Soft Matter

View Article Online

PAPER View Journal | View Issue

Competing polymer–substrate interactions

mitigate random copolymer adsorption†
Published on 22 August 2018. Downloaded by Bogazici Universitesi on 10/29/2022 12:40:47 PM.

Cite this: Soft Matter, 2018,

14, 7204 a
Mary J. B. Davis, Biao Zuoab and Rodney D. Priestley *ac

Annealing a supported polymer film in the melt state, a common practice to relieve residual stresses and
erase thermal history, can result in the development of an irreversibly adsorbed nanolayer. This layer of
polymer chains physically adsorbed to the substrate interface has been shown to influence thin film
properties such as viscosity and glass transition temperature. Its growth is attributed to many
simultaneous interactions between individual monomer units and the substrate stabilizing chains against
desorption. A better understanding of how these specific polymer–substrate interactions influence the
growth of the adsorbed layer is needed, particularly given how strongly the properties of geometrically-
confined polymeric systems are impacted by interfaces. Here, we use homopolymers and random
copolymers of styrene and methyl methacrylate to form adsorbed layers and examine the influence of
chemical composition and the resulting polymer–substrate interactions on adsorbed layer growth and
structure. Ellipsometric measurements reveal a non-monotonic trend between composition and
thickness of the adsorbed layers that is inconsistent with the behavior normally exhibited by random
copolymers, being intermediate to their respective homopolymers. We examine this trend in terms of
plateau thickness and growth kinetics at two different annealing temperatures and propose a
Received 12th July 2018, mechanism for how different polymer–substrate interactions combine to influence adsorption when
Accepted 13th August 2018 copolymer films are annealed. By introducing compositional heterogeneity, this mechanism extends the
DOI: 10.1039/c8sm01433g study of irreversible adsorption to complex chemistries and provides for a more general understanding
of how annealing should be accounted for in the proper selection and processing of polymer thin films
rsc.li/soft-matter-journal for applications in nanotechnology.

Introduction studies have shown that free surfaces act to greatly enhance
polymer mobility and reduce the glass transition temperature
Polymers confined to nanometre length scales are of rising (Tg).19,24,25 Additionally, attractive interactions between poly-
importance as technology is driven to increasingly smaller mer and substrate result in a reduction in mobility and increase
dimensions. In confined geometries, polymers have been shown in Tg in supported films.23,26 These effects, originating at
to exhibit dynamic,1–3 mechanical,4–8 and thermal9–13 properties interfaces, have been shown to penetrate tens of nanometres
much different than their corresponding bulk values. One such into a film, thus forming gradients in local properties.2,11 As a
geometry, a polymer thin film, has many industrial applications result of the direct competition of interfacial influences, the
in membrane,14,15 electronic,16,17 and drug delivery18 technologies magnitude, direction, and onset of observed deviations from
and has proven to be an excellent model system to study the bulk behaviour with film thickness vary depending on polymer
influence of interfaces on material properties. These interfaces, composition9,26,27 and interaction strength with supporting
whether they be with air (free surface),12,19,20 another polymer,21,22 substrates.23,28 For example, polystyrene (PS), which weakly
or a solid substrate,23 are largely believed to contribute to the interacts with a silicon substrate, exhibits large negative devia-
observed changes in the properties of confined polymers. Many tions from its bulk Tg below 40 nm film thickness11,12 due
to enhanced mobility at the free surface.29,30 Alternatively,
a
Department of Chemical and Biological Engineering, Princeton University, silicon substrates have hydrogen-bonding interactions with
Princeton, NJ 08544, USA. E-mail: [email protected] poly(methyl methacrylate) (PMMA) and poly(2-vinylpyridine)
b
Department of Chemistry, Key Laboratory of Advanced Textile Materials and (P2VP) films that create positive deviations in film Tg with
Manufacturing Technology of the Education Ministry, Zhejiang Sci-Tech University,
reduced thickness, below a critical value.28 However, eliminating
Hangzhou 310018, China
c
Princeton Institute for the Science and Technology of Materials, Princeton
the potential for hydrogen bonding, by removing or coating the
University, Princeton, NJ 08544, USA substrate, allows the free surface effect to dominate, resulting in
† Electronic supplementary information (ESI) available. See DOI: 10.1039/c8sm01433g negative deviations in Tg for PMMA.23,31

7204 | Soft Matter, 2018, 14, 7204--7213 This journal is © The Royal Society of Chemistry 2018
 View Article Online

Paper Soft Matter

Random copolymers exhibit thickness-dependent Tg inter- and a ‘‘loosely-bound layer’’.39,40 Jiang et al. studied the growth of
mediate to their respective homopolymers,9,32 which allows for PS, PMMA, and P2VP flattened layers further, observing an increase
the control of thin film Tg by tuning composition. Park et al. in hp with increasing polymer–substrate interaction strength.39
observed random copolymer confinement behaviour reflective The growth of PS adsorbed layers from the melt state is
of composition by ellipsometry.32 P(S-r-MMA) with a majority S commonly believed to follow two kinetic regimes: fast growth at
component showed negative deviations from bulk Tg, and the short times, corresponding to chains pinning to a largely
magnitude of these deviations decreased with higher MMA uncovered substrate, and slow surface-coverage limited growth
content. They described these deviations with a parallel- at long times, as chains adsorb through initially attached
type additive interaction parameter, in which the polymer chains and rearrange. Jiang et al. further described the fast
Published on 22 August 2018. Downloaded by Bogazici Universitesi on 10/29/2022 12:40:47 PM.

component with lower substrate interaction strength (styrene) growth regime at early stages of adsorption as a race between
dominated confinement behaviour. They found this relation zipping down of tightly-bound chains and diffusion-limited
held for P2VP-co-PS as well. Mundra et al. also measured a adsorption of loosely-bound chains.39 This two-stage mecha-
composition-dependent Tg-confinement effect in P(S-r-MMA) nism is consistent with the bimodal growth reported for chains
by intrinsic fluorescence, showing a more evenly-balanced adsorbing from solution.39,55 Housmans et al. found both these
influence from component polymers.9 stages to be dependent on annealing temperature and polymer
Annealing polymer thin films in the melt state is often an molecular weight.50 Simavilla et al. later proposed that annealing
essential step in their preparation in order to remove thermal temperature impacts the growth of homopolymer adsorbed
history, solvent, and stresses introduced during processing.33 layers from two distinct contributions: thermal fluctuations of
However, in addition to its role in erasing thermal history, the polymer chain and polymer–substrate interaction strength.53
annealing above Tg can result in the growth of an irreversibly Changes in polymer chain fluctuation with temperature depend
adsorbed nanolayer.34–40 This layer of adsorbed polymer chains on the polymer’s thermal activation energy of segmental motion
has been shown to grow in thickness and evolve in dynamic (Ea), which was reflected by an Arrhenius temperature depen-
properties with annealing time,38 while impacting film proper- dence of adsorption rate. Individual segmental interaction
ties such as diffusion,41 mobility,41–43 viscosity,44 and Tg.38,45–47 strength with the substrate was on the order of BkBT, consistent
Recently, we used a fluorescence bilayer technique to selectively with findings in adsorption from solution studies.56 Simavilla
measure Tg of exposed and buried PS adsorbed layers.48 These et al. argued that this temperature dependence was much lower
geometries enabled isolation of the impact of the free surface than that of Ea, and therefore annealing temperature (Tann)
on adsorbed layer Tg and revealed that it was suppressed at should have a much larger impact on kinetics than hp. They
later stages of growth due to stronger chain adsorption at the demonstrated that hp did not change significantly with Tann for
substrate dominating over free surface effects. Perez-de-Eulate several homopolymer systems.
et al. further demonstrated the ability of irreversible adsorption Despite the well-established influence of polymer–substrate
to largely erase a dramatic free surface effect27 through the interactions and the growing appreciation for the role of
recovery of bulk Tg in poly(tert-butylstyrene) (PtBS) thin films irreversible adsorption in determining confined film proper-
with extended annealing.49 ties, considerable investigation is still needed into how these
Since the development of adsorbed nanolayers has been influences are interrelated. While there has been limited
closely associated with deviations from bulk behaviour in thin work examining polymer–substrate interactions as related to
films,38,41,44,48,49 understanding the mechanism and kinetics of homopolymer adsorbed layer growth,39,51,53 there has been no
their growth gives insight into how annealing can impact con- examination of how different polymer–substrate interactions
fined polymer properties. Adsorbed layers have been isolated combine in the growth of random copolymer adsorbed layers.
through solvent washing and experimentally characterized in Extending the study of irreversible adsorption to polymers
terms of thickness,38,48,50 density,39,40 surface topography,39,48 with heterogeneous compositions is critical in determining
and Tg.38,45,48 These measurements enabled elucidation of its generality and application to more complex systems.
their equilibrium thickness,48,51–53 architecture,39,40 and growth Here, we study model systems of PS, PMMA, and P(S-r-MMA)
mechanism.39,50,53,54 With increased annealing time, adsorbed spanning the full range of composition. This set of polymers is
layers gradually thicken toward a quasi-equilibrium, or plateau, advantageous fundamentally because it allows for the compar-
thickness.51 Experiments and simulations on homopolymer ison between two different polymer–substrate interactions: van
systems have shown that this thickness (hp) scales with chain der Waals in the case of PS and hydrogen bonding in the case of
length following random walk statistics,51,52 with the scaling PMMA.57 Additionally, P(S-r-MMA) brushes are often used to
factor depending on polymer–substrate interactions.51,53 Chains modify substrates for PS-b-PMMA block copolymer films,58,59
adsorb in a combination of trains (consecutive segments which are of particular interest in lithography.16 Through the
adsorbed to the substrate), loops (unadsorbed stretches with use of this polymer series, we investigate the impact of specific
bound segments on either end), and tails (unadsorbed sections polymer–substrate interactions on adsorbed layer growth and
with only one adsorbed end).40 Koga and co-workers focused on how they compete in the irreversible adsorption of random
the characterization of adsorbed chain conformations resulting copolymers. We find that, unlike with their Tg confinement
from different stages of growth and identified two distinct behaviour (also strongly tied to interfacial interactions), the
sublayers: a tightly-bound, densely-adsorbed ‘‘flattened layer’’ growth metrics of P(S-r-MMA) adsorbed layers are not intermediate

This journal is © The Royal Society of Chemistry 2018 Soft Matter, 2018, 14, 7204--7213 | 7205
 View Article Online

Soft Matter Paper

to those of PS and PMMA. Rather, hp is dictated by the degree of a TA Instruments Q2000 DSC, measured as the onset of Tg
compositional heterogeneity and adsorption kinetics are heav- upon second heating at 10 K min!1 after cooling at 2.5 K min!1.
ily biased towards those of PMMA. We use these results to Rg for PS and PMMA were determined from empirical relations,
propose a copolymer adsorption mechanism in which differ- described in the ESI.†
ences in adsorption density and polymer–substrate attraction
combine to create a barrier to growth with increasing composi- Adsorbed layer preparation
tional heterogeneity. Polymer films 200 " 20 nm thick were spin-coated from dilute
solution onto piranha-treated silicon (wafers were immersed in
70 : 30 H2SO4 : H2O2 for B30 minutes and rinsed with deionized
Experimental methods
Published on 22 August 2018. Downloaded by Bogazici Universitesi on 10/29/2022 12:40:47 PM.

H2O). Films were then annealed at a given temperature (Tann)

Polymer synthesis and characterization under 10!2 Torr vacuum for set lengths of time (tads) to promote
Polymers used in this study were synthesized via free radical the growth of adsorbed layers. When annealing films relative to
polymerization of styrene (S) and methyl methacrylate (MMA), their bulk Tg, Tann was: 423 K for PS, P(S-r-MMA) (75 : 25), and
both purchased from Sigma-Aldrich. Inhibitors were removed P(S-r-MMA) (48 : 52); 430 K for P(S-r-MMA) (27 : 73); and 438 K
by running these monomers through a column packed with for PMMA. Those temperatures will be referred to as Tg + 45 K
aluminium oxide prior to their use. Relative feed ratios of S and throughout the remainder of the text. Adsorbed layers were
MMA were selected for desired copolymer composition, using then isolated through a series of three rinsing and 10 minute
reactivity ratios of 0.49 for both S and MMA.60 Reactions were soaking steps in toluene. Fresh toluene was used for each step.
conducted without additional solvent at 348 K under nitrogen These isolated adsorbed layers were then dried overnight
with precise concentrations of benzoyl peroxide (Sigma-Aldrich) at room temperature under vacuum to remove any residual
initiator to control molecular weight. Copolymer chain composi- solvent before characterization.
tional uniformity was maintained by restricting the reactions to
o15% conversion to avoid drift, and it was confirmed by compar- Adsorbed layer thickness measurement
ing the composition of polymers synthesized with different reaction All adsorbed layer characterization was conducted at room
times (see ESI†). Synthesized polymers were dissolved in temperature. Adsorbed layer thickness was determined using a
toluene (Fisher Chemical) and precipitated in methanol (Fisher Woollam M2000 Spectroscopic Ellipsometer. Each sample was
Chemical) three to four times to remove unreacted monomer, measured at a 651 angle at 3 different locations near the centre of
as confirmed by 1H-NMR. the film, each 0.3 cm apart. D and C data for wavelength range
Table 1 summarizes the characterization data obtained for 400–1680 nm were fit to a Cauchy model for transparent films.
polymers in this study. Copolymers are referred to by their Cauchy model parameters were fixed and determined for each
molar composition (X : Y) where X is the mol% S and Y is polymer by averaging fitted measurements of several thicker
the mol% MMA. Composition was determined from 1H-NMR films, as has been shown previously.48,50 Native oxide layer
spectra by comparing aliphatic and aromatic proton signatures thickness was measured on bare substrates prior to spin-coating
(see ESI†). Molecular weight was measured via size exclusion and fixed for each sample when measuring adsorbed layer thick-
chromatography in tetrahydrofuran (THF) against PS standards. ness, thus reducing the likelihood of correlation between
Weight average molecular weight (Mw) was determined for each measurement of the native oxide layer and adsorbed layer. The
polymer from light scattering data analysed using composition- ability of these models to accurately report thickness of adsorbed
weighted dn/dc values;61 homopolymer values were 0.0818 mL g!1 layers was verified through comparisons with atomic force micro-
for PMMA62 and 0.189 mL g!1 for PS.63 Dispersity, Ð, was deter- scopy measurements (see ESI†).
mined from differential refractive index data and used to calculate
Mn. The Mw of all polymers were within 10% of 108 kg mol!1 in
order to reduce the effect of molecular weight on our results. Tg Results and discussion
was determined by Differential Scanning Calorimetry (DSC) on Bulk and confined Tg of P(S-r-MMA)
The copolymers synthesized for this investigation exhibited an
Table 1 Specifications for random copolymers synthesized via free asymmetric trend in bulk Tg with composition, as shown by the
radical polymerization
orange circles in Fig. 1. P(S-r-MMA) exhibited Tg within 2 K of
Mole fraction Mw Tg Rg PS until its MMA weight fraction, wMMA, increased above B0.5.
Polymer name MMA, xMMA wMMA (kg mol!1) Ð (K) (nm) The Tg of copolymers with wMMA 4 0.5 increased rapidly and
PS 0 0 109 1.8 378 9.2 approximately linearly with wMMA until reaching that of PMMA.
P(S-r-MMA) (75 : 25) 0.25 0.24 99 1.7 376 This data set was supplemented by additional P(S-r-MMA) samples,
P(S-r-MMA) (48 : 52) 0.52 0.51 109 2.0 377 synthesized via the same method with different monomer feed
P(S-r-MMA) (27 : 73) 0.73 0.72 105 2.0 385
PMMA 1 1 119 2.2 393 9.2 ratios, to clarify the trend with composition. P(S-r-MMA) is a known
exception to the Fox equation,64 shown by the orange line in Fig. 1,
Composition was measured via 1H-NMR. Molecular weight and Ð were
measured via SEC. Tg was measured via DSC. Rg was calculated from Mw exhibiting a clear negative deviation. Other consistent data sets,
following empirical relations described in the ESI. taken from the literature, are shown by the green triangles32

7206 | Soft Matter, 2018, 14, 7204--7213 This journal is © The Royal Society of Chemistry 2018
 View Article Online

Paper Soft Matter

Examining the irreversible adsorption of P(S-r-MMA) presents

the opportunity to compare the impact of polymer–substrate
interactions on adsorption and confined Tg as well as infer how
annealing could impact confinement behaviour in copolymer
thin films.

Composition effects on adsorbed layer plateau thickness

As described previously, irreversible adsorption has been shown
to progress with annealing time, and PS, P(S-r-MMA), and PMMA
Published on 22 August 2018. Downloaded by Bogazici Universitesi on 10/29/2022 12:40:47 PM.

adsorbed layers followed the standard growth pattern. Fig. 2

plots adsorbed layer thickness for each polymer as a function of
time annealed (tads) at Tg + 45 K prior to isolation through
solvent leaching. The considerably thinner adsorbed layers
made by the copolymers, reaching a minimum in P(S-r-MMA)
(48 : 52), were immediately evident and were quantified by hp.
The differences in hp between these polymers, despite similar
Mw, indicate a composition-driven effect. However, the decrease
of hp with increasing heterogeneity, indicated by the non-
Fig. 1 Glass transition temperatures of a series of P(S-r-MMA) polymers monotonic trend with composition shown in Fig. 3A, provides
deviate from the Fox equation64 and show an asymmetric trend with an interesting contrast to the composition-dependent deviations
composition. Polymers synthesized via free radical polymerization are of confined Tg.9,32 This difference in behaviour suggests that the
shown in orange circles. Closed circles represent polymers used in this
close association of adsorption with Tg in homopolymers38,48,49
study; open circles are from additional polymers synthesized by the same
method as described above that are included to show a more complete does not hold for copolymers and that copolymer hp is impacted
composition trend. Literature values from Park et al.32 (green triangles) by additional factors that disrupt its direct increase with
and Zhang and Blum65 (purple squares) are included and also display an polymer–substrate interactions (i.e., MMA content).53 We note
asymmetric trend. that, unlike when comparing adsorbed layer thickness (hads) for
a given polymer,69 the relative thicknesses of these adsorbed
layers cannot be taken to be directly reflective of adsorbed
and purple squares.65 This large deviation from the Fox equation
amount (since the proportionality between adsorbed amount
has been studied in greater detail elsewhere and was attributed
and hads varies for each polymer). This adsorbed amount is
to contributions from dyad and triad sequences along the
associated with changes in polymer properties with annealing.38
copolymer chains.32,66,67 These sequences are particularly impor-
It is important to note that P(S-r-MMA) (27 : 73) and PMMA
tant in PMMA homopolymers, as tacticity has been shown to
were annealed at higher absolute temperature to maintain the
have a large impact on its bulk Tg and confinement behavior.‡68
constant difference between Tann and bulk Tg. Johnson et al.
Despite the influence of these chain sequences on the Tg of bulk
reported a temperature dependence of segmental interactions,
P(S-r-MMA), they do not appear to impact the Tg of P(S-r-MMA)
stronger with MMA than PS,56 which might imply that this
thin films, which reported Tg deviations proportional to their
higher Tann was solely responsible for the non-monotonic trend
composition.9,32 Dominated by interfacial influences rather than
chain sequences, P(S-r-MMA) confined Tg was attributed to the
balance between S and MMA contributions to specific inter-
actions at the substrate (van der Waals for S and hydrogen-
bonding from MMA) and free surface effects.9,32
Irreversible adsorption is also an interface-driven phenomena
and thus should not be strongly impacted by the chain
sequences that dictate bulk Tg. Changes in Tg of PS38,45,48 and
PtBS49 homopolymer films with annealing have been strongly
correlated with the growth of adsorbed layers, but this has not
been extended to systems with competing substrate interactions.

‡ Tacticity-dependent confinement behavior has been attributed to the degree of

hydrogen bonding with the substrate allowed by the relative proximity of carbonyl
groups to the native oxide layer.68 From preliminary experiments and studies of
spin-coated PMMA on Al,76 we expect tacticity also influences PMMA’s affinity for
irreversible adsorption. The PMMA used here has a 4 : 37 : 59 isotactic : atactic : Fig. 2 The growth of PS, P(S-r-MMA), and PMMA irreversibly adsorbed
syndiotactic triad ratio, as determined by 1H-NMR (see ESI†), which is consistent layers at Tg + 45 K reveals a non-monotonic trend with composition.
with what is expected for PMMA synthesized via free radical polymerization Thickness is measured by ellipsometry, and error bars represent "1
in bulk.77,78 standard deviation from the average thickness of replicate samples.

This journal is © The Royal Society of Chemistry 2018 Soft Matter, 2018, 14, 7204--7213 | 7207
 View Article Online

Soft Matter Paper

Published on 22 August 2018. Downloaded by Bogazici Universitesi on 10/29/2022 12:40:47 PM.

Fig. 3 (A) Plateau thickness, hp, and (B) linear, v, and (C) logarithmic, P, growth rates of PS, P(S-r-MMA), and PMMA adsorbed layers created at Tann = Tg +
45 K (open symbols) and Tann = 443 K (closed symbols). Stars in (A) represent hp for P(S-r-MMA) (73 : 27) and PMMA created at Tann = 423 K. hp was
calculated by averaging hads (tads 4 15 h).

Table 2 Quantification of adsorbed layer growth

Tg + 45 K
Polymer hp (nm) a b (nm!1) v (nm h!1) h0 (fit)a (nm) tcross (nm) P hcross (nm)
PS 6.0 " 0.6 0.65 " 0.07 0.25 " 0.03 0.8 " 0.23 1.0 " 0.5 5 2.4 " 0.4 4.2 " 0.1
P(S-r-MMA) (75 : 25) 3.4 " 0.5 — 0.14 " 0.02 0.30 " 0.04 1.2 " 0.2 6 2.9 " 0.3 1.8 " 0.1
P(S-r-MMA) (48 : 52) 2.9 " 0.5 — 0.13 " 0.02 0.11 " 0.01 1.09 " 0.05 10 1.5 " 0.3 2.15 " 0.05
P(S-r-MMA) (27 : 73) 4.1 " 0.4 — 0.18 " 0.02 0.2 " 0.1 1.8 " 0.4 6 1.1 " 0.8 3.2 " 0.5
PMMA 5.9 " 0.2 0.62 " 0.02 0.25 " 0.01 0.34 " 0.06 1.7 " 0.2 7 1.6 " 0.4 5.0 " 0.2

443 K
Polymer hp (nm) a b (nm!1) v (nm h!1) h0 (fit) (nm) tcross (nm) P hcross (nm)
PS 7.0 " 0.5 0.76 " 0.05 0.29 " 0.02 3.1 " 0.3 0 " 0.1 2 2.8 " 0.5 4.2 " 0.4
P(S-r-MMA) (75 : 25) 5.9 " 0.6 — 0.25 " 0.03 0.5 " 0.2 1.8 " 0.4 6 3 " 1 4.2 " 0.6
P(S-r-MMA) (48 : 52) 5.56 " 0.07 — 0.24 " 0.01 0.66 " 0.08 1.3 " 0.2 5 2.1 " 0.2 4.15 " 0.09
P(S-r-MMA) (27 : 73) 5.8 " 0.6 — 0.25 " 0.03 1.2 " 0.2 1.4 " 0.1 2 0.7 " 0.6 5.0 " 0.5
PMMA 6.3 " 0.5 0.68 " 0.05 0.27 " 0.02 2.6 " 0.7 1.4 " 0.4 2 0 " 0.7 7.1 " 0.4
Several methods for quantifying the growth patterns for PS and PMMA adsorbed layers exist in the literature and have been applied here to
homopolymer and copolymer data sets where possible. Plateau thickness, hp, was determined from averaging adsorbed layer thickness after 15 h
annealing. The values for a and b are scaling factors for Rg and N1/2, respectively, with hp. v and P are the linear and logarithmic growth rates
determined by fitting data to the model predicted by Housmans et al.50 h0 and hcross are the corresponding fitted thicknesses at tads = 0 and tcross.
a
The initial thickness at tads = 0 (h0) was fit to the data and fell within error of experimentally-measured thicknesses without annealing.

in hp. However, as indicated by the stars in Fig. 3A, annealing of PS and degree of polymerization (N) of PMMA by the
P(S-r-MMA) (27 : 73) and PMMA at 423 K (the same Tann as the following scaling relationships:51,52
other polymers) still showed an increase in hp for xMMA 4 0.5,
hp(PS) = aRg (1)
confirming that the non-monotonic trend was not simply due
to different Tann. hp(PMMA) = bN1/2 (2)
To better evaluate and understand the composition-
dependent adsorption behaviour indicated by hp, we first com- where a and b are scaling constants. We calculated values of a
pare the homopolymer extremes. Despite different interaction and b for PS and PMMA based on our experimental hp values
strengths with the substrate resulting from their mechanisms of and Rg and N derived from measured Mw and Mn, respectively.
interaction, PS and PMMA adsorbed layers have hp values within aPS was within error of that previously calculated for identically-
error of each other (listed in Table 2). We calculated hp as synthesized PS,§48 and bPMMA was within range of the b = 0.21 "
the average thickness after 15 h annealing time, and it was 0.05 nm!1 scaling reported by Durning et al. for PMMA adsorbed
consistent with hp estimated from saturated exponential fits to layers.52 Not only were aPS and bPMMA, listed in Table 2, in line
the data54 (see ESI† for details and comparison). While adsorp-
§ aPS calculated is higher than the a = 0.47 reported by Fujii et al. for mono-
tion from the melt has been studied for both PS and PMMA,
disperse PS.51 This difference is likely due to the broader distribution of chain
simultaneous study for a direct comparison of their growth has lengths in our polymer not being fully captured by average Rg and disproportion-
been limited to their flattened sublayers.39 Experimentally, ally contributing to hp. Since hp increases with molecular weight,51 it follows that
plateau thickness has been related to radius of gyration (Rg) longer chains likely increase hp, resulting in a larger a.

7208 | Soft Matter, 2018, 14, 7204--7213 This journal is © The Royal Society of Chemistry 2018
 View Article Online

Paper Soft Matter

with those reported previously, but they were within error of effects, with longer chains in the distribution leading to faster
calculated values of aPMMA and bPS, respectively. The similarities adsorption. Although lower molecular weight chains are
in hp and its scaling with Rg and N for PS and PMMA confirmed expected to preferentially adsorb in melt systems due to smaller
that chain length plays a larger role than polymer–substrate entropic penalties,71,72 they would not exclude all high mole-
interactions in determining hp for homopolymers, consistent cular weight chains. Regardless, the similar Mw and Ð of our
with theoretical arguments.52 P(S-r-MMA) adsorbed layers did polymers insured that the derived growth rates were sufficient
not follow this chain-length controlled behaviour. As shown in to compare polymer growth kinetics as a function of composi-
Table 2, their b values were almost 50% lower than those of PS tion, without molecular weight effects.
and PMMA, indicating that the largely chemistry-independent Although the kinetics of PMMA adsorption from the melt
Published on 22 August 2018. Downloaded by Bogazici Universitesi on 10/29/2022 12:40:47 PM.

scaling does not extend to copolymers. This discrepancy merited has been studied less extensively, it follows a bimodal adsorp-
an examination of the homopolymers in search of behaviour that tion pattern from solution55 that is consistent with the two-
was not captured by the comparison of their hp values. stage mechanism proposed for melt adsorption of PS, and
Adsorbed layer architecture and chain conformation studies eqn (3) has been used to determine its linear growth rate.53
provided further insight into differences in PS and PMMA Adsorption rates, v and P, are plotted against composition as
adsorbed layers, which were used to better understand copolymer the open symbols in Fig. 3B and C. The slower adsorption of
behaviour. Superimposing the structure of the tightly-bound PMMA relative to PS is consistent with both simulation and
flattened components reported by Jiang et al.39 onto our adsorbed experimental results. Both Linse and Kallrot’s bead-spring
layers, we inferred that the PMMA layer was likely comprised of a model70 and the study of flattened layers by Jiang et al.39 argued
larger fraction of tightly adsorbed chains (i.e., a thicker flattened that chains with stronger substrate-interaction strength adsorb
layer). Following this logic, PS and PMMA adsorbed layers reached more tightly to the substrate, serving as a kinetic barrier to
similar hp but for different reasons: PMMA formed a denser additional adsorption and thereby slowing growth.
adsorbed layer (due to its thicker flattened layer) with more chains As with plateau thickness, the kinetic trends observed for homo-
adsorbed, while PS formed an adsorbed layer with more loosely polymers could not be directly applied to copolymers, and adsorp-
adsorbed chains and an increased formation of loop and tail tion rate did not slow with increased MMA content. However, while
structures. This is consistent with the different scaling relation- hp increased symmetrically with increasing compositional homo-
ship between hads and adsorbed amount in different polymers.69 geneity, kinetics indicated a more biased trend. Apart from
Simulations reported by Linse and Kallrot further supported this (48 : 52), the other copolymers were both within error of PMMA’s
picture, with increasing polymer–substrate interaction strength linear growth rate, and polymers with xMMA 4 0.5 displayed
yielding higher amounts of adsorbed chains and contact points logarithmic growth rates within error of PMMA. These kinetic
per chain.70 Their bead-spring model and Monte Carlo simula- commonalities between P(S-r-MMA) and PMMA suggest a
tions also predicted a higher concentration of trains and a lower similar growth mechanism and a shared kinetic barrier.
concentration of tails in polymer chains with higher interaction
strength. Sensitivity to annealing temperature
Comparison of adsorbed layer growth at multiple annealing
Effects of composition on growth kinetics temperatures provided a means to elucidate the relative
Adsorption kinetics proved another useful metric by which to temperature-sensitivity of hp and kinetics, which enabled
compare the impact of polymer–substrate interaction strength on the isolation of different contributions to adsorption: relative
irreversible adsorption of PS and PMMA. While the secondary, polymer–substrate interaction strength and activation energy of
slower adsorbed layer growth is generally accepted to proceed segmental motion.53 Films were all annealed at 443 K (well
logarithmically,39,50 the initial, fast growth regime has been above Tg of all polymers) prior to adsorbed layer isolation. Fig. 4
described to follow either a linear50,53,54 or power law39,40 function. illustrates the changes in the resulting adsorbed layer growth
Our data showed better agreement with linear growth at short patterns, with corresponding hp and kinetic parameters repre-
times, following the model proposed by Housmans et al.:50,53 sented by the solid symbols in Fig. 3. The change in annealing
8 temperature (DTann) varied for each polymer, depending on its
< ht¼ 0 þ vt
> t o tcross Tg, and is indicated in Fig. 4 caption. It is clear that an increase
hads ðtÞ ¼ t (3) in Tann led to an overall increase in adsorption kinetics,
>
: hcross þ P log t 4 tcross
tcross consistent with previous studies on PS.50 Interestingly, copolymer
adsorbed layers also exhibited an increase in hp, which is contrary
where hads is the adsorbed layer thickness at annealing time t, v to the results of Simavilla et al. for homopolymers.53
and P are the linear and logarithmic growth rates, and tcross is a As indicated in Fig. 3A, plateau thicknesses of PS and PMMA
crossover time. These kinetic parameters, determined by fitting did not show large changes with Tann, reflecting the expected
eqn (3) to the data in Fig. 2, are reported in Table 2 with details behaviour for weakly-temperature dependent polymer–substrate
of the fitting procedure given in the ESI.† The linear growth rate interaction strength (BkBT).50,53 By contrast, P(S-r-MMA) showed
(v = 0.7 nm h!1) that we measured for PS was 2–3 times higher significant increases in hp, reaching scaling values within error
than that reported by Housmans et al. for monodisperse PS of bPS and bPMMA. This change in copolymer hp with Tann indicates
with the same Mw.50 We attribute this difference to dispersity a different governing mechanism for copolymer growth where

This journal is © The Royal Society of Chemistry 2018 Soft Matter, 2018, 14, 7204--7213 | 7209
 View Article Online

Soft Matter Paper

hp at 443 K (see ESI†), confirming that the low hp observed at

423 K was not due to slow kinetics masking a later plateau.
Increasing Tann resulted in an overall increase in linear growth
rate, with homopolymers exhibiting the largest degree of change.
While the increases in v were approximately equal for PS and
PMMA, DTann was 20 K for PS and only 5 K for PMMA. The
increased temperature sensitivity of v in PMMA was consistent
with its higher activation energy of segmental motion (Ea) relative
to PS53,73 and corresponds to a larger increase in molecular
Published on 22 August 2018. Downloaded by Bogazici Universitesi on 10/29/2022 12:40:47 PM.

motion with temperature, following Arrhenius kinetics.50,53 The

linear growth rate of P(S-r-MMA) adsorbed layers increased with
MMA content, despite a decreasing DTann, indicating an increase
in Ea and temperature-sensitivity with MMA content. Logarithmic
growth rates stayed approximately constant with DTann in PS and
S-rich copolymers, and they decreased in PMMA and MMA-rich
copolymers. We attribute this to the accelerated linear growth
of MMA-rich polymers quickly filling surface sites and reducing
opportunities for chain diffusion and/or rearrangement,
thereby slowing P.

Proposed P(S-r-MMA) adsorption mechanism

P(S-r-MMA) possesses characteristics of both PS and PMMA,
which manifest as intermediate Tg in bulk systems as well as
confined Tg in thin films. However, as shown above, these shared
characteristics do not translate directly to adsorbed layer thick-
ness. The illustrations in Fig. 5 depict proposed differences in PS
and PMMA adsorbed layer chain configuration that combine in
their random copolymers. Consistent with experimental observa-
tions and simulation results,39,70 PS (left) adsorbs more loosely
due to its relatively weak interaction with the substrate. While still
forming a tightly-bound flattened layer,74,75 it has been previously
shown to be thinner,39 thus allowing more chains to loosely
Fig. 4 Adsorbed layer growth after annealing films at 443 K reveals adsorb. These PS chains would also be more able to rearrange
composition-dependent changes from those annealed at Tg + 45 K.
at later stages of adsorption, which is reflected in their higher P.
Thickness plateaus for Tg + 45 K adsorbed layers are shown by the lines
for comparison. (A) PS, DTann = 20 K; (B) P(S-r-MMA) (75 : 25), DTann = 20 K; The overall PS adsorbed layer would therefore have a higher
(C) P(S-r-MMA) (48 : 52), DTann = 20 K; (D) P(S-r-MMA) (27 : 73), DTann = concentration of loops and tails compared to that of PMMA
13 K; (E) PMMA, DTann = 5 K. (right), which has more tightly adsorbed chains in the form of
trains due to stronger polymer–substrate interactions.
In a P(S-r-MMA) adsorbed layer (centre), these tendencies for
small increases in segmental interaction strength with the strong and weak chain adsorption from MMA and S contribu-
substrate can impact hp. tions combine to increase packing frustration and prevent
Annealing copolymers up to 60 h at Tg + 45 K did not show additional chain adsorption. The stronger MMA interactions
an increase in adsorbed layer thickness to within range of their dictate chain conformations close to the substrate, leading to

Fig. 5 Diagram of proposed differences between PS, PMMA, and P(S-r-MMA) adsorbed layers. PS exhibits loser adsorption, with more loops and tails.
PMMA has tighter chain adsorption, but also many loosely-adsorbed chains with only a few contact points. P(S-r-MMA) has tighter bound chains than PS,
giving it less large loops and tails, but it also does not have the loosely-bound chains in gaps in the tightly bound layer that PMMA does.

7210 | Soft Matter, 2018, 14, 7204--7213 This journal is © The Royal Society of Chemistry 2018
 View Article Online

Paper Soft Matter

tighter adsorption and a thicker flattened layer that increases the chains dictate copolymer adsorption rates, as in PMMA. Annealing
barrier for additional adsorption, slowing kinetics consistent at a higher temperature increased kinetics and enabled the recovery
with the similarities in kinetics with PMMA. Chain packing of copolymer hp scaling consistent with homopolymers.
frustration from different interaction strengths at the substrate These insights provided a picture of P(S-r-MMA) adsorption
could also contribute to kinetic barriers. Simultaneously, the S where tighter adsorption and weaker average individual inter-
fraction contributes toward weaker average segmental attraction action strengths resulting from compositional heterogeneity
to the substrate, which, combined with a thicker flattened layer, combine to limit the adsorption of loosely-bound chains and
does not allow many chains to penetrate to the surface and give thinner adsorbed layers. P(S-r-MMA) formed thicker
loosely adsorb. With limited adsorption of loosely-bound chains, tightly-bound layers at the substrate than PS, which served
Published on 22 August 2018. Downloaded by Bogazici Universitesi on 10/29/2022 12:40:47 PM.

P(S-r-MMA) adsorbed layers would not greatly exceed the thick- as a barrier to additional adsorption of chains with weaker
ness of their flattened layer, thus forming adsorbed layers than average segment-substrate interactions than PMMA. Raising Tann
their homopolymer components. In PMMA, individual segments increased the segmental attraction to the substrate sufficiently
have interactions strong enough to diffuse through denser, to overcome the growth barrier imposed by the tightly-bound
tightly-bound chains to adsorb. In PS, weaker-adsorbing chains layer and allowed for additional chains to be adsorbed.
are less encumbered by the flattened layer, and chains form Differences between confined Tg and adsorption trends
more loops and tail configurations. This competition between suggest that polymer–substrate interactions are not sufficient
tight and loose chain adsorption increases with increasing to predict adsorption of compositionally-complex systems. Their
compositional heterogeneity. Unbalanced compositions in reduced adsorbed layer thickness implies that P(S-r-MMA) films
P(S-r-MMA) (27 : 73) and (75 : 25) allowed MMA and S units to are less susceptible to annealing-induced property changes,
dominate the adsorption mechanism, respectively, resulting providing copolymerization as a route to reducing processing
in larger hp. However, the nearly symmetric composition of effects on thin film behaviour. However, in order to directly
P(S-r-MMA) (48 : 52) led to the most dramatic competition of connect our results to changes in film properties, additional
influences and hp was thinner accordingly. The different struc- information is needed to determine precisely how hads translates
tures of copolymer adsorbed layers explain proportionality to adsorbed amount in copolymers. Work should also be under-
differences between hads and adsorbed amount. This likely taken to determine if copolymers with similar polymer–substrate
accounts for the different trends between deviations from bulk interactions show inhibited adsorption.
Tg under confinement9,32 and hp with composition.
Finally, the temperature-sensitivity of copolymer adsorption
is consistent with this proposed mechanism. The large change Conflicts of interest
in copolymer hp at elevated Tann indicated a change in polymer– There are no conflicts to declare.
substrate interaction strength and supports the idea of limited
loosely bound chain adsorption. In copolymers, this small
increase in BkBT interactions allows for the penetration of Acknowledgements
additional chains through the tightly-bound layer to bind with
the substrate. However, the fact that P(S-r-MMA) (27 : 73) did R. D. P. acknowledges the support of the National Science
not show a change in hp when created at Tann = 423 K implies Foundation (NSF) Materials Research Science and Engineering
that a sufficiently high Tann is needed to overcome the barrier to Center Program through the Princeton Center for Complex
adsorption imposed by tightly-bound chains. The relatively Materials (DMR-1420541) and the AFOSR through a PECASE
small change in copolymer v demonstrated that the thermal Award (FA9550-15-1-0017).
fluctuation of chains played a less important role in limiting
adsorption than interaction energy with the substrate. Notes and references
1 J. Pye, K. Rohald, E. Baker and C. Roth, Macromolecules,
Conclusions 2010, 43, 8296.
2 R. Priestley, C. Ellison, L. Broadbelt and J. Torkelson,
Through the examination of adsorbed layer growth in PS, Science, 2005, 309, 456.
PMMA, and a series of their random copolymers, we gained a 3 C. Zhang, V. Boucher, D. Cangialosi and R. Priestley,
clearer picture of how different polymer–substrate interactions Polymer, 2013, 54, 230.
combine to influence irreversible adsorption in systems with 4 C. Stafford, C. Harrison, K. Beers, A. Karim, E. Amis,
compositional heterogeneity. Our results revealed that the M. VanLandingham, H. Kim, W. Volksen, R. Miller and
irreversible adsorption of P(S-r-MMA) faces additional barriers E. Simonyi, Nat. Mater., 2004, 3, 545.
to growth than either of its component homopolymers and 5 J. Torres, C. Wang, B. Coughlin, J. Bishop, R. Register,
is not intermediate to their two behaviours. This differed R. Riggleman, C. Stafford and B. Vogt, Macromolecules,
substantially from the blended properties usually displayed 2011, 44, 9040.
by random copolymers, which extend to their confined Tg 6 C. Stafford, B. Vogt, C. Harrison, D. Julthongpiput and
behavior.9,32 Kinetic analysis confirmed that tightly binding R. Huang, Macromolecules, 2016, 39, 5095.

This journal is © The Royal Society of Chemistry 2018 Soft Matter, 2018, 14, 7204--7213 | 7211
 View Article Online

Soft Matter Paper

7 M. Zhang, S. Askar, J. Torkelson and L. Brinson, Macro- 41 N. Jiang, L. Sendogdular, X. Di, M. Sen, P. Gin, M. Endoh,
molecules, 2017, 50, 5447. T. Koga, B. Akgun, M. Dimitriou and S. Satija, Macromole-
8 B. Vogt, J. Polym. Sci., Part B: Polym. Phys., 2018, 56, 9. cules, 2015, 48, 1795.
9 M. Mundra, C. Ellison, R. Behling and J. Torkelson, Polymer, 42 C. Rotella, S. Napolitano, L. De Cremer, G. Koeckelberghs
2006, 47, 7747. and M. Wübbenhorst, Macromolecules, 2010, 43, 8686.
10 S. Kim, S. Hewlett, C. Roth and J. Torkelson, Eur. Phys. J. E: 43 J. Xu, Z. Liu, Y. Lan, B. Zuo, X. Wang, J. Yang, W. Zhang and
Soft Matter Biol. Phys., 2009, 30, 83. W. Hu, Macromolecules, 2017, 50, 6804.
11 C. Ellison and J. Torkelson, Nat. Mater., 2003, 2, 695. 44 T. Koga, N. Jiang, P. Gin, M. Endoh, S. Narayanan, L. Lurio
12 J. Keddie, R. Jones and R. Cory, Europhys. Lett., 1994, 27, 59. and S. Sinha, Phys. Rev. Lett., 2011, 107, 225901.
Published on 22 August 2018. Downloaded by Bogazici Universitesi on 10/29/2022 12:40:47 PM.

13 J. Forrest, K. Dalnoki-Veress, J. Stevens and J. Dutcher, Phys. 45 S. Napolitano, C. Rotella and M. Wübbenhorst, ACS Macro
Rev. Lett., 1996, 77, 2002. Lett., 2012, 1, 1189.
14 H. Park, J. Kamcev, L. Robeson, M. Elimelech and B. Freeman, 46 N. Jiang, M. Sen, M. Endoh, T. Koga, E. Langhammer,
Science, 2017, 356, 1137. P. Bjöörn and M. Tsige, Langmuir, 2018, 34, 4199.
15 E. Jackson and M. Hillmyer, ACS Nano, 2010, 4, 3548. 47 S. Sun, H. Xu, J. Han, Y. Zhu, B. Zuo, X. Wang and W. Zhang,
16 C. Bates, M. Maher, D. Janes, C. Ellison and C. Willson, Soft Matter, 2016, 12, 8348.
Macromolecules, 2014, 47, 2. 48 M. Burroughs, S. Napolitano, D. Cangialosi and R. Priestley,
17 C. Hardy and C. Tang, J. Polym. Sci., Part B: Polym. Phys., Macromolecules, 2016, 49, 4647.
2013, 51, 2. 49 N. Perez-de-Eulate, M. Sferrazza, D. Cangialosi and
18 S. Yang, J. Yang, E. Kim, G. Jeon, E. Oh, K. Choi, S. Hahn S. Napolitano, ACS Macro Lett., 2017, 6, 354.
and J. Kim, ACS Nano, 2010, 4, 3817. 50 C. Housmans, M. Sferrazza and S. Napolitano, Macromolecules,
19 O. Bäumchen, J. McGraw, J. Forrest and K. Dalnoki-Veress, 2014, 47, 3390.
Phys. Rev. Lett., 2012, 109, 55701. 51 Y. Fujii, Z. Yang, J. Leach, H. Atarashi, K. Tanaka and
20 S. Kim and J. Torkelson, Macromolecules, 2011, 44, 4546. O. Tsui, Macromolecules, 2009, 42, 7418.
21 R. Baglay and C. Roth, J. Chem. Phys., 2015, 143, 111101. 52 C. Durning, B. O’Shaughnessey, U. Sawhney, D. Nguyen,
22 R. Baglay and C. Roth, J. Chem. Phys., 2017, 146, 203307. J. Majewski and G. Smith, Macromolecules, 1999, 32, 6772.
23 J. Keddie, R. Jones and R. Cory, Faraday Discuss., 1994, 98, 219. 53 D. Simavilla, W. Huang, P. Vandestrick, J. Ryckaert, M. Sferrazza
24 J. Sharp and J. Forrest, Phys. Rev. Lett., 2003, 91, 235701. and S. Napolitano, ACS Macro Lett., 2017, 6, 975.
25 M. Ediger and J. Forrest, Macromolecules, 2014, 47, 471. 54 D. Simavilla, A. Panagopoulou and S. Napolitano, Macromol.
26 R. Priestley, M. Mundra, N. Barnett, L. Broadbelt and Chem. Phys., 2017, 219, 1700303.
J. Torkelson, Aust. J. Chem., 2007, 60, 765. 55 H. Schneider, P. Frantz and S. Granick, Langmuir, 1996,
27 C. Ellison, M. Mundra and J. Torkelson, Macromolecules, 12, 994.
2005, 38, 1767. 56 H. Johnson, J. Douglas and S. Granick, Phys. Rev., 1993,
28 C. Roth, K. McNerny, W. Jager and J. Torkelson, Macro- 70, 3291.
molecules, 2007, 40, 2568. 57 H. Johnson and S. Granick, Macromolecules, 1990, 23, 3367.
29 K. Paeng, R. Richert and M. Ediger, Soft Matter, 2011, 8, 819. 58 P. Mansky, Y. Liu, E. Huang, T. Russell and C. Hawker,
30 Z. Fakhraai and J. Forrest, Science, 2008, 319, 600. Science, 1997, 275, 1458.
31 C. Roth, A. Pound, S. Kamp, C. Murray and J. Dutcher, Eur. 59 S. Pujari, M. Keaton, P. Chaikin and R. Register, Soft Matter,
Phys. J. E: Soft Matter Biol. Phys., 2006, 20, 441. 2012, 8, 5358.
32 C. Park, J. Kim, M. Ree, B. Sohn, J. Jung and W. Zin, Polymer, 60 M. Coote, L. Johnston and T. Davis, Macromolecules, 1997,
2004, 45, 4507. 30, 8191.
33 J. Chung, T. Chastek, M. Fasolka, H. Ro and C. Stafford, ACS 61 W. Bushuk and H. Benoit, Can. J. Chem., 1958, 36, 1616.
Nano, 2009, 3, 844. 62 D. Christie, R. Register and R. Priestley, ACS Cent. Sci., 2018,
34 A. Panagopoulou and S. Napolitano, Phys. Rev. Lett., 2017, 4, 504.
119, 097801. 63 A. Burns and R. Register, Macromolecules, 2016, 49, 269.
35 C. Rotella, S. Napolitano and M. Wübbenhorst, Macro- 64 C. Brazel and S. Rosen, Fundamental Principles of Polymeric
molecules, 2009, 42, 1415. Materials, Wiley, Hoboken, NJ, 3rd edn, 2012.
36 S. Napolitano, Non-equilibrium Phenomena in Confined Soft 65 B. Zhang and F. Blum, Macromolecules, 2003, 36, 8522.
Matter: Irreversible Adsorption, Physical Aging, and Glass 66 G. Liu, Z. Meng, W. Wang, Y. Zhou and L. Zhang, J. Phys.
Transition at the Nanoscale, Springer, Cham, 2015. Chem. B, 2008, 112, 93.
37 S. Napolitano, S. Capponi and B. Vanroy, Eur. Phys. J. E: 67 G. Liu, L. Zhang, Y. Yao, L. Yang and J. Gao, J. Appl. Polym.
Soft Matter Biol. Phys., 2013, 36, 1. Sci., 2003, 88, 2891.
38 S. Napolitano and M. Wübbenhorst, Nat. Commun., 2011, 2, 260. 68 Y. Grohens, L. Hamon, G. Reiter, A. Soldera and Y. Holl, Eur.
39 N. Jiang, J. Shang, X. Di, M. Endoh and T. Koga, Macro- Phys. J. E: Soft Matter Biol. Phys., 2002, 8, 217.
molecules, 2014, 47, 2682. 69 D. Simavilla, W. Huang, C. Housmans, M. Sferrazza and
40 P. Gin, N. Jiang, C. Liang, T. Taniguchi, B. Akgun, S. Satija, S. Napolitano, ACS Cent. Sci., 2018, 4, 755.
M. Endoh and T. Koga, Phys. Rev. Lett., 2012, 109, 265501. 70 P. Linse and N. Kallrot, Macromolecules, 2010, 43, 2054.

7212 | Soft Matter, 2018, 14, 7204--7213 This journal is © The Royal Society of Chemistry 2018
 View Article Online

Paper Soft Matter

71 J. Van der Gucht, N. Besseling and G. Fleer, Macromolecules, 75 M. Sen, N. Jiang, M. Endoh, T. Koga, A. Ribbe, A. Rahman,
2002, 35, 6732. D. Kawaguchi, K. Tanaka and D. Smilgies, Macromolecules,
72 J. Van Der Gucht, N. Besseling and G. Fleer, Macromolecules, 2018, 51, 520.
2004, 37, 3026. 76 Y. Grohens, M. Brogly, C. Labbe and J. Schultz, Polymer,
73 L. Manring, Macromolecules, 1988, 21, 528. 1997, 38, 5913.
74 M. Sen, N. Jiang, J. Cheung, M. Endoh, T. Koga, 77 Y. Isobe, K. Yamada, T. Nakano and Y. Okamoto, J. Polym.
D. Kawaguchi and K. Tanaka, ACS Macro Lett., 2016, Sci., Part A: Polym. Chem., 2000, 38, 4693.
5, 504. 78 R. Kirste, W. Kruse and K. Ibel, Polymer, 1975, 16, 120.
Published on 22 August 2018. Downloaded by Bogazici Universitesi on 10/29/2022 12:40:47 PM.

This journal is © The Royal Society of Chemistry 2018 Soft Matter, 2018, 14, 7204--7213 | 7213