Fargo.
Igor Erukhimovich (Russia)
Nesmeyanov Institute of Organoelement Compounds, Russian
Academy of Sciences, Moscow, and
Physics Department, Moscow State
University.
Kenji Fukunaga (Japan) Polymer
Laboratory, Ube Industries, Ltd.,
Chiba.
Ian W Hamley (United Kingdom) Department of Chemistry,
University of Reading.
Takeji Hashimoto (Japan)
Advanced Science Research Centre,
Japan Atomic Energy Agency,
Ibaraki.
Michael C Holmes (United
Kingdom) Centre for Materials
Science, Department of Physics,
Astronomy and Mathematics,
University of Central Lancashire,
Preston.
Takashi Honda (Japan) Japan
Chemical Innovation Institute,
and Department of Organic and
Polymeric Materials, Tokyo Institute
of Technology.
Zhi-Feng Huang (USA, Canada)
Department of Physics and Astronomy, Wayne State University,
Detroit and McGill Institute for
Advanced Materials and Department
of Physics, McGill University,
Montreal.
Toshihiro Kawakatsu (Japan)
Department of Physics, Tohoku
University, Sendai.
XII
List of Contributors
Johan R C van der Maarel (Singapore) Department of Physics,
National University of Singapore.
Nadezhda P Shusharina (USA)
Department of Chemistry, University
of North Carolina at Chapel Hill.
Johan T Padding (The Netherlands) Computational Biophysics,
Faculty of Science and Technology,
University of Twente, Enschede.
Larisa A Tsarkova (Germany)
Physikalische Chemie II, Universität
Bayreuth.
Ignacio Pagonabarraga (Spain)
Departament de Física Fondamental,
Universitat de Barcelona.
Jorge Viñals (Canada) McGill
Institute for Advanced Materials
and Department of Physics, McGill
University, Montreal.
Danilo Roccatano (Germany)
School of Engineering and Science,
Jacobs University Bremen.
Michael Rubinshtein (USA)
Department of Chemistry, University
of North Carolina at Chapel Hill.
Thomas P Russell (USA)
Polymer Science and Engineering Department, University of
Massachusetts, Amherst.
Wiebke F C Sager (Germany)
IFF-Soft Matter, Forschungszentrum
Jülich.
G J Agur Sevink (The Netherlands) Leiden Institute of Chemistry,
Leiden University.
Julian C Shillcock (Germany)
Theory Department, Max Plank
Institute of Colloids and Interfaces,
Potsdam.
Jiayu Wang (USA) Polymer
Science and Engineering Department, University of Massachusetts,
Amherst.
Qiang Wang (USA) Department
of Chemical and Biological Engineering, Colorado State University, Fort
Collins.
Ting Xu (USA) Department of
Materials Science and Engineering,
University of California, Berkeley.
Zhou Yang (United Kingdom)
School of Materials, The University
of Manchester.
Andrei V Zvelindovsky (United
Kingdom) Centre for Materials
Science, Department of Physics,
Astronomy and Mathematics,
University of Central Lancashire,
Preston.
The latest development
of the weak segregation theory
of microphase separation in block copolymers
Igor Erukhimovich
1. Introduction.
One of the most interesting phenomena occurring in copolymer systems is so-called
order-disorder transition (ODT) or microphase separation, i.e. formation of ordered
morphologies possessing symmetry of a crystal lattice,1-5 which occurs with changing
(typically decreasing) temperature T. With further decrease of T the ODT is often followed
by various order-order transitions between the different ordered morphologies1. Obviously,
the physical reason for this ordering is competition between the short-range segregation and
long-range stabilization tendencies. More precisely, with decreasing T the energy gain upon
local segregation grows as compared to the loss of the translational entropy accompanying
such segregation whereas the immiscible blocks can not separate fully because of their
covalent bonding. As a result, an ordered pattern of alternating domains, which are filled
preferably by the monomers of the same sort, arises. Block copolymer melts with different
structural and interaction parameters are known to form different morphologies at the orderdisorder and order-order transitions so that the ultimate goal of the theory is to determine
the symmetry and geometry of the most stable ordered phases for a copolymer melt or blend
given its composition, architecture and temperature as well as predict the thermodynamic,
scattering and mechanical properties of the phases.
The milestones in achieving this goal would be i) a general understanding of the
relationships between the block copolymer architecture and the symmetry of the occurring
ordered morphologies as well as scenarios of the order-disorder and order-order transitions;
ii) a quantitative description and calculation of the stable ordered phases structure; and iii)
calculation of various parameters of the phases. Therewith, having in mind the huge variety
of the block copolymer architecture the chemists could synthesize, it is very important that
the theory would be capable to provide an express information (even though a semiquantitative one) about new copolymers rather than explain the already observed systems
only. Putting it by words of Reimund Stadler, “the large number of a priori imaginable
combinations in ABC triblock copolymers makes it necessary that theory and experiment
have to be closely linked. By such a close feedback it may be avoided that the development
of new materials is merely based on accident.”
1
In what follows we use the terms “morphology”, “ordered phase” and “symmetry class” as synonyms.
1
The first ever attempt to describe theoretically the microphase separation transition was
done by Meier6 within so-called strong segregation approximation (SSA). According to the
SSA, the alternating domains forming the ordered morphologies are just microphases,
which consist of the corresponding chemically homogeneous blocks, the width d of the
inter-domain transient layer (domain interface) being rather narrow as compared to the
domain size L. Therefore, the SSA is based on the idea that the ODT is controlled by a
balance between an entropic loss due to confinement of the polymer blocks within (or
outside of) the domains (micelles) and energetic gain (as compared to the uniform state)
under formation of these micelles. Since within the SSA the unlike monomers contact and
interact within the domain interfaces only, the energetic gain is basically determined by the
values of the total domain interface and interface (surface) free energy. The SSA, which
was further elaborated by Helfand et al.7-10 and, finally, by Semenov et al.,11-14 results
naturally (in the limit of zero interface width) in the idea that the morphology of the ordered
phases is fully determined by the requirement that the total inter-domain surface is minimal.
Very transparent and appealing from the geometrical point of view, the minimal surface
approach is rather popular among mathematicians and experimentalists.15-18 However, the
condition of the narrow interface d<<L is practically never fulfilled in the real block
copolymer melts above the glass transition temperature Tg . E.g., in the binary AB diblock
copolymers the condition d<<L fulfills for χ ≥ 100 ,19 where χ = χN is the reduced Flory
interaction parameter20, N is the total degree of polymerization of the diblock copolymer
chain. In case the Flory parameter χ is related to the temperature T via the simple
relationship 2χ = θ T , θ being the Flory temperature, the condition holds for rather low
temperatures T ≤ θN 100 ~ TODT 10 , where TODT is the ODT temperature calculated by
Leibler21 in the opposite so-called weak segregation approximation (WSA). Besides, the
minimal surface approach disregards a long-range contribution11 due to micelles’ ordering.
The WSA, on which this review is focused, is related to the situation at the very onset
of ordering. The physical idea the WSA is based on was first stated by Landau22 as early as
in 1937 as a toy model of the 2nd order phase transition accompanied by an explicit
(
F {Φ ( r )}
)=1
(
ΔF r, {Φ ( r )}
) dr .
symmetry change. In the contemporary designations the original Landau Hamiltonian reads:
Φ ( r1 ) c ( r1 − r2 ) Φ ( r2 ) d r1 d r2 + ∫
2∫
(1.1)
Here the Fourier transform c ( q ) = ∫ d r c ( r ) exp ( iqr ) has a minimum at a finite q = q* ,
T
T
the order parameter Φ is assumed to be scalar for a while, the specific excess free energy of
the system in a point r
2
(
ΔF r, {Φ ( r )}
) = α Φ3 ( r ) + β Φ 4 ( r )
3!
T
(1.2)
4!
is determined by the value Φ(r) in the same point only and T is the temperature measured in
the energetic units (the Boltzmann constant kB = 1).
the Hamiltonian (1.1) is in a state described by a profile {Φ ( r )} with a probability
As consistent with general principles of statistical physics,23 the system described by
(
( (
)
) )
w {Φ ( r )} ~ Z −1 exp − F {Φ ( r )} T ,
(1.3)
corresponding functional integral over all profiles {Φ ( r )} :
the normalization constant Z being the partition function, which can be written as the
( (
) )
Z = ∫ exp − F {Φ ( r )} T δΦ ( r )
(1.4)
Thus, the total free energy of the system and any observable quantity a, which, evidently, is
just a thermodynamic average, i.e. the average taken with the probabilistic measure (1.3),
( (
) )
F = −T ln Z = −T ln ∫ exp − F {Φ ( r )} T δΦ ( r )
read
a ({Φ ( r )} ) exp ( − F ({Φ ( r )} ) T ) δΦ ( r )
) ∫ exp − F Φ r T δΦ r
∫ ( ({ ( )}) ) ( )
(
a = a {Φ ( r )} =
(1.5)
(1.6)
Φ (r ) = Φ (r )
The most important observables are, of course, the average order parameter
(1.7)
S ( r1 − r2 ) = ( Φ ( r1 ) − Φ ( r1 ) ) ( Φ ( r2 ) − Φ ( r2 ) )
and the correlation function
(
) ( (
as well as the scattering factor
G ( q ) = ∫ S r1 − r 2 exp i q r1 − r 2
) ) dr1dr 2
(1.8)
V,
(1.9)
where V is the total volume of the system.
For τ>0, a minimum (at least, a metastable one) of the virtual free energy (1.1) is
Φ (r ) ≡ 0 .
provided by the order parameter profile
(1.10)
Then Φ ( r ) = 0 , whereas to calculate the correlation function and scattering factor one
can keep only the first non-vanishing (quadratic) terms in the expansion of the Hamiltonian
(1.1) appearing in the integrals (1.5)-(1.8). In this approximation, which is referred to as the
random phase approximation (RPA), the desired expressions read:
3
(
⎡
G ( q ) = 1 c ( q ) ≈ ⎢ τ + C q 2 − q*2
⎣
) ⎥⎦
2 ⎤ −1
,
(1.11a)
value of the minimum τ = c ( q* ) , which plays role of a reduced temperature, measuring
where the second approximate equality holds near the minimum of the function c(q). If the
how close is the spatially uniform (disordered) system to the loss of its stability, is small
enough, then the correlation function reads (in an approximation valid for τ
(
(
) ( r τC )
) ( q* τC )
⎧sin ( q*r ) exp − r τ C ( 2q* )
⎪
S (r ) ∼ ⎨
⎪cos ( q*r ) exp − r τ C ( 2q* )
⎩
d =3
d =1
.
q*4 )
(1.11b)
The behavior of the weakly ordering systems described by the correlation function
(1.11b), which is shown in Fig. 1,.is rather different from that of the simple liquids2. Since
the correlation function is proportional to the response function,23,24 the equation (1.11b)
implies that any perturbation arising in the weakly ordering systems is propagating as a
harmonic wave (spherical, cylindrical or plane one depending on the perturbation’s
dimensionality), which oscillates with the period L = 2π q* and decays on the distance
rc ~ q* C τ . The waves’ amplitude is infinitely increasing (within the RPA) as τ → 0 .
Figure 1. Typical behavior of the plane perturbations in the weakly ordering systems. The curves of
different colors correspond to different values of the reduced (positive) temperature τ, the values of τ are
decreasing and the system is approaching ODT in the downward direction.
2
S ( r ) ∼ exp ( − r rc ) r , where rc is the correlation radius.
Remember that the correlation function of the simple liquids is described by the Ornstein-Zernike
correlation function
4
At τ ≤ 0 the perturbations become infinite. Were there no other restrictions, it would
mean that the spatially homogeneous state becomes absolutely unstable with respect to the
waves as the critical ones) and the RPA is not valid anymore. In this case Φ ( r ) = 0
growth of the order parameter waves with the wave lengths L (henceforth we refer to such
provided by a finite thermodynamically equilibrium order parameter profile Φ ( r ) ≠ 0 .
corresponds to a maximum of the virtual free energy, whereas the minimum of the latter is
In fact, however, the local order parameter cannot exceed some finite value (e. g., for
diblock copolymer melts Φ is a linear function of the volume fraction of the monomers A),
which implies that with increase of the fluctuations they become strongly correlated even in
the disordered state, the correlation being resulted in a decrease of the correlation function
S(r) and, thus, a stabilization of the disordered (spatially homogeneous) state as compared
to the RPA results. Quantitatively this effect is addressed and explained by so-called
Brazovskii (Hartree) fluctuation corrections25 to the mean field approximation, which are
shown to transform the order-disorder transition into the 1st order one even at the critical
point.
new equilibrium order parameter profile Φ ( r ) ≠ 0 could be neglected, which implies
In the mean field approximation, i.e. under assumption that the fluctuations around the
calculation of the integral (1.4) via the deepest descent method within the so-called pre-
(
F = −T ln Z = F {Φ ( r )}
)
exponential accuracy, we arrive at the well-known expression for the free energy
(1.12)
profile Φ ( r ) is expected to be build by all the critical waves basically. However, since the
It follows from the presented discussion that at the onset of ordering the equilibrium
critical waves are coupled via the non-linear excess free energy (1.2), the critical waves
both interfere and generate some new waves (so-called higher harmonics). As a result, only
some discrete sets of these waves survive, each stable set of the standing order parameter
If the symmetry of a morphology is that of a spatial lattice ℜ, Φ ( r ) , generally, is an
waves being corresponded to a crystal lattice.
infinite series in the Fourier harmonics corresponding to the set of the points of the lattice
ℜ-1 reciprocal to ℜ:22
Φ (r ) =
∑
qi ∈ℜ
−1
A ( qi ) exp i ( q i r + ϕi )
(1.13)
But close to the critical point τ = 0, α = 0 , where α is the coefficient appearing in the
cubic term of the Landau expansion (1.2), the coupling generating the higher harmonics is
5
small, so that one can keeps in the expansion (1.13) the main (or primary) harmonics only,
Φ ( r ) = AΨ ( r ) , Ψ ( r ) =
∑
exp i ( q i r + ϕ ( qi ) )
i.e. those whose wave numbers are equal to the critical value q* :
qi = q*
(1.14)
Then arising of the spatially periodic order parameter could proceed (in the mean field
approximation) as the 2nd order phase transition or the 1st order phase transition close to the
2nd order one, which situation is referred to as the weak crystallization.22,26,27
It was shown21,28,29 basing on the universal (Gaussian) conformational behavior of the
long polymer blocks that the Landau instability does occur in block copolymer systems.
E.g., for the binary incompressible AB copolymers the function c(q), which appears in (1.1)
g AA ( q ) + g BB ( q ) + 2 g AB ( q )
and according to (1.11a) is just the inverse scattering intensity, was shown21,28-30 to read
c (q) =
g AA ( q ) g BB ( q ) − ( g AB ( q ) )
2
− 2χ ,
(1.15)
where g ( q ) = gij ( q ) is so-called structure matrix (see section 3), which is determined by
the macromolecules’ architecture only. The plots c(q) for molten diblock copolymer which
were first calculated by Leibler,21 are presented n Fig. 2 for various χ (temperatures). It is
easy to see that the main prediction of the theory in the disordered state is that the only
change of the shape of the curve c ( q ) ~ I −1 ( q ) is its downward shift as the whole.
correspond to the values of χ = 2 ( n − 1) .
Figure 2. The curves c(q) labeled by numbers n
Figure 3. Inverse intensity for different temperatures.
The curves correspond to T of 240,220,200,180,160
and 100 oC (from top to bottom). (Fig. 5a of ref 31).
As is seen from Fig. 3, where the experimental data31 for I −1 ( q ) are presented, this
prediction holds very well for T >180 oC, i.e. until systems stays in the disordered state. It is
worth to notice that the way the experimental data are presented in Fig. 3 is the most natural
from theoretical point of view (it provides a direct opportunity to identify the region of the
disordered state as the whole) but extremely rare in the experimental literature (I hardly can
present any other example).
6
Moreover, it was shown by Leibler21 via a strict microscopic consideration that the
Landau toy model perfectly describes the phase behavior of not too asymmetric molten
diblock copolymers AnBm. His seminal theory of microphase separation in diblock
copolymers20 became a real paradigm for both building the phase diagrams of the ordered
phases’ stability within the WSA given the phenomenological coefficients in the Landau
expansion of the free energy of the weak segregating systems in powers of an order
parameter and microscopic calculation of the coefficients for block copolymers with a given
architecture. In particular, Leibler21 found that the thermodynamically stable ordered
morphologies for diblock copolymers are the body-centered cubic lattice (BCC) and the
structures possessing hexagonal (HEX) and lamellar (L) symmetries, the sequence of the
transitions being the disordered phase (DIS) – BCC- HEX – L (see Fig. 2). Further we refer
to the phases BCC, HEX and L as the conventional or classic ones. The phase diagrams
with the same topology were found for a variety of molten AB block copolymers with
different architectures32-36 (see Fig. 4).
Figure 4. The conventional phase diagrams of the molten diblock copolymers An Bm 21 (left) and star
block copolymers ( An ) ( Bm ) 35,36 (right). The phase transition line DIS-BCC, BCC-HEX and HEX-L
4
4
are plotted by the blue, green and red colors, respectively.
For some special sets of parameters, all the phase transitions lines (DIS-BCC, BCCHEX and HEX-L) merge at the critical point where the 2nd order phase transition from the
disordered to lamellar phase occurs. For AnBm diblock copolymers the critical point
corresponds to the symmetric diblock copolymer in case the repeated units of both blocks
have the same excluded volumes v and Kuhn lengths a. Later Fredrickson and Helfand37
incorporated the Brazovskii fluctuation corrections into the Leibler theory and showed that
7
they considerably shift the ODT towards higher temperatures as compared to the mean field
WST.
It is worth to notice that according to the WSA the ODT in block copolymers shares a
common physical background (the Landau weak crystallization) with various physical
phenomena like the blue phases appearance in liquid crystals,38,39 charge-density waves
generation upon addition of an ionic solute to a solvent in its critical region40,41 and
microphase separation in weakly charged polyelectrolyte solutions,42,43 the polymer specific
features of this approach being appeared at the stage of microscopic calculation of the
Landau expansion coefficients only. The WSA provides also description of a rather special
type of ordering predicted within the WSA also for random copolymers.44-51 Thus, the weak
segregation theory of microphase separation in block copolymer systems provides a unique
opportunity to test the general phenomenological concepts of the statistical theory of solidliquid transition via a rigorous microscopic consideration.
Unfortunately, the region of the WSA applicability corresponds to a rather narrow
vicinity of the critical point. Besides, the WST employs the so-called first harmonics
approximation we discuss in more detail below and it is often believed52 that within this
approximation “the predictions about ordered structures are limited to classical phases of
lamellar, hexagonal, and body-centered cubic structures, and consequently the possibility of
other structures such as bicontinuous structures, e.g., double gyroid, is excluded.” The
double gyroid (G) phase mentioned here is an important phase characterized by Ia 3d
space group symmetry, which was first discovered in lipid-water and surfactant systems53,54
and has been attracted much interest during the last decade due to bi-continuous
morphology characteristic of this phase. So, during the last decade the so-called self
consistent field theory (SCFT) by Matsen,55-57 which is free of these shortcomings, became
dominant in understanding the behavior of the ordering block copolymer systems.
The SCFT, which is considerably more polymer-specific than the WST, is based on the
Edwards58-Helfand7-10 idea that inhomogeneities in the density profiles of the chemically
different polymer repeating units are caused by some effective (self-consistent) fields,
which, in turn, are themselves determined by the arising density profiles. The new powerful
trick elaborated by Matsen and Schick55,56 was to seek for the desired density profiles and
the self-consistent fields as some series in the eigenfunctions of the corresponding diffusion
equation with due regard for their space symmetry. In contrast to the WST, which involves
only few (1 to 12, depending on the lattice symmetry) primary harmonics, the SCFT series
involve many hundreds of the eigen functions, and, thus, provide a much broader region of
the SCFT applicability. It enabled Matsen and Schick55,56 to success in building the SCFT
phase diagrams of molten AB diblock and star copolymers (see Fig. 3) revealing the stable
8
G phase in a reasonable agreement with experiment.59,60 A specific feature of these phase
diagrams the plane
( f , χ ) is existence of two triple points
f A = f1t < f crit , T = T1t and
f A = f 2t > f crit , T = T2t , where three phases HEX, G and L coexist. (As usual, fA =n/N is
the composition of the A monomers, n,m and N=n+m are the total numbers of the A and B
units per block copolymer chain and the total degree of polymerization of the chain.)
Therewith, the conventional sequence DIS-BCC-HEX-L and non-conventional one DISBCC-HEX-G-L hold for compositions within and out of the interval ( f1t , f 2t ) , respectively.
The SCFT has been successfully applied to describe the stable ordered phases in various
block copolymer systems (see, e.g., refs 61-63 and the references in other Chapters of the
book) and it became dominant in understanding the behavior of the ordering block
copolymer systems. Accordingly, the WST has been considered for some time as a sort of
old-fashioned and outdated technique.
In fact, however, the areas of expertise of the SCFT and WST are rather complementary
than overlapping. In particular, it is worth to notice that the experimental59 and SCFT55
( f , χ)
phase diagrams are in a qualitative agreement only and there is a notable upward shift of the
experimental phase transition lines in the plane
as compared to the SCFT ones. The
shift is due to the Brazovskii-Fredrickson-Helfand (BFH) fluctuation corrections25-27,37
neglected within the SCFT, which are far from being minor. However, these corrections are
easily incorporated into the WST37,64-71 and it is within the WST that this upward shift of the
phase transition lines was explained and quantitatively described3.
The application of the WST requires calculation of cumbersome expressions for the socalled higher structural correlators; still, the SCFT is no less technically involved, whereas
the corresponding numerical calculations are much more time consuming than those needed
for the WST (the same is valid for other numerical methods discussed in the book). Besides,
unlike the SCFT, a considerable part of the calculations necessary to build the phase
diagrams within the WST can be done analytically. All these advantages enabled the WST
analysis73-76 of the ternary ABC block copolymers, which resulted in understanding of many
peculiar properties of these systems. In particular, it is easy to incorporate into the
framework of the WST the effects of non-Flory interactions, which resulted in development
of the WST theories of the ODT in reversibly associating52,77-80 and compressible81-83 block
3
Besides, the WST with due regard for so-called fluctuation caused q* -renormalization67 provided
explanation of a shift of the scattering peak location towards lower values of q with temperature decrease,
which is noticeable in Fig. 3 and was first reported in ref 72.
9
copolymers as well as the WST analysis of the ODT phase diagrams with the low critical
ordering temperature.84,85
At last, it was demonstrated via a general WS analysis26,76,86 that under certain conditions
the most stable phases around the critical point are not necessarily the classic ones. Instead,
some others cubic phases such as gyroid G, simple cubic (SC), face-centered cubic (FCC),
so-called BCC2 also called single or alternated gyroid, (we refer to all the phases but BCC,
HEX, and L as the nonconventional ones) are stable, whereas the classic phases are
metastable only. Moreover, the phase diagrams of the ternary ABC block copolymers (both
linear and miktoarm) build via a generalization76 of the Leibler WST were found to reveal
all the aforementioned non-conventional stable phases. The list of the non-conventional
phases shown to be stable within the WST (we describe the phases in more detail in the
Appendix) is to be appended by a new phase belonging to the symmetry class Fddd (O70)
and called orthorhombic. The phase, which was found by Bates et al.87-90 in the ternary
linear ABC block copolymers, is strongly degenerate to allow weak segregation as
discussed by Morse et al.91,92 Remarkably, the phase was shown to be stable within the
WST even in a vicinity of the critical point92 (see Fig. 5).
Figure 5. The phase diagram of the molten diblock copolymers92 including the Fddd phase. Full lines are
the phase boundaries calculated from the WST. Points denoted by + are the lamellar-Fddd boundary
points and ◊ are the Fddd-hexagonal boundary points calculated using the numerical SCFT method. (Fig.
3 of ref 92).
One more recent achievement of the WST is discovery of the so-called “structure-instructure” morphologies or two-scale microphase separation. It was found experimentally
by ten Brinke, Ikkala et al.93 who studied self-assembling supramolecular structures in
poly(4-vinylpyridine)-block-polystyrene (P4VPb-PS) diblock copolymer with side chains
(e.g., pentadecylphenol, PDP) attached by hydrogen bonds to the P4VP block. The idea of
the resulting morphologies is given by the cartoon and TEM micrograph presented in Fig. 6
10
and Fig. 7. Another (and quite different) example of 2-scale morphologies was found by
Goldacker,94 Abetz and Stadler in blends of polystyrene-b–polybutadiene-b–poly(tert-butyl
methacrylate S33B34T33 (total M = 1.6 ×105) and polystyrene-b–poly(tert-butyl methacrylate
S47T53 (total M = 1.03 ×105), which contains 10 wt % triblock copolymer (see Fig. 8).
Figure 7. Transmission electron micrograph of PS-blockP4VP(NPD)1.0, where nominally one nonadecylphenol
(NPD) has been hydrogen bonded with each pyridine group.
The long period of the alternating PS (light grey) and
P4VP(NPD)1.0 (dark grey) lamellae equals LD~ 550 Å. The
number-averaged molecular masses of the P4VP and PS
blocks were 49,500 and 238,000 daltons, respectively. The
P4VP-(NPD)1.0 lamellae are further ordered into alternating
lamellae of nonpolar nonadecyl tails of NPD molecules and
polar P4VP backbones. The long period of this structure is
LC ~ 40 Å. The two sets of lamellar structures are, as
expected, mutually perpendicular. (Fig. 4 of ref 93).
Figure 6. Schematic illustration of the self-organized
structures of PS-block-P4VP(MSA)1.0(PDP)1.0. The local
structures are indicated; macroscopically, the samples are isotropic. (A) Alternating PS layers and layers
consisting of alternating onedimensional slabs of P4VP(MSA)1.0 and PDP for T< TODT (14). (B)
Alternating two-dimensional PS and disordered P4VP(MSA)1.0(PDP)1.0 lamellae for TODT < T < TOOT. (C)
One-dimensional disordered P4VP(MSA)1.0(PDP)x (with x<<1) cylinders within the three-dimensional
PS-PDP medium for T > TOOT. (Fig. 3 of ref 93).
Figure 8. Transmission electron micrograph of the mixture of
copolymers S33B34T33 and S47T53; dark-grey and light-grey bands
correspond to ST lamellas, and black bands refer to B layers. (Fig. 1 of
ref 97).
11
Despite an apparent visual difference of the structures shown in Fig. 6 and Fig. 7, it was
shown95-97 that within the WST both systems possess a common feature. Namely, for these
systems the function c(q) appearing in (1.1), (1.11a) and (1.15) is rather sensitive to details
*
*
and q = qmax
,
of the block copolymer architecture and could have two minima4 at q = qmin
*
*
and Lmin = 2π qmax
being corresponded to the bigger and
the lengths Lmax = 2π qmin
smaller characteristic scales of the morphologies (see Fig. 9).
Figure 9. The q-dependence of the inverse scattering intensity for the poly (A)m-block-poly(A-graft-B)n
polymer shown in Figs. 6,7 and analyzed within the RPA in ref 95. Here the curves labeled I, II, III, IV
and V correspond to (n,m) = (21, 3), (20, 5.5), (20, 5.69), (20, 6) and (1, 44), respectively (n is the
number of the side chains per the comb-like block, m = N A d , d and N A being the degrees of
polymerization of the side chains and homopolymer A block, respectively. (Fig. 3 of ref 95).
h
h
A similar behavior was found also for some block copolymer solutions,98,99 which
stimulates to seek some properly designed block copolymer solutions capable to serve as
photonic crystals. The WST phase diagrams of the two-length-scale morphologies were
build100-103 and supported by the SCFT calculations104-107 of the phase diagrams in a broader
temperature interval. Such a two-scale behavior is closely related to formation of the nonconventional morphologies,76,100-103 and expected to be, along with using the multicomponent block copolymers, a new rather efficient route towards tuning the phase
behavior of the self-assembling block copolymer nanostructures.
It is to provide a better understanding of the latest advancements in the WST, which is
the purpose of the rest of the Chapter. Since the basic features of the WST in the binary AB
block copolymers are well described in the original papers as well as reviews,1-3 in what
follows we skip the derivations as well as discussion of the BFH fluctuation corrections and
4
Note that Landau himself has strongly emphasized21 that “it is absolutely improbable” that the
function c(q) vanishes simultaneously in more than one minimum and, thus, the disordered state
becomes unstable simultaneously for the order parameter waves with different wave lengths. More
precisely, the two-scale instability corresponds to a strongly degenerate situation. But the wealth of
the parameters controlling the block copolymer fluctuation behavior makes it possible to realize
such a degenerate and “improbable” situation in some properly designed systems.
12
focus only on the latest results, which are obtained in the mean field approximation. The
subsequent presentation is organized as follows. I consider a simple weak segregation
model enabling us to understand the physical and mathematical bases, which cause the nonconventional phases’ stability in section 2. Here I introduce also the 2nd shell harmonics
approximation, which gives an example of important distinction between the strongly and
weakly fluctuating fields. A generalization of the Leibler WST to the multi-component
block copolymer systems, which turns out thanks to the broader application of the
distinction between the strongly and weakly fluctuating fields, is outlined in section 3.
Some applications of the generalized WST to the ternary ABC block copolymers are
described in section 4. In Conclusion I summarize my feeling of the current state and most
urgent problems of the WST. At last, in the Appendix the most typical conventional and
non-conventional weakly segregated ordered morphologies are described in detail.
2. The WST and the non-conventional phases’ stability.
Non-locality (forth vertex angle dependence) effects. The stability of the nonconventional morphologies in weakly segregated systems is determined by the degree and
character of non-locality of their free energy as a functional of a specified profile of the
corresponding order parameter (we refer to such functional as the virtual free energy).
Indeed, let us start with a generalization of the original Landau Hamiltonian (1.1), which is
to take into account that for polymer systems the specific excess free energy (1.2) is not
(
) ∑ n1! ∫ Γn ( r1 − r,..rn − r ) ∏ Φ ( ri ) d ri
n =3,4
i =1
local. Namely, it takes the form
ΔF r, {Φ ( r )} =
n
(2.1)
where some continuous functions Γ3 ( R1 , R 2 , R 3 ) , Γ 4 ( R1 , R 2 , R 3 , R 4 ) describe how
{Φ ( r )} in a vicinity of the point r rather than on the local value of Φ(r) at the very point r
much the specific excess free energy at the point r depends on the whole profile of the
Γ3 ( R1, R 2 , R 3 ) = α δ ( R1 ) δ ( R 2 ) δ ( R 3 ) ,
only. Obviously, the expressions (1.2) and (2.1) coincide in the limit
Γ 4 ( R1, R 2 , R 3 , R 4 ) = β δ ( R1 ) δ ( R 2 ) δ ( R 3 ) δ ( R 4 ) .
The meaning of the non-locality becomes clearer when rewriting the free energy (1.1), (2.1)
in the Fourier-representation:
13
F =∫
Δ Fn =
c ( q ) Φq dq
2
( 2π )
3
2
+ ΔF3 + ΔF4 ,
(2.2)
n Φ q dq
⎛ n
⎞
( i) i .
1
δ
Γ
q
q
,
,
q
(
..
)
⎜
⎟
∑
∏
i
n
n
1
∫
⎜
⎟
n ! ⎝ i =1 ⎠
( 2 π )3
i =1
(2.3)
Hereafter, we refer to functions and their Fourier transforms as the same functions in rand q-representations, respectively, and distinguish them only by the choice of the letters
used to denote their arguments; this convention is not expected to cause any
misunderstandings due to the context. It is worth to notice that the functions Γn appearing in
the cubic and quadric free energy contributions (2.1a) depend on the structure of the system.
Substituting characteristic of the WST expression (1.14) for the equilibrium order
parameter into the mean field free energy (1.12) with due regard for expressions (1.1),
(
) (α
)
(2.1)-(2.3) and minimizing the final expression with respect to the amplitude A we get20
ΔF
fℜ ( τ ) = ℜ =
VT
2
3 α ℜ + 9α ℜ
− 32τβ ℜ
3
212 β3ℜ
ℜ
(
2
− 9αℜ
− 32τβℜ
)
Here we introduced the cubic and quadric vertices
αℜ =
k −3 / 2
3!
∑
q1 + q 2 + q 3 = 0,
q1 = q 2 = q 3 = q*
ℜ
( 3)
Cℜ = ∑ 3 cos Ω j k 3 / 2
βℜ =
=
k2
4!
q1 +q 2 +q3 +q 4 =0,
q1 = q 2 = q3 = q 4 = q*
λ0 ( 0 )
4k
∑
+
(2.4)
Γ 3 ( q1 , q 2 , q 3 ) exp i ∑ i =1 ϕi = γ (1) Cℜ ,
3
(
Γ 4 ( q1 , q 2 , q3 , q 4 ) exp i ∑ i =1 ϕi
k ∑ λ 0 (hi ) + ∑
ℜ
λ
4
2
( h1, h2 ) cos Ω j
4
(2.5)
)
( 4)
k
(2.6)
.
In (2.5), (2.6) the phases Ω (j3) , Ω (j4 ) are the algebraic sums of the phase shifts ϕ appearing in (1.14)
for triplets and non-coplanar quartets of the vectors {q i } involved in the definitions of α and β,
respectively. Besides, we used the designations and parameters of Leibler:20
γ (h) = Γ3 (q1 , q 2 , q 3 ), q12 = q 22 = q*2 , q32 = (q1 + q 2 )2 = hq*2 ,
λ(h1 , h2 ) = Γ4 (q1 , q 2 , q 3 , q 4 ),
(2.7)
q i = q* , i = 1,..4
(2.8)
14
h1 = ( q1 + q2 )
q∗2 , h2 = ( q1 + q3 )
2
2
q∗2 , h3 = 4 − h1 − h2 = ( q1 + q 4 )
λ 0 (h) = λ ( 0, h ) = Γ 4 (q, −q, p, −p), h = ( q + p )
The symbol
∑
ℜ
n
2
2
q∗2 .
(2.9)
q*2 .
(2.10)
implies summation over all sets of n vectors for given morphology ℜ
appearing in (2.5), (2.6). The first summation in (2.6) is over all pairs of non-collinear
vectors q i and q j , 2k is the number of the vectors in the reciprocal space belonging to the
The phase transition lines τℜ1 / ℜ2 ( γ (1) ) between the morphologies ℜ1 and ℜ2 are
coordination sphere with the radius q*.
(
)
(
)
fℜ1 τ, ( γ (1) ) = fℜ2 τ, ( γ (1) ) .
determined by the equation
(2.11)
In particular, if the cubic vertex vanishes due to the symmetry for both morphologies ℜ1 and
ℜ2 the phase transition line between the morphologies is determined by the equation
βℜ1 = βℜ2
(2.11a)
function λ (h1 , h2 ) , which appears in the expression (2.6) for the fourth vertex βℜ. Following
The topology of thus obtained phase diagrams is influenced by the explicit form of the
Leibler20 we refer to the h-dependence of the vertex λ as the angle one since the values of
the parameters hi depend on the angles between the vectors qi.
For diblock copolymers the angle dependence is rather weak,20 which enabled
Fredrickson and Helfand34 to propose the following commonly accepted approximation:
λ ( h1 , h2 ) ≈ λ ( 0, 0 ) = λ 0 ( 0 ) .
(2.12)
However, the approximation (2.12) was shown71 not to stay true for any polymer systems.
crystallizing systems we approximate the function λ (h1 , h2 ) as follows:71,81
To estimate effects of this angle dependence on the phase diagram of the weakly
3
3
⎛ 3δ ⎛
⎞⎞
λ(h1 , h2 , h3 ) = ∑ f (hi ) = λ 0 ⎜⎜1 − ⎜ 4 2 − ∑ hi2 ⎟ ⎟⎟.
i =1
i =1
⎠⎠
⎝ 32 ⎝
(2.13)
λ (h1 , h2 ) in powers of hi, the positive (negative) sign of δ corresponds to a disadvantage
The approximation (2.13) keeps only the first non-constant term in the expansion of
The resulting phase diagrams in the plane (τ, γ (1)) are sets of parabolas
(advantage) of the lamellar structure as compared to all other ones.
τℜ1 / ℜ2 = 9τℜ1 / ℜ2 γ 2 (1) ( 32λ 0 ) ,
(2.14)
15
where the reduced temperatures τℜ1 / ℜ2 depend only on the angle dependence strength δ.
(
)
The reduced phase diagram in the plane τℜ1 / ℜ2 , δ demonstrated71,81 that increase of the
strength of the model angle dependence (2.13) results in increase of stability of the G phase
(as compared to HEX and L) and that of the various non-conventional cubic phases as
compared to L. Therewith, all competing phases are to be taken into account. As shown in
Fig 10, the stability of the newly discovered orthorhombic phase Fddd (see refs 81-86 and
Appendix) turns out to reveal a non-monotonous behavior with increase of δ, the region of
the Fddd phase stability being turned out as big as that of the G phase.
a)
b)
τ eff
Figure 10. The reduced phase diagram for the model angle dependence (2.13) in the plane (the reduced
temperature
- the angle dependence strength δ). a) the phase diagram76 calculated without taking
into account that the orthorhombic phase Fddd could exist; b) the phase diagram calculated with due
regard for the phase Fddd.
It was claimed,71 basing on the reduced phase diagram shown in Fig. 10a, that the
conventional phase transition sequence DIS-BCC-HEX-L occurs only for δ<δ0=0.362.
Now, after the Fddd phase is discovered, one can assert that this sequence never holds.
Instead, the following non-conventional sequences occur:
i) the sequence5 DIS-BCC-HEX-Fddd-L for δ<δ12 , δ12 = 4/9;
ii) the sequence DIS-BCC-HEX-Fddd-SG for δ1 > δ > δ12, δ1 = 0.61005;;
iii) the sequence DIS-BCC-HEX-G-Fddd-SG for δ2 > δ > δ1, δ2 = 0.61684;
ii) the sequence DIS-BCC-HEX-G-SG for δ23 > δ > δ2, δ23 = 2/3;
iii) the sequence DIS-BCC-HEX-G-FCC for δ3 > δ > δ23, δ3 = 0.822;
5
For δ=0 and for the angle dependence characterizing the Leibler molten diblock copolymers this
sequence was found first by Morse et al.86,87
16
iv) the sequence DIS-BCC-G-FCC for δ34 > δ > δ3, δ34 = 5/6;
v) the sequence DIS-BCC-G-SC for δ45 > δ > δ34, δ45 = 0.891;
vi) the sequence DIS-BCC-SC for δ2 > δ > δ45, δ2 = 4/3;
vii) the sequence DIS-BCC for δlim > δ > δ2, δlim = 1.538;
At last, βBCC vanishes at δ = δlim , which is characteristic of so-called tricritical point where
the 2nd order phase transition line terminates and the 1st order phase transition line starts. For
δ > δlim the original weak segregation approximation does not hold anymore and the next
(5th, 6th etc.) terms of the Landau expansion are to be taken into account.
The approximation (2.13) could seem to be too academic but it provides a good idea of
the phase behavior of some real systems. In particular, the reduced phase diagram of the
molten ternary linear ABC block copolymers with the non-selective middle block (see Fig.
11a) calculated by the author71 before the Fddd phase was discovered is rather similar to
that presented in Fig.9a, the composition of the middle block playing the role of the angle
dependence strength parameter δ. Thus, the region of the Fddd phase stability for the
ternary linear ABC block copolymers with the non-selective middle block is expected to
neighbor with that for the G and SG phases in qualitative agreement with experiment.84,85
Remarkably, the phase behavior of the molten ternary star (miktoarm) ABC block
copolymers with one of the arms non-selective with respect to two other ones differs
strongly both from that of molten diblock copolymers and the linear ABC block copolymers
middle block (see Fig. 11b).
a)
b)
Figure 11. The reduced phase diagrams76 along the critical line f A = f C = f B
2 calculated with due
regard for the actual angle dependence of the ternary ABC block copolymers in the plane (the nonselective block composition fB, the reduced temperature τ eff ) for linear (a) and miktoarm (b) ABC block
copolymers. The regions of the stability of the disordered state and body-centered cubic, hexagonal,
lamellar, double gyroid, single gyroid, face-centered cubic, simple cubic and BCC3 lattices are labeled by
the numbers 0, 1, 2, 3, 4, 5, 6, 7 and 8 respectively.
17
The physical origin of the non-conventional cubic phases’ stability for the ternary linear
ABC block copolymers with a long non-selective middle block is obvious: short strongly
incompatible side blocks prefer to aggregate into small micelles rather than into thin layers,
it is this tendency, which is “coded” into the angle dependence of the effective non-local
vertices Γ 4 (q1, q 2 , q3 , q 4 ) and becomes apparent already at the very onset of ordering
when segregation is still weak.
The 2nd shell harmonics approximation. Thus, the first harmonics approximation
(1.14) along with due regard for the angle dependence of the 4th vertex provides a
reasonable explanation of the stability of the non-conventional phases. However, during the
last decade some authors71,108-110 queried reliability of the approximation (1.14). In
particular, Hamley and Podnek71 suggested that the gyroid morphology existence is due to
anomalously large (by modulus) but negative contribution of the second harmonics with
h = q 2 q*2 = 4 / 3 characteristic for the G morphology to the total free energy.
Of course, this suggestion by itself is not sufficient to explain the G phase stability in
diblock copolymers since the second harmonics with h = 4 3 are characteristic also for the
BCC3 and G2 morphologies, which all belong to the G family. Moreover, it has no relation
to the problem of the G stability at the very critical point, which depends only on the
strength of the angle dependence of the forth vertex as shown above. However, if the angle
dependence is not strong enough to provide the G stability at the critical point, we are to
deal with two closely related problems: i) which phases are stable at the triple points (if
any) existing near the critical point; and ii) which factors determine location of the triple
points. In general, contribution of many harmonics (rather than that of the only second ones)
determines location of the triple point in question, which is shown via direct calculation by
Matsen and Schick.51,52 But it is natural to expect that only certain finite number of the
higher harmonics is really relevant if the triple points are close enough to the critical one.
To single out such selected higher harmonics, instead of the first harmonics approximation (1.14)
Φ (r ) = Ψ (r ) + ∑ ψh (r ) ,
we chose the trial function as follows:47,76
(2.15)
h≠1
where
Ψ (r ) = A
∑
exp i (q i r + ϕ i )
qi ∈ℜ −1 , qi2 = q*2
ψ h (r ) =
and
∑
qi ∈ℜ ,
−1
qi2 = hq*2
aqi exp i ((q i r ) + ϕ(q i ) )
(2.16)
(2.17)
18
are the sums of the main and higher harmonics, respectively. After substituting the trial
function (2.15)-(2.17) into (2.2), (2.3) the virtual free energy takes the form
F = Fmain + Fhigh + Fcoupling ,
where Fmain = F
({Ψ ( r )}) is the contributions of the dominant harmonics (2.16), ΔFcoupling is
(2.18)
that due to coupling between the higher dominant and harmonics, generated by the cubic
term of the original Hamiltonian:
( 3)
ΔFcoupling
VT
=
A02
∑ γ ( q1, q2 , −q1 − q2 ) aq1+q2 exp ( i ( φ1 + φ2 ) )
2 q ∈ℜ−1 , q2 =q 2
i
(2.19)
*
i
( q ( h −1) + τ) a
and the contribution of the higher harmonics is determined as follows:
Fhigh =
VT
2
qi∈ℜ
∑
−1
, qi2 = hq*2
2
4
*
qi
2
.
(2.20)
All other terms we skipped in (2.18) are irrelevant. Indeed, minimization of the free
energy (3.31) with respect to the complex amplitudes aq , q 2 q*2 = h ≠ 1 , gives
aq = −
where
((
A02 γ ( h ) ∑ q exp i φi + φ j
2
2
q*4 ( h − 1) + τ
))
(2.21)
Σ q means summation over all pairs of the main harmonics given the condition
qi + q j + q = 0
(2.22)
According to (2.21), (2.22) the second harmonics induced by coupling (2.19) belong to all
the coordination spheres of the corresponding reciprocal lattice, radius of which does not
exceed the doubled radius of the dominant coordination sphere, their amplitudes being of
the 2nd order of magnitude with respect to the main harmonics amplitudes A0. The number
of the coordination spheres of different radii satisfying eq (2.22) depends on the lattice
symmetry and varies from 1 (for the LAM) to 8 (for the G) and 10 (for the BCC3). We refer
∑ q exp ( i ( φi + φ j ) )
to all of the corresponding harmonics as the 2nd shell harmonics. For the lattices with nonzero phase shifts like SG, G and G2, the factor
appearing in (2.21)
and, thus, the corresponding higher harmonics vanish identically, which gives a natural
derivation of the extinction rules111 within the WS theory.
( )
Substituting (2.21) into eqs (2.18)-(2.20) results in the final expression for the free
(
)
energy including the higher harmonics contribution up to order of
ΔF = VT τ A02 + αℜ A03 + βℜ A04 ,
O A04 :
(2.23)
19
where βℜ is related via eqs (2.6), (2.8)-(2.10) to the forth vertex Γ 4 renormalized with due
regard for the 2nd shell higher harmonics:
Γ 4 (q1 , ..,q 4 ) = Γ 4 (q1 , ..,q 4 )
− ( B ( q1 , .q 2 ; q3 ,q 4 ) + B ( q1 , .q3 ; q 2 ,q 4 ) + B ( q1 , .q 4 ; q 2 ,q3 ) )
B ( q1 , .q 2 ; q3 ,q 4 ) =
γ2 ( h)
q*4 ( h − 1)
, p = −q1 − q 2 = q3 + q 4 , h =
2
τ is omitted in the definition (2.25) since τ
(2.24)
p2
(2.25)
q*2
q*4 is a condition of the WSA validity (see
Appendix 2)..
It is worth to remember now that the reduced phase diagrams build in Fig.9 and Fig.10
describe the stability of the weakly segregated phases in the very critical point only, i.e. in
the limit γ→0. It follows from eqs. (2.24), (2.25) that the phase transition lines are affected
by the actual dependence γ(h) in a finite vicinity of the critical point where the cubic vertex
is small but finite.
The advantages and limits of the 2nd shell harmonics approximation. The phase
diagram of the molten diblock copolymer calculated in ref 71 with taking into account the
2nd shell harmonics contribution into the 4th vertex of the virtual free energy as described
above is presented in Fig. 12. One can estimate both the advantages and deficiencies of our
approximation comparing it with those of Leibler21 and Matsen and Schick.55
a)
b)
Figure 12. The phase diagram of the diblock copolymer in the 2nd shell harmonics approximation (without
taking into account the Fddd phase existence). The designations of the phases are the same as in Figure
11. a) comparison with the Leibler21 phase diagram (shown by the dashed lines); b) extrapolation to the
region of comparatively high values of χ , which demonstrates possibility of the G2 lattice (labeled by the
number 9) stability.
20
As is seen in Fig. 12a, as far as the conventional phases is concerned our phase diagram
would almost coincide with that of Leibler21 precisely approaching the latter in the vicinity
of the critical point. The only difference would be some broadening of the BCC phase
stability region (basically at cost of the HEX phase) with increase of the diblock copolymer
asymmetry. The situation changes drastically as soon as we include into the list of
competing phases those of the G family, which were not taken into account in the original
paper.21 All three phases of the family described in section 2 become stable when the
asymmetry f − f c increases and the 2nd shell harmonics effect is taken into account. It is
worth to notice that when we calculate the free energy of the ordered phases within the
conventional 1st harmonics approximation the phases of the G family turn out to be
metastable only. Therewith, in our approximation the triple point LAM-HEX-G is located at
χ = 10.88 , which is rather close to the result f=0.452, ~
χ = 11.14 obtained by
f=0.462, ~
Matsen and Schick55 within the SCFT using much more harmonics. Comparing the
presented numerical results for the triple point we conclude that our 2nd shell harmonics
approximation somewhat overestimates the effect of the higher harmonics but, nevertheless,
is in a reasonable agreement with the SCFT results obtained using the whole series of the
higher harmonics.55
( χDIS-BCC ( f ) → −∞ ) when f → f
Another clear evidence of such an overestimation is that the phase transition line DISBCC3 falls down sharply
3
di
= 0.4183. At the first
sight this result seems meaningless, but, in fact, its physical meaning is rather clear. It could
be understood by analogy with that of the spinodal of block copolymers with respect to
microphase separation. The latter was defined21,28,29 as the line (surface) where the inverse
scattering factor appearing in the quadratic term of the free energy (2.2) vanishes.
Accordingly, the correlation function (1.11b) calculated within the RPA diverges here and,
account the fluctuation corrections25-27,37,64-71 shows that the function S ( r ) stays finite and
thus, the uniform state of the systems becomes absolutely unstable. However, taking into
the uniform state stays stable (at least metastable) even beyond the RPA spinodal. Thus, the
latter should be now understood as a crossover line between the regions with different
temperature scaling of the correlation radius and the exact border of the region where the
RPA does not hold even qualitatively.
Quite similarly, the sharp falling down of the phase transition line DIS-BCC3 when f →
fdi is shown76 to be determined by the fact that the minimal quadric vertex β ℜ = min β ℜ
changes the sign in the point f = f di due to the 2nd shell harmonics renormalization of the
21
vertex, therewith ℜ = BCC 3 for molten diblock copolymer. So, β BCC 3 ( f ) < 0 for f < fdi
and, therefore, the expansion of the Landau Hamiltonian in powers of the order parameter Φ
up to the 4th term only becomes inapplicable. As in the spinodal case, the unphysical
divergence of the leading term is to be removed by including into the expansion the terms
of the higher order than that causing the divergence. In our case it means to take into
account at least the terms of the 5th and 6th powers in Φ as well as the 3rd shell harmonics
contributions. The corresponding generalization of the WST is expected to smooth (not
eliminate!) the sharp phase transition line DIS-BCC3 (in general, DIS - ℜ ) shown in Figure
12. It is natural to refer to the line β ℜ = 0 as the WS border line since beyond it the higher
harmonics effect becomes so important that the system could not be described properly
even within the 2nd shell harmonics approximation of the WS theory.
Two more interesting features of the modified WS phase diagram shown in Figure 12 are
the phase transition lines G – BCC3 at f=0.4343 and G – G2 (see Fig. 12b) situated at
relatively high values of ~
χ . It is important to stress that the WS theory can not claim
responsibility for prediction of precise location of both these phase transition lines. Indeed,
they lay too far from the critical point so that the stability of the phases BCC3 and G2 could
be only an artifact of the WST extrapolation beyond its validity region. Nevertheless, these
phase transition lines are interesting as indications of the fact that the stability of the double
gyroid phase G is caused by a moderate development of the 2nd shell harmonics only,
whereas a further increase of the degree of segregation results in increase of stability of
other cubic phases (in our case, BCC3 and G2) at cost of the G phase.
The described features of the phase diagrams of molten diblock copolymers are
characteristic of binary AB block copolymers with various architectures as is exemplified
by the phase diagrams of molten symmetric triblock and trigraft AmBnAm copolymers we
calculated within the 2nd shell harmonics approximation (see Fig. 13).
a)
b)
22
Figure 13. The phase diagrams of the molten symmetric triblock (a) and trigraft (b) AmBnAm copolymers
in the 2nd shell harmonics approximation (without taking into account the Fddd phase existence). The
designations of the phases are the same as in Figure 11.
Summarizing, the WST in the 2nd shell harmonics approximation provides a rather
reasonable accuracy in locating the triple point HEX-G-LAM and interesting (much less
reliable, though) hints as to stability of some other non-conventional cubic phases.
3. The WST applications to the multi-component block copolymer systems.
Φ i (r ) = φ i (r ) − φi of the partial local volume fractions φi ( r ) = vρi ( r ) of the repeated units
The order parameter for these systems is the n-component vector of the local deviations
(monomers) of the i-th sort (here v is the excluded volume supposed to be the same for all
sorts of monomers) from their values φi averaged over the whole volume of the system. As
common for all polymer systems consisting of flexible macromolecules, their virtual free
({φi ( r )}) + F * ({φi ( r )}) .
energy takes the form112,113
F = Fstr
(3.1)
where the first term is so-called structural free energy Fstr corresponding to the entropy of
the inhomogeneous ideal system of the copolymer macromolecules under consideration
with certain specified spatial profiles of the volume fractions of the repeated units of the i-th
WS theory the free energy (3.1) is evaluated by its expansion in powers of Φ i ( r ) :
sort and the second is the interaction contribution we discuss in more detail later. Within the
(
)
F {φi ( r )} = F
({φi }) + ΔF2 + ΔF3 + ΔF4 + ... .
(3.2)
with the contributions ΔFn defined as follows:
ΔF2 1 Γij ( q ) Φ i ( q ) Φ j ( −q ) dq
,
= ∫
3
2
T
( 2π )
(3.3)
n Φ (q ) dq
ΔFn 1 ⎛ n
⎞ )
αi
i
i
= ∫ δ⎜ ∑ q i ⎟ Γα(n1 ,..,
, n = 3,4
α n (q 1 ,.., q n ) ∏
3
T
n! ⎝ i =1 ⎠
(2π)
i =1
(3.4)
In the disordered state the thermodynamically average values of the fluctuations Φ i ( r )
n
Φ i ( 0 ) Φ j ( r ) exp ( − ΔF ({Φ l ( r )} ) T ) ∏ l =1 δΦ l ( r )
∫
Sij ( r ) = Φi ( 0 ) Φ j ( r ) =
n
exp ( − ΔF ({Φ l ( r )} ) T ) ∏ l =1 δΦ l ( r )
are zero. Then the only observable quantities are the correlation functions
(3.5)
and their Fourier transforms (scattering factors) Gij ( q ) = ∫ Sij ( r ) exp ( iqr ) dr.
23
Random Phase Approximation. Let us assume that the fluctuations are small and keep in
the free energy expansion (3.2) the quadratic contribution (3.3) only. Then, as first shown
by the author114 (see also refs 115-117), the matrix G = Gij ( q ) can be expressed in terms
of two independent matrices characterizing the connectivity and interactions effects:
G −1 = Γ = g −1 − c ,
(3.7)
where the matrices g-1 and -c are contributions to the matrix Γ appearing in (3.3) from the
structural and energetic addendums, respectively (see eq (3.1)), matrices G-1 and g-1 are
inverse to the matrices G and g, respectively, and the so-called structural matrix g is
defined114-116 as follows:
(S )
g = gij ( q ) , gij ( q ) = ∑ nS γ ij ( q ) .
(3.8)
S
In the second of the definitions (3.8) nS is the number density of the (macro)molecules with
structure S, summing up is carried out over all the structures S present in the system
(S )
including the monomers, and the molecular form-factors γ ij ( q ) read
(
( )) S ,
(S )
γ ij ( q ) = ∑ exp iq r ( li ) − r n j
(3.9)
where r(li) is the vector-radius of the l-th repeated unit of the sort j and the symbol <...>s
implies averaging over all Gaussian conformations of the macromolecule S. For monomers,
(1)
obviously, γ ij ( q ) = δij .
The matrix g is just the matrix of the correlation functions for the ideal polymer system
i.e. the system with the same structure as that under study but with no interactions between
(
cij ( r ) = −δ2 F * {φl ( x )}
) ( δρi ( 0 ) δρ j ( r ) )
their repeated units, whereas
(3.10)
is the matrix of the direct correlation functions, which is well known in the theory of simple
liquids.118
The uniform (disordered) phase stays thermodynamically stable (at least, metastable)
with respect to micro- or macrophase separation when the quadratic term (3.3) is positive
definite:29
min Λ ( q ) = Λ ( q* ) > 0
(3.11)
where Λ(q) is the minimal of the eigenvalues λi ( q ) , i = 1,.., n of the matrix Γ of the rank n
and the wave number q* of the critical order parameter waves is the location of the absolute
minimum of the function Λ(q). Therewith
24
q*2 > 0
(3.12)
is the condition that it is micro- rather than macrophase separation, which occurs after the
uniform state becomes unstable. Accordingly, the spinodal line (surface), which delineates
the region in the space of the structural and interaction parameters of the system under study
where the spatially uniform (disordered) state is absolutely unstable within the RPA, reads29
min Λ ( q ) = Λ ( q* ) = 0,
(3.13)
The interaction term F* is naturally determined assuming that it does not depend on the
polymer structure of the system. (This natural requirement is, in fact, rather subtle. Strictly
speaking, it is correct only when the Lifshitz number Li = v a3 , which plays the role of the
Ginzburg parameter, is small (here a is the Kuhn length). If Li is not small then the effective
monomer-monomer interaction is strongly influenced by the correlation between the
neighboring (along the chain) monomers and renormalized accordingly. It is this
phenomenon for which Khokhlov119,120 coined the term “quasimonomers” and which is
quantitatively addressed by the PRISM theory.121,122 Henceforth we believe that the
interaction term F* describes the properly renormalized interaction.) In the simplest case of
the compressible Flory-Huggins lattice model (when some cells of the lattice are not
(
F * {φi ( r )}
)=
occupied by any monomer repeating units) the interaction term F* reads
T
⎞
dr ⎛
1
⎜
⎟,
−
φ
−
φ
+
χ
φ
φ
r
r
r
r
1
ln
1
(
)
(
)
(
)
(
)
(
)
(
)
∑
ij
i
j
∫v⎜
⎟
2 i≠ j
⎝
⎠
(3.14)
where φ ( r ) = ∑ i =1 φi ( r ) is the total volume fraction of all sorts of monomers. Then the
l
respectively. In the incompressibility limit 1 − φ ( r ) → 0 the first addendum in (3.14)
addendums in (3.14) describe the compressibility and van der Waals interaction effects,
(
)
F * {φi ( r )} = (T 2 ) ∫ ∑ χij φi ( r ) φ j ( r ) dr v .
vanishes and we get the well known quadratic expression62,117
(3.15)
i≠ j
Φ i ( r ) are determined by the structural entropic term only. But in general (beyond the
Thus, in this limit the higher vertices in the virtual free energy (3.1) expansion in powers of
incompressible Flory-Huggins model) the enthalpic (interaction) contribution into the free
energy could affects the cubic and quadric terms either as found for molten diblock
copolymers taking into account the “quasimonomer” renormalization123,124 and using the
equation of state model.81-83 It is worth to notice that the characteristic scales of the
structural and interaction contributions into the higher vertices are rather different (of the
order of the macromolecule and monomer size, respectively). Thus, the main source of the
25
strong angle dependence (if any) of the 4th vertex, which may lead to stability of the nonconventional phases as discussed in section 2, is expected to be the structural contribution
determined by so-called higher structural correlators as shown in detail in refs 21,47,76,117.
The critical points in the multi-component block copolymer systems and the strongly
and weakly fluctuating fields. Violation of the (meta)stability condition (3.11) is sufficient
but not necessary to guarantee crystal ordering. Typically, a finite order parameter profile
(1.13) arises via discrete 1st order phase transition when the condition (3.11) still holds and
the disordered phase is at least metastable. However, the span of this profile decreases when
the relative magnitude of the cubic terms decreases and, finally, the ordering transforms into
a continuous 2nd order phase transition (within the mean field approximation only!) at the
critical point where the cubic terms vanish. Thus, the WST certainly holds (at least in the
same sense as the SCFT does) in a vicinity of the critical point(s).
It is easy to locate the critical point(s) (if any) for the scalar WST describing the ODT
in the incompressible binary block copolymers, where only one cubic term exists (see eq
(1.2)). But the virtual free energy of the n-component block copolymer systems contains
~n3 cubic terms (that of incompressible does (n-1)3 ones), so that it could appear62 that it is
hardly possible at all to apply the WST to many-component systems. Nevertheless,47,76 it is
possible to reduce consideration of the n-component systems to that presented in section 2
for those with a scalar order parameter via distinguishing the minimal Λ(q) and all other
(positive) eigenvalues λ i (q ) (i=2,…n) of the matrix Γ appearing in (3.3) as well as the
corresponding eigenvectors Ei (q ) and ei( s ) (q ) . The projections of the vector order parameter
Ψ (q ) = E i (q ) Φ i (q ), ϕ s (q ) = ei(s ) (q ) Φ i (q ), s = 2,.., n.
(3.16)
into the corresponding eigenvectors play the role of the strongly and weakly fluctuating fields,
respectively, the account of the latter being carried out similarly to treating of the 2nd shell
harmonics in section 2. As the result, one gets the effective scalar free energy of the weakly
segregated multi-component block copolymer systems in the form (2.23), where the effective
cubic vertex αℜ is defined by expression (2.5) with
( 3)
Γ3 ( q1 , ..,q3 ) = Γ α ,α
1
2 ,α 3
(q1 , ..,q3 )Eα1 ( q1 ) Eα 2 ( q 2 ) Eα3 ( q3 ) .
(3.17)
All other cubic terms either renormalize the effective the quadric vertex βℜ described by a
cumbersome expression71 we skip here for brevity or contribute to the terms of the order of
magnitude Ο(Ψ5) and thus exceed the accuracy of the WST.
26
4. The WST predicted peculiarities in the multi-component block copolymer systems.
The RPA structure factors in the disordered state. To calculate the structure factor
matrix G for the Flory-Huggins model we find the matrix c for finite compressibility from
(
eqs (3.10), (3.14):
cij ( q ) = v (1 − φ )
−1
)
+ χij 2 ,
(4.1)
and, finally, take the incompressibility limit 1 − φ ( r ) → 0 . In this (or equivalent) way the
substitute (4.1) into the r.h.s. of the general RPA equation (3.7), invert the resulting matrix
structure factor matrices G(q) were found and the spinodal conditions analyzed73-76,97-99 for
some ternary ABC block copolymer systems. Werner and Fredrickson74 studied the
spinodal conditions (3.11), (3.12) for molten linear and comb-like ABC (monodisperse and
statistical) block copolymers and found the spinodal lines as well as the q* dependence on
the values of three independent Flory parameters χ AB , χ AC , χ BC . They found that increase
of one of these χ-parameters (given two others are fixed) could result in a non-monotonous
ordering tendency called the reentrant ODT. A similar result was found by the author et al.73
Namely, the spinodal ODT temperature Ts in the linear ABC block copolymers, in which
one of blocks is much shorter and more incompatible that two others, changes nonmonotonously with increase of the short block incompatibility, the minimal value of Ts and
the period of the arising ordered structure being less than those for the corresponding
diblock copolymer without any third strongly incompatible block. Thus, the peculiar
situation that “more incompatibility results in less segregation” is explained by an
additional entropic loss related to confinement of the ABC macromolecules in a lattice with
the smaller periodicity.
Two-length-scale behavior in molten ABC. Cochran, Morse and Bates75 considered
in more detail the scattering behavior of the linear ABC triblock copolymer melts and found
that tuning the values of the architecture, scattering contrast and interaction parameters one
achieves a reasonable agreement between the two-peak profile of the SAXS indicatrix
observed in the disordered poly(isoprene-b-styrene-b-dimethylsiloxane) (ISD)125 and the
RPA structure factor, the height of one of the peaks being increased when approaching the
ODT (see Fig. 13). The authors concluded that “the RPA structure factor is representative
of the true structure in disordered ABCs” and attributed such a two-peak profile “to the
natural existence of multiple length scales in ABCs”.
27
Figure 14. Computed RPA scattering of a model
compositionally symmetric, N=100 ISD block
copolymer, at six temperatures approaching the
spinodal limit for three different choices of the
scattering contrast vector c. The temperatures shown in
each instance are 1.0324 Ts , 1.0171 Ts , 1.0121 Ts ,
1.0072 Ts , 1.0020 Ts , and 1.0002 Ts , where Ts =
302.4 K. Temperature decreases with increasing peak
intensity. (Fig. 1 of Ref 75).
An additional insight into the nature of such an unusual behavior is provided by the
RPA analysis76 for symmetric An BmCn block copolymers with non-selective middle block
( χ AB = χ BC = χ ). For such a symmetric ABC the RPA free energy (3.3) reads
ΔF2 =
1
4N
∫ ( 2π )3 ( λ + ( Φ + ( r ) )
d 3q
2
)
+ λ− (Φ− (r )) ,
2
λ + = b ( q ) + χ AC − 4χ, Φ + ( r ) = Φ A ( r ) + Φ C ( r ) , .
λ − = a ( q ) − χ AC , Φ − ( r ) = Φ A ( r ) − Φ C ( r )
where χ = χN , χ AC = χ AC N ,
(4.2)
N = m + 2n and the functions a(q) and b(q) defined in ref
76 both for linear and miktoarm ABC depend on the reduced squared wave number
Q = q 2 a 2 N 6 only.
28
It turns out that both the character of the weakly segregated morphology occurring in
such a symmetric system and the very possibility to describe it within the WS theory
χ AC , ~
χ ) is
depends crucially on the values of the interaction parameters. Namely, the plane (~
~ ,χ
~ ) = 0 and λ (~
~
divided by the lines λ − (χ
AC
+ χ AC , χ ) = 0 into i) the stability region
( λ − > 0, λ + > 0 ), where the fluctuations of both order parameters Φ + (r ), Φ − (r ) are finite
and the uniform state is stable (or at least metastable) with respect to these fluctuations; ii)
the AC-modulation region ( λ − < 0, λ + > 0 ), where the uniform (disordered) state is
unstable with respect to formation of certain profile Φ − (r ) ≠ 0 , the order parameter Φ + (r )
iii) the B-modulation region ( λ + < 0, λ − > 0 ), where the
uniform state is unstable with respect to formation of certain profile Φ + (r ) = −Φ B (r ) ≠ 0 ,
being weakly fluctuating;
the order parameter Φ − (r ) being weakly fluctuating; and iv) the region ( λ − < 0, λ + < 0 ),
Φ + (r ) and Φ − (r ) .
where the uniform state is unstable with respect to fluctuations of both order parameters
Figure 15. The spinodal behavior of the symmetric ternary ABC copolymers:
a) the fA-dependences of the reduced squared critical wave numbers q- (curves 1) and q+ (curves 2) for
both the linear (solid) and miktoarm (dashed) symmetric ABC copolymers;
b) the classification of the spinodal instability regions in the (χ, χ AC ) -plane. The solid lines satisfy eqs
(4.2a), (4.2b) for fA = 0.245, the numbers 0, 1, 2 and 3 label the stability, AC-, B-modulation and twolength-scale regions, respectively; the dashed lines describe the temperature evolution of the systems with
I)
I ) ~ (I )
χ (I ) and ~
χ AC χ > k 0 (a), χ AC χ = k 0 (b) and χ AC χ < k 0 (c), where k 0 = χ (AC
χ (AC
,χ
c) the fA-dependences of the coordinates χ AC (curves 1) and χ
(I )
are the coordinates of the point of intersection of the solid lines;
d) the fA-dependences of the ratio k = χ AC χ
(I )
(I )
(I )
(curves 2) for both the linear (solid)
and miktoarm (dashed) symmetric ABC copolymers;
for both the linear (solid) and miktoarm (dashed)
symmetric ABC copolymers.
~ ,χ
~ ) are the straight lines
The lines λ + = 0 and λ − = 0 in the plane (χ
AC
(( ) (q ) + (~g ) (q )) ,
= min a(q ) = a(q ), a(q ) = (~
g ) (q ) − (~
g ) (q ) ,
χ AC − 4χ = − min b(q ) = −b(q + ), b(q ) = min ~
g −1
χ AC
−
−1
−1
11
−1
11
12
12
(4.3a)
(4.3b)
29
the critical wave numbers q+ and q- characterizing the periods of the profiles Φ + (r ) and
Φ − (r ) , respectively, being the locations of the absolute minima of the function a(q) and
b(q), sought within the semiaxis 0 < q 2 < ∞ . The lines (4.3a) and (4.3b) intersect in the
I
χ = χ( ) = ( a ( q− ) + b ( q+ ) ) 4
point with the co-ordinates
( I)
χ AC = χ AC = a ( q− ) ,
(4.4)
As is seen from Fig. 14a, the values of the reduced squared critical wave numbers q+ and
q- for both the linear and miktoarm ABC copolymers are rather different.
A typical separation of the plane (~
χ AC , ~
χ ) into the regions with different types of the
I)
spinodal instability is shown in Fig. 14b. The dependences of the coordinates χ (AC
, χ (I ) and
I)
the ratio k = χ (AC
χ (I ) of f are plotted in Figs. 14c,d, respectively. As seen from Fig. 14c, the
solid curves 1 and 2 do intersect and for the compositions corresponding to the intersection
point the ABC should reveal the 2-scale behavior we discussed in the Introduction, which is
in a good agreement with the results of ref 75.
Two-length-scale behavior in blends ABC and AC. Another way to realize the 2-scale
behavior is to blend the triblock ABC and diblock AC copolymers. The author et al.97
analyzed the spinodal stability of such blends within the RPA, the famous Hildebrand
approximation for the χ-parameters being used for simplicity:
(
χij = v δi − δ j
)
2
( 2T ) ,
(4.5)
where δi is the solubility parameter of the i-th component supposed to be temperatureindependent. It is convenient73 to take as the two independent interaction parameters
characterizing the ternary systems in the approximation (4.5) the following ones:
χ = χ AC = v ( δ A − δC )
2
( 2T ) ,
x = ( 2δ B − δ A − δC ) ( δ A − δC ) .
(4.6)
χAC characterizes incompatibility of the side blocks in the ABC triblock copolymer whereas
the selectivity parameter x describes how much is the middle block B selective with respect
to the side blocks.
Remarkably, in this approximation the spinodal condition (3.11) takes simple form73
τ = ( 2χ AC N ) max W ( Q, x )
−1
(4.7)
where N is the total degree of polymerization of the triblock copolymer and the function W
depends on the reduced squared wave number Q = q 2 a 2 N 6 , selectivity x and structural
parameters involved in the definition of the structural matrix g(Q) as described in refs 73,
97. For the blends of ABC and AC the function W is shown97 to have one or two maxima at
30
Q = Qmin and Q = Qmax depending on the blend structural parameters and selectivity and,
thus, to be capable of revealing two-length-scale behavior. A new interesting feature of
these blends is that Qmin can reduce to zero6.
parameters of the ABC/AC blends is provided by so-called phase portrait in the ( φ ABC , x )
Information concerning localization of various modes of instability in the space of
plane as shown in Fig. 15.
a)
b)
Figure 16. The phase portraits of the AN C N / AhN BmN ChN for h=2; a) m=4 and b) m=10.
The region corresponding to the macro- and microphase separation instabilities are labeled by the
numbers 1 and 2, respectively. a) the only Lifshitz line is plotted; b) the bold solid and dashed lines are
stable and metastable Lifshitz lines, respectively, within the region confined by the dotted thin lines the
function W(Q) has two maxima; on the thin short dash line both maxima are equal and the system reveals
two-length-scale behavior.
For the blends where the middle block of ABCs is not long enough the function W(Q)
has the only maximum Q = Qmax , which could be located either on the boundary of the
interval 0 ≤ Q < ∞ of the permissible values of Q, i.e. at Qmax = 0 (in this case the blend
would undergo macrophase separation at low temperatures) ) or within the interval, i.e. at
Qmax > 0 (in this case the blend would undergo microphase separation at low
temperatures). The line separating these two regions is referred to as the Lifshitz line (see
Fig. 15a). The Lifshitz curve has the two vertical asymptotes at φ ABC = 0 and φ ABC = φL
( φL = 0.2 in Fig. 15a). Thus, the macrophase separation of the ABC/AC blend may occur
6
The situation is reminiscent of that found by Holyst and Schick126 who carried out the RPA analysis of
symmetric ternary mixture of A and B homopolymers and AB diblock copolymer. They discovered that
in some situations one of the components of the matrix of the structure factors Gij ( q ) for this mixture
could exhibit two equal maxima. However, this situation could be observed in the disordered phase only
and both maxima never diverge simultaneously in the A/B/AB mixture how it occurs in the ABC/AC
blend.
31
only in the interval 0 < φABC < φL , For the mixtures with φABC > φL the ODT (microphase
separation) is only possible. For the longer middle blocks B the function W(Q) could have
two maxima and, accordingly, the Lifshitz line splits into two lines corresponding to
reducing to zero of location of the stable or metastable maximum. A typical phase portrait
for this situation is presented in Fig. 15b.
An important prediction made in ref 97 is that it could be possible to provide occurring
of regular superstructures reminiscent of the pattern shown in Fig. 8 via a subtle tuning of
the selectivity, composition φABC and structural parameters of the ABC/AC blends.
A similar RPA analysis was carried out for solutions of the di- tri- and regular polyblock
copolymers in non-selective solvents,98 where also the conditions for two-length-scale
The critical lines for the ABCs. For simple temperature dependence χ AC (T ) = Θ AC (2T ) ,
behavior were found.
χ(T ) = Θ (2T ) the states of a ternary ABC system with different temperatures are located on
~ ,χ
~ ) . As shown in Fig. 14b, the
χ AC = k ~
χ , k = Θ AC Θ in the plane (χ
the straight line ~
AC
system leaves the stability region crossing either the line (4.3a) or (4.3b) depending on the
value of k. In the first case (B-modulation), which occurs, e.g., for ABA copolymer
( χ AC = 0 ), the effective cubic vertex reads
(
)
(3 )
(3 )
(3 )
(3 )
γ (1) = 2 −3 / 2 Γ111
+ Γ222
+ 3Γ112
+ 3Γ221
,
Γijk(3 ) (1) = Γijk(3 ) (q1 , q 2 , q 3 ),
(4.8a)
q i = q* , q1 + q 2 + q 3 = 0 . The straightforward calculation as
consistent with refs 32,33,36,73 shows that there is the only critical point where the cubic
vertex (4.8a) vanishes for symmetric triblock (miktoarm) copolymer An Bm An . The point is
located at f B = 0.49 ( f B = 0.557 ). On the contrary, in the second case (AC-modulation)
the cubic vertex vanishes identically for symmetric copolymer with any composition of the
(
non-selective block since it reads
)
( 3)
( 3)
( 3)
( 3)
γ (1) = 2−3/ 2 ⎜⎛ Γ111 − Γ 222 + 3 Γ 221 − Γ112 ⎞⎟
⎝
⎠
(4.8b)
So, the ternary ABCs belonging to the AC-modulation class are expected to undergo much
smoother ODT than those belonging to the B-modulation class.
In the Hildebrand approximation the symmetry assumption χ AB = χ BC = χ holds if the
middle block is non-selective with respect to both side blocks (x=0), which occurs, e.g., for
poly(isopren-b-styrene-b-2-vinylpyridine) triblock copolymers.127 But the continuous ODT
transition in the ternary block copolymers occurs not only for f A = f C and x=0. The
32
critical lines were build by the author76 via numerical solving the equation γ (1, f A , f C ) = 0
for different values of 0<x<1 (see Fig. 16).
Figure 17. The critical lines for the linear ABCs in
the Hildebrand approximation for different values of
the selectivity x. The symmetric bold lines correspond
to non-selective middle block (x = 0), the critical lines
labeled by the numbers 1, 2, 3, 4 and 5 correspond to
the values of the selectivity parameter x = 0.01, 0.1,
0.3, 0.5 and 0.8, respectively. The dashed lines cb and
ab are the critical lines for x = ±1 .
Remarkably, the line f A = f C is not the only critical line even for x=0. Another critical
line is the curve ac, which is rather close to the straight line ac. For x ≠ 0 the critical lines
consist of two branches. In the limit x → 1 one of the branches, which corresponds to the
case fC → 0.5,
χ AC − χ BC
χ AC + χ BC , approaches the diblock copolymer critical
line ab whereas another branch, which corresponds to AB copolymer with a short strongly
interacting C block,73 approaches Bc.
The WST border lines and the phase diagrams of the ABC triblock copolymers. As
discussed in sections 2 and 3, taking into account the higher harmonics and other weakly
fluctuating fields results in a renormalization of the effective 4th vertex β of the effective
free energy, and, eventually, in vanishing β with moving off the critical lines. Therewith,
the line β ( f A , fC ) = 0 has meaning of the WST border line, i.e. a crossover line confining
the region beyond which the WST does not hold even qualitatively. In Fig. 17 the WST
border lines calculated within the Hildebrand approximation for the linear and miktoarm
ABC triblock copolymers are presented.76 It is worth to make two remarks here. First, as is
seen from Fig. 17b, the WST border line can intersect the critical line as shown in Fig. 17b.
The intersection point is expected76 to be the tricritical point where the line of the 2nd order
phase transitions transforms into that of the 1st order ones. Thus, in general, a judgment on
the validity of the RPA (and even critical point) analysis in each case requires the full WST
analysis including finding of the 4th vertex. Second, the WST validity region in the
composition triangle is far from being negligibly small, especially for linear ABCs. In
33
particular, in the interval 0.42 ≤ f B ≤ 0.58 the WST phase diagram can be build for any
asymmetry of the side blocks (see Fig.18).
a)
b)
Figure 18. The maps describing the WST application to the melts of the ABC a) linear and b) miktoarm
triblock copolymers with the nonselective middle block (x = 0) within the Hildebrand approximation. The
critical and WS border lines are shown by solid and short dashed lines, respectively. Note that the WST
border line intersects the critical line f A = fC for miktoarm ABC triblock copolymers.
Figure 19. TheWST phase diagram of the molten
linear ABC block copolymers calculated
f B = 0.55 within the Hildebrand approximation
in the plane ( σ, χ ) , where σ = f A ( f A + fC ) is
the asymmetry parameter. The designations of the
phases are the same as in Figure 11.
It is worth to stress a rather broad (as compared to that for diblock copolymers) region
(4) of the double gyroid (G) phase stability, which in this case can be predicted basing on
the WST only. We skip here many other phase diagrams build and discussed in ref 76 the
basic information of which is already given above in Fig. 11.
5. Conclusion.
Even this brief survey of the current development of the weak segregation theory brings
forward, hopefully, the idea that the theory is still far from being exhausted its potential. As
I tried to show in the review, the WST is capable to describe and predict non-conventional
phases in multi-component block copolymer systems like gyroids (both single and double),
face-centered and simple cubic phases as well as rather non-conventional new types of two34
scale-length ordering. The WST easily incorporates various ideas of the short-range
thermodynamic interactions between the repeated units of polymer components. The main
advantage of the WST is its capability to state analytically and solve both analytically and
numerically the problems of optimization of various properties of the ordered block
copolymers and identifying the most promising structures of multi-component block
copolymers worth to be synthesized by chemists, which could be too special to find them by
chance.
Still, the WST has some rather natural and important problems to be solved, in which
case its predictive capacity could much increase. First of all I mean extension of the WST to
the level of the 3rd shell approximation. Such an extension seems to be a natural limit for the
Landau-like theory, which would correspond to the Landau expansion up to the 6th order in
powers of the order parameter and thus provide an opportunity to describe the ODT near the
tricritical point and the WST border line discussed above. Thereupon it is worth to mention
here the work,51 where the authors have shown that inclusion of the 6th order terms in the
Landau expansion of the free energy of polydisperse copolymers results in considerable
corrections to the previous results. The next problem is still to improve our understanding of
the fluctuation effects. In particular, the closely related problems of the interplay between
the short- and long range fluctuation corrections128,129 and theoretical explanation of the
observed deviations (see ref 31 and Fig. 3) of the inverse structure factor from the RPA
predictions are still open issues. One more important problem is to study coupling in the
ordered phases between the density fluctuations and transversal sound waves (shear
fluctuations), which until now has been considered130,131 only in the disordered state and is
expected132 to be especially important in the weak segregation regime where the shear
modulus is small.133,134
But the most important problem is, of course, to establish a prompt and reliable feedback
between the theoreticians, experimentalists and industry since there is no good physics,
which could not generate a good technology, and there is no good technology which would
be not based (at least, implicitly) on a good physics.
To conclude, I thank my colleagues and friends Volker Abetz, Henk Angerman, Kurt
Binder, Gerrit ten Brinke, Monica Olvera de la Cruz, Andrey Dobrynin, Alexander
Grosberg, Jean-Francois Joanny, Albert Johner, Alexei Khokhlov, Ludwik Leibler, Marcus
Müller, Alexander Semenov, Friederika Schmidt and Reimund Stadler†, whose feedback so
much helped me in my work.
35
The BCC family is determined by the set {q i } consisting of 12 vectors whose relative
Appendix 1. The basic weakly segregated morphologies.
directions are given, e.g., by the six vectors listed below and the same vectors taken with
(
= (q
)
2 )(0, − 1, − 1),
the opposite sign:
q 1 = q*
qI
*
2 (0,1, − 1),
(
= (q
q 2 = q*
q II
*
)
2 )(−1, 0,− 1, ),
2 (−1, 0,1, ),
(
= (q
q 3 = q*
q III
*
)
2 )(−1, − 1, 0)
2 (1, − 1, 0),
20
The vectors could be visualized as the edges of octahedron or tetrahedron24 or via their
planar mapping shown in Figure A1. The arrow circuits in Fig. Aa correspond to the
equalities q I = q II + q3 , q II = q III + q1 , q III = q I + q 2 , q1 + q 2 + q3 = 0 .
a)
b)
Figure A1. The planar mapping of the vectors characterizing the symmetry of the main harmonics a) for
the BCC family and b) for the orthorhombic lattice Fddd (see explanations in the text).
1. For the conventional BCC all phases ϕ appearing in the definition (1.14) of the basic
function Ψ are zero and the vertices read
α BCC = 8 γ 6 3 / 2 , β BCC = [λ 0 (0 ) + 8λ 0 (1) + 2λ 0 (2 ) + 4λ (1,2,1)] 24 .
(A1.1)
2. If the phases corresponding to three vectors q1, q2 and q3, which form a base of the
tetrahedron, and those of three non-coplanar vectors qI, qII and qIII (the thin and solid lines,
respectively, in Figure Xa) are equal to π/2 and 0, respectively, we get a lattice first
3)
discussed in ref 12 and called there the BCC2. For this lattice Ω (ABC
= 3 π 2,
3)
3)
3)
4)
4)
(4)
Ω (ABS
= Ω (ACS
= Ω (BCS
= π 2, Ω (ABCSA
= Ω (BCASB
= Ω CABSC
= π and the basic function and
α BCC 2 = 0 , β BCC 2 = [λ 0 (0 ) + 8λ 0 (1) + 2λ 0 (2 ) − 4λ (1,2 )] 24 .
the vertices read
(A1.2)
The BCC2 phase could be shown76 to possess symmetry of the I 4132 space group
(No.214), which is non-centrosymmetric and closely related to so-called single gyroid
surface. Thus, it seems to be the simplest (and the only up to now) cubic non36
centrosymmetric morphology that could be described (and for some case predicted) within
the WS theory.
The BCC2 belongs to the class of morphologies like L, FCC and SC we refer to as the
degenerate ones because for them the cubic vertex (2.5) identically equals zero due to the
ΔFℜ = − τ 2 (4β ℜ ).
symmetry reasons. For these degenerate morphologies the free energy (2.4) reads21
(A1.3)
Thus, the most stable degenerate morphology is that having the least value of the quadric
vertex. For references, we present the expressions for the vertices also for FCC L and SC
βFCC = ⎡⎣ λ 0 ( 0 ) + 6λ 0 ( 4 3) − 2λ ( 4 3, 4 3) ⎤⎦ 16 ,
βL = λ 0 ( 0 ) 4 , βSC = ⎡⎣ λ 0 ( 0 ) + 4λ 0 ( 2 ) ⎤⎦ 12.
morphologies:
(
= (q
= (q
= (q
)
6 )(− 2,−1,−1),
6 )(+ 2,+1,−1),
6 )(+ 2,−1,+1),
(
= (q
= (q
= (q
6 (− 2,+1,+1), q 02 = q*
)
6 )(+ 1,+2,−1),
6 )(− 1,−2,−1),
6 )(− 1,+2,+1),
(A1.4)
(
= (q
= (q
= (q
6 (+ 1,−2,+1), q 03 = q*
)
6 )(+ 1,−1,+2 ),
6 )(− 1,+1,+2 ),
6 )(− 1,−1,−2 ).
6 (+ 1,+1,−2 ),
The G family is determined by 2⋅12 main harmonics given by the 12 vectors
q 01 = q*
q11
q 21
q 31
*
*
*
q12
q 22
q 32
*
*
*
q13
q 23
q 33
*
*
*
(A1.5)
Figure A2. The planar mappings of the vectors characterizing the symmetry of the main harmonics for the
G family (left); and of the edges of the regular icosahedron (right).
a) the vectors depicted by bold lines have zero phases for all three morphologies G, G2 and BCC3, those
depicted by dashed and thin solid lines have phases equal to π only for the double gyroid (G) and both for
G and G2, respectively (see the definitions (2.19a) and (2.19b);
b) the edges depicted by thin and bold lines correspond to the vectors to be removed and properly rotated
to transform the icosahedron into the G cell.
and 12 opposite ones. The planar mapping of the vectors is shown in Figure A2a. For
comparison, in Figure A2b the planar mapping of the regular icosahedron is shown.
37
Obviously, the set of vectors (A.5) (and the opposite ones) is obtained via a deformation of
the regular icosahedron, which involves removing 6 of 30 edges of the icosahedron and the
proper rotations of the remaining edges, the resulting polyhedron being the Wigner-Zeitz
cell of the corresponding crystal lattice. It is this relationship between the G family and the
icosahedron symmetry which causes the famous 10 spot SAXS pattern observed in the
gyroid phase.
1. The morphology arising if all the phases ϕi are set to equal zero we call the BCC3. It is
just the ordinary BCC but the fact that the dominant harmonics correspond here to the 3rd
(rather than the 1st!) co-ordination sphere.
α BCC3 = γ 33/ 2 ,
βBCC3 = ⎣⎡ Λ1 + 4 ⎣⎡ λ (1 3, 2 3 ) + λ ( 2 3,5 3 ) ⎦⎤ + 2λ ( 2 3, 2 3 ) ⎦⎤ 48,
(
Λ1 = λ 0 ( 0 ) + 2 λ 0 ( 4 3 ) + 2 ( λ 0 (1 3 ) + λ 0 ( 2 3 ) + λ 0 (1) + 2λ 0 ( 5 3 ) )
)
(A1.6)
2. The trial function (1.14) with the main harmonics (A.5) and the phase choice
ϕ12 = ϕ23 = ϕ31 = ϕ01 = ϕ02 = ϕ03 = 0,
ϕ21 = ϕ32 = ϕ13 = ϕ11 = ϕ22 = ϕ33 = π.
corresponds to the bi-continuous gyroid (G) or double gyroid morphology having the
α G = γ 33 / 2 , β G = [Λ 1 − 2Λ 2 − 4λ (1 3 , 2 3)] 48 ,
symmetry Ia 3 d .
Λ 2 = 2λ (2 3 , 5 3) − λ (2 3 , 2 3). (A1.7)
It is seen from (A.6), (A.7) that due to the symmetry of the BCC3 and G lattices their
cubic vertices are identical and, therefore, the BCC3 - G phase transition line (surface) is
determined by equation
β G − β BCC3 = 0
(A1.8)
3. The trial function (1.14) with the main harmonics (A.5) and the phase choice
ϕ21 = ϕ32 = ϕ13 = π, ϕ12 = ϕ23 = ϕ31 = ϕ01 = ϕ02 = ϕ03 = ϕ11 = ϕ22 = ϕ33 = 0
(A1.9)
corresponds to the morphology of the symmetry I 4 3d we refer to as the G2. For this
(
)
α G 2 = γ 2 ⋅ 33 / 2 , β G 2 = [Λ 1 + 2Λ 2 − 4λ (1 3 , 2 3)] 48 .
morphology
α HEX = 2 γ 3 3 / 2 , β HEX = [λ 0 (0 ) + 4λ 0 (1)] 12 .
(A1.10)
For reference, we give here also the vertices for the HEX morphology20
(A.11)
The orthorhombic lattice (space symmetry group O70 or Fddd) is generated by the 2⋅4
e1 = q* ( a, b, c ) , e 2 = q* ( a, −b, −c ) , e3 = q* ( − a, b, −c ) , e 4 = q* ( − a, −b, c )
main harmonics given by the 4 vectors
(A1.12)
38
where q* = q*
a 2 + b 2 + c 2 , the values of the periods a, b, c of the reciprocal lattice are,
in general, all different and the choice of the phases is ϕ01 = π, ϕ02 = ϕ03 = ϕ04 = 0.
Thus, in general, the lattice belongs to the class of the degenerate morphologies and the
corresponding 3rd and 4th vertices read
(
3
⎡
α 0ortho = 0 , β0ortho = ⎢ λ 0 ( 0 ) + 2∑ λ 0 ( hi ) − 2λ h1 , h 2
⎢⎣
i =1
where the designations hi = 4ai2
(a
2
1
)
)⎥⎥
⎤
⎦
(A1.13)
16 ,
+ a22 + a32 are introduced. In particular, for the case
a = b = c the orthorhombic lattice becomes just the FCC one. If a 2 = 2b 2 = 2c 2 then two
second harmonics ( 0, ±2b, 0 ) and ( 0, 0, ±2c ) also belong to the first coordination sphere of
the reciprocal lattice, which now contains (along with the main harmonics) 2⋅6 vectors and
At last, if a 2 = 4b 2 = 12c 2 then two second harmonics b1 = e1 − e 2 = q* ( 0, 2b, 2c ) and
could be easily checked to correspond to the BCC lattice in the co-ordinate space.
b 2 = e 4 − e3 = q* ( 0, −2b, 2c ) as well as the forth harmonic c = b1 + b 2 = q* ( 0, 0, 4c ) also
belong to the first coordination sphere with the radius q*2 = a 2 + b 2 + c 2 = 16c 2 . The planar
mapping of the vectors is shown in Figure Xd.
In this special case only those harmonics, which have the same symmetry rather than
belong to the same coordination sphere of the reciprocal lattice, have equal amplitudes, so
( ( ) ( ))
= 2 ( cos ( e r ) + cos ( e r ) + cos ( e r ) − cos ( e r ) ) , Ψ = 2 cos ( cr ) .
that the trial function (1.14) reads
Φ ( r ) = AΨ A ( r ) + BΨ B ( r ) + C Ψ C ( r ) , Ψ B = 2 cos b 2 r − cos b 1 r ,
ΨA
2
3
4
1
(A1.15)
C
The choice of the signs in (A.15) is determined by equivalent requirements to satisfy the
F = VT min f ( A, B, C ) ,
Fddd symmetry and provide the minimal free energy, which takes the form
f ( A, B, C ) = τ ( 4 A2 + 2 B 2 + C 2 ) − 4γ A2 B − 2γ B 2C
where
(
)
+ λ ( 0 ) + 2 ( λ (1 4 ) + λ ( 3 4 ) + λ (1) − λ (1 4,3 4 ) ) A4
+ ( λ (1) + (1 2 ) λ ( 0 ) ) B 4 + ( λ ( 0 ) 4 ) C 4 + 4λ (1,3 2 ) A2 BC
+4 ( λ (1) + λ ( 3 2 ) ) A2 B 2 + 4λ ( 3 2 ) A2C 2 + 2λ (1) B 2C 2 .
(A1.16)
Taking approximation (2.13) for the 4th vertices λ and minimizing the free energy (A16)
numerically with respect to all three amplitudes A, B, C results in the reduced phase
diagram shown in Fig. 2b.
39
References and notes
1. Bates, F.S.; Fredrickson, G.H. Annu. Rev. Phys. Chem. 1990, 41, 525.
2. Erukhimovich. I.Ya.; Khokhlov A.R. Polymer Sci. 1993, 35, 1522.
3. Binder, K. Advances in Polymer Science 1974, 112, 181.
4. Bates, F.S.; Fredrickson, G.H. Physics Today 1999 (February), 33, 32.
5. Abetz, V. Assemblies in Complex Block Copolymer Systems, in: Supramolecular
polymers, A. Ciferri ed., M. Dekker, New York 2000.
6. Meier, D.J. J. Polymer Sci.: Part C, 1968, 26, 81.
7. Helfand E., Macromolecules, 1975, 8, 552.
8. Helfand E., Wasserman Z. R. Macromolecules, 1976, 9, 879.
9. Helfand E., Wasserman Z. R. Macromolecules, 1978, 11, 961.
10. Helfand E., Wasserman Z. R. Macromolecules, 1980, 13, 994.
11. Semenov A.N. Sov. Phys. JETP, 1985, 61, 733.
12. Semenov A.N., Macromolecules, 1993, 26, 2273.
13. Olmsted, P.D.; Milner, S.T. Macromolecules 1998, 31, 4011.
14. Likhtman. A.E.; Semenov. A.N., Europhys. Lett., 2000, 51, 307.
15. Thomas, E.L.; Anderson, D.M.; Henkee, C.S.; Hoffman, D. Nature, 1988, 334,
598.
16. Hyde, S.T. Current Opinion in Solid State & Materials Science, 1996, 1, 653.
17. Wohlgemuth, M.; Yufa, N.; Hoffman, J.; Thomas, E.L. Macromolecules 2001, 34,
6083.
18. Garstecki, P.; Hołyst, R. Macromolecules 2003, 36, 9181, 9191.
19. Witten, T.A.; Leibler, L.; Pincus, P.A. Macromolecules, 1990, 23, 824.
20. Flory, P. J. Principles of Polymer Chemistry; Cornell University Press: Ithaca,
NY, 1953.
21. Leibler, L. Macromolecules 1980, 13, 1602.
22. Landau, L.D. Zh. Exp. Theor. Fiz. 1937, 7, 627.
23. Landau, L. D.; Lifshitz, E. M. Statistical Physics, 3rd ed.; Pergamon Press:
Oxford, 1980; Vol. 1.
24. de Gennes, P. G. Scaling Concepts in Polymer Physics; Cornell University Press:
Ithaca, NY, 1979.
25. Brazovskii, S. A. Sov. Phys. JETP 1975, 41, 85.
26. Brazovskii, S.A.; I.E. Dzyaloshinski and A.R. Muratov, Sov.Phys. JETP 1987, 66,
625.
27. Kats E. I.; Lebedev V. V.; Muratov A. R. Phys. Rep. 1993, 228, 1.
40
28. de Gennes, P.G. Faraday Disc.Chem.Soc. 1979, 68, 96.
29. Yerukhimovich, I.Ya. Polymer Sci. U.S.S.R. 1982, 24, 2223.
30. Yerukhimovich, I.Ya.; Letichii, B.A. Polymer Sci. U.S.S.R. 1979, 21, 1394.
31. Bartels, V.T.; Abetz, V.; Mortensen, K.; Stamm, M. Europhysics Letters 1994,
27, 371.
32. Mayes, A. M.; Olvera de la Cruz, M. J. Chem. Phys. 1989, 91, 7228.
33. Dobrynin, A.V.; Erukhimovich, I.Ya. Vysokomol.Soyed., 1990, 32B, 9, 663.
34. Dobrynin, A.V.; Erukhimovich, I.Ya. Vysokomol.Soyed., 1990, 32B, 10, 743.
35. Dobrynin, A.V.; Erukhimovich, I.Ya. Vysokomol.Soyed., 1990, 32B, 11, 852.
36. Dobrynin, A.V.; Erukhimovich, I.Ya. Macromolecules 1993, 26, 276.
37. Fredrickson, G.H.; Helfand, E. J. Chem. Phys. 1987, 87, 697.
38. Brazovskii, S.A.; Dmitriev, S.G. Sov. Phys. JETP, 1975, 42, 497.
39. Wright, D.C.; Mermin, N.D. Rev. Mod. Phys.,1989, 61, 385.
40. Nabutovskii, V.M.; Nemov N.A.; Peisakhovich, Yu.G. Sov.Phys.JETP, 1980, 52,
1111.
41. Høye, J.S.; Stell, G. J.Phys.Chem., 94, 7899, (1990)
42. Borue, V.Yu. Erukhimovich, I.Ya. Macromolecules, 1988, 21, 3240.
43. Joanny, J.-F.; Leibler, L. J. Phys. (Paris) 1990, 51, 545.
44. Shakhnovich, E. I.; Gutin, A. M. J. Phys. (France) 1989, 50, 1843.
45. Panyukov, S. V.; Kuchanov, S. I. JETP Lett. 1991, 54, 501; J. Phys. II (France)
1992, 2, 1973.
46. Fredrickson, G. H.; Milner, S. T.; Leibler, L. Macromolecules, 1992, 25, 6341.
47. Erukhimovich, I. Ya.; Dobrynin, A. V. Macromol. Symp. 1994, 81, 25.
48. Angerman, H.J.; ten Brinke, G.; Erukhimovich, I. Macromolecules, 1996, 29,
3255.
49. Angerman, H.J.; ten Brinke, G.; Erukhimovich, I. Macromolecules, 1998, 31,
1958.
50. Panyukov, S.V.; Potemkin, I.I. Phys. Rev. E, 1998, 57, 6902.
51. Kuchanov, S. I.; Panyukov, S. V. J. Phys. Condensed Matter, 2006, 18, L43.
52. Huh, J.; W.H. Jo. Macromolecules, 2004, 37, 3037.
53. Luzzati, V.; Spegt, P.A. Nature 1967, 215, 701.
54. Luzzati, V. et al., Nature 1968, 220, 485.
55. Matsen, M.W.; Schick, M. Phys.Rev.Lett. 1994, 72, 2660.
56. Matsen, M.W.; Schick, M. Macromolecules 1994, 27, 6761.
57. Matsen, M; Bates, F. Macromolecules, 1996, 29, 1091.
58. Edwards, S.F. Proc. Phys. Soc. 1966, 88, 265.
41
59. Hajduk, D.A.; Harper, P.E.; Gruner, S.M.; Honeker, C.; Kim, G.; Thomas, E.L.
Macromolecules 1994, 27, 4063.
60. Forster, S.; Khandpur, A.K.; Zhao, J.; Bates, F.S.; Hamley, I.W.; Ryan, A.J.; Bras,
W. Macromolecules 1994, 27, 6922.
61. Matsen, M. J. Chem. Phys. 1995, 103, 3268.
62. Matsen, M. J. Chem. Phys. 1998, 108, 785.
63. Wickham, R.A.; Shi, A.-C. Macromolecules, 2001, 34, 6487.
64. Mayes, A. M.; Olvera de la Cruz, M. J. Chem. Phys. 1991, 95, 4670.
65. Dobrynin, A.V.; Erukhimovich, I.Ya. J. Phys. II Fr. 1991, 1, 1387.
66. Olvera de la Cruz, M. Phys. Rev. Lett. 1991, 67, 85.
67. Barrat, J.L.; Fredrickson, G.H. J. Chem. Phys. 1991, 95, 1281.
68. Dobrynin, A.V.; Erukhimovich, I.Ya. Polymer Sci. U.S.S.R. 1992, 33, 1012.
69. Erukhimovich, I.Ya.; Dobrynin, A.V. Macromolecules, 1992, 25, 4411.
70. Floudas, G., Pispas, S.; Hadjichristidis, N.; Pakula, T.; Erukhimovich, I.
Macromolecules, 1996, 29, 4142.
71. Hamley, I.W.; Podnek, V.E., Macromolecules, 1997, 30, 3701.
72. Almdal, K.; Rosedale, J.H.; Bates, F.S.; Wignall, G.D.; Fredrickson, G.H. Phys.
Rev. Lett. 1990, 92, 6255.
73. Erukhimovich, I.Ya.; Abetz, V.; Stadler, R. Macromolecules 1997, 30, 7435.
74. Werner, A.; Frederickson, G. H. J. Polym. Sci., Part B: Polym. Phys. 1997, 35,
849.
75. Cochran, E.W.; Morse, D.C.; Bates F.S. Macromolecules 2003, 36, 782.
76. Erukhimovich, I., Eur. Phys. J. E 2005, 18, 383.
77. Tanaka, F.; Ishida, M.; Matsuyama, A. Macromolecules 1991, 24, 5582.
78. Dormidontova, E.; ten Brinke, G. Macromolecules 1998, 31, 2649.
79. Angerman, H.A; ten Brinke, G. Macromolecules 1999, 32, 6813.
80. Kudlay, A.N.; Erukhimovich, I.Ya.; Khokhlov, A.R. Macromolecules, 2000, 33,
5644.
81. Hino, T.; Prausnitz, J.M. Macromolecules 1998, 31, 2636.
82. Cho, J. Macromolecules, 2000, 33, 2228.
83. Cho, J. J. Chem. Phys. 2003, 119, 5711.
84. Yeung, C.; Desai, R.C.; Shi, A.-C.; Noolandi, J. Phys. Rev. Lett. 1994, 72, 1834.
85. Cho, J.; Kwon, Y.K. J. of Polymer Sci: Part B: Polymer Physics, 2003, 41, 1889.
86. Erukhimovich, I.Ya. JETP Letters 1996, 63, 459.
87. Bailey, T.S.; Hardy, C.M.; Epps, T.H.; Bates, F.S. Macromolecules, 2002, 35,
7007.
42
88. Epps, T.H.; Cochran, E.W.; Hardy, C.M.; Bailey, T.S.; Waletzko, R.S.; Bates,
F.S. Macromolecules, 2004, 37, 7085.
89. Epps, T.H.; Cochran, E.W.; Bailey, T.S.; Waletzko, R.S.; Hardy, C.M.; Bates,
F.S. Macromolecules, 2004, 37, 8325.
90. Cochran, E.W.; Bates, F.S. Phys. Rev. Lett. 2004, 93, 087802. .
91. Tyler, C.A.; Morse, D.C. Phys. Rev. Lett. 2005, 95, 208302.
92. Ranjan, A.; Morse,D.C. Phys. Rev. E. 2006, 74, 011803.
93. Ruokolainen, J.; Mäkinen, R.; Torkkeli, M.; Mäkelä, T.; Serimaa, R.; ten Brinke,
G.; Ikkala, O. Science, 1998, 280, 557.
94. Goldacker, T., Doctoral Dissertation, Bayreuth: Univ. Bayreuth, 1999.
95. Nap, R.; Kok, C.; ten Brinke, G.; Kuchanov, S. Eur. Phys. J. E. 2001, 4, 515.
96. Erukhimovich, I.Ya., in: Proceedings of the Int. Conf. on Advanced Polymers and
Processing, ICAPP 2001 Yonezawa, Iwakura, K., Ed., p. 122-128..
97. Erukhimovich, I.; Smirnova, Yu.;Abetz, V. Polymer Science Series A, 2003 45,
1093.
98. Tarasenko, S.; Erukhimovich, I. Polymer Science Series A 2005, 47, 229.
99. Tarasenko, S.; Erukhimovich, I. Polymer Science Series A 2006, 48, 207.
100.
Smirnova, Yu.; ten Brinke, G; Erukhimovich, I. Polymer Science Series A
2005, 47, 430.
101.
Smirnova, Yu.; ten Brinke, G; Erukhimovich, I. J. Chem. Phys. 2006, 124,
054907.
102.
Pichugin, V.E.; Kuchanov, S.I. J. Stat. Mech. P07009 (2005)
103.
Pichugin, V.E.; Kuchanov, S.I. ten Brinke, G. e-polymers, 2006
104.
Nap, R.; ten Brinke, G. Macromolecules 2002, 35, 952.
105.
Nap, R.; Erukhimovich, I.; ten Brinke, G. Macromolecules 2004, 37, 4296
106.
Nap, R.; Sushko, N.; Erukhimovich, I.; ten Brinke, G. Macromolecules 2006,
39 (in press).
107.
Kriksin, Yu.A.; Khalatur, P.G.; Smirnova, Yu.G.; Erukhimovich, I.; ten
Brinke, G. (in preparation).
108.
Milner, S.T.; Olmsted, P.T., J. Phys. II, 1997, 7, 249.
109.
Marques, C.; Cates, C.M. Europhys. Lett. 1990, 13, 267.
110.
Olvera de la Cruz, M.; Mayes, A.M.; Swift, B.W. Macromolecules 1992, 25,
944.
111.
International Tables for Crystallography; Kluwer Publishing: Dortdrecht, The
Netherlands, 1996, Vol.B.
112.
Lifshitz, I.M.; Grosberg, A.R.; Khokhlov, A.R. Rev. Mod. Phys.1978, 50, 683.
113.
Dobrynin, A.V.; Erukhimovich, I.Ya. Sov. Phys. JETP, 1991, 72, 751.
43
114.
Yerukhimovich, I. Ya. Polymer Science U.S.S.R. 1979, 21, 470.
115.
Benoit, H.C.; Benmouna, M. Polymer, 1984, 25, 1059.
116.
Higgins, J.S.; Benoit, H.C. Polymers and Neutron Scattering. Oxford:
Clarendon Press, 1996.
117.
Hong, K.M.; Noolandi, J. Macromolecules 1983, 16, 1083.
118.
Hansen, J.P.; McDonald, I.R. Theory of Simple Liquids, Academic Press,
London, 1986.
119.
Khokhlov, A.R. J. Physique, 1977, 38, 845.
120.
Khokhlov, A.R. Polymers, 1978, 19, 1387.
121.
Schweizer, K.S.; Curro, J.G. Adv. Polym. Sci. 1994, 116, 319.
122.
Schweizer, K.S.; Curro, J.G. Advances in Chemical Physics, edited by I.
Prigogine and S. A. Rice , 1997, vol. XCVIII, Chap. 1.
123.
Stepanov, S. Macromolecules, 1995, 28, 8233.
124.
Guenza, M.; Schweizer, K. J. Chem. Phys., 1997, 106, 7391.
125.
Hardy, C. M.; Bates, F. S.; Kim, M.-H.; Wignall, G. D. Macromolecules 2002,
35, 3189-3197.
126.
Holyst, R.; Schick, M. J. Chem. Phys. 1992, 96, 7728.
127.
Mogi, Y.; Nomura, M.; Kotsuji, H.; Ohnishi, K.; Matsushita,Y.; Noda, I.
Macromolecules 1994, 27, 6755.
128.
Maurer, W.W.; Bates, F.S.; Lodge, T.P.; Almdal, K.; Mortensen, K.;
Fredrickson, G.H. J. Chem. Phys. 1998, 108, 2989.
129.
Kudlay, A.; Stepanow, S. J. Chem. Phys. 2003, 118, 4272.
130.
Doi, M.; Onuki, A. J. Phys. II France 1992, 2, 1631.
131.
Erukhimovich, I.Ya.; Kudryavtsev, Ya.V. Eur. Phys. J. E 2003, 11, 349.
132.
Landau, L.D.; Lifshitz, E.M. Electrodynamics of continuous media 1984
Pergamon Press; Oxford, New York, Beijing, Frankfurt.
133.
Amundsen, K.; Helfand, E. Macromolecules 1993, 26, 1324.
134.
Kawasaki, K.; Ohta, T. Physica 1986, 139A, 223.
44