You are on page 1of 20

Nonlinear rheological behavior of multiblock copolymers under large amplitude

oscillatory shear
Zhijun Nie, Wei Yu, and Chixing Zhou

Citation: Journal of Rheology 60, 1161 (2016); doi: 10.1122/1.4961483


View online: https://doi.org/10.1122/1.4961483
View Table of Contents: https://sor.scitation.org/toc/jor/60/6
Published by the The Society of Rheology

ARTICLES YOU MAY BE INTERESTED IN

New measures for characterizing nonlinear viscoelasticity in large amplitude oscillatory shear
Journal of Rheology 52, 1427 (2008); https://doi.org/10.1122/1.2970095

Dynamic wall slip behavior of yield stress fluids under large amplitude oscillatory shear
Journal of Rheology 61, 627 (2017); https://doi.org/10.1122/1.4982704

Large amplitude oscillatory shear of hard-sphere colloidal dispersions: Brownian dynamics simulation and
Fourier-transform rheology
Journal of Rheology 60, 1041 (2016); https://doi.org/10.1122/1.4955433

Viscoelasticity of entangled random polystyrene ionomers


Journal of Rheology 60, 1031 (2016); https://doi.org/10.1122/1.4955432

Rheology of fumed silica/polydimethylsiloxane suspensions


Journal of Rheology 61, 205 (2017); https://doi.org/10.1122/1.4973974

On the use of continuous relaxation spectra to characterize model polymers


Journal of Rheology 60, 1115 (2016); https://doi.org/10.1122/1.4960334
Nonlinear rheological behavior of multiblock copolymers under
large amplitude oscillatory shear

Zhijun Nie, Wei Yu,a) and Chixing Zhou

Advanced Rheology Institute, Shanghai Key Laboratory of Electrical Insulation and Thermal Ageing,
Department of Polymer Science and Engineering, Shanghai Jiao Tong University, Shanghai 200240,
People’s Republic of China

(Received 31 January 2016; final revision received 30 July 2016; published 23 September 2016)

Abstract
The nonlinear responses of olefin multiblock copolymers (OBCs) melts were investigated under large amplitude oscillatory shear (LAOS)
using Fourier-Transform rheology in this work. Time-temperature superposition (TTS) was examined using linear viscoelastic functions, as
well as nonlinear viscoelastic functions like the intrinsic Q factor (Q3,0) and the dynamic moduli of the third harmonic (G033;0 and G0033;0 ). It is
found that the intrinsic Q factor (Q3,0) exhibits similar TTS behavior as those of storage moduli and loss moduli. However, the deviation
from typical liquidlike behavior of homogeneous melt in Q3,0 is much more evident than that in the storage moduli, indicating more sensitive
of Q3,0 to the microstructure than linear viscoelastic functions. In addition, we suggested a new plot, the intrinsic phase angle at third har-
monic (d3,0) versus the intrinsic complex moduli at third harmonic (jG33;0 j), which is named as the vGP3 plot. Using the vGP3 plot, the much
weaker mesophase separation can be distinguished from the homogeneous state successfully. The high sensitivity of vGP3 to the failure of
TTS makes it a new candidate in studying the thermorheological behaviors of complex fluids. Finally, the frequency dependency of intrinsic
nonlinearity under LAOS was modeled by combining the molecular stress function (MSF) model and emulsion model. The MSF model can
describe the experiment data of linear dynamic moduli and Q3,0 quite satisfactorily for homogeneous OBCs. For heterogeneous melt, the
MSF model and the simple emulsion model account for the homogeneous contribution and the interfacial contribution, respectively. The pre-
dictions of the combined model agree quite well with the frequency dependency of the experimental data for both linear storage moduli and
C 2016 The Society of Rheology. [http://dx.doi.org/10.1122/1.4961483]
intrinsic Q factor (Q3,0). V

I. INTRODUCTION time is no longer purely sinusoidal, and such behavior can no


longer be described in terms of the storage moduli and the loss
The rheological properties are usually classified by the
moduli because of higher harmonic contributions. Therefore,
extent of deformation. Under small deformation, the linear
the dynamic moduli are no longer sufficient beyond this limit
viscoelastic experiment gives the structural information under
because they are defined as a measure of the elastic and viscous
equilibrium state [1]. As the deformation increases, nonlinear
contributions only within the linear viscoelastic limit. In the
responses may imply the mechanism of structural evolution
past, several methods have been proposed to analyze LAOS
under flow [2]. The nonlinear viscoelastic experiment has
data. One of the popular methods is the graphical approach,
been designed and performed in various ways, among which
which involves plotting the stress signal as a function of strain
large amplitude oscillatory shear (LAOS) is of special interest
or strain rate resulting in elastic or viscous Lissajous curves,
because both strain amplitude and frequency can be con-
respectively [7–12]. Although this method provides an intuitive
trolled independently and it does not involve any sudden
representation of the nonlinear behavior as the shape of the
jump in speed or position [3,4]. It has been shown that the
closed loops starts to deviate from an ellipsoidal shape, it is dif-
LAOS characterization may provide much more insights for
ficult to extract the underlying physics with this method. Then,
distinguishing the structural differences [5]. For example, lin-
ear and star-branched polystyrene solutions behaved similarly a generalization of linear viscoelasticity from a geometrical
under small amplitude oscillatory shear (SAOS) condition point of view was introduced to facilitate the interpretation
and no differences occurred under nonlinear step-shear condi- LAOS data. This method consists of an orthogonal stress
tion, whereas significant differences could be monitored decomposition (SD) of the total stress response signal into elas-
under LAOS condition [6]. Thus, exploiting LAOS tests to tic contribution and viscous contribution [12–14]. Actually, the
investigate and quantify the nonlinear viscoelastic behavior SD [12,13] or the generalized SD [14] can be regarded as a
of complex fluids have been the focus of considerable special case of geometric average of Lissajous curves [15],
research both theoretically and practically. from which the mean stress versus strain (rate) curve and the
In a strain-controlled test, when the strain amplitude in an stress versus mean strain (rate) curve in a shear cycle can be
oscillatory shear flow is large, the stress curve as a function of defined. Several moduli and viscosities have been defined from
the mean stress-strain (rate) curves [12] and the stress-mean
strain (rate) curves [15]. These quantities are continuous func-
a)
Author to whom correspondence should be addressed; electronic mail: tions of the strain (or stress) amplitude, and exhibit gradual
wyu@sjtu.edu.cn change from linear regime to nonlinear regime.
C 2016 by The Society of Rheology, Inc.
V
J. Rheol. 60(6), 1161-1179 November/December (2016) 0148-6055/2016/60(6)/1161/19/$30.00 1161
1162 NIE, YU, AND ZHOU

In addition to the SD method, another frequently used different from the traditional monodisperse systems. Li et al.
quantitative technique to analyze LAOS behavior is Fourier- found that the polydispersity can increase the microdomain
Transform rheology [16–18]. The Fourier-Transform rheol- size, and the domain spacing (102–158 nm) in polydisperse
ogy converts the response signal (the stress data in strain diblock or multiblock olefin multiblock copolymers (OBCs)
controlled experiment) in the time domain into frequency is three to five times larger than that in equivalent monodis-
spectra, in which the response signal is expressed as a perse systems [45,46]. As the much larger domain size, the
Fourier series capturing the higher harmonics. Among the separation in polydisperse OBCs is often called “mesophase
higher harmonics, the relative intensity of the third harmonic separation” instead of the “microphase separation,” which
(I3/1) is the most important in quantifying nonlinearity. In was commonly used in monodisperse systems [47–50]. In
fact, on a log-log plot, I3/1 of response stress signal shows a our previous work, we have discussed how the mesophase
linear dependence on the applied strain amplitude at low and separation affects the linear rheological behavior of
medium oscillatory strain amplitudes, reflecting the underly- ethylene-octene multiblock copolymers (OBCs), and how to
ing scaling behavior [19]. This motivated the definition of define the mesophase separation temperature using a new
medium amplitude oscillatory shear (MAOS) fulfilling the method based on the two-dimensional mechanical spectrum
scaling laws at the region between SAOS and LAOS, where [48]. The phase diagram of OBCs in terms of vN with v the
the shear strain amplitude ranges of this region depends on Flory-Huggins interaction parameter and N the degree of
the material and the excitation frequency [20,21]. The polymerization had also been discussed in detail [48]. It has
MAOS region takes advantage of the fact that only the scal- been found that multiblock and polydispersity cause great
ing behavior of the higher harmonic contributions needs to decrement in the critical vN as compared to the monodis-
be considered, and it avoids problems occurring in LAOS persed, diblock (or triblock) copolymers. In this paper, we
such as edge fracture and long time to reach steady state. In will focus our attention on the nonlinear rheological behavior
the MAOS region, a quadratic scaling on strain amplitude of of OBCs under LAOS. With the nonlinear coefficients, we
I3/1 for linear polymer has been observed experimentally systematically investigate nonlinearity from the Fourier-
[20,22,23] and is predicted by different rheological constitu- Transform rheology for OBCs.
tive equations [24–30]. Motivated by such relationship,
Hyun and Wilhelm [21] defined a new nonlinear coefficient
Q3  I3=1 =c20 with c0 the strain amplitude, as well as the II. EXPERIMENTAL SECTION
so-called zero-strain nonlinearity or intrinsic nonlinearity
Q3;0 ¼ limc0 !0 Q3 as the limiting value of Q3 at small shear A. Materials
strain amplitudes. It was noticed that the intrinsic nonlinear- The OBCs synthesized via the chain-shuttling technology
ity Q3,0 as a function of frequency is a sensitive indicator of were provided as pellets by The Dow Chemical Company
relaxation processes in monodisperse linear and comb poly- [51]. Six low octene content OBCs with different block
styrene melts [21], and is more sensitive to detect the effect structure were selected in this work. The detailed informa-
of nanoparticles than the linear viscoelastic properties in tion on molecular structure is given in Table I. The octene
polymer nanocomposites [31–33]. In contrast to the moduli content (C8) and block structure of OBCs were determined
and viscosities defined from Lissajous curves [12,15], Q3 is a by 13C NMR, the molecular weight (Mw) and polydispersity
purely nonlinear parameter and contains no information index (PDI ¼ Mw/Mn) were measured by high temperature
of linear viscoelasticity even when the strain amplitude gel permeation chromatography (GPC). Details about the
becomes zero (i.e., Q3,0). In contrast to the great attentions characterization have been reported in our previous publica-
that have been paid on the intensities at higher harmonic, the tions [48,52]. These six OBCs are identified by the codes
phase angles at higher harmonic were less studied [19, Hx/Mx/Lx, where H/M/L reflects the relatively high/medium/
34–37]. It has been reported that the phase of the harmonics low content of hard block, and the subscripts x are numbers
can indicate shear-thinning/thickening and strain-hardening/ that increase with the total molecular weight. The main
softening response within a cycle [19,38]. differences between the six OBCs are the total hard block
It is well known that block copolymer melts may undergo content and the molecular weight. The difference between
microphase separation due to the incompatibility of distinct the octene content in hard segment and soft segment (DC8) is
blocks. The morphology of the ordered phases can be tuned similar in all six samples, which implies similar interaction
by the block copolymer molecular weight and composition. between hard segment and soft segment in these samples.
Most previous studies of block copolymer melts under Then, the phase separation behavior is mainly determined by
LAOS have focused on the microphase orientation in lamel- the molecular weight and the hard block content.
lar diblock or triblock copolymer melts [39–44]. In these
early works, it has been demonstrated via small angle X-ray B. Measurements
scattering (SAXS) that the lamellar block copolymers melt
can be aligned in different ways under LAOS conditions 1. Rheology
[39–41]. Subsequently, it is straight forward to apply All rheological measurements were performed on OBCs
Fourier-Transform rheology to the study of lamellar align- melts using a strain-controlled rotational rheometer (ARES-
ment of block copolymers to determine the degree of nonlin- G2, TA Instruments). Parallel-plate geometries with diame-
ear response during this process [42–44]. In polydisperse ter of 8, 25, and 40 mm were used. Unless stated otherwise,
multiblock copolymer, the phase behaviors could be quite all measurements were carried out using 25 mm stainless
NONLINEAR RHEOLOGY OF MULTIBLOCK COPOLYMERS 1163

TABLE I. Molecular parameters of OBCs.

Sample code MW (kg/mol) PDI fhard (wt. %) total C8 (mol. %) TMST (  C) Mesophase separation

L01 79.8 1.97 11.0 13.7 163.5 Extremely weak


L02 126.5 2.06 11.0 14.0 >200 Strong
M01 61.7 1.83 25.0 11.7 153.8 Extremely weak
M02 82.6 2.15 25.0 12.0 160.4 Medium
M03 156.6 1.97 25.0 11.5 >200 Very strong
H01 77.6 1.92 35.0 9.6 154.0 Medium

steel parallel-plate fixtures. The disklike samples with thick- In


jGnn jðx; c0 Þ ¼ ; ðn ¼ 1; 3; 5; :::Þ; (4)
ness about 1 mm were made by compression molding at cn0
180  C under a pressure of 10 MPa. In order to evaluate ther-
mal stability of OBCs, dynamic time sweeps were performed and the real and imaginary parts of the nth harmonic
at 1 rad/s. The storage moduli of OBCs stay constant and
deviate from the initial value no more than 5% up to the time G0nn ðx; c0 Þ ¼ jGnn j cos ðdn Þ;
for the subsequent rheological measurements. For M01, it G00nn ðx; c0 Þ ¼ jGnn j sin ðdn Þ; ðn ¼ 1; 3; 5; :::Þ: (5)
can keep thermal stability more than 10 h at 135  C and 2 h
at 180  C. For other OBCs, the duration time of thermal Because of the asymptotic dependence of In=1 on c0n1
stability is longer than M01. LAOS experiments were per- [14,30], the so-called zero-strain nonlinearity or intrinsic
formed on OBCs melts at a fixed frequency with prescribed nonlinearity Q3,0 as the limiting value of Q3 (¼ I3=1 =c20 ) at
strain amplitude within a temperature range of 135–180  C. small shear strain amplitudes has been suggested [21].
Samples were allowed to equilibrate for at least 10 min Similarly, we can also define the intrinsic complex moduli
before collecting data. For the raw data acquisition, the TRIOS jGnn;0 ðxÞj and the intrinsic phase angle dn;0 ðxÞ as the limit-
software (TA Instruments) with a sampling rate of 1024 pts/s ing values of both moduli jGnn j and phase angle dn ,
was used. In order to obtain reliable raw stress data, at least respectively
eight cycles were applied on the samples and only the last
five cycles were collected for a given strain amplitude. The jGnn;0 ðxÞj ¼ lim jGnn ðx; c0 Þj;
Lissajous curves (stress vs strain) of different cycles should c0 !0

be overlapped to ensure that both the stress amplitude and dn;0 ðxÞ ¼ lim dn ðx; c0 Þ; ðn ¼ 1; 3; 5; :::Þ: (6)
c0 !0
the phase angle do not change with time. The stress data as a
function of time were analyzed using Fourier-Transform rhe-
Then, the intrinsic nonlinearity Qn;0 equals to the ratio
ology. All LAOS tests were repeated for at least three times
between jGnn;0 j and the complex moduli jG11;0 j
to ensure the repeatability of the data.
The shear stress under nonlinear oscillatory shear from In=1 jGnn;0 ðxÞj
Fourier-Transform rheology is described as follows [2]: Qn;0 ðxÞ ¼ lim ¼ ; ðn ¼ 3; 5; :::Þ: (7)
c0 !0 cn1
0 jG11;0 ðxÞj
X
1
rðtÞ ¼ I2kþ1 sin ðð2k þ 1Þxt þ d2kþ1 Þ; (1) A typical result for M01 is shown in Fig. 1. The depen-
k¼0 dencies of complex moduli and phase angle at fundamental
frequencies on strain amplitude are shown in Figs. 1(a) and
where I2kþ1 denotes the stress amplitude of the ð2k þ 1Þth 1(b), respectively. Plateaus can be seen clearly at low strain
harmonic, and d2kþ1 is the corresponding phase angle. Only amplitudes, from which the limiting values jG11;0 j and d1;0
the odd harmonics are considered in Eq. (1). The relative are defined. The complex moduli and phase angle at third
intensities (In=1 ) can be defined as follows harmonic versus the relative intensity I3=1 are shown in Figs.
1(c) and 1(d), respectively. Plateaus in moduli and phase
In=1 ¼ In =I1 ; ðn ¼ 3; 5; :::Þ: (2)
angle at small I3=1 can also be seen clearly, which looks simi-
lar to that under fundamental frequency. Then, the limiting
The general expression Eq. (1) can be further expanded
values jG33;0 j and d3;0 can be defined from these plateaus,
as [2]
respectively. The dependence of the complex moduli on I5=1
X
1 X
p at fifth harmonic is shown in Fig. 1(e), from which jG55;0 j
rðtÞ ¼ cp0 ½G0np ðx; c0 Þ sinðnxtÞ can be defined. The limiting value of phase angle under fifth
p¼1 n¼1 harmonic cannot be obtained in the present experiments due
p odd n odd
to low signal-to-noise ratio for the fifth harmonic.
þ G00np ðx; c0 Þ cosðnxtÞ; (3) Due to the inhomogeneous shear deformation in parallel-
plate geometry, the nonlinear quantities determined above
where c0 is the amplitude of strain signal. From the Eqs. (1) are apparent values. For comparison with constitutive mod-
and (3), we indicate here the complex moduli of the nth els, the apparent nonlinear functions have to be converted to
harmonic the true nonlinear function. According to Wagner [53], the
1164 NIE, YU, AND ZHOU

FIG. 1. The dependence of jG11 j (a) and d1 (b) on strain amplitude, the dependence of jG33 j (c) d3 (d) on I3=1 , and the dependence of jG55 j (e) on I5=1 for M01
at 135  C.

true value of Qtrue


n;0 can be connected with the apparent one onto silicon wafers. The thin films were annealed for 1 h at
QApp
n;0 by various temperatures and then quenched into liquid nitrogen.

3 þ n App
Qtrue
n;0 ðxÞ ¼ Qn;0 ðxÞ; ðn ¼ 3; 5; :::Þ: (8)
4
III. NONLINEAR VISCOELASTICITY
Such relation is also applicable for G033;0 ðxÞ, G0033;0 ðxÞ, A. Time-temperature superposition
G055;0 ðxÞ, and G055;0 ðxÞ.
For a wide class of materials such as linear polymer melts
and solutions, changes in temperature do not alter the shape
2. Atomic force microscopy of the linear viscoelastic functions (such as dynamic moduli)
Atomic force microscopy (AFM) experiments were per- as a function of frequency or time. The frequency range of
formed with MultiMode Nanoscope (Digital Instrument, linear viscoelastic functions can then be extended by shifting
Inc.) in a tapping mode. The AFM was operated under ambi- along the frequency or time axis, which is the well-known
ent conditions with a silicon cantilever probe tips. The values time-temperature superposition (TTS) principle [1]. Success
for normal spring constant and the tip radius of curvature are of TTS is also known as the thermorheological simple
15 N m1 and 10 nm, respectively. Specimens for AFM behavior, which usually appears in the absence of heteroge-
imaging were spin-coated from 1 wt. % xylene solution dried neity. In copolymers and polymer blends, failure of TTS is
NONLINEAR RHEOLOGY OF MULTIBLOCK COPOLYMERS 1165

often associated with the appearance of concentration fluctu- OBC samples. The TTS holds successfully for dynamic
ation or phase separation [48,54]. Although there are a lot of moduli (G011;0 and G0011;0 ) in the experimental temperature
reports on the TTS principle based on the linear viscoelastic range. For L01 and M02, which have slightly higher molecu-
properties in polymer multiphase systems [48,54–56], the lar weight than M01, the shifted moduli almost superpose on
study of TTS in nonlinear flow (especially LAOS) is still those of M01. It indicates the success of TTS and similar
rare. relaxation behavior of three samples (M01, M02, and L01).
We first illustrate the TTS behavior in linear viscoelastic- Therefore, it is hard to infer any sign of mesophase transition
ity. The G011;0 and G0011;0 for OBCs under different tempera- in these OBC samples at least from the shifted dynamic mod-
tures are shown in Fig. 2. All the data have been shifted to uli. Actually, the failure of TTS in M02 has been observed in
the reference temperature Tref ¼ 135  C. Because the change a normalized Cole–Cole plot [48], where the weak difference
in the density of OBCs at different temperatures is very in storage moduli at low frequencies could be amplified. As
small, the vertical shift factor bT varies little over the experi- molecular weight increases further, the positive deviation of
mental temperature range and is omitted here. The horizontal G011;0 from those of M01 in the terminal regime becomes
shift factor aT is determined by superposing G011;0 and G0011;0 gradually obvious, which could be ascribed to the additional
at higher frequencies. Moreover, additional horizontal shift contribution from concentration fluctuation (or interface)
based on the molecular weight is adopted to compare all the due to mesophase separation. Moreover, failure of TTS can
OBCs simultaneously. For polymers with high molecular be seen clearly and becomes more obvious as the molecular
weight, the effect of molecular weight (Mw ) is significant weight increases from H01 to L02, and M03. Both the
only in the terminal regime due to the relaxation of whole enhanced moduli and the failure of TTS imply that meso-
chain, where the terminal relaxation time kterminal satisfies phase transition is likely to happen in M03, L02, and H01. It
kterminal  g0  Mwn with g0 , the zero shear viscosity and n, can be concluded that failure of TTS, thus appearance of
the power law exponent. It means that the terminal behavior mesophase separation, can happen by either increasing the
of linear polymers with different molecular weight can be molecular weight (from M01 to M02, and M03 with similar
superposed if the dynamic moduli are shifted along the angu- hard block content) or increasing the hard block content
lar frequency axis using the molecular shift factor (from L01 to M02, to H01 with similar molecular weight)
aM, which is defined as aM ¼ kterminal =kterminal;ref ¼ g0 =g0;ref [48].
¼ ðMw =Mw;ref Þn . But such shifting will not work for the rub- Next, we turn to the TTS in nonlinear regime. In litera-
bery and glassy regimes since they are not affected by the ture, it has been reported that TTS under LAOS works for
length of polymer chains. In this work, the only interested some simple fluids [21,57]. In this work, we only focus on
thing is the terminal behaviors of OBCs, which are due to the MAOS region, where the nonlinear coefficient Q3 has a
the overlapped relaxations of polymer chains and phase- constant value, i.e., the intrinsic nonlinearity Q3,0 [see Eq.
separated domains. The molecular shift factor aM is then (7) and Fig. 1]. Q3,0 of OBCs under different temperatures
adopted to exclude the influence of molecular weight on the are shown in Fig. 3. All the data have been shifted to the ref-
terminal behaviors, which helps to illustrate the difference in erence temperature Tref ¼ 135  C. The horizontal shift factors
the phase separation in OBCs. The power law exponent n is (aT) from linear data are adopted to shift the frequency in
taken as 3.4 for OBCs [48]. The terminal slope of 2 for G011;0 Fig. 3. For homogeneous OBCs such as M01, Q3,0(x) under
and 1 for G0011;0 in the log-log plot can be seen clearly for different temperatures can be superposed perfectly, implying
M01, which has the lowest molecular weight among six that TTS works successfully for Q3,0 in the experimental

FIG. 2. Shifted G011;0 (a) and G0011;0 (b) for OBCs under different temperatures. aT is the temperature shift factor (c) with the reference temperature
Tref ¼ 135  C. The solid line in (c) is the fitting curve by Arrhenius equation. aM is the molecular weight shift factor with the molecular weight of M03 as the
reference molecular weight Mw, ref ¼ Mw, M03.
1166 NIE, YU, AND ZHOU

FIG. 3. Shifted Q3,0 for OBCs under different temperatures.

temperature range. The terminal slope of 2 in Q3,0 is seen


clearly in the log-log plot, which is consistent with the obser-
vations in homopolymers [21] as well as model predictions
[14,30]. For M02 and L01, the changes of Q3,0(x) from that
of M01 are not obvious, although a very weak increment of
Q3,0 at low frequency is seen in L01. Superposition of Q3,0
for M02 and L01 is also acceptable. Therefore, TTS can be
regarded as applicable in the MAOS regime for these sam-
ples. As molecular weight continues to increase, both the
deviation of Q3,0 from that of M01 (or L01) and the failure
of TTS can been found for M03 (or L02). Similarly, as hard
block content increases, clear deviations of Q3,0 from that of
L01 to H01 can also be seen. It should be pointed out that
the degree of deviation from the reference in homogeneous
state (M01 and L01) in Q3,0 is more significant than that in
G011;0 and G0011;0 for all mesophase-separated samples.
Especially for the sample with highest molecular weight
(M03), the deviation in G011;0 is only shoulderlike, but a clear
plateau can be seen in Q3,0, which indicates that the addi-
tional contribution of concentration fluctuation (or interface)
to the intrinsic nonlinearity is much evident than to the linear
moduli. However, the extent of failure of TTS is similar in
both linear dynamic moduli and intrinsic nonlinearity Q3,0.

B. Effect of molecular weight


As shown above, the failure of TTS in Q3,0 or in G011;0 and
G0011;0becomes more and more obvious as the molecular
weight of OBCs increases. However, the TTS behavior
between Q3,0 and G011;0 (or G0011;0 ) is somewhat not compara- FIG. 4. TTS of G033;0 and G0033;0 for OBCs with similar hard block content but
different molecular weights. Positive data are represented by solid symbols,
ble. The phase angle of fundamental frequency makes G011;0 and hollow symbols denote negative value. The reference temperature is
a more sensitive factor for TTS than jG11;0 j, while Q3,0 135  C.
depends only on the ratio between stress amplitudes of dif-
ferent harmonics. It would be interesting if the real and found for these samples. For M01, G033;0 is negative over the
imaginary parts of G33;0 instead of Q3,0 are taken into consid- experimental frequency range, and G0033;0 is positive at high
eration. In fact, these quantities have been rarely studied, frequency and changes to negative at about 1.5 rad/s under
especially their dependencies on oscillatory frequency. the reference temperature. The superposition of jG033;0 j and
Figure 4 shows jG033;0 j and jG0033;0 j for OBCs with similar jG0033;0 j seems good even for those negative G0033;0 data at low
hard block content but different molecular weights (M01, frequencies. For M02, whose molecular weight is higher
M02, M03). Very different frequency dependency can be than M01, similar dependency on frequency is found except
NONLINEAR RHEOLOGY OF MULTIBLOCK COPOLYMERS 1167

that the whole curves are shifted to lower frequency. structures in polymer blends [59]. In the vGP3 plot, the phase
However, the superposition of data under different tempera- angle curves of M01 under different temperatures can still
tures is not as good as that for M01. Clear failures of TTS superpose on each other. However, those of M02 and M03
are seen at high frequency for G033;0 and for G0033;0 at low fre- fail to overlap, and the deviation becomes more distinct as
quency. For M03 with the highest molecular weight, G033;0 the molecular weight increases. For M02, the vGP3 plots at
becomes positive at high frequency and negative at low fre- 165 and 180  C superpose quite well, and deviations appear
quency, indicating the phase angle of third harmonic is quite at 150  C at low jG33;0 j. The temperature range where TTS
different from that of M01 and M02. Moreover, the failure starts to fail is consistent with our previous results [48]. The
of TTS is clearly seen over a wide frequency range below trend of vGP3 for M03 as temperature changes is similar to
about 3 rad/s in G033;0 . that of vGP1, i.e., d3;0 under all temperatures cannot super-
It has been illustrated that G033;0 and G0033;0 are very sensi- pose and decreases as temperature decreases. The high sensi-
tive parameters to justify the TTS principle. It is believed tivity of vGP3 to the failure of TTS makes it a new candidate
that such sensitivity is ascribed to the phase angle. In linear in studying the thermorheological behaviors of complex
viscoelasticity, the plot of phase angle (d1;0 ) against the cor- fluids.
responding absolute value of the complex shear moduli
(jG11;0 j) had been proved to be a sensitive evaluation of TTS C. Effect of hard block content
without the necessity of curves shifting. Such plot is known
as vGP plot (called vGP1 plot thereafter) and has been used Next, we compare the nonlinear behaviors for OBCs with
to evaluate the TTS behavior related to the liquidliquid similar molecular weight but different hard block content,
phase separation in polymer blends [58,59]. Similarly, we i.e., L01, M02, and H01. The TTS of these samples is shown
can plot the phase angle of third harmonic (d3;0 ) against the in Fig. 6 for jG033;0 j and jG0033;0 j. The frequency dependency of
third harmonic complex moduli (jG33;0 j), and name it the G033;0 and G0033;0 for these samples is quite similar due to their
vGP3 plot, while the corresponding plot at fundamental fre- close molecular weight. Well superposition of both G033;0 and
quency is called vGP1 plot. The vGP1 plot (d1;0  jG11;0 j) G0033;0 can be seen in L01, which has the lowest hard block
and vGP3 plot (d3;0  jG33;0 j) of OBCs with similar hard content among them. As the hard block content increases,
block content but different molecular weights are shown in obvious failures of superposition can be found at low fre-
Fig. 5. In the vGP1 plot, the phase angle curves of M01 and quencies for G0033;0 and at high frequencies for G033;0 . The
M02 under different temperatures superpose, indicating that vGP1 plot (d1;0  jG11;0 j) and vGP3 plot (d3;0  jG33;0 j) of
the vGP1 plot is unable to identify the difference between the three OBCs are shown in Fig. 7. Three samples show very
phase behaviors of two samples. The phase angle d1;0 of close vGP1 plot, which implies that the vGP1 plot is hard to
M03 is smaller than those of M01 and M02 under the same distinguish these samples. For L01 and M02, data from dif-
complex moduli, and vGP1 plots under different tempera- ferent temperatures superpose well, and slightly failure of
tures cannot superpose successfully, which implies the fail- TTS is seen in H01. For the nonlinear behavior, the vGP3
ure of TTS. An inflection point can be seen in d1;0 for M03 plot for three samples is well separated, indicating the effect
at about 70 , which has been ascribed to the phase-separated of different hard block contents.

FIG. 5. vGP1 plot (d1;0  jG11;0 j) and vGP3 plot (d3;0  jG33;0 j) of OBCs with similar hard block content but different molecular weights.
1168 NIE, YU, AND ZHOU

distribution, vGP1 plot is irrelevant to the molecular weight


and composition of block copolymer. However, their differ-
ences in vGP3 plot are very clear. Both the increase of
molecular weight and the increase of hard block content can
reduce d3;0 . The decrease of d3;0 can be more significant if
there is mesophase separation. It has been reported that the
vGP1 plot is a sensitive method to detect the structural
change due to the failure of TTS [60,61]. Our results demon-
strate that the vGP3 plot can do the same with even higher
sensitivity. Moreover, its sensitivity to the variation of
molecular weight is also higher, while the effect of molecu-
lar weight in the vGP1 plot is only evident near the plateau
moduli. Furthermore, the sensitivity of the vGP3 plot to the
change of block content in copolymer has not been resolved
by any other rheological methods. In fact, the rather close in
the vGP3 plots of M01 and L01 manifests the role of molecu-
lar weight and hard block content, i.e., M01 has a smaller
molecular weight but higher hard block content as compared
to L01.

IV. THEORETICAL ANALYSIS


A. MSF model for homogeneous melt
As shown above, the most significant characteristic of
OBCs in heterogeneous state is the greatly enhanced intrinsic
nonlinearity Q3,0 as compared to those in homogeneous state
(Fig. 3). To model such behavior, we separate the contribu-
tions from homogeneous matrix and from interfaces, which
are described by the molecular stress function (MSF) model
and simple emulsion model, respectively. One important
thing is that the strain rate varies everywhere in the continu-
ous matrix and inside the dispersed domains. The distribu-
tion of local strain rate has been suggested in different
models. The strain rate distribution in Palierne model [62] is
valid for any viscoelastic components but only in the linear
regime, i.e., under small deformation, which is not applica-
ble to nonlinear oscillatory shear. If the components are
assumed to obey much simpler constitutive law (like
Oldroyd B model), our previous works [63,64] have shown
that such strain rate distribution satisfies a complex integro-
differential equation, which makes the nonlinear analysis
intractable. If further assumption on the Newtonian behavior
of components is made, the strain rate distribution becomes
much simpler [65–67], but definitely very different from
those viscoelastic cases. In all these treatments, the available
constitutive models are unable to give suitable form of local
strain rate for viscoelastic components, which is applicable
FIG. 6. TTS of jG033;0 j and jG0033;0 j for OBCs with similar molecular weight for nonlinear analysis. Therefore, it is necessary to find other
but different hard block contents. Positive data are represented by solid sym-
bols, and hollow symbols denote negative value. The reference temperature
approaches. In this work, we adopted the simplest assump-
is 135  C. tion, i.e., only the average macroscopic strain rate in the
sample is considered, and all the nonlinear effects due to the
The vGP1 plot (d1;0  jG11;0 j) and vGP3 plot (d3;0 local strain rate distributions are ignored. Actually, such
 jG33;0 j) of OBCs at 180  C are compared in Fig. 8. Those assumption gives the zeroth-order approximation on the non-
in the phase-separated regime are shown as solid symbols; linear effect in both Newtonian and viscoelastic systems.
otherwise, they are shown as hollow symbols. In the vGP1 This assumption would be suitable when the nonlinear effect
plot, the phase angle curves of L01, M01, M02, and H01 are is weak, e.g., in the MAOS regime.
quite close. Because these samples are in the homogeneous For homogeneous melt, we adopt the MSF model to
regime of phase diagram and close molecular weight describe the nonlinear behavior, where the approach by
NONLINEAR RHEOLOGY OF MULTIBLOCK COPOLYMERS 1169

FIG. 7. vGP1 plot and vGP3 plot of OBCs with similar molecular weight but different hard block contents.

X gi
Wagner is followed here [53]. In a single integral constitu- m_ ðt  t0 Þ ¼
ð
e tt

=ki
: (10)
tive equation of a homogeneous melt, the shear stress i
ki
rhomo
xy ðtÞ in simple shear can be given by
ðt Sxy ðt; t0 Þ in Eq. (9) is a measure of strain, and can be
expanded to the power of strain according to the MSF model
rhomo
xy ðtÞ ¼ _  t0 ÞSxy ðt; t0 Þdt0 ;
mðt (9)
1
Sxy ¼ f 2 SDE 3 5
xy ¼ c  ða  bÞc þ lc : (11)
_  t0 Þ is the linear viscoelastic memory function,
where mðt
which can be expressed in a discrete relaxation spectrum The coefficient a is related to the molecular orientation,
with relaxation moduli gi and relaxation time ki and b is related to the molecular stretching. High order

FIG. 8. Comparisons of the vGP1 plot and the vGP3 plot of OBCs at 180  C. OBCs in homogeneous state are shown in hollow symbols, and in heterogeneous
state are shown in solid symbols.
1170 NIE, YU, AND ZHOU

coefficient l is related to both processes. In shear flow, a which come from the discrete relaxation spectrum, and are also
sinusoidal shear strain input c ¼ c0 sin ðxtÞ gives known as the generalized Maxwell model. G0 homo
33;0 is obtained as

X 9gi k4i x4 ð1  k2i x2 Þ


cðt; t0 Þ ¼ c0 ½sin ðxtÞ  sin ðxt0 Þ; (12) homo
G0 33;0 ðxÞ ¼ ða  bÞ ;
i ð1 þ k2i x2 Þð1 þ 4k2i x2 Þð1 þ 9k2i x2 Þ
where x is the angular frequency and c0 is the maximum (15)
strain. Substituting Eqs. (11) and (12) into Eq. (9), it is easy which can be expressed in terms of G0 homo
as 11;0
to obtain the leading values of moduli in Eq. (3). The moduli
 
for the fundamental frequency are 0 homo 3 0 homo 0 homo ð 1 0 homo
G 33;0 ðxÞ¼ ðabÞ G 11;0 ðxÞG 11;0 2xÞ þ G 11;0 ð3xÞ :
4 3
homo
X gi x2 k2 (16)
G0 11;0 ðxÞ ¼ i
; (13)
1 þ x 2 k2
i i The same expression holds for G00 homo 0 homo
33;0 when G 11;0 is
00 homo
X gi xki replaced by G 11;0 . Equations (15) and (16) are the same as
homo
G00 11;0 ðxÞ ¼ ; (14) those suggested by Wagner et al. [53]. G0 homo
55;0 can also be
2 2
i 1 þ x ki obtained as

homo
X 225gi k6i x6 ð1  15k2i x2 þ 8k4i x4 Þ
G0 55;0 ðxÞ ¼ l ; (17)
i ð1 þ k2i x2 Þð1 þ 4k2i x2 Þð1 þ 9k2i x2 Þð1 þ 16k2i x2 Þð1 þ 25k2i x2 Þ

which can be expressed in terms of G0 homo


11;0 as considered as a free parameter, a value of b ¼ 0.115 is found
 to fit the data quantitatively. Although the experimental data
homo 5 homo homo of G011;0 , G0011;0 , Q3;0 , and Q5;0 are consistent with the MSF
G0 55;0 ðxÞ ¼ l G0 11;0 ðxÞ  2G0 11;0 ð2xÞ
16 model predictions, the MSF model predictions deviate from

0 homo ð 0 homo ð 1 0 homo the experimental data of jG033;0 j and jG0033;0 j. Similar results
Þ Þ
þ2G 11;0 3x  G 11;0 4x þ G 11;0 5x : ð Þ
5 can be obtained by comparing with the Giesekus model’s
(18) predictions [dashed lines in Figs. 10(a) and 10(b)]. It should
be mentioned that both the strain tensor and the memory
Once again, the same expression holds for G00 homo
55;0 when function in Eq. (9) can be dependent on the strain amplitude,
G0 homo
11;0 is replaced by G00 homo
11;0 . It is noticed that G0 homo
33;0 ,
and it is the general case where time and strain are insepara-
ble. The simplified approach [Eqs. (10) and (11)] adopted by
G 33;0 , G0 homo
00 homo
55;0 , and G00 homo
55;0 can all be expressed as combina- Wagner et al. [49] expands asymptotically the evolution
tions of linear viscoelastic moduli G0 homo 00 homo
11;0 and G 11;0 at angu- equation for chain stretch with strains, while the tube
lar frequencies nx (n ¼ 1  5). But it should be stressed that
these relations connecting linear viscoelastic moduli and non-
linear moduli are model dependent. As an example shown in
the Appendix A, G0 homo 00 homo
33;0 and G 33;0 depends not only on
G0 homo 00 homo
11;0 and G 11;0 at angular frequencies x, 2x, 3x but also
on the derivatives at x in Giesekus model.
According to Wagner [53], MSF model can describe qual-
itatively the general shape of the experimental data. The dif-
ference of the nonlinearity parameters a and b determines
the level of the intrinsic nonlinearity, while its general shape
depends solely on the linear-viscoelastic spectrum as deter-
mined in the SAOS regime. On the basis of the experimen-
tally determined storage moduli and loss moduli, discrete
relaxation spectra with partial moduli and relaxation times of
M01 were obtained using the RSI Orchestrator software
(Fig. 9). In Fig. 10, experimental data and modeling results
of G011;0 , G0011;0 , jG033;0 j, jG0033;0 j, Q3;0 , and Q5;0 are presented
for M01. If the parameter a ¼ 5/21 is kept constant according
to the orientation tensor, and the stretch parameter is FIG. 9. Discrete relaxation spectrum for M01 at 135  C.
NONLINEAR RHEOLOGY OF MULTIBLOCK COPOLYMERS 1171

between the predictions of the MSF model and the experi-


ment data for jG033;0 j and jG0033;0 j implies that neglecting the
tube diameter relaxation may attribute to such deviations.

B. Simple emulsion model for heterogeneous melt


For heterogeneous melt, the phase-separated domains
give additional contribution to the dynamic moduli. As
compared to the macro-phase-separated polymer blends,
modeling the rheology of microphase (or mesophase)-
separated block copolymers is still a challenge. To sim-
plify the problem, the morphology of phase-separated
OBCs is assumed to have the droplet-matrix morphology,
which can be described by the droplet radius R and the
interfacial tension C. Such assumption can be validated by
direct observation on the phase-separated domains of
OBCs (see below for details). Then the emulsion rheologi-
cal models may be used to describe the nonlinear behavior
of heterogeneous melt.
Preliminary work has been done [68] using the emulsion
constitutive model [69]. Under LAOS flow, the equilibrium
shape of a spherical domain is assumed to be an ellipsoid
instead of a sphere under SAOS. The analysis starts from the
expansion of the normalized morphology tensor M ~ around a
spherical shape

~ ¼ d þ Ca0 ½M1 cos ðxst Þ þ M2 sin ðxst Þ


M
þ Ca20 ½M0 þ M3 cos ð2xst Þ þ M4 sin ð2xst Þ
þ Ca30 ½M5 cos ðxst Þ þ M6 sin ðxst Þ
þ M7 ðxÞ cos ð3xst Þ þ M8 ðxÞ sin ð3xst Þ
þ Ca40 ½M9 þ M10 ðxÞ cos ð2xst Þ þ M11 cos ð2xst Þ
þ M12 cos ð4xst Þ þ M13 sin ð4xst Þ
þ Ca50 ½M14 cos ðxst Þ þ M15 sin ðxst Þ
þ M16 cos ð3xst Þ þ M17 sin ð3xst Þ
þM18 cos ð5xst Þ þ M19 sin ð5xst Þ þ OðCa60 Þ; (19)

where d is the unit tensor representing the shape of a sphere,


Ca0 is the capillary number defined as Ca0 ¼ xsc0 with the
emulsion time s ¼ gm R=C. gm is the shear viscosity of
matrix. Mi (i ¼ 0  19) are symmetric tensors. For M1 , M2 ,
M5  M8 , and M14  M19 , the only nonzero component is
the 12 (and 21) component. For other Mi , diagonal compo-
nents are the only nonzero components. Yu et al. [68] only
expand the morphology tensor M ~ to the second order of Ca0 ,
which underestimates the Q3;0 (for details, see Appendix B).
Substituting such expansion [Eq. (19)] into the evolution
FIG. 10. Comparisons between experimental data and model predictions on ~ [Eq. (B4)], and letting the coefficient of the
equation of M
G011;0 , G0011;0 , jG033;0 j, jG0033;0 j, Q3;0 , and Q5;0 for M01 at 135  C. Experimental
jG033;0 j, jG0033;0 j, and Q3;0 are multiplied by a factor of 1.5, and Q5;0 are multi- corresponding trigonometric functions at the same order of
plied by a factor of 2 according to Eq. (8). The solid lines are predictions of Ca0 at the both side of equation, we can obtain 20 algebraic
MSF model with a ¼ 5=21, b ¼ 0:115, and l ¼ 0:007. The dash lines are equations which can be solved analytically to determine Mi
predictions of Giesekus model [Eq. (A10)] with ag ¼ 0:5.
(i ¼ 0  19). Then the determined expansion of morphology
diameter relaxation at higher strain is neglected. They dem- tensor [Eq. (19)] is substituted into the interfacial stress
onstrated that model predictions agree quite well with the equation [Eq. (B2)] to obtain the interfacial contribution of
experiments in Q3;0 for linear polystyrene [49], which is also dynamic moduli. The interface contribution to linear visco-
found in our work [Fig. 10(b)]. However, the inconsistency elastic moduli can be expressed
1172 NIE, YU, AND ZHOU

int 2Kx2 s2 f 2 for the interfacial contribution. Since M03, L02, and H01
G0 11;0 ¼  2 2 2 2  ; (20) samples show significant deviation from the homogeneous
3 x s þ f1
state both in linear and nonlinear behaviors (Figs. 2 and 3),
int 2Kxsf1 f 2 comparisons between the above combined model and experi-
G00 11;0 ¼  2 2 2 2  : (21) ments will be made only on these samples. In Fig. 11, we
3 x s þ f1

These expressions are the same as those obtained by Yu


et al. [69], where the coefficients K, f1 , and f2 are given in
the Appendix B. The interfacial contributions of the dynamic
moduli at the third harmonic are
    
int Kx4 s4 f22 f14 6þf22 17x2 s2 f12 f22 þ6x4 s4 1þf22
G0 33;0 ¼  2    ;
6 x2 s2 þf12 4x2 s2 þf12 9x2 s2 þf12
(22)

int Kx3 s3 f1 f22½f14 þx2 s2 f12ð107f22Þþx4 s4ð11þ17f22Þ


G00 33;0 ¼ ;
6ðx2 s2 þf12 Þ2 ð4x2 s2 þf12 Þð9x2 s2 þf12 Þ
(23)

which can be rewritten in terms of linear viscoelastic moduli


as

int
int 2  f22 0 int f 2 dG0 11;0 ðxÞ
G0 33;0 ðxÞ ¼ G 11;0 ðxÞ  2
16 8 d ln x
1  f22 0 int 2  f22 0 int
 G 11;0 ð2xÞ þ G 11;0 ð3xÞ: (24)
8 48

Equation (24) also holds for G00 int 0 int


33;0 ðxÞ when G 11;0 ðxÞ is
int
replaced by G00 11;0 ðxÞ. The interfacial moduli at the fifth har-
monic are also given in Appendix B.
For heterogeneous melt, the total dynamic moduli can be
written as the sum of the interfacial contribution and that
from the homogeneous contribution

int homo
G0nn;0 ðxÞ ¼ G0 nn;0 ðxÞ þ G0 nn;0 ðxÞ; ðn ¼ 1; 3; 5; :::Þ; (25)

int homo
G00nn;0 ðxÞ ¼ G00 nn;0 ðxÞ þ G00 nn;0 ðxÞ; ðn ¼ 1; 3; 5; :::Þ: (26)

The moduli separation in the above equations [Eqs. (25)


and (26)] has been widely used in polymer blends [69,70].
Actually, for monodispersed diblock or triblock copolymer,
the relaxation of individual chains is fully coupled with the
relaxation of concentration fluctuation, while the two relaxa-
tion processes can be well separated in polymer blends.
Because of the polydispersed multiblock architecture, the
phase behaviors of OBCs could be quite different from the
traditional monodispersed systems. The domain spacing
(102–158 nm) in polydispersed diblock or multiblock OBCs
is three to five times larger than that in equivalent monodis-
persed systems [45,46]. Moreover, the sharp drop in G0 at the
mesophase transition temperature for monodispersed block
copolymer is lost for OBCs, while a gradual transition is
observed [48]. These observations indicate that the relaxation
behavior in polydispersed multiblock copolymer is more like FIG. 11. Comparisons between experimental data and model predictions on
G011;0 ðxÞ, jG33;0 ðxÞj, and jG55;0 ðxÞj. Experimental data of jG33;0 ðxÞj and
polymer blends instead of monodispersed block copolymer.
jG55;0 ðxÞj have been multiplied by 1.5 and 2, respectively. The parameters
The MSF model will be adopted for the homogeneous for simple emulsion model are C ¼ 0:05 mN=m and p ¼ 1. Drop size distri-
contribution, and the simple emulsion model will be adopted butions are shown in Fig. 12.
NONLINEAR RHEOLOGY OF MULTIBLOCK COPOLYMERS 1173

compare the model predictions on G011;0 , jG33;0 j, and jG55;0 j


with the experimental data. The interfacial tension between
phase-separated domains is taken as C ¼ 0:05 mN=m, which
is a quite low value implying the good affinity between the
hard block and the soft block. The viscosity ratio is taken as
p ¼ 1. The total volume fraction of phase-separated domains
is calculated from the weight fraction of hard block content
(Table I), where the densities of hard block rich domain and
soft block rich domain are taken as 0.92 and 0.87 g/cm3,
respectively. It is seen in Fig. 11 that the model can predict the
elastic moduli quite well for H01 and L02, where the deviation
from the homogeneous state (M01) is weak. Good agreement
can also be found in jG33;0 j for H01 and L02. The model pre-
dictions are somewhat smaller than experiments in both G011;0
and jG33;0 j for M03. It might be attributed to the high concen-
tration of phase-separated domains in M03. The simple emul-
sion model considers only the deformation of spherical
domains and neglects the interactions among them [71]. For FIG. 12. Domain size distributions of M03, L02, and H01 that have been
the moduli at fifth harmonic jG55;0 j, quantitative agreements adopted in the calculations in Fig. 11.
between experiments and model predictions are not satisfac-
tory. However, the predictions of the combined model illustrate and the larger cylindrical domains are formed due to inevita-
a clear trend, i.e., the deviation of moduli from the homoge- ble crystallization during quench which connects the spheri-
neous state becomes more and more evident at higher har- cal domains. It is clear that the domain size grows slightly as
monic. Such trend is consistent with the experimental
observations. Moreover, the agreement between the predictions
on jG33;0 j in phase-separated system and in homogeneous sys-
tem at high frequency implies that the calculation using the
average strain rate is reasonable at least in MAOS regime.
One important parameter in the model is the volume
fraction of phase-separated domains. For M03 and L02, it is
assumed that all hard blocks are phase separated from the
soft blocks since the reference temperature 135  C is much
lower than the mesophase separation temperatures of them
(>200  C, Table I, [48]). For H01, since the reference tem-
perature is just below its phase-separation temperature, it is
expected that only a part of hard blocks separate and then the
volume fraction in emulsion model is taken as a fitting
parameter. The best fit for H01 as shown in Fig. 11 assumes
that 10% of total hard block forms phase-separated domains.
As a result of model prediction, only partial phase separation
happens in H01, which implies further annealing or shear
flow may affect the process of mesophase separation of H01,
as it does in partially miscible polymer blends [72]. The
effect of flow on the phase separation of multiblock copoly-
mers will be discussed in our next work.
In the predictions, it is also assumed that the domain size
is not monodispersed. Discrete distributions are adopted to
make the model predictions close to the experimental results,
and the domain size distribution is shown in Fig. 12. The vol-
ume average radius of domains from Fig. 12 is 27.1, 23.9,
and 31.3 nm for H01, L02 and M03, respectively. To justify
these values of domain size, the morphology of OBCs under
different annealing temperatures was studied by AFM, and
some of M01 and M03 are shown in Fig. 13. It is seen that
M01 is homogeneous under 135 and 180  C, which is consis-
tent with rheological results. For M03, the phase-separated FIG. 13. The AFM phase diagrams of OBCs over a scan size of 1  1 lm.
The samples were annealed for 1 h after removing thermal history at 200  C
domains are clearly seen, where both spherical and cylindri- for 5 min, followed by rapidly quenching into liquid nitrogen to suppress the
cal domains can be observed. It is believed that the smaller crystallization. (a) M01 at 135  C, (b) M01 at 180  C, (c) M03 at 135  C, (d)
spherical domains are the original hard block-rich domains, M03 at 150  C, (e) M03 at 165  C, (f) M03 at 180  C.
1174 NIE, YU, AND ZHOU

the annealing temperature decreases. Such phenomenon is National Basic Research Program of China (973 Program
due to the increasing quenching depth as temperature No. 2011CB606005).
decreases. Moreover, it is also found that the domain size is
quite nonuniform, ranging from about 9 to about 44 nm.
APPENDIX A: ASYMPTOTIC ANALYSIS
Such observation makes it reasonable to choose the domain
OF GIESEKUS MODEL
size distribution in the model.
In the multimode Giesekus model, the stress is the sum of
V. CONCLUSION the contributions from each mode
The nonlinear response of OBC melts has been investigated X
r¼ rk : (A1)
under LAOS using Fourier-Transform rheology. For homoge-
k
neous OBCs such as M01, Q3,0(x) under different tempera-
tures can be superposed perfectly, implying that TTS works The stress tensor rk in each mode satisfies
successfully for Q3,0 in the experimental temperature range.
The terminal slope of 2 in Q3,0 is seen clearly in the log-log r 1 ag
plot, which displays a relaxation process of disentanglement of rk þ rk þ rk rk ¼ 2gk D; (A2)
kk kk gk
polymer chains such as linear viscoelastic properties. For
mesophase-separated OBCs such as M03, both the deviation where the coefficient ag is assumed to be identical in all
from typical terminal behavior and the failure of TTS can been modes. Normalizing the time by the amplitude of shear rate
found. The degree of the deviation from the reference of c_ 0 ¼ c0 x and the stress by gk , Eq. (A2) becomes
homogeneous state in Q3,0 is more significant than that in G011;0
and G0011;0 for all mesophase-separated samples. r 1 ag
~k þ
r ~k þ
r ~k r
r ~
~ k ¼ 2D; (A3)
We also suggested a new plot, the vGP3 plot, i.e., the Wik Wik
intrinsic phase angle at third harmonic versus the intrinsic
complex moduli at third harmonic. The high sensitivity of where Wik ¼ kk c_ 0 ¼ kk c0 x is the Weissenberg number of
vGP3 to the failure of TTS makes it a new candidate in the kth mode. The stress can be expanded as
studying the thermorheological behaviors of complex fluids. 0 1 0 1
Both the increase of molecular weight and the increase of 0 r12;1 0 r11;2 0 0
hard block content can reduce d3;0 . The decrease of d3;0 can B C B C
~ k ¼ Wik @ r12;1
r 0 0 A þ Wi2k @ 0 r22;2 0 A
be more significant if there is mesophase separation. Using
0 0 0 0 0 r33;2
the vGP3 plot, the weaker mesophase separation can be suc- 0 1
cessfully distinguished from the homogeneous state. 0 r12;3 0
B C
The nonlinear response of homogeneous OBCs and hetero- þ Wi3k @ r12;3 0 0 A; ðA4Þ
geneous OBCs could be described by combining the MSF 0 0 0
model and the emulsion model. For homogeneous OBCs,
although the experimental data of G011;0 , G0011;0 , Q3;0 , and Q5;0 where
are consistent with the MSF model predictions, the MSF
model predictions deviate somewhat from the experiment data r12;1 ¼ a1 sin ð~t =c0 Þ þ b1 cos ð~t =c0 Þ; (A5)
of jG033;0 j and jG0033;0 j. For heterogeneous melt, the MSF model
was adopted for the homogeneous contribution, and the simple rii;2 ¼ ci þ di sin ð2~t =c0 Þ þ ei cos ð2~t =c0 Þ; i ¼ 1; 2; 3 (A6)
emulsion model was adopted for the interfacial contribution.
The predictions of the combined model illustrate a clear trend, r12;3 ¼ a3 sin ð~t =c0 Þ þ b3 cos ð~t =c0 Þ þ f3 sin ð3~t =c0 Þ
i.e., the deviation of moduli from the homogeneous state
þ g3 cos ð3~t =c0 Þ: (A7)
becomes more and more evident at higher harmonic. Such
trend is consistent with the experimental observations.
The coefficient in Eqs. (A5)–(A7) can be solved by
substituting Eq. (A4) into Eq. (A3). Then, dynamic moduli
ACKNOWLEDGMENTS
can be obtained from the stress expansion [Eq. (A4)]. The
The authors thank the support from the National Natural moduli of the fundamental frequency are the same as Eqs.
Science Foundation of China (No. 21474063) and the (12) and (13). The dynamic moduli at the third harmonic are

X ag gk k4 x4 ð16ag  21 þ ð30  68ag Þk2 x2 þ 3ð17 þ 4ag Þk4 x4 Þ


G033;0 ¼ k k k
; (A8)
k 4ð1 þ k2k x2 Þ2 ð1 þ 4k2k x2 Þð1 þ 9k2k x2 Þ

X ag gk k3 x3 ð2ag  3 þ 48ð1  ag Þk2 x2 þ ð33 þ 46ag Þk4 x4  18k6 x6 Þ


G0033;0 ¼ k k k k
: (A9)
k 4ð1 þ k2k x2 Þ3 ð1 þ 4k2k x2 Þð1 þ 2 2
9kk x Þ
NONLINEAR RHEOLOGY OF MULTIBLOCK COPOLYMERS 1175

Eq. (A8) can be rewritten in terms of G011;0 as



9 0 1 dG011;0 ðxÞ 3 0 3 0
G033;0 ðxÞ ¼ ag G ðx Þ þ  G11;0 ð2xÞ þ G11;0 ð3xÞ
16 11;0 8 d ln x 4 16
 0 2 0

2 13 0 1 dG 11;0 ð x Þ 1 d G 11;0 ð x Þ 0 3 0
þ ag  G11;0 ðxÞ   þ G11;0 ð2xÞ  G11;0 ð3xÞ : (A10)
16 2 d ln x 8 d ln x2 16

The same expression holds for G0033;0 when G011;0 is replaced by G0011;0 .

APPENDIX B: ASYMPTOTIC ANALYSIS OF EMULSION MODEL


Yu et al. [68] have suggested an emulsion model, which includes a time evolution equation of morphology tensor M [Eq.
(B1)] and a stress expression [Eq. (B2)], which depends on M

@M f1 3I3
¼ f 2 ðM D þ D M Þ þ ðX M  M X Þ  M d ; (B1)
@t s I2

2f2 K 2
rint ¼ I1 M  M M  I2 d : (B2)
I2 3

I2 and I3 are the second and third scalar invariants of M, D and X are the deformation rate and vorticity tensors, respec-
tively, and d is the unit tensor. The coefficients in these equations are expressed as

40ð p þ 1Þ 5 6C ð p þ 1Þð 2p þ 3Þu


f1 ¼ ; f2 ¼ ; K¼ ; (B3)
ð 2p þ 3Þð 19p þ 16Þ 2p þ 3 5R 5ð p þ 1Þ  ð 5p þ 2Þu

where p is the ratio between the viscosity of droplet phase (gd ) and that of matrix phase (gm ), u is the volume fraction of dis-
persed phase. Normalizing the morphology tensor by R2 , time by the emulsion time s, the deformation rate and vorticity ten-
sors by shear rate amplitude c_ 0 , we obtain the dimensionless evolution equation [69]

~ 
@M ~ ~ ~ ~ ð ~ ~ ~ ~ Þ ~ 3I3
ð Þ
¼ Ca0 f2 M D þ D M þ Ca0 X M  M X  f1 M  d : (B4)
@~t I2

Substituting the expansion of morphology tensor [Eq. (19)] into Eq. (B4), we can obtain a series equations from the coeffi-
cients of the corresponding trigonometric functions at the same order of Ca0 . For example, to the first order of Ca0 , we have

M1 xs ¼ M2 f1 ; (B5)

M2 xs ¼ M1 f1 þ 2Df2 : (B6)

To the second order of Ca0 , we have

1 1 1
2M3 xs ¼ ðD M2 þ M2 DÞf2 þ ðX M2  M2 XÞ  M4 f1 þ ½tr ðM1 M2 Þ  4tr ðM4 Þf1 I; (B7)
2 2 6

1 1 1  
2M4 xs ¼ ðD M1 þ M1 DÞf2 þ ðX M1  M1 XÞ  M3 f1 þ tr ðM1 M1 Þ  tr ðM2 M2 Þ  8tr ðM3 Þ f1 I; (B8)
2 2 12

1 1 1  
0 ¼ ðD M1 þ M1 DÞf2 þ ðX M1  M1 XÞ  M0 f1 þ tr ðM1 M1 Þ þ tr ðM2 M2 Þ  8tr ðM0 Þ f1 I; (B9)
2 2 12

where trðÞ denotes the trace of a tensor. To the third order of Ca0 , we have

1 1
M5 xs ¼ ðD M4 þ M4 DÞf2 þ ðX M4  M4 XÞ  M6 f1 ; (B10)
2 2

1 1
M6 xs ¼ ðD M3 þ M3 D þ 2D M0 þ 2M0 DÞf2 þ ðX M3  M3 X þ 2X M0  2M0 XÞ  M5 f1 ; (B11)
2 2
1176 NIE, YU, AND ZHOU

1 1
M7 xs ¼ ðD M4 þ M4 DÞf2 þ ðX M4  M4 XÞ  M8 f1 ; (B12)
2 2
1 1
M8 xs ¼ ðD M3 þ M3 DÞf2 þ ðX M3  M3 XÞ  M7 f1 : (B13)
2 2

Similar equations can be obtained up to the fifth order of Ca0 . Solving Eqs. (B5)–(B13) together with those at the
fourth order and fifth order of Ca0 , All Mi in Eq. (19) can be solved. The nonzero components of Mi ði ¼ 0  4Þ have
been given by Yu et al. [69]. Those components of Mi ði ¼ 5  8Þ, which are relevant to the third harmonic, are written
as
      
f1 f2 f14 9 þ 6f22 þ x2 s2 f12 18 þ 17f22  x4 s4 9 þ f22
m5;12 ¼  3   ; (B14)
12 x2 s2 þ f12 4x2 s2 þ f12

      
xsf2 f14 18 þ 11f22 þ 3x2 s2 f12 12 þ 11f22 þ 2x4 s4 9 þ 5f22
m6;12 ¼  3   ; (B15)
12 x2 s2 þ f12 4x2 s2 þ f12

      
f1 f2 f14 3 þ 2f22 þ x2 s2 f12 30  13f22 þ 3x4 s4 11  13f22
m7;12 ¼  2    ; (B16)
12 x2 s2 þ f12 4x2 s2 þ f12 9x2 s2 þ f12

    
xsf2 f14 18 þ 11f22 þ 17x2 s2 f12 f22  18x4 s4 1 þ f22
m8;12 ¼  2    : (B17)
12 x2 s2 þ f12 4x2 s2 þ f12 9x2 s2 þ f12

The nonzero components of Mi in higher order expansions ði ¼ 9  19Þ will not be listed here due to the lengthy expres-
sions. Then, the stress tensor can be obtained by substituting the expansion Eq. (19) into Eq. (B2). G0 int 00 int 0 int
11;0 , G 11;0 , G 33;0 , and
00 int 0 int 00 int
G 33;0 are shown in Eqs. (20)–(23). G 55;0 and G 55;0 can be written as

int Kx6 s6 f22  14      


G0 55;0 ¼ 3f1 45  17f22 þ 3f24 þ x2 s2 f112 945 þ 3987f22  1138f24  6x4 s4 f110 2025  4773f22 þ 505f24
216z
   
þ 2x6 s6 f18 16065  46329f22 þ 49814f24 þ x8 s8 f16 30645  374607f22 þ 206621f24
     
 9x10 s10 f14 405 þ 38041f22 þ 48130f24 20x12 s12 f12 513 þ 2523f22 þ 5242f24 þ 1440x14 s14 3 þ 24f22 þ 13f24 ;
(B18)

int Kx5 s5 f1 f22  14    


G00 55;0 ¼ 9f1  3x2 s2 f112 231  258f22 þ 55f24 þ 2x4 s4 f110 1728  2340f22 þ 1525f24
216z
   
þ 2x6 s6 f18 1575  54414f22 þ 20443f24  9x8 s8 f16 3665 þ 19128f22 þ 6068f24
   
þ x10 s10 f14 56079 þ 78102f22  486625f24 þ 2x12 s12 f12 19431 þ 132696f22 þ 52561f24
 
þ 72x14 s14 137 þ 1646f22 þ 1177f24 ; (B19)

z ¼ ðx2 s2 þ f12 Þ5 ð4x2 s2 þ f12 Þ2 ð9x2 s2 þ f12 Þð16x2 s2 þ f12 Þð25x2 s2 þ f12 Þ: (B20)

If the components are purely Newtonian fluid, which do not contribute to the nonlinearity of blends, the nonlinear
behavior of sample is solely determined by the interface. The interfacial dominated nonlinear parameters can be expressed
as

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

2 2
jGint G0 int þ G00 int
33;0 j
33;0 33;0
Qint
3;0 ¼ ¼ r
ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
; (B21)
jGint
11;0 j 2 2
G0 int
11;0 þ G00 int
11;0
NONLINEAR RHEOLOGY OF MULTIBLOCK COPOLYMERS 1177

References
[1] Ferry, J. D., Viscoelastic Properties of Polymers (Wiley, New York,
1980).
[2] Dealy, J. M., and K. F. Wissbrun, Melt Rheology and Its Role in
Plastics Processing: Theory and Applications (Chapman and Hall,
New York, 1990).
[3] Giacomin, A. J., and J. M. Dealy, Techniques in Rheological
Measurement (Chapman and Hall, London, 1993).
[4] Hyun, K., S. H. Kim, K. H. Ahn, and S. J. Lee, “Large amplitude oscil-
latory shear as a way to classify the complex fluids,” J. Non-
Newtonian Fluid Mech. 107, 51–65 (2002).
[5] Hyun, K., M. Wilhelm, C. O. Klein, K. S. Cho, J. G. Nam, K. H. Ahn,
S. J. Lee, R. H. Ewoldt, and G. H. McKinley, “A review of nonlinear
oscillatory shear tests: Analysis and application of large amplitude
oscillatory shear (LAOS),” Prog. Polym. Sci. 36, 1697–1753 (2011).
FIG. 14. Comparisons between the prediction of dynamic moduli from Eqs. [6] Neidh€ ofer, T., S. Sioula, N. Hadjichristidis, and M. Wilhelm,
(B1) to (B2) using numerical solutions and the analytical solutions, Eqs. “Distinguishing linear from star-branched polystyrene solutions with
(20)–(23) and (B18)–(B19). The model parameters are: gd ¼ 1000 Pa s,
Fourier-Transform Rheology,” Macromol. Rapid Commun. 25,
gm ¼ 2000 Pa s, R ¼ 1 lm, C ¼ 2 mN=m, u ¼ 0:1.
1921–1926 (2004).
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi [7] Li, W. H., H. Du, G. Chen, S. H. Yeo, and N. Guo, “Nonlinear visco-
2 2 elastic properties of MR fluids under large amplitude oscillatory
int G 0 int þ G 00 int
jG55;0 j 55;0 55;0 shear,” Rheol. Acta 42, 280–286 (2003).
Qint
5;0 ¼ int ¼ r ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2 : (B22) [8] Debbaut, B., and H. Burhin, “Large amplitude oscillatory shear and
jG11;0 j int
2
int
0
G 11;0 þ G 11;0 00
Fourier-transform rheology for a high-density polyethylene: Experiments
and numerical simulation,” J. Rheol. 46, 1155–1176 (2002).
[9] Kallus, S., N. Willenbacher, S. Kirsch, D. Distler, T. Neidh€ ofer, M.
Validations of the analytical solution are shown in Figs.
Wilhelm, and H. W. Spiess, “Characterization of polymer dispersions
14 and 15. Nice agreement between analytical solution and
by Fourier transform rheology,” Rheol. Acta 40, 552–559 (2001).
numerical solution is seen in Fig. 14 for all dynamic moduli
[10] Chopra, D., D. Vlassopoulos, and S. G. Hatzikiriakos, “Nonlinear rhe-
except that slight overestimations on jG0 int 00 int
55;0 j and jG 55;0 j can ological response of phase separating polymer blends: Poly(styreneco-
be seen at high frequency. The agreement between numerical
maleic anhydride)/poly(methyl methacrylate),” J. Rheol. 44, 27–45
solution and analytical solution is also excellent for Qint 3;0 . (2000).
However, the prediction of Yu et al. [68] underestimates
[11] Hyun, K., J. G. Nam, M. Wilhellm, K. H. Ahn, and S. J. Lee, “Large
Qint
3;0 over the whole frequency range, which is due to that the amplitude oscillatory shear behavior of PEO-PPO-PEO triblock copol-
morphology tensor was only expanded to the second order of ymer solutions,” Rheol. Acta 45, 239–249 (2006).
Ca0 . The analytical prediction of Qint5;0 is quite accurate at [12] Ewoldt, R. H., A. E. Hosoi, and G. H. McKinley, “New measures for
low frequency, but weakly underestimates at high frequency characterizing nonlinear viscoelasticity in large amplitude oscillatory
plateau regime. shear,” J. Rheol. 52, 1427–1458 (2008).
[13] Cho, K. S., K. Hyun, K. H. Ahn, and S. J. Lee, “A geometrical inter-
pretation of large amplitude oscillatory shear response,” J. Rheol. 49,
747–758 (2005).
[14] Yu, W., P. Wang, and C. Zhou, “General stress decomposition in non-
linear oscillatory shear flow,” J. Rheol. 53, 215–238 (2009).
[15] Yu, W., Y. Du, and C. Zhou, “A geometric average interpretation on
the nonlinear oscillatory shear,” J. Rheol. 57, 1147–1175 (2013).
[16] Wilhelm, M., “Fourier-transform rheology,” Macromol. Mater. Eng.
287, 83–105 (2002).
[17] Wilhelm, M., P. Reinheimer, and M. Ortseifer, “High sensitivity
Fourier-transform rheology,” Rheol. Acta 38, 349–356 (1999).
[18] Wilhelm, M., D. Maring, and H. W. Spiess, “Fourier-transform
rheology,” Rheol. Acta 37, 399–405 (1998).
[19] Neidh€ ofer, T., M. Wilhelm, and B. Debbaut, “Fourier-transform rheol-
ogy experiments and finite-element simulations on linear polystyrene
solutions,” J. Rheol. 47, 1351–1371 (2003).
[20] Hyun, K., E. S. Baik, K. H. Ahn, S. J. Lee, M. Sugimoto, and K.
Koyama, “Fourier-transform rheology under medium amplitude oscil-
latory shear for linear and branched polymer melts,” J. Rheol. 51,
FIG. 15. Comparisons between the prediction of Qintn;0 from Eqs. (B1) to
1319–1342 (2007).
(B2) using numerical solutions and the analytical solutions, Eqs. (B20)
and (B21). The prediction on Q3,0 of Yu et al. [69] is shown as the dash [21] Hyun, K., and M. Wilhelm, “Establishing a new mechanical nonlinear
line. coefficient Q from FT-Rheology: First investigation of entangled
1178 NIE, YU, AND ZHOU

linear and comb polymer model systems,” Macromolecules 42, alignment in a lamellar diblock copolymer: Interplay of frequency,
411–422 (2009). strain amplitude, and temperature,” Macromolecules 29, 875–884
[22] Hyun, K., K. Ahn, S. Lee, M. Sugimoto, and K. Koyama, “Degree of (1996).
branching of polypropylene measured from Fourier-transform [41] Wiesner, U., “Lamellar diblock copolymers under large amplitude
rheology,” Rheol. Acta 46, 123–129 (2006). oscillatory shear flow: Order and dynamics,” Macromol. Chem. Phys.
[23] Abbasi, M., N. G. Ebrahimi, and M. Wilhelm, “Investigation of the 198, 3319–3352 (1997).
rheological behavior of industrial tubular and autoclave LDPEs under [42] Oelschlaeger, C., J. S. Gutmann, M. Wolkenhauer, H.-W. Spiess, K.
SAOS, LAOS, transient shear, and elongational flows compared with Knoll, and M. Wilhelm, “Kinetics of shear microphase orientation and
predictions from the MSF theory,” J. Rheol. 57, 1693–1714 (2013). reorientation in lamellar diblock and triblock copolymer melts as
[24] Pearson, D. S., and W. E. Rochefort, “Behavior of concentrated poly- detected via FT-Rheology and 2D-SAXS,” Macromol. Chem. Phys.
styrene solutions in large amplitude oscillating shear fields,” J. Polym. 208, 1719–1729 (2007).
Sci., Part B: Polym. Phys. 20, 83–98 (1982). [43] Meins, T., N. Dingenouts, J. K€ ubel, and M. Wilhelm, “In situ
[25] Sim, H. G., K. H. Ahn, and S. J. Lee, “Three-dimensional dynamics rheodielectric, ex situ 2D-SAXS, and Fourier transform rheology
simulation of electrorheological fluids under large amplitude oscilla- investigations of the shear-induced alignment of poly(styrene-b-1,
tory shear flow,” J. Rheol. 47, 879–895 (2003). 4-isoprene) diblock copolymer melts,” Macromolecules 45,
[26] Nam, J. G., K. Hyun, K. H. Ahn, and S. J. Lee, “Prediction of normal 7206–7219 (2012).
stresses under large amplitude oscillatory shear flow,” J. Non- [44] Meins, T., K. Hyun, N. Dingenouts, M. Fotouhi Ardakani, B. Struth,
Newtonian Fluid Mech. 150, 1–10 (2008). and M. Wilhelm, “New insight to the mechanism of the shear-induced
[27] Liu, J., W. Yu, W. Zhou, and C. Zhou, “Control on the topological macroscopic alignment of diblock copolymer melts by a unique and
structure of polyolefin elastomer by reactive processing,” Polymer 50, newly developed Rheo-SAXS combination,” Macromolecules 45,
547–552 (2009). 455–472 (2012).
[28] Hyun, K., W. Kim, S. J. Park, and M. Wilhelm, “Numerical simulation [45] Li, S., R. A. Register, B. G. Landes, P. D. Hustad, and J. D. Weinhold,
results of the nonlinear coefficient Q from FT-Rheology using a single “Crystallization in ordered polydisperse polyolefin diblock copoly-
mode pom-pom model,” J. Rheol. 57, 1–25 (2013). mers,” Macromolecules 43, 4761–4770 (2010).
[29] Bharadwaj, N. A., and R. H. Ewoldt, “The general low-frequency pre- [46] Li, S., R. A. Register, J. D. Weinhold, and B. G. Landes, “Melt and
diction for asymptotically nonlinear material functions in oscillatory solid-state structures of polydisperse polyolefin multiblock copoly-
shear,” J. Rheol. 58, 891–910 (2014). mers,” Macromolecules 45, 5773–5781 (2012).
[30] Bharadwaj, N. A., and R. H. Ewoldt, “Constitutive model fingerprints [47] Park, H. E., J. M. Dealy, G. R. Marchand, J. Wang, S. Li, and R. A.
in medium amplitude oscillatory shear,” J. Rheol. 59, 557–592 (2015). Register, “Rheology and structure of molten olefin multiblock copoly-
[31] Lim, H. T., K. H. Ahn, J. S. Hong, and K. Hyun, “Nonlinear viscoelas- mers,” Macromolecules 43, 6789–6799 (2010).
ticity of polymer nanocomposites under large amplitude oscillatory [48] He, P., W. Shen, W. Yu, and C. Zhou, “Mesophase separation and rhe-
shear flow,” J. Rheol. 57, 767–789 (2013). ology of olefin multiblock copolymers,” Macromolecules 47, 807–820
[32] Salehiyan, R., Y. Yoo, W. J. Choi, and K. Hyun, “Characterization of (2014).
morphologies of compatibilized polypropylene/polystyrene blends [49] Jin, J., J. Du, Q. Xia, Y. Liang, and C. C. Han, “Effect of mesophase
with nanoparticles via nonlinear rheological properties from FT- separation on the crystallization behavior of olefin block copolymers,”
Rheology,” Macromolecules 47, 4066–4076 (2014). Macromolecules 43, 10554–10559 (2010).
[33] Salehiyan, R., H. Y. Song, W. J. Choi, and K. Hyun, “Characterization [50] Liu, G., Y. Guan, T. Wen, X. Wang, X. Zhang, D. Wang, X. Li, J.
of effects of silica nanoparticles on (80/20) PP/PS blends via nonlinear Loos, H. Chen, K. Walton, and G. Marchand, “Effect of mesophase
rheological properties from fourier transform rheology,” separation and crystallization on the elastomeric behavior of olefin
Macromolecules 48, 4669–4679 (2015). multi-block copolymers,” Polymer 52, 5221–5230 (2011).
[34] Schlatter, G., G. Fleury, and R. Muller, “Fourier transform rheology of [51] Arriola, D. J., E. M. Carnahan, P. D. Hustad, R. L. Kuhlman, and T. T.
branched polyethylene: Experiments and models for assessing the mac- Wenzel, “Catalytic production of olefin block copolymers via chain
romolecular architecture,” Macromolecules 38, 6492–6503 (2005). shuttling polymerization,” Science 312, 714–719 (2006).
[35] Vittorias, I., M. Parkinson, K. Klimke, B. Debbaut, and M. Wilhelm, [52] He, P., B. Chen, W. Yu, and C. Zhou, “Liquid-solid transition in meso-
“Detection and quantification of industrial polyethylene branching phase separated olefin multiblock copolymers during crystallization,”
topologies via Fourier-transform rheology, NMR and simulation using RSC Adv. 5, 40607–40619 (2015).
the Pom-pom model,” Rheol. Acta 46, 321–340 (2007). [53] Wagner, M. H., V. H. Rol on-Garrido, K. Hyun, and M. Wilhelm,
[36] Vittorias, I., and M. Wilhelm, “Application of FT rheology to indus- “Analysis of medium amplitude oscillatory shear data of entangled lin-
trial linear and branched polyethylene blends,” Macromol. Mater. Eng. ear and model comb polymers,” J. Rheol. 55, 495–516 (2011).
292, 935–948 (2007). [54] Haley, J., and T. Lodge, “Failure of time-temperature superposition in
[37] Liu, J., W. Yu, and C. Zhou, “Evaluation on the degrading behavior of dilute miscible polymer blends,” Colloid Polym. Sci. 282, 793–801
melt polyolefin elastomer with dicumyl peroxide in oscillatory shear (2004).
flow by Fourier transform rheology,” Polymer 49, 268–277 (2008). [55] Han, C. D., and J. K. Kim, “On the use of time-temperature superposi-
[38] Poulos, A. S., J. Stellbrink, and G. Petekidis, “Flow of concentrated tion in multicomponent/multiphase polymer systems,” Polymer 34,
solutions of starlike micelles under large amplitude oscillatory shear,” 2533–2539 (1993).
Rheol. Acta 52, 785–800 (2013). [56] Yu, W., R. Li, and C. Zhou, “Rheology and phase separation of poly-
[39] Patel, S. S., R. G. Larson, K. I. Winey, and H. Watanabe, “Shear orien- mer blends with weak dynamic asymmetry,” Polymer 52, 2693–2700
tation and rheology of a lamellar polystyrene-polyisoprene block (2011).
copolymer,” Macromolecules 28, 4313–4318 (1995). [57] Vananroye, A., P. Leen, P. van Puyvelde, and C. Clasen, “TTS in
[40] Gupta, V. K., R. Krishnamoorti, Z. R. Chen, J. A. Kornfield, S. D. LAOS: Validation of time-temperature superposition under large
Smith, M. M. Satkowski, and J. T. Grothaus, “Dynamics of shear amplitude oscillatory shear,” Rheol. Acta 50, 795–807 (2011).
NONLINEAR RHEOLOGY OF MULTIBLOCK COPOLYMERS 1179

[58] Gurp, M. V., and J. Palmen, “Time-temperature superposition for poly- [66] Frankel, N. A., and A. Acrivos, “The constitutive equation for a dilute
meric blends,” Rheol. Bull. 67, 5–8 (1998). emulsion,” J. Fluid Mech. 44, 65–78 (1970).
[59] Li, R., W. Yu, and C. Zhou, “Phase behavior and its viscoelastic [67] Chen, Q., W. Yu, and C. Zhou, “Rheological properties of immiscible
responses of poly(methyl methacrylate) and poly(styrene-co-maleic polymer blends under parallel superposition shear flow,” J. Polym.
anhydride) blend systems,” Polym. Bull. 56, 455–466 (2006). Sci., Part B: Polym. Phys. 46, 431–440 (2008).
[60] Trinkle, S., and C. Freidrich, “Van Gurp-Palmen-plot: A way to char- [68] Yu, W., M. Bousmina, M. Grmela, and C. Zhou, “Modeling of oscilla-
acterize polydispersity of linear polymers,” Rheol. Acta 40, 322–328 tory shear flow of emulsions under small and large deformation fields,”
(2001). J. Rheol. 46, 1401–1418 (2002).
[61] Trinkle, S., P. Walter, and C. Friedrich, “Van Gurp-Palmen plot II: [69] Yu, W., M. Bousmina, M. Grmela, J.-F. Palierne, and C. Zhou,
Classification of long chain branched polymers by their topology,” “Quantitative relationship between rheology and morphology in
Rheol. Acta 41, 103–113 (2002). emulsions,” J. Rheol. 46, 1381–1399 (2002).
[62] Palierne, J. F., “Linear rheology of viscoelastic emulsions with interfa- [70] Yu, W., W. Zhou, and C. Zhou, “Linear viscoelasticity of polymer
cial tension,” Rheol. Acta 29, 204–214 (1990). blends with co-continuous morphology,” Polymer 51, 2091–2098
[63] Yu, W., M. Bousmina, C. Zhou, and C. L. Tucker, “Theory for drop (2010).
deformation in viscoelastic systems,” J. Rheol. 48, 417–438 (2004). [71] Yu, W., C. Zhou, and Y. Xu, “Rheology of concentrated blends with
[64] Yu, W., C. Zhou, and M. Bousmina, “Theory of morphology evolution immiscible components,” J. Polym. Sci., Part B: Polym. Phys. 43,
in mixtures of viscoelastic immiscible components,” J. Rheol. 49, 2534–2540 (2005).
215–236 (2005). [72] Xu, Y., C. Huang, W. Yu, and C. Zhou, “Evolution of concentration
[65] Cox, R. G., “The deformation of a drop in a general time-dependent fluctuation during phase separation in polymer blends with viscoelastic
fluid flow,” J. Fluid Mech. 37, 601–623 (1969). asymmetry,” Polymer 67, 101–110 (2015).

You might also like