See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/271519957

Advanced rheological characterization of soft colloidal model systems

Article in Journal of Physics: Condensed Matter · October 2012

DOI: 10.1088/0953-8984/24/46/464102

CITATIONS READS

13 553

1 author:

Sudipta Gupta
Louisiana State University
33 PUBLICATIONS 583 CITATIONS

SEE PROFILE

All content following this page was uploaded by Sudipta Gupta on 31 July 2015.

The user has requested enhancement of the downloaded file.

Home Search Collections Journals About Contact us My IOPscience

Advanced rheological characterization of soft colloidal model systems

This article has been downloaded from IOPscience. Please scroll down to see the full text article.

2012 J. Phys.: Condens. Matter 24 464102

(http://iopscience.iop.org/0953-8984/24/46/464102)

View the table of contents for this issue, or go to the journal homepage for more

Download details:
IP Address: 134.94.112.45
The article was downloaded on 01/11/2012 at 14:49

Please note that terms and conditions apply.

IOP PUBLISHING JOURNAL OF PHYSICS: CONDENSED MATTER
J. Phys.: Condens. Matter 24 (2012) 464102 (9pp) doi:10.1088/0953-8984/24/46/464102

Advanced rheological characterization of

soft colloidal model systems
S Gupta, S K Kundu, J Stellbrink, L Willner, J Allgaier and D Richter

Jülich Centre for Neutron Science (JCNS-1) and Institute for Complex Systems (ICS-1),
Forschungszentrum Jülich, D-52425 Jülich, Germany

E-mail: j.stellbrink@fz-juelich.de

Received 20 April 2012, in final form 11 September 2012

Published 31 October 2012
Online at stacks.iop.org/JPhysCM/24/464102
Abstract
The complex flow behavior of polymer-based soft colloidal model systems was investigated
using steady and oscillatory shear to prove new concepts for advanced rheological
characterization. In the very dilute regime we investigated high molecular weight
polybutadiene star polymers to quantify the internal relaxation time arising from the polymeric
nature of these ultra-soft colloids. The observed shear-induced brush deformation is
interpreted in terms of the internal Zimm time τz . The observed dependence of τz on matrix
viscosity can be explained by shrinkage of the star polymer due to an increasing
incompatibility with increasing matrix molecular weight.
The influence of the polymeric nature on the characteristic structural relaxation time in
the concentrated regime was investigated using non-linear rheology following Wyss et al
(SRFS) (2007 Phys. Rev. Lett. 98 238303). Here we used star-like block copolymer micelles to
systematically tune the ‘softness’ of the colloids by variation of the block ratio. A master
curve with proper scaling parameters could be generated independent of the degree of
colloidal ‘softness’. However, the obtained strain-rate independent structural relaxation time
τ0 was not observed in the linear regime. In addition, a high frequency discrepancy was clearly
found in all our experimental data. Both reflect the shortcomings of the SRFS approach.

1. Introduction time of the colloidal system as a whole. A distinction between

single and collective behavior in this ‘sea-of-blobs’ regime is
The complex flow behavior of polymer-based soft colloids [2], no longer possible. The general relation between microscopic
e.g. microgels [1], star polymers [3, 4], block copolymer structure and macroscopic flow properties is therefore still a
micelles [5] or polymer-grafted nanoparticles [6], and their special point of interest [7–9].
corresponding mixtures is an interesting and still challenging Whereas for theory and simulations such a relation is
topic in today’s literature. This holds both for dilute solutions fairly easily obtainable [10, 11], experimentally this is still
where the microscopic structure of individual colloidal a big challenge. Recently combined in situ structural and
particles is affected by external shear and for concentrated rheological characterization (using small angle neutron or
solutions where particle interactions are affected. Here in x-ray scattering; rheo-SANS/rheo-SAXS) turned out to be
particular the glassy region of the phase diagram has been a very effective method to investigate the non-equilibrium
intensively studied in recent years. We have to emphasize structure–property relationship [5, 12–14]. Unfortunately,
that the two regimes are intimately linked. The characteristic such experiments cannot be performed on a routine basis
relaxation time of the individual soft colloid arising from due to the need for large scale facilities such as synchrotron
its internal polymeric structure changes with increasing sources or nuclear reactors. Therefore, there is a strong need
concentration. Finally, when polymeric coronae of different to derive structural information from rheology alone. For soft
soft colloids overlap in the concentrated regime, these colloids linear frequency sweep experiments fail to provide
additional constraints affect crucially the structural relaxation adequate information regarding the structural relaxation of

0953-8984/12/464102+09$33.00 1 c 2012 IOP Publishing Ltd Printed in the UK & the USA
J. Phys.: Condens. Matter 24 (2012) 464102 S Gupta et al

Table 1. Molecular weight characterization: PBO:PEO block ratio used calculated in terms of repetitive units, with Mw /Mn as the
polydispersity index [28], as determined by 1 H-NMR and SEC. δpoly is the average density of the PBO–PEO block copolymer. The
aggregation number f and the micellar radius Rm are determined by SANS [27].
Mw (PBO) Mw (PEO) PBO:PEO
Diblock copolymer (g mol−1 ) (g mol−1 ) block ratio Mw /Mn δpoly (g cm−3 ) f Rm (Å)
hPBO2.5–dPEO50 2 570 53 300 1:37 1.05 1.298 176 ± 2 467.5 ± 2
hPBO10–hPEO50 10 500 52 100 1:8 1.02 1.163 500 ± 10 616.4 ± 10

the system. In this regard, to probe the structural relaxation 36 mN m−1 compared to 46 mN m−1 for the previously
of the system in an elaborate way we adapt a relatively new used poly(ethylene–propylene) PEP [26]), which allows even
approach as proposed by Wyss et al [1]. This approach is more precise ‘fine-tuning’ of the micellization behavior.
further benefited by more advanced rheological techniques The investigated block copolymers hPBO10–hPEO50 and
and analysis schemes, e.g. Fourier transform rheology [15, hPBO2.5–dPEO501 are labeled according to their block
1, 16], which has been applied to several colloidal systems. molecular weights in kg mol−1 and characterized by their
However, this field is changing rapidly right now, and a block ratios (PBO:PEO), which are calculated using the ratio
broader database covering well defined model systems will of the number of repetitive units for the PBO and PEO blocks:
certainly be helpful. hPBO10–hPEO50 (PBO:PEO 1:8) and hPBO2.5–dPEO50
Here we present a detailed rheological study to test these (PBO:PEO 1:37); for details see table 1.
new concepts using well defined soft colloidal model systems. Samples were prepared by mixing directly the desired
To study single particle brush deformation in the dilute regime amounts of polymer and solvent (filtered twice using an
we used a high molecular weight 16-arm polybutadiene (PB) Anotop membrane filter of 20 nm pore size). Dilute solutions
star polymer, Mw,arm = 215 000 g mol−1 [17], dispersed in were then put into an electronic shaker to homogenize
low molecular weight linear PB oligomer matrix with varying at 20 ◦ C, whereas the concentrated gel-like solutions were
molecular weight, Mw = 800, 1580 and 2740 g mol−1 , around heated to 40 ◦ C, for 5 h, with continuous shaking, as all the
the entanglement molecular weight, Me ≈ 2000 g mol−1 . samples run into a frozen state, which macroscopically shows
Such a regular star polymer can be seen as the limiting non-flowing behavior. We did not increase the temperature
ultra-soft colloid [18]. For studying the effect of external above 50 ◦ C, as it was found [26] that a long heating procedure
shear on particle interactions in the glassy regime we used at higher (70 ◦ C) temperature causes degradation of the PEO
kinetically frozen block copolymer micelles [19] in aqueous chain, probably due to the high oxygen content of PEO.
solution, which have been recently introduced as an excellent However, degradation was checked by SEC measurements of
colloidal model system showing tunable softness [20, 21]. the polymer Mw after the experiments.
The aggregation number or functionality f of the formed
2. Experimental details star-like micelles was independently determined by small
angle neutron scattering in dilute solution to be f = 500 ± 10
2.1. Synthesis for PBO:PEO 1:8 and f = 176 ± 2 for PBO:PEO 1:37 [27],
and thereby PBO:PEO 1:8 is considerably less soft than the
Star polymers. All PB polymers were prepared using ‘living’ PBO:PEO 1:37 system [20], since f determines crucially the
anionic polymerization. The synthesis of star polymers was interaction potential.
accomplished by a linking reaction of ‘living’ PB arms and
chlorosilane dendrimers as has been described by Roovers 2.2. Rheology
et al [22]. Molecular characteristics for matrix polymers are
800, 1580 and 2740 g mol−1 , respectively, as determined Steady shear. Steady shear flow curves have been measured
by SEC and 1 H-NMR end group analysis. The molecular for all solutions using a stress-controlled AR-G2 rheometer
weight for a PB star polymer with 16 arms is 3.45 × (TA Instruments, USA) in parallel plate geometry. We used
106 g mol−1 . Star polymer solutions with concentrations c ≈ tools made from stainless steel with a diameter of 40 mm and
0.004 g ml−1 in the oligomeric matrix were prepared by first the gap between the parallel plates was 0.3 mm. The sample
dissolving the PB star polymer as well as the PB oligomer volume required to fill the shear cell was 0.4 ml. We employed
separately in tetrahydrofuran (THF). These solutions were a maximum shear rate of γ̇ = 1000 s−1 for the present
mixed afterwards under gentle shaking to ensure complete experiment. Temperature control to 22 ◦ C was achieved by a
dissolution. After THF removal by high vacuum, we obtained circulating fluid bath thermostat (Julabo, Germany).
finally homogeneous transparent solutions.
Oscillatory shear. All dynamic–mechanical experiments were
Block copolymers. The well defined block copolymer performed using a strain-controlled ARES-G2 rheometer
poly(butylene oxide)–poly(ethylene oxide) (PBO–PEO) was (TA Instruments, USA). All measurements were made at
synthesized by an anionic polymerization technique, de- a temperature T = 20 ± 0.1 ◦ C, set via a Peltier plate.
scribed in [23]. Here PBO and PEO are the hydropho-
bic and hydrophilic moieties in the diblock copolymer. 1 Here the pre-factor ‘h’ stands for a fully protonated and pre-factor ‘d’
PBO [23–25] is slightly less hydrophobic (interfacial tension stands for a fully deuterated block.

2
J. Phys.: Condens. Matter 24 (2012) 464102 S Gupta et al

Table 2. Results from Carreau fit of steady shear flow curves. of even tiny amounts of star polymer to the matrix, c ≈
a 0.004 g ml−1 ≈ 0.1c∗ . For all matrix molecular weights it was
Mw
(g mol−1 ) ηm b (Pa s) ηc (Pa s) τc (s) also found that initially at low shear rates solution viscosity
was independent of shear rate, but finally shear thinning, i.e. a
800 0.073 ± 0.004 0.083 ± 0.003 3.73×10−3 ±0.02 shear rate-dependent viscosity η(γ̇ ), was clearly observed.
1580 0.18 ± 0.009 0.21 ± 0.01 5.71×10−3 ±0.02
Now we obtained the polymer contribution ηstar to the solution
2740 0.65 ± 0.01 0.67 ± 0.02 1.84×10−2 ±0.02
viscosity by subtracting matrix viscosity ηm from solution
a Matrix molecular weight. viscosity η using
b Matrix viscosity.
c Solution viscosity. ηstar (γ̇ ) = η(γ̇ ) − ηm . (1)
Finally, we calculated the reduced polymer viscosity
Cone-plate geometry has been used; all measurements were normalized by zero shear viscosity by using
also verified and cross-checked with different cone-plate ηstar (γ̇ )
geometries with cone angle less than 0.01 rad, and ηnorm (γ̇ ) = (2)
with varying diameters (2R) such as 20, 25, 40 and ηstar (γ̇ = 0)
50 mm, depending on the concentration of the sample. To to compare the flow curves from different matrix molecular
prevent solvent evaporation a small amount of dodecane weights. Now, experimental flow curves, i.e. reduced viscosity
(low viscosity fluid) is deliberately used during every ηnorm as a function of shear rate γ̇ , measured under steady
measurement, in a solvent trap setup, present with each shear using parallel plate geometry, are shown in figure 1(a).
geometry. Here we compared data of dilute solutions of PB star polymers
with fixed star functionality f = 16 and concentration c ≈
3. Results and discussion 0.004 g ml−1 dispersed in different PB oligomer matrices.
Only at low shear rates is a Newtonian plateau, i.e. a shear
3.1. Brush deformation in dilute solutions rate independent viscosity, observed, whereas the viscosity
decreases with increasing shear rate after a critical shear rate
Dilute solutions of linear polymer coils in a good solvent γ˙c (shear thinning). It is also found from this figure that γ˙c
exposed to simple shear flow in steady state are the decreases with increasing matrix molecular weight. The flow
easiest case of shear-induced phenomena, but nevertheless curves of the PB star polymer solutions are analyzed using a
experimentally difficult to observe due to the required high well known Carreau function [30] of the form
shear rates γ̇ to overcome the internal relaxation time τc η0
of the polymer [29]. Only at high Weissenberg numbers η(γ̇ ) = 1−b
(3)
[1 + (γ̇ /γ˙c )a ] a
Wi = γ̇ τc > 1 does a shear effect become experimentally
observable. In dilute solution the characteristic internal where η0 is the zero shear viscosity, and γ˙c is the critical shear
relaxation time of a linear polymer is given by the Zimm rate indicating the onset of shear thinning in units of s−1 .
time τz = (ηm R3g )/(KB T), with ηm the solvent viscosity (1−b) is the power law exponent (in the shear thinning region)
and Rg the radius of gyration of the polymer [10]. and describes the dependence of the viscosity η on shear
Therefore, a combination of high molecular weights and high rate γ̇ in the shear thinning region, and the dimensionless
viscosity solvent is the best experimental approach to study parameter a represents the width of the transition region
shear-induced phenomena. Branching crucially influences the between the constant Newtonian plateau observed at low shear
macroscopic flow properties of polymer melts; therefore, a rate and the asymptotic power law decrease of the viscosity
detailed microscopic understanding is highly interesting. Such found at high shear rates. Commonly, values of a equal to
knowledge can be obtained starting from the shear-induced 2 are observed. The relaxation time, τc , is calculated by
deformation of the individual polymer as observed in dilute τc = 1/γ˙c . The solid lines shown in figure 1(a) are the fits
solution. of the Carreau function for all flow curves. It is seen from
So, we used low molecular weight PB oligomers as figure 1(a) that all the flow curves are described well by
solvent (matrix) to clarify the shear-induced phenomena for equation (3). The obtained critical shear rate decreases with
star polymers (ultra-soft colloids) in dilute solution. We varied increasing matrix molecular weight, which is expected as an
the oligomer molecular weight around the entanglement increasing viscosity of the matrix naturally slows down the
molecular weight of PB, Me ≈ 2000 g mol−1 , to investigate dynamics of the dissolved star polymer. The variation of the
the effect of the matrix. We have measured the viscosity as relaxation time τc with the viscosity of the matrix polymer
a function of shear rate γ̇ for pure matrix (PB oligomers) is shown in figure 1(b). The relaxation time can be scaled
and the star polymer solutions (PB star polymer dispersed qualitatively with the viscosity of the matrix polymer, giving
a power law ∼ηm 0.8±0.05 ; see figure 1(b). From the equation
in oligomeric matrix) separately. It was found that, for the
pure matrix, the viscosity ηm increases with increasing matrix for the Zimm time a linear dependence ∼ηm is expected,
molecular weight, and also viscosity ηm was independent whereas in figure 1(a) a substantial slowing down is observed
of shear rate within the experimental window (Newtonian on crossing Me , i.e. between Mw = 1580 g mol−1 and
behavior); see table 2. For star polymer solutions, it was Mw = 2740 g mol−1 ; re-scaling to the dependence on matrix
found that viscosity increases substantially upon addition viscosity as shown in figure 1(b) reveals a rather smooth

3
J. Phys.: Condens. Matter 24 (2012) 464102 S Gupta et al

Figure 1. Steady shear results for a 16-arm PB star polymer dispersed in PB oligomer matrix with varying molecular weight,
c = 0.004 g ml−1 , T = 22 ◦ C. (a) Steady shear flow curves, i.e. reduced viscosity ηnorm versus shear rate. (b) Variation of the relaxation time
from the Carreau fit, τc , as a function of matrix viscosity.

continuous behavior. The observed weaker dependence of τc periodic stress response alone. For small strain amplitudes
on matrix viscosity ηm most probably results from a size γ0 the shear stress response is a simple harmonic function,
change of the star polymer. An osmotic pressure-induced oscillating with the fundamental frequency dictated by the
shrinkage in the radius of gyration is observed for star externally applied strain field. Complex fluids exhibit complex
polymers dispersed in molecular solvents. Such an increasing non-linear rheological responses [35, 36], and as such have
osmotic pressure results either from an increasing star been of interest to experimentalists and theoreticians for
polymer concentration [31] or from addition of linear chains many decades. The question arises of how to probe the
(depletion) [32]. Here, in the case of a polymeric matrix as complex viscoelastic response of such soft materials in
solvent, the observed shrinkage results most probably from the most appropriate way. So a need for a generalized
an increasing incompatibility between the star polymer and new approach to probe the structural relaxation of the soft
linear matrix with increasing matrix molecular weight. A colloidal glasses is evident. Usually, frequency dependent
minor decrease in Rg can already have a considerable effect, measurements, as shown in figure 2(a) for star-like micelles
since τz scales as ∼R3g . Therefore, quantitative interpretation formed by hPBO10–hPEO50 (block ratio 1:8, f = 500) and
of rheological data without knowing structural details is hPBO2.5–dPEO50 (block ratio 1:37, f = 176) in the glassy
extremely difficult, but advanced rheological characterization regime, were performed in the linear viscoelastic regime at a
before a rheo-SANS/rheo-SAX experiment is a helpful tool to constant strain amplitude γ0 ≤ 1%.
explore the parameter range and prepare the final experiment. The most interesting feature is in the star-like regime
Nevertheless, from this rheological study, we have for the block ratio 1:8 and for 1:37 in figure 2(a), which
successfully obtained the shear effect, i.e. the onset of shows remarkably similar behavior. The storage modulus G0
the non-linear regime, by increasing the viscosity of the is only very weakly dependent on frequency, whereas the
matrix polymer in the present experimental conditions. It loss modulus G00 exhibits a shallow minimum. At even lower
would be very interesting to extend experiments well into frequencies, there should be a cross-over from solid-like to
the non-linear regime to quantify the power law slope. In liquid-like behavior, which is signaled by a pronounced peak
particular, comparison with theoretical results [33, 34] would in G00 (ω), reflecting the generic structural relaxation. This
be helpful to understand the microscopic origin of shear
effect is manifested on looking closely into two different
effects. However, achieving high shear rates is experimentally
block ratios 1:8 and 1:37, measured under exactly the same
extremely challenging and will therefore be the subject of
experimental conditions at volume fraction φ = ∼4φ ∗ for
future work.
both block ratios 1:8 and 1:37, at a strain amplitude of
γ0 = 1%, as shown in figure 2(a). They are fitted with
3.2. Oscillatory shear: strain-rate–frequency superposition the continuous relaxation spectrum (CRS) as obtained in
(SRFS) in the concentrated regime equations (6) and (7). The generalized Maxwell’s model,
equations (4) and (5), could be represented as the discrete
A standard method to probe the viscoelastic character
relaxation spectrum [37]
of a material is to measure the time dependent stress
response to an externally applied oscillatory shear field. X ω2 τi2
The simplicity of oscillatory shear experiments presents G0 (ω) = Gi (4)
i 1 + ω2 τi2
distinct practical advantages when compared to other flow
protocols and thus establishes a systematic method for the X ωτi
G00 (ω) = Gi (5)
rheological characterization of a material on the basis of the i 1 + ω2 τi2

4
J. Phys.: Condens. Matter 24 (2012) 464102 S Gupta et al

Figure 2. (a) Frequency sweep results for constant γ0 = 1% at φ = 12% ∼ 4.5φ ∗ for block ratio 1:8 and at φ = 10.4% ∼ 3.85φ ∗ for block
ratio 1:37, fitted with continuous relaxation spectrum (CRS). (b) Corresponding relaxation time distribution: the relaxation mode H(τ ) (Pa)
as a function of the relaxation time τ . (c) Corresponding strain sweep data for ω = 0.1 rad s−1 for block ratio 1:8 and 1:37 as shown.

where τi is the relaxation time of the ith element, given by G00 (γ0 ), as a function of strain amplitude γ0 at constant ω =
τi = ηi /Gi . For an infinite number of elements in the Maxwell 0.1 rad s−1 , for block ratio 1:8 and 1:37. The pronounced
model [37], resulting in a continuous spectrum in which peak (figure 2(c)) in the loss modulus is a remarkably
each infinitesimal contribution to rigidity in the Maxwell robust feature of soft glassy materials [38–41]; this universal
element [37] is associated with relaxation times lying in the behavior can be described by MCT (mode coupling theory)
range between ln τ and ln τ +d ln τ (in logarithmic time scale), based approaches and is also captured by SGR (soft glassy
one can obtain the corresponding expressions for G0 and G00 rheology) [42] and other phenomenological models. The peak
for a continuous relaxation spectrum, as represents a maximum energy dissipation, where it is expected
that a large number of cages break causing an irreversible
ω2 τ 2
Z ∞
G0 (ω) = H(ln τ ) d(ln τ ) (6) particle rearrangement [43]. However, to probe the structural
−∞ 1 + ω2 τ 2 relaxation of the system in an elaborate way we have to adopt
ωτ
Z ∞
a relatively new approach.
G00 (ω) = H(ln τ ) d(ln τ ) (7)
−∞ 1 + ω2 τ 2 Various methods have been proposed to describe the
where H(ln τ ) is the relaxation spectrum, which provides a structural relaxation of the system [44–46]. However, for
continuous function of relaxation time τ rather than a discrete low-frequency applicability, it is interesting to inquire if
set. The corresponding relaxation time distribution, plotting the rheological behavior of our soft colloidal model system
the relaxation mode H(τ ) (Pa) as a function of the relaxation can be explained following the generalized new argument
time τ (s), is shown in figure 2(b). It clearly elucidates proposed by Wyss et al [1] when we tune systematically
the fact that, both at lower and at higher frequencies, there the softness of the interaction potential. We can establish
exists a more probable peak in the relaxation spectrum, a pre-defined scaling concept and, therefore, can obtain
which significantly emphasizes the possibility of two different detailed information on the strain-rate dependence of the
relaxation process, at low and at high frequency. However, slow relaxation process, under considerable assumptions. In
these two relaxations are not accessible by a normal frequency this context we can also perform the non-linear viscoelastic
sweep experiment within the linear rheological limit, as has measurements, which contain useful information on the slow
been shown in figures 2(a) and (b), since the relaxation relaxation dynamics of our system.
frequencies are often much too low/high to be accessed by The structural relaxation of the soft colloidal glasses is
standard linear rheological measurements. Figure 2(c) shows not accessible by a normal frequency sweep test. Usually,
the general strain sweep response (LAOS test) G0 (γ0 ) and frequency dependent measurements as shown in figure 2(a)

5
J. Phys.: Condens. Matter 24 (2012) 464102 S Gupta et al

Figure 3. Frequency dependent G0 and G00 for different strain-rate amplitudes γ̇0 varying from 0.01 to 5 s−1 (a) for block ratio PBO:PEO
1:8 and (b) for the block ratio PBO:PEO 1:37.

were performed in the linear viscoelastic regime at a constant similarities/differences between the viscoelastic responses for
strain amplitude γ0 ≤ 1%. An alternative time scale can the two samples PBO:PEO 1:8 and 1:37. These could be
be obtained, as proposed by Wyss et al [1] through SRFS. manifested by rescaling all the elastic moduli in terms of both
Instead of the strain amplitude γ0 , SRFS is based on the frequency and amplitude as shown in figures 4(a) and (b),
strain-rate amplitude γ̇0 = ωγ0 , as a natural scale, obtained respectively. The solid line represents the fitted data using
by rheological techniques. They proposed a γ̇0 dependent the single mode (i = 1) discrete relaxation spectrum (DRS)
structural relaxation time scale as (equations (4) and (5)), to obtain the cross-over point. Both
the storage (G0 ) and loss (G00 ) modulus are scaled [1] using
1 1
≈ + K γ̇0ν (8)
τ (γ̇0 ) τ0 ω
 
1
G∗scaled (ω) = G∗ (9)
where τ (γ̇0 ) and τ0 are the γ̇0 dependent and independent a(γ̇0 ) b(γ̇0 )
structural relaxation times respectively, and K and ν are where a(γ̇0 ) and b(γ̇0 ) are the vertical and horizontal shift
constants to be determined from the experimental scaling factors, respectively. Now, all the data in figures 3(a) and
conditions. (b) can be superimposed onto a single master curve, with
All the experiments were performed for both block ratios respect to γ̇0 = 0.01 s−1 , as shown in figures 4(a) and (b).
1:8 and 1:37 with a volume fraction around ≈4φ ∗ , where φ ∗ is Figure 5(a) shows the variation of the scaling factors a(γ̇0 ) and
the overlap volume fraction. Figure 3(a) shows the frequency b(γ̇0 ) as a function of strain-rate amplitude γ̇0 . Interestingly,
dependent measurement of G0 (ω) and G00 (ω) for different for both samples, i.e. PBO:PEO 1:8 and 1:37, they obey the
strain-rate amplitudes γ̇0 varying over more than two orders same scaling laws. The value of a(γ̇0 ) remains constant (≈1)
of magnitude from 0.01 to 5 s−1 . It has been observed that, over the entire range of γ̇0 accessed, whereas the frequency
for each of the curves, the system behaves like a viscoelastic scaling factor b(γ̇0 ) depends strongly on γ̇0 . b(γ̇0 ) shows a
liquid (G00 > G0 ) at smaller angular frequencies ω, followed power law dependence at large strain rates as b(γ̇0 ) ∝ γ̇00.9 ,
by a cross-over from liquid to solid-like viscoelastic response, which gives ν = 0.9 in equation (8). At low strain rates it
finally leading to a plateau of G0 , and the elastic behavior asymptotically approaches unity, which suggests that in this
dominates. The loss modulus G00 shows a pronounced peak regime the imposed strain rate is sufficiently small that it no
√ and then at higher ω shows
just after the cross-over point, longer affects the viscoelastic behavior. In addition, we also
a linear increase with a ω dependence for the sample show new data from a recently investigated, slightly different
PBO:PEO 1:8; however, it is not so evident for the block block copolymer PEP:PEO 1:25 with f = 120 [7], which also
ratio 1:37. The viscoelastic response depends strongly on γ̇0 . follows the same scaling law. Therefore, a master curve with
With increasing γ̇0 , the entire frequency dependent response proper scaling parameters can be generated independent of
is shifted towards higher ω. We investigate the system the degree of colloidal ‘softness’. This is consistent with the
non-linearity based on Fourier-transform analysis of the input scaling behavior as predicted by Wyss et al [1] in equation (8),
strain signal as described elsewhere [15, 16]. In this regard, as with b(γ̇0 ) = τ0 /τ (γ̇0 ) ∝ γ00.9 . The fitted and calculated
a measure for non-linearity it has been observed that the ratio results for the different relaxation times from figure 4(a)
of the third to the first harmonic intensity I(3ω)/I(ω) [15] of are shown in table 3 for PBO:PEO 1:8. Similarly, those for
the stress response from the rheometer is always about 20% PBO:PEO 1:37 from figure 4(b) are shown in table 4, where
or less at the cross-over points. At the peak value of G00 , this ωc and Gc are the cross-over frequency and elastic modulus
ratio is about 12% or less. It is to be noted that the varying respectively. A final mean value of τ0 = 10.72 ± 0.8 s was
strain amplitude γ0 is always less than or equal to the yield obtained for PBO:PEO 1:8 from table 3 and τ0 = 12.63±0.8 s
strain [27] of the system. was obtained for PBO:PEO 1:37 from table 4.
Generation of the master curve. The next step would be the As shown in figure 3(a), the high frequency
√ viscoelastic
creation of the SRFS master curve to visualize the basic response in micellar solution exhibits a ω-dependence for

6
J. Phys.: Condens. Matter 24 (2012) 464102 S Gupta et al

Figure 4. Master curve obtained from constant rate measurement γ̇0 fitted with the discrete relaxation spectrum (DRS), at the cross-over
point (a) for block ratio PBO:PEO 1:8 and (b) for block ratio PBO:PEO 1:37, respectively.

Figure 5. For PBO:PEO 1:8 and 1:37 as shown in the legend compared with the PEP:PEO 1:25 system (hPEP1hPEO20) with f = 120 [7].
(a) The amplitude and frequency scaling factors a(γ̇0 ) and b(γ̇0 ) respectively,
√ as a function of strain-rate amplitude γ̇0 , which gives ν = 0.9
and K = 60.78 ± 0.012 in equation (8). (b) Loss modulus G00 (ω) with c ω subtracted, as a function of ω/b(γ̇0 ), for different γ̇0 as
indicated in the legend.

Table 3. For PBO:PEO 1:8, fitted, rescaled∗ and calculated∗∗ results from the DRS model and experimental scaling as described in the text
for different γ̇0 with b(γ̇0 ) = τ0 /τ (γ̇0 ) from equation (8).
γ̇0 (s−1 ) τ (γ̇0 ) (s) ωc (rad s−1 ) Gc (Pa) b(γ̇0 )∗ a(γ̇0 )∗ τ0∗∗ (s)
0.01 9.13 ± 1 0.091 1644.6 1 1 9.13
0.05 2.78 ± 0.6 0.373 1731.5 4.073 1.049 11.32
0.1 1.47 ± 0.07 0.693 1730.4 7.58 1.05 11.14
0.5 0.38 ± 0.03 2.77 1561.83 30.45 0.95 11.57
1 0.18 ± 0.03 5.45 1629.3 59.802 0.99 10.76
5 0.037 ± 0.01 25.59 1888.5 281.95 1.147 10.43

Table 4. For PBO:PEO 1:37, fitted, rescaled∗ and calculated∗∗ results from the DRS model and experimental scaling as described in the
text for different γ̇0 with b(γ̇0 ) = τ0 /τ (γ̇0 ) from equation (8).
γ̇0 (s−1 ) τ (γ̇0 ) (s) ωc (rad s−1 ) Gc (Pa) b(γ̇0 )∗ a(γ̇0 )∗ τ0∗∗ (s)
0.01 11.36 ± 2 0.079 564 1 1 11.36
0.05 2.25 ± 0.8 0.44 606.18 5.56 1.07 12.55
0.1 1.11 ± 0.6 0.86 618.18 10.93 1.09 12.18
0.5 0.27 ± 0.08 3.62 580.16 45.83 1.02 12.72
1 0.15 ± 0.04 6.73 584.5 85.253 1.03 13.38
5 0.031 ± 0.03 28.43 642.1 359.87 1.13 13.61

PBO:PEO 1:8, which was identified as the viscous flow is used for each data set, from the loss modulus G00 (ω), and
along randomly oriented slip planes [47]. This results in a all the data are rescaled onto a single master curve, by shifting
√
contribution that scales as ω, which is same for all γ̇0 . So the ω axis by ω/b(γ̇0 ) as shown in figure 5(b). Here the results
√
we subtract this fixed value c ω, where the same value of c from PBO:PEO 1:8 and 1:37 are compared. The striking

7
J. Phys.: Condens. Matter 24 (2012) 464102 S Gupta et al

similarities among the shapes of the resulting relaxation peaks not observed so clearly by Wyss et al [1] in their experiments.
of G00 (ω) are clearly visible as the strain-rate amplitude is Therefore our present study clearly corroborates the general
varied. criticism of the SRFS approach.
Therefore, this approach clearly elucidates the presence
of a structural relaxation peak with different structural
4. Conclusions
relaxation times as shown in tables 3 and 4, for PBO:PEO 1:8
and PBO:PEO 1:37 respectively, and can be converged to the
natural relaxation time τ0 of the micelles using a constitutive In summary, we investigated two well defined soft colloidal
scaling law as proposed by Wyss et al in equation (8). This model systems by rheology to test new experimental concepts.
approach allows us to directly probe the structural relaxation, In the dilute regime we used a high molecular weight 16-arm
at a frequency too low to be easily accessible by standard PB star polymer dispersed in low molecular weight linear
rheological experiments. Different soft materials exhibit such PB oligomer matrix. We varied systematically the matrix
a low-frequency structural relaxation [1]. molecular weight, Mw = 800, 1580 and 2740 g mol−1 ,
However, there was substantial criticism on the SRFS around the entanglement molecular weight of PB. Steady
approach raised in [48] that needs to be discussed. From shear flow curves show a pronounced shear thinning behavior
figure 5(a) one can extrapolate to γ̇0 = 0 to give b(γ̇0 = starting at characteristic shear rate, decreasing with increasing
0) = 1.93 for PBO:PEO 1:8 and 1.78 for the PBO:PEO 1:37 matrix molecular weight and matrix viscosity. The observed
sample. In this case, one can obtain τ0 (γ̇0 = 0.1) = 17.6 s weaker dependence on matrix viscosity, as theoretically
and 20.22 s for PBO:PEO 1:8 and 1:37, respectively. This expected, can be explained by a size change of the PB star
can give a corresponding angular frequency (ω0 = 2π/τ0 ) of polymer. Increasing the matrix molecular weight results in
0.35 rad s−1 and 0.31 rad s−1 for PBO:PEO 1:8 and 1:37, an increasing incompatibility between star polymer and linear
respectively. However, no such relaxation process is visible matrix that leads to a shrinkage in the radius of gyration of the
from the linear frequency sweep data in figure 2(a). Therefore, star polymer.
the main claim of the SRFS approach to derive quantitatively In the glassy regime we used kinetically frozen block
the characteristic structural relaxation time τ0 of the system copolymer micelles in aqueous solution as a tunable model
by use of equation (8) fails clearly. This shortcoming of system for soft colloids [12]. We varied systematically the
SRFS becomes even more severe when we keep in mind softness of the colloidal interaction potential by changing
that the non-linearity of our systems is comparable to the the ratio between hydrophobic and hydrophilic blocks. This
limit that was proposed as the basic assumption to derive directly transforms into a change in the micellar functionality
the SRFS approach [1]. The non-linearity of our systems is (or aggregation number) from f = 120, 174, and 500 [50]
less than 20%, which we obtain from Fourier-transform based that governs the interaction potential [18]. We used this
analysis as described by Ewoldt et al [16]. However, this 20% tunable system to investigate the validity of a relatively new
non-linearity seems to be already large enough to drive the
rheological approach proposed by Wyss et al [1]: the SRFS.
slow structural relaxation process at the time scale of the
Following the SRFS a master curve with proper scaling
imposed strain rate. This means that at this point, instead of
parameters was obtained independent of colloidal softness.
observing the natural structural relaxation of the system, we
Moreover, we observe that for all samples irrespective of their
are probing the one defined by the system non-linearity.
softness the power law dependence of the scaling parameter
Furthermore, the scaling rule excludes the high frequency
b(γ̇0 ) ∝ γ̇00.9 is identical and obeys the same power law
region; it is applicable only in the low-frequency regime.
As predicted by Wyss et al [1] in equation (8), under the for the scaling parameters as predicted by Wyss in the
assumption that the effect of the vertical shift factor is SRFS approach [1]. Finally, we could successfully derive a
negligible, they also proposed that structural relaxation time τ0 from the master curve and scaling
parameters, which again only slightly depends on the colloidal
ω
 
softness. However, for all systems under study, irrespective
G (ω, γ0 ) = GR
∗ ∗
+ G∗HF (ω) (10)
b(γ̇0 ) of their softness, the quantitative value of τ0 derived from
SRFS does not describe rheological data obtained in the linear
where G∗R (ω) describes the structural relaxation process and
G∗HF (ω) describes the high frequency contribution, which is regime, although the non-linearity of our systems, ≤20%,
independent √ of strain amplitude, under the assumption that as obtained from FT analysis, is well in the validity regime
G00HF (ω) ≈ c ω. This is clearly illustrated in figure 5(b), for the SRFS given by Wyss et al. Therefore, in conclusion,
where, despite having a nice overlap for different γ̇0 in our present systematic study confirms the crucial criticism of
the structural relaxation pattern, at high frequency there is SRFS to fail to give quantitative results as raised in [48].
no convincing overlap in the master curve. This mismatch In general, interpretation of rheological data without
increases with the increase in γ̇0 . This is because, as the strain knowing structural details is extremely difficult, but recent
rate becomes large, the scaling approximation in equations (8) new rheological characterization approaches are really helpful
and (10) does not hold any more. This discrepancy at higher for a more detailed understanding of the flow properties of
frequencies in figure 5(b) is indeed supported by the work of soft colloids. Nevertheless, the ultimate goal will still be
Fuchs et al [49], which suggests a rate-dependent change of to perform rheo-SANS/rheo-SAXS experiments to derive a
the high frequency part of the relaxation. This was however proper structure–property relationship.

8
J. Phys.: Condens. Matter 24 (2012) 464102 S Gupta et al

Acknowledgment [22] Zhou L L and Roovers J 1993 Macromolecules 26 963

[23] Allgaier J, Willbold S and Chang T 2007 Macromolecules
40 518
We acknowledge the Deutsche Forschungsgemeinschaft
[24] Ströbl M 2008 Diploma Thesis Universität Regensburg
(DFG) for financial support in the framework of the [25] Nicolai T, Colombani O and Chassenieux C 2010 Soft Matter
Transregio Sonderforschungsbereich TR6 (Teilprojekt A2). 6 3111–8
[26] Laurati M, Stellbrink J, Lund R, Willner L, Zaccarelli E and
Richter D 2007 Phys. Rev. E 76 041503
References [27] Gupta S 2012 PhD Thesis Universität Münster
[28] Rubinstein M and Colby R 2003 Polymer Physics (Oxford:
[1] Wyss H M, Miyazaki K, Mattsson J, Hu Z, Reichman D R and Oxford University Press)
Weitz D A 2007 Phys. Rev. Lett. 98 238303 [29] Lindner P et al 1988 Colloid Polym. Sci. 266 886
[2] Likos C N 2006 Soft Matter 2 478–98 [30] Carreau P J 1972 Trans. Soc. Rheol. 16 99–127
[3] Rogers S A, Vlassopoulos D and Callaghan P T 2008 Phys. [31] Stellbrink J, Allgaier J, Monkenbusch M, Richter D, Lang A,
Rev. Lett. 100 128304 Likos C N, Watzlawek M, Löwen H, Ehlers G and
[4] Erwin B M, Vlassopoulos D, Gauthier M and Cloitre M 2011 Schleger P 2000 Prog. Colloid Polym. Sci. 115 88–92
Phys. Rev. E 83 061402 [32] Wilk A et al 2010 Eur. Phys. J. E 32 127–34
[5] Stellbrink J, Lonetti B, Rother G, Willner L and [33] Gnazzoli F and Raffaini G 1999 Macromol. Theory Simul.
Richter D 2008 J. Phys.: Condens. Matter 20 404206 8 234–46
[6] Anderson B J and Zukoski C F 2009 Macromolecules [34] Huang C, Winkler R G, Sutmann G and Gompper G 2010
42 8370–84 Macromolecules 43 10107–16
[7] Renou F, Stellbrink J and Petekidis G 2010 J. Rheol. 54 1219 [35] Dhont J C G, Gompper G and Richter D 2002 33th IFF Winter
[8] Rogers S A, Callaghan P T, Petekidis G and School, Soft Matter: Complex Materials on Mesoscopic
Vlassopoulos D 2010 J. Rheol. 54 133–58 Scale (Jülich: Forschungszentrum Jülich)
[9] Rogers S A, Erwin B M, Vlassopoulos D and Cloitre M 2011 [36] Macosko W M 1994 RHEOLOGY Principles, Measurements
J. Rheol. 55 435–58 and Applications (New York: Wiley)
[37] Ferry J D 1980 Viscoelastic Properties of Polymers
[10] Ripoll M, Winkler R G and Gompper G 2006 Phys. Rev. Lett.
(New York: Wiley)
96 188302
[38] Mason T and Weitz D A 1995 Phys. Rev. Lett. 75 2770
[11] Singh S P and Winkler R 2011 private communication [39] Mason T G, Bibette J and Weitz D A 1995 Phys. Rev. Lett.
[12] Stellbrink J, Rother G, Laurati M, Lund R, Willner L and 75 2051
Richter D 2004 J. Phys.: Condens. Matter 16 S3821–34 [40] Robertson C G and Wang X 2005 Phys. Rev. Lett. 95 075703
[13] Förster S, Konrad M and Lindner P 2005 Phys. Rev. Lett. [41] Crassous J J et al 2008 J. Chem. Phys. 128 204902
94 017803 [42] Sollich P, Lequeux F, Hébraud P and Cates M E 1997 Phys.
[14] Cola E D, Moussaı̈d A, Sztucki M, Narayanan T and Rev. Lett. 78 2020
Zaccarelli E 2009 J. Chem. Phys. 131 144903 [43] Petekidis G, Moussaid A and Pusey P N 2002 Phys. Rev. E
[15] Wilhelm M J and Dealy J M 1999 Rheol. Acta 38 349–56 66 051402
[16] Ewoldt R, Hosoi A and McKinley G 2008 J. Rheol. [44] Knauss W G and Emri I J 1981 Comput. Struct. 13 123–8
52 1427–58 [45] Bernstein B and Shokooh A 1980 J. Rheol. 24 189–211
[17] Stellbrink J et al 2002 J. Appl. Phys. A 74 (suppl.) S355–7 [46] Cho K S and Kim S Y 2003 J. Appl. Polym. Sci. 89 2400–11
[18] Likos C N, Löwen H, Watzlawek M, Abbas B, Jucknischke O, [47] Liu A J, Ramaswamy S, Mason T, Gang H and
Allgaier J and Richter D 1998 Phys. Rev. Lett. 80 4450–3 Weitz D A 1996 Phys. Rev. Lett. 76 3017
[19] Lund R 2004 PhD Thesis Universität Münster [48] Erwin B M, Rogers S A, Cloitre M and Vlassopoulos D 2010
[20] Laurati M 2005 PhD Thesis Universität Münster J. Rheol. 54 187
[21] Lonetti B, Camargo M, Stellbrink J, Likos C N, Zaccarelli E, [49] Fuchs M and Cates M E 2002 Phys. Rev. Lett. 89 248304
Willner L, Lindner P and Richter D 2011 Phys. Rev. Lett. [50] Laurati M, Stellbrink J, Lund R, Willner L, Richter D and
106 228301 Zaccarelli E 2005 Phys. Rev. Lett. 94 195504

View publication stats