on a Mobility Gradient of Polymer Chains near an Impenetrable Solid

Mar 6, 2015 - Effect of CO2 on a Mobility Gradient of Polymer Chains near an. Impenetrable Solid. Naisheng Jiang,. †. Levent Sendogdular,. †. Xiao...
0 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/Macromolecules

Effect of CO2 on a Mobility Gradient of Polymer Chains near an Impenetrable Solid Naisheng Jiang,† Levent Sendogdular,† Xiaoyu Di,† Mani Sen,† Peter Gin,† Maya K. Endoh,§ Tadanori Koga,*,†,§,‡ Bulent Akgun,∥,⊥ Michael Dimitriou,∥ and Sushil Satija∥ †

Department of Materials Science and Engineering and ‡Chemical and Molecular Engineering Program, Stony Brook University, Stony Brook, New York 11794-2275, United States § Department of Chemistry, Stony Brook University, Stony Brook, New York 11794-3400, United States ∥ Center for Neutron Research, National Institute of Standards and Technology, Gaithersburg, Maryland 20899, United States ⊥ Department of Chemistry, Bogazici University, Bebek, Istanbul 34342, Turkey S Supporting Information *

ABSTRACT: We report a mobility gradient of polymer chains in close proximity of a planar solid substrate in compressed carbon dioxide (CO2) gas. A series of bilayers composed of bottom hydrogenated polystyrene (h-PS) and top deuterated PS (d-PS) layers were prepared on Si substrates. A high-pressure neutron reflectivity (NR) technique was used to study the diffusive motion at the h-PS/dPS interface as a function of the distance from the substrate interface. The results reveal that the interdiffusive chain dynamics gets strongly hindered compared to the bulk when the distance from the substrate is less than 3Rg (Rg is the radius of polymer gyration of the h-PS). At the same time, by utilizing rapid quench of CO2 and subsequent solvent leaching, we reveal the presence of the CO2-induced polymer adsorbed layer on the substrate. We postulate that loop components in the adsorbed polymer chains provide a structure that can trap the neighboring polymer chains effectively, hence reducing the chain mobility in the close vicinity of the solid substrate even in the presence of the effective plasticizer.

I. INTRODUCTION A polymer thin film on top of a solid is of vital importance in many technologies (e.g., coating, adhesion, corrosion) as well as new emerging nanotechnologies such as organic photovoltaics, semiconductor chips, and biosensors.1 A spin-coating process is a well-established technique to prepare homogeneous polymer thin films on planar substrates. But, it is also known that this rapid solvent evaporation process results in nonequilibrium stressed conformations of polymer chains on substrates, and such residual stress causes film instability2−4 and changes in properties of polymer thin films.5,6 In order to achieve full relaxation of the residual stress and equilibration of the chain conformations, prolonged thermal annealing (at temperatures far above the bulk glass transition temperature, Tg) compared to bulk reputation times7 is typically required.4,5,8 On the other hand, such thermal annealing expedites irreversible polymer adsorption even onto weakly attractive solids via physisorption.9 Based on the approach proposed by Guiselin,10 which combines thermal annealing at T ≫ Tg and subsequent solvent leaching, several research groups have shown the formation of very thin polymer adsorbed layers with less than 1Rg thick (Rg is the radius of polymer gyration) on planar solid surfaces.11−21 From the thermodynamic point of view, the chains adsorption can be in principle drawn by a counterbalance between the conformational entropy loss during the transition from a © 2015 American Chemical Society

randomly coiled state to the adsorbed state and the energy gain of attached segments to the solid surface in the total free energy.9 Because of many (∼N1/2) physical contacts of polymer repeating units with a solid surface,10 it has been believed that the adsorbed chains are nearly immobile in air, i.e., no thermal expansion18,20,22,23 and no interdiffusion17 with the free chains even at temperatures far above the bulk Tg of a polymer. Moreover, many studies have shown long-range perturbations associated with the adsorbed layer in viscosity,17 chain diffusion,24−27 and crystalline structures,19,21 which compete against the opposite contributions from a surface “mobile” layer at the air/polymer interface.28−31 It is postulated that chain entanglements between the adsorbed chains and free polymer chains in a matrix are responsible for the long-range perturbations through the so-called “reduced mobility interface layer”32,33 that acts as a “transition zone” to ensure continuity in the mobility profile from the adsorbed layer to the bulk. Because of recent considerable attention to solvent vapor annealing in the context of controlling orientations of block copolymer microdomain structures on solids,34 the fundamental, but unsolved, question arises: Are the long-range Received: December 23, 2014 Published: March 6, 2015 1795

DOI: 10.1021/ma502591x Macromolecules 2015, 48, 1795−1803

Article

Macromolecules

care should be taken when handing it) for at least 30 min, subsequently rinsed with deionized water thoroughly, and followed by submersion in an aqueous solution of hydrogen fluoride (HF) to remove a native oxide (SiO2) layer. However, it should be noted that a thin SiO2 layer (about 1.3 nm in thickness) was reproduced even just after hydrofluoric acid etching, as reported by Shin and co-workers.52 We also confirmed that the water contact angle of the HF-etched Si substrate (covered with the 1.3 nm thick SiO2 layer) is 82 ± 1° (i.e., a hydrophobic surface) which enabled us to prepare stable PS thin films on the substrates during the CO2 annealing. Hereafter, we assign the HF-etched Si substrate covered with the 1.3 nm thick SiO2 layer as the “HF-etched Si”. For the interdiffusion NR experiments, h-PS (Mw = 650 kDa, Mw/ Mn = 1.1, Pressure Chemical Co., hereafter assigned as 650khPS) and five different molecular weights of d-PS (Mw = 90, 120, 334, 676, and 1323 kDa, Mw/Mn < 1.1, Polymer Source Inc.) were used to prepare bilayer films. Bottom h-PS thin films with different thicknesses (0.6Rg, 1Rg, 2Rg, 3Rg, and 4Rg thick) were prepared by spin-coating PS/ toluene solutions onto the HF-etched Si substrates. The thicknesses of the spin-cast PS thin films were measured by an ellipsometer (Rudolf Auto EL-II). The d-PS overlayers, whose thicknesses were fixed to 3Rg, were then floated onto the bottom h-PS layers from a bath of deionized water. The resultant bilayer films were dried at 50 °C under vacuum for 24 h to remove residual solvent and water molecules trapped before the interdiffusion experiments. We confirmed that no interdiffusion at the h-PS/d-PS interface occurred under the mild condition. In order to illuminate the interdiffusive property of the adsorbed layer, we also prepared bilayers of bottom 650khPS adsorbed layers and top 3Rg thick d-PS (Mw = 334 kDa, assigned as 334kdPS) layers. To prepare the adsorbed layer, we reproduced the established protocol for PS:18,20 Approximately 50 nm thick spin-cast films prepared on the HF-etched Si substrates were annealed at 150 °C (∼ Tg + 50 °C) for prolong time (2−5 days) under vacuum below 10−3 Torr; the films were then leached in baths of a fresh good solvent (toluene) at room temperature until the resultant film thickness remained constant.10,12 It should be noted that the irreversibly adsorbed polymer layer is composed of the two different nanoarchitectures: flattened chains that constitute the inner higher density region of the adsorbed layer and loosely adsorbed polymer chains that form the outer bulk-like density region.8 It is possible to prepare the lone PS flatten layer by tuning the leaching conditions,18,20 but in this study we focus only on the adsorbed layers composed of the two layers to prevent any artificial effects due to the heterogeneous surface of the lone PS flattened layer. After the leaching process, the PS adsorbed layers were dried in a vacuum oven to remove any excess solvent trapped in the films, and the thickness was measured at room temperature by using X-ray reflectivity or the ellipsometer before the NR experiments. We also prepared a single d-PS (Mw = 676 kDa, assigned as 676kdPS) adsorbed layer on the Si substrate to investigate the swelling behavior in CO2 by using NR. II-2. In Situ Neutron Reflectivity (NR). Specular NR measurements were performed on the NG-7 horizontal reflectometer at the National Institute of Standards and Technology, Center for Neutron Research. The wavelength (λN) of the neutron beams was 0.47 nm with ΔλN/λN = 2.5%. Details of high-pressure NR experiments including a custom-made high-pressure cell have been described elsewhere.35 The temperature and pressure stabilities during the NR measurements were within an accuracy of ±0.1 °C and ±0.2%, respectively. The isothermal swelling experiments (T = 36 °C) of the 676kdPS adsorbed layer were conducted with elevated pressures up to P = 17.5 MPa. The d-PS adsorbed layer was exposed to CO2 for up to 4 h prior to data acquisition to ensure the equilibrium swelling. The scattering length density (SLD) of CO2, which varies from 0.0004 × 10−4 to 2.5× 10−4 nm−2 in the pressure range of 0.1 < P < 17.5 MPa at T = 36 °C, were calculated based on the density of CO2 obtained by the equation of state.53 The NR data were obtained by successively increasing pressure and then slowly decreasing pressure. Since the background scattering from a pure CO2 phase increases dramatically near the critical point,35,36 we measured the scattering from the pure

perturbations associated with the adsorbed layer facilitated or diluted in the presence of solvent molecules? To answer the question, we chose interdiffusion experiments at the chemically identical polymer/polymer interfaces in the presence of CO2 molecules using high-pressure neutron reflectivity (NR) technique.35 Polystyrene (PS) was used since (i) a wide variety of monodisperse hydrogenated and deuterated PS are commercially available, (ii) the long-range perturbations on Si substrates have been characterized in melts,17,24,25 (iii) the protocol to prepare the adsorbed layer has been established,18,20 and (iv) the swelling35−39 and interdiffusion40,41 behaviors of polymer thin films in CO2 have been reported. It is known that sorption of CO2 molecules into polymers plays a role as a diluent or plasticizer for glassy polymers by significantly lowering the glass transition temperature (Tg) and hence enhancing the chain mobility.42−45 Especially, several research groups have shown that the excess sorption of CO2 molecules takes place within the narrow temperature and pressure regimes near the critical point of CO2, known as the “density fluctuation ridge”,46 resulting in the anomalous swelling of supported polymer thin films regardless of a choice of polymers.35−38,41,47−50 Furthermore, it has been shown that the excess sorption of CO2 molecules leads to the excess interdiffusion at the hydrogenated and deuterated PS interface.41 In the present study, a series of bilayer films composed of bottom hydrogenated PS (h-PS) spin-cast films of different film thicknesses ranging from 0.6Rg to 4Rg thick and deuterated PS (d-PS) overlayers of the fixed thickness (i.e., 3Rg thick) were prepared on Si substrates. The interdiffusion processes at the density fluctuation ridge condition of CO2 (T = 36 °C and P = 8.2 MPa) were investigated by in situ NR. The NR results reveal that the effective diffusion constants (Deff) of the polymer chains depend on the distance from the substrate. However, we also found that the critical threshold (∼3Rg thick from the substrate interface), above which the substrate effect can be negligible, is much less than those reported in the PS melts (10Rg thick25 or 25Rg thick51 from the substrate interface). In order to provide further insight into the gradient of segmental diffusion in the close proximity of the substrate, we also investigated the polymer adsorption via the CO2 annealing. Spin-cast h-PS thin films (without prior thermal annealing treatment) were exposed to CO2 at the ridge condition and subsequently quenched to air to preserve the swollen structures of the PS chains via the vitrification process.41 The exposed films were then leached with toluene thoroughly to remove unadsorbed chains and characterized by ex situ X-ray reflectivity/ellipsometry experiments. The results are intriguing to show the formation of the CO2-induced polymer adsorbed layer of about 7 nm in thickness on the substrate regardless of the original dry film thickness. Hence, we may conclude that the robust entanglements between the irreversibly adsorbed polymer chains and neighboring unadsorbed chains result in the critical threshold even in the presence of the effective plasticizer, while the free chains in the polymer/solvent matrix are freely entangled/disentangled with the neighboring unadsorbed chains in CO2.

II. EXPERIMENTAL SECTION II-1. Sample Preparation. Si wafers were precleaned using a piranha solution (i.e., a mixture of H2SO4 and H2O2; caution: a piranha solution is highly corrosive upon contact with skin or eyes and is an explosion hazard when mixed with organic chemicals/materials; extreme 1796

DOI: 10.1021/ma502591x Macromolecules 2015, 48, 1795−1803

Article

Macromolecules fluid phase (i.e., the long-range density fluctuations) for each pressure condition. The NR data corrected for the background scattering were analyzed by comparing the observed reflectivity curves with the calculated ones based on model SLD profiles with three fitting parameters for each layer: film thickness, SLD, and roughness between the layers represented as a Gaussian function.54 The SLD profiles were subsequently converted into the corresponding polymer volume fraction profiles. Assuming that the concentration of the mixture is homogeneous through the entire film, the SLD value of the polymer/ CO2 system is defined by

SLDmix = SLDpolymer × ϕ(z) + SLDCO × (1 − ϕ(z))

(1)

where SLDmix is the SLD value of the mixture at a distance z from the substrate, SLDpolymer and SLDCO are the pure component SLDs of the polymer and CO2, respectively, and ϕ(z) is the volume fraction of the polymer at a distance z from the substrate. The density of CO2 dissolved in the polymer was taken to be 0.956 g/cm3 since the molar volume of CO2 within the polymer can be much different from the molar volume of the bulk CO2.55 To ensure conservation of mass, the volume fraction profiles were calculated such that the same amount of the polymer chains remained at all solvent concentrations including in the dry state. The interdiffusion experiments were conducted at the CO2 density fluctuation ridge condition (i.e., P = 8.2 MPa and T = 36 °C) where the interdiffusion process under the isothermal condition is maximized due to the excess absorption of CO2 molecules.41 We confirmed that the interdiffusion process is much slower than the data acquisition time (∼80 min) for each NR run. The exposure times used in the present NR study correspond to elapsed times at the middle of each reflectivity experiment after we set to the ridge condition. II-3. X-ray Reflectivity (XR). Before the NR experiments, we characterized the adsorbed layers developed by the thermal annealing and solvent leaching processes by means of XR. The XR experiments at room temperature were performed at the X10B and X20A beamlines of the National Synchrotron Light Source, Brookhaven National Laboratory. The specular reflectivity was measured as a function of the scattering vector in the perpendicular direction, qz = 4π sin θ/λ, where θ is the incident angle and λ is the X-ray wavelength (λ = 0.087 nm at X10B and λ = 0.118 nm at X20A, which are equivalent to the X-ray energy of 14.2 and 10.5 keV, respectively). The XR data were fit by using a standard multilayer fitting routine for a dispersion value (δ in the X-ray refractive index) in conjunction with a Fourier transformation method, a powerful tool to obtain detailed structures for low X-ray contrast polymer multilayers.56,57 In addition, as will be discussed later, the adsorbed layers derived from the CO2 exposed thin films were also characterized by XR.

Figure 1. (a) Representative NR profiles for the 676kdPS adsorbed layer at the four different pressures under the isothermal condition of T = 36 °C. The solid lines correspond to the best fits to the data based on the corresponding concentration profiles shown in (b). z corresponds to the distance from the SiO2 interface.

III. RESULTS AND DISCUSSION III-1. Swelling Experiments. We start with the unique swelling behavior of the adsorbed layer prepared by the thermal annealing in CO2. Figure 1a shows representative NR profiles for the 676kdPS adsorbed layer in CO2 at four different pressures under the isothermal condition (T = 36 °C): P = 0.1 MPa (air), P = 5.8 MPa, P = 8.2 MPa, and P = 10.4 MPa. The thickness of the adsorbed layer, which was initially 12 nm (which is equivalent to 0.6Rg thick), slightly increases and reaches the maximum value of 13 nm at around P = 8.2 MPa, i.e., the density fluctuation ridge. Figure 1b shows the volume fraction profile of the 676kdPS obtained from the best fits. Since the solvent quality of CO2 for PS is poor even at the ridge,37 the simple boxlike shape profile could be approximated to represent the volume fraction profiles of the swollen adsorbed layer. The linear dilation (Sf) of the adsorbed layer in the direction normal to the surface was then calculated by the equation Sf = (L − L0)/L0, where L and L0 are the measured thicknesses of the swollen and unswollen adsorbed layer, respectively. Figure 2 summarizes the pressure dependence of

Figure 2. Pressure dependences of the linear dilation (Sf) of the 676kdPS adsorbed layer and spin-cast film. Both thicknesses in the dry state were 12 nm thick.

the Sf values. We confirmed that the repeated pressurization and depressurization processes exhibit the same swelling isotherm. The evidence of the excess swelling maximum is consistent with previous experimental data that the anomalous swelling occurs not only at the polymer−CO2 interface but also at the polymer−substrate interface.48,58,59 We previously found that the anomalous swelling of the entire d-PS thin films with the scaled initial thickness of 1.2Rg < L0 < 8Rg is approximated by the exponential function36 Sf (T = 36 °C, P = 8.2 MPa) = 0.11 + 0.37 exp[−0.53L0 /R g ] (2) 1797

DOI: 10.1021/ma502591x Macromolecules 2015, 48, 1795−1803

Article

Macromolecules Hence, the equation along with the present NR results implies that the first (constant) term in the right-hand side of eq 2 is attributed to the excess swelling at the substrate interface (i.e., the adsorbed layer), which is independent of L0, while the second term is originated from a surface effect at the CO2/ polymer interface.35−38,41,47−50 According to a report by Jia and McCarthy,60 CO2 interacts with a hydroxyl group on a Si substrate, screening the polymer−substrate interactions. However, as they also pointed out, when a strong interaction between a polymer and substrate is present, the screening effect of CO2 is not fully achieved.60 Hence, it may be reasonable to suppose that the interaction between PS and the HF-etched Si is attractive enough.14 We are currently performing NR experiments using different polymer−substrate systems to provide further insight into the screening effect of CO2. At the same time, the experimental data suggest that the excess absorption of CO2 molecules would facilitate polymer adsorption on the substrate rather than chain stretching away from the substrate. To address this hypothesis, we also prepared an as-cast 676kdPS film of the same thickness as the adsorbed layer (12 nm) on the HF-etched Si substrate and measured the swelling behavior in CO2. As shown in Figure 2, the entire swelling isotherm including near the ridge condition is in good agreement with that of the adsorbed layer. This indicates that the polymer chains tend to increase the segment−solid contacts so as to overcome the conformational entropy loss in the total free energy61 in CO2, transforming the 0.6Rg thick interfacial region into the irreversibly adsorbed layer. Below, we further discuss the formation process of the adsorbed layer via the CO2 annealing. III-2. Adsorbed Layers Formed via the CO2 Annealing. In order to verify the polymer adsorption near the substrate induced by CO2, we attempted NR experiments for 676kdPS spin-cast thin films (20−50 nm in thickness) at the CO2 ridge condition. However, as mentioned above, the intensity at qz > 0.1 nm−1, where we can attain the information about the adsorbed layer, was completely overwhelmed by the scattering intensity from pure CO2.36 In order to overcome this difficulty, we instead performed ex situ experiments by utilizing subsequent rapid evaporation of CO2, which induces the vitrification of the polymer and allows us to preserve the swollen structures even after the depressurization.41 We prepared 650khPS spin-cast films of the 0.6Rg (i.e., 12 nm) and 2.3Rg (i.e., 50 nm) thick on the HF-etched Si substrates. In order to illuminate the effect of CO2 on polymer adsorption, these spin-cast films were not annealed at a high temperature such that the polymer chains were hardly bound to the substrate. As shown in the Supporting Information, there is a very limited amount of the polymer chains adsorbed onto the HF-etched Si substrate (the surface coverage is less than 10%). This set of data can be used as a reference for the CO2-induced polymer adsorption. The spin-cast films (without prior thermal annealing treatment) were then exposed to CO2 at the ridge condition (T = 36 °C and P = 8.2 MPa) with different exposure times and subsequently quenched to air (within 10 s) to preserve the structures of the swollen PS chains. The exposed films were then leached with fresh toluene thoroughly to remove unadsorbed chains. Figure 3 shows a representative XR profile of the PS adsorbed layer formed after the CO2-annealing of 72 h. The best fit (the solid line in Figure 3) to the data was obtained by using a four-layer (a Si substrate, a SiO2 layer, and two PS layers with different densities) dispersion model which was

Figure 3. Representative XR curve of the PS adsorbed layer formed after via the CO2-annealing at T = 36 °C and P = 8.2 MPa at tco = 72 h. The solid line corresponds to the best-fit to the data based on the density profile against z from the SiO2 surface shown in the inset. The dispersion value was converted into the density of the film relative to the bulk by using the measured dispersion (δ) value of the bulk (δbulk = 1.14 × 10−6 with the X-ray energy of 14.2 keV).

determined based on the corresponding Fourier transformation (FT) of the XR profile. The detailed XR analysis along with the FT method has been described elsewhere.18 The corresponding density profile of the adsorbed layer was shown in the inset of Figure 3. Hence, the XR results clearly show that the PS adsorbed layer formed via the CO2-annealing is composed of a 2.0 nm thick inner higher density layer (∼10% higher density than the bulk) and a 4.8 nm outer bulklike density layer. This two-layer formation is consistent with the one formed via the thermal annealing process.18,20 Figure 4 shows the total thickness of the adsorbed layers (had) measured by XR as a function of the CO2 annealing time (tCO). From the figure we can see the adsorbed layers extracted from both the spin-cast films exhibit a power-law growth (had ∝

Figure 4. Growth curves of the 650khPS adsorbed layers derived from the two different spin-cast films (0.6Rg and 2.3Rg thick) as a function of tco at the ridge condition. As a comparison, the growth curve of the adsorbed layers prepared via thermal annealing at T = 150 °C was also plotted as a function of the annealing time. The dotted lines and the arrows correspond to the best fits of the power-law growth and crossover times (tc) (see the text for details), respectively. 1798

DOI: 10.1021/ma502591x Macromolecules 2015, 48, 1795−1803

Article

Macromolecules tβCO) with β = 0.68 ± 0.05 at the early stage of the adsorption kinetics before the crossover time (tc = 2 h) when the substrate surface is fully covered and then transfers into a sluggish growth during the late stage.20 Thus, it is likely that the overall adsorption kinetics in CO2 including tc is independent of the original film thickness. In addition, we confirmed that the resultant adsorbed layers after t c are featureless (i.e., homogeneous) (see Supporting Information), indicating more polymer chains adsorbed on the Si substrates during the CO2 annealing. Furthermore, as a comparison, we also prepared the adsorbed layers from the same 650khPS spin-cast films (with 2.3Rg thick) via thermal annealing at T = 150 °C under vacuum. The details of the adsorbed layer formation via thermal annealing have been described elsewhere.20 As shown in Figure 4, the comparison highlights the characteristics of the adsorbed layers formed via the CO2 annealing: the much stronger powerlaw growth before tc (α = 0.22 ± 0.05 for the thermal annealing) and much smaller tc value (tc = 10 h for the thermal annealing). These facts are indicative of the plasticization effect of CO2 at the polymer/CO2 interface,48,58,59 accelerating the chain mobility. However, at t > tc when a “reeling-in” process of the partially adsorbed chains governs,62 the growth of the adsorbed layer via the CO2 annealing is very limited, resulting in the much thinner final thickness (8.0 ± 0.5 nm) compared to that via the thermal annealing (12.0 ± 0.5 nm). At this moment, it is not conclusive whether the differences in the adsorption process between the CO2 annealing and thermal annealing are inherent because the screening effect of CO2 may promote some detachments of the surface−segment contacts. This reduction would cause an additional decrease in the desorption energy of the adsorbed chains,63 expediting a washout of the incompletely adsorbed chains during the toluene leaching. We are currently using different (less strong) solvents to examine the difference in more detail. III-3. Diffusion Experiments. We next focus on the interdiffusion process at the interface between the adsorbed layer and the overlayer in CO2. We chose the ridge condition of T = 36 °C and P = 8.2 MPa for all the interdiffusion experiments to maximize the plasticization effect of CO2. Figure 5a shows representative NR results for the bilayer of the bottom 650khPS adsorbed layer (0.6Rg thick) and the top 3Rg thick (L0 = 28 nm) d-PS (Mw = 120 kDa) film at two different exposure times. From the figure we can see significant changes in the higher order fringes of the NR profiles, indicating the broadening of the interfacial width between the two layers. Based on the best-fits to the data shown in Figure 5a, the rootmean-square (RMS) roughness (σ) between the two layers was determined to be 2.8 ± 0.4 nm for 0.5 h exposure and 4.1 ± 0.4 nm for 12.7 h exposure. Figure 5b shows the time evolution of the SLD profiles of the bilayers (highlighted near the polymer/ polymer interface) obtained from the best fits. Hence, deuterated polymer segments do diffuse into the bottom hydrogenated adsorbed layer, resulting in a slow broadening of the interface. Figure 6 summarizes σ vs the CO2 exposure time (tco) for different Mw of the top d-PS films. Hence, we find that σ scales linearly with tco0.5, implying that diffusion between the two layers would follow the Fickian diffusion7 with an diffusion coefficient (D), D = σ2/2tco. However, it should be noted that the Fickian diffusion is valid only for diffusion distances much larger than the size of Rg.64 Hence, the diffusion coefficient determined from the fits is not the bulk diffusion constant but can be considered as an effective segment diffusion constant

Figure 5. (a) Observed (circles) and calculated (solid lines) NR profiles of the bilayer of the bottom 650khPS adsorbed layer and the top 3Rg thick d-PS (Mw = 120 kDa) overlayer at tco = 0.5 h (red) and 12.7 h (blue) at the ridge condition. Representative SLD profiles obtained from the best fits to the NR data are shown in (b) as a function of the CO2 exposure time.

Figure 6. RMS roughness for the bilayers of the bottom 650khPS adsorbed layer and top 3Rg thick d-PS overlayers plotted as a function of the square root of tco. The solid lines correspond to the best fits to the data based on the Fickian law described in the text.

(Deff), providing a gauge of the interdiffusion rate. According to previous diffusion experiments using bilayers of bottom 0.4Rg poly(methyl methacrylate) (PMMA) spin-cast film and a top ∼80 nm thick PMMA film in melts,65 the time dependence of σ ∼ t1/8 at the interface was reported even after the bulk reputation time. Lin and co-workers claimed that this is 1799

DOI: 10.1021/ma502591x Macromolecules 2015, 48, 1795−1803

Article

Macromolecules

considering monomer−solid contacts (∼ N1/2 per chain, N: the degree of polymerization) that restrict the chain mobility near the interface and modify the friction force from the bulk. This model may be applicable here with the fact that N1/2 contacts are characteristic of the adsorbed layers.14 However, there is still a slight discrepancy in the exponent. Further experiments are needed to clarify whether the N−1 dependence is rather general regardless of a choice of third components. Nevertheless, it is interesting to address that simulation results showed the lateral diffusion coefficient (D∥) of isolated polymer chains adsorbed on a surface is scaled as D∥ ∼ N−1 with and without the presence of solvent molecules.70−72 In addition, experimentally, adsorbed DNA diffusion on cationic phospholipid bilayers in the fluid phase73 and the PS diffusion in confined slit pores (∼2 nm) with attractive surface sites showed the same N−1 dependences.74 In parallel with the interdiffusion experiments using the adsorbed layers, we also prepared bilayers composed of bottom 0.6Rg thick 650khPS spin-cast films, while the d-PS overlayers (3Rg thick) were kept same. Note the thickness of the spin-cast films is identical to that of the adsorbed layer. On the basis of the aforementioned swelling experiments, it is expected that the CO2 annealing turns the 0.6Rg thick spin-cast film into the irreversibly adsorbed layer. As shown in Figure 7, the NR results demonstrate that the bilayers composed of the 0.6Rg thick spin-cast films exhibit the nearly same Deff values (red circles) as those of the bilayers composed of the bottom adsorbed layers at the ridge condition of CO2. Thus, the experimental results support our hypothesis that the CO2 annealing facilitates the polymer chain adsorption onto the substrate, transforming the limited interfacial region in close proximity of the substrate into the irreversibly adsorbed layer. To further explore how far the effect of the adsorbed layer on the hindered interdiffusion can propagate into the film interior, we measured the interdiffusion processes at the polymer/ polymer interface as a function of the distances from the substrate interface. The four bilayers composed of the same top 3Rg thick 334kdPS overlayers on top of the bottom 650khPS spin-cast films with 1−4Rg thick were studied by NR. The Deff values for these bilayers are plotted in Figure 8 as a function of the initial film thickness of the bottom 650khPS films scaled by Rg. From the figure we can see that the Deff value at the interface distanced 1Rg from the substrate is nearly identical to that at the interface distanced 0.6Rg from the substrate (which

attributed to a large number of segment contacts with the substrate in the bottom 0.4Rg PMMA film.65 Similar time dependence of σ ∼ t1/8 was also reported at the interface between a bottom h-PS brush layer and a d-PS overlayer.66 Hence, we assume that the recovery of σ ∼ tco0.5 scaling in the present data is attributed to the plasticization effect of CO2 at the polymer−substrate interface, accelerating the retarded mobility of the adsorbed polymer chains. Figure 7 plots the Deff values (blue circles) as a function of Mw of the top d-PS layer on a log−log scale. For a comparison

Figure 7. Molecular weight dependence of the effective segment diffusion constant (Deff) for the bilayers composed of the 3Rg thick dPS overlayers on top of (i) the 650khPS adsorbed layer (blue circles) and (ii) the 0.6Rg thick spin-cast film (red circles). The molecular weights correspond to those of the top d-PS. For a comparison, the Deff values for the bilayers of the 3Rg thick d-PS overlayers on top of 3Rg thick h-PS spin-cast films are also plotted (black dots).41

purpose, we also plot the diffusion constants of the polymer chains at the h-PS/d-PS interface distanced 3Rg from the substrate (black squares) where the normal Fickian diffusion is recovered at the ridge condition of CO2.41 Hence, it was found that the Deff values are almost 2 orders of magnitude smaller than the bulk values. Besides, we can see the unusual power law behavior (Deff ∼ Mwα) with α = −1.0 ± 0.1. This is quite different from the aforementioned results at the interface distanced 3Rg from the substrate (α ∼ −2.0 ± 0.1), which is consistent with the scaling for entangled polymer melts above Tg,7 or PS chains in concentrated solutions of organic solvents.67,68 Hence, while the plasticization effect of CO2 does enhance the mobility of the adsorbed polymer chains, the segmental diffusion in close proximity to the solid substrate would be still hindered compared to the free chains. The power law exponent of α = −1.0 reminds of the beadfriction-controlled Rouse behavior.69 However, based on the fact that all of the polymer systems examined here exceed the entanglement molecular weight of PS swollen by a low molecular weight solvent (i.e., Me = 18000/c,40 where c is the PS concentration, which is estimated to be at most 90% in the present study from Figure 1b), the Rouse dynamics may be ruled out. Zheng and co-workers reported similar significant suppression in the PS interdiffusion process (about 2 orders of magnitude smaller than the bulk) in the vicinity of the Si interface in melts.24,25 In addition, they also reported the power exponent of −1.5 for the scaling of the diffusion constants against Mw. They proposed that the unusual scaling behavior can be still explained by the reptation theory7 by further

Figure 8. Deff values for the bilayers composed of the 3Rg thick 334kdPS overlayers on top of the 650khPS spin-cast films with the different scaled film thicknesses of 650khPS. 1800

DOI: 10.1021/ma502591x Macromolecules 2015, 48, 1795−1803

Article

Macromolecules

IV. CONCLUSION In the present study, we focused on the effects of CO2 on the long-range perturbations in the chain mobility developed near the polymer−substrate interface. Polystyrene was chosen as a model polymer, and high-pressure neutron reflectivity (NR) experiments were utilized to study the diffusive chain dynamics at the chemically identical polymer/polymer interface as a function of the distance from the substrate interface. The results are intriguing to show that the diffusion dynamics at the distance of 0.6Rg thick from the substrate strongly hindered. In addition, the power law of the interdiffusion rates vs Mw shows the unusual exponent of −1, which deviates from the reptation theory7 and previous experimental results in concentrated solutions of CO240,41 or organic solvents.67,68 However, we found that the hindered chain dynamics at the polymer/ polymer interface recovers the bulk diffusion when the distance is beyond at least 3Rg, indicating that the long-range perturbations way beyond the characteristic size of the unperturbed chain reported in the melts25,51 vanish in CO2. In order to further understand the mechanism of the retarded chain dynamics, we focused on CO2-induced polymer adsorption on the substrate. It turns out that the plasticization effect of CO2 facilitates polymer adsorption rather than chain stretching away from the substrate, transforming the limited interfacial region in close proximity of the substrate into the irreversibly adsorbed layer. Hence, we may draw the conclusion that the entanglements between the adsorbed chains and neighboring unadsorbed chains remains robust even in the presence of CO2, known as an effective plasticizer for polymers. It is anticipated that the formation of loops in the adsorbed chains would provide a structure with which unadsorbed polymer chains can entangle effectively,77,78 thereby giving rise to this cohesion strength. Further studies on adhesive properties of multiply bound systems compared to those of traditional singly bound polymer structures (polymer brushes) are currently in progress. The microscopic insight into the polymer−solid interface in the presence of the third component would also impact the development of new polymer nanoscale devices such as block polymer lithography via solvent annealing.34

corresponds to the irreversibly adsorbed polymer region), while the Deff value at the interface distanced 2Rg from the substrate increases by a factor of about 2. Remarkably, a sharp transition in Deff takes place at the distance between 2Rg and 3Rg from the substrate: the Deff value increases by a factor of 10 at the interface distanced 3Rg from the substrate and seems to reach a plateau value of Deff ∼ 1.8 × 10−18 cm2/s. As mentioned above, the diffusion of the PS chains at the polymer/polymer interface distanced 3Rg from the substrate is reasonably described by the normal Fickian diffusion (i.e., the bulk diffusion) at the same CO2 condition.41 Hence, we estimate the upper limit of the transition zone to be between 2Rg and 3Rg thick from the substrate, which is significantly “diluted” compared to the reported values for PS melts at T ≫ Tg25,51 (>10Rg thick). It may be interesting to address previous interdiffusion process enabled by absorbed solvent.75 Thompson and co-workers used ex situ 3He+ nucleation analysis for the resultant interdiffusion on d-PS/h-PS bilayers after exposed to cyclohexane vapor. Their results showed that the diffusion coefficients (on the assumption of the Fickian diffusion) increase with increasing the distance from the substrate before reaching a plateau value. However, the range of the distance where the increase was observed is 20−60Rg from the substrate interface, which is even much longer range perturbations than the aforementioned ones in the PS melt films.25,51 Further NR experiments using different organic solvents or solvent vapors are needed to discuss the length scale of the transition zone in the presence of small molecules. Since the 0.6Rg thick region from the substrate corresponds to the adsorbed layer, the effective transition zone in the mobility profile from the adsorbed layer to the bulk is estimated to be about 2Rg thick. Therefore, it is reasonable to suppose that the neighboring unadsorbed chains directly interacting with the adsorbed chains are the origin of the transition zone formed in CO2. The present findings thus reveal that the cohesion strength between the adsorbed chains and neighboring unadsorbed chains remains robust even in the presence of the effective plasticizer, while the polymer chains located at the distance of at least 3Rg from the substrate are freely entangled/ disentangled together. We postulate that this unusual cohesion strength is attributed to the formation of loops in the adsorbed polymer chains. According to a self-consistent-field study of chain conformational statistics at the solid−polymer melt interface,76 the fraction of loops per an adsorbed chain is predicted to be about 30% for the polymer used in this study (Mw = 650 kDa), while that of tails per an adsorbed chain is estimated to be about 60%. Such a loop conformation would provide a structure with which free or unadsorbed polymer chains can entangle effectively,77,78 improving interfacial adhesion between multiple phases compared to those modified with singly grafted chains.79−81 However, it is also suggested that the entanglements take place only when the density of the loops is sufficiently low; when the loop density increases, the interfacial properties approach to those of singly grafted polymer brushes.78 At this point, it is difficult to quantify the critical loop density below which the strong cohesion is achieved. Further interdiffusion experiments using telechelic polymers, which allow us to control the ratios of doubly bound polymer loops to singly bound polymer chains by changing molecular weight,78 will be performed to address the question.

■

ASSOCIATED CONTENT

S Supporting Information *

Figure S1. This material is available free of charge via the Internet at http://pubs.acs.org.

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected] (T.K.). Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We acknowledge Steve Bennett and Jean Jordan-Sweet for the XR measurements. T.K. acknowledges partial financial support from NSF Grants (CMMI-084626 and CMMI-1332499). Use of the National Synchrotron Light Source was supported by the U.S. Department of Energy, Office of Science, Office of Basic Energy Sciences, under Contract DE-AC02-98CH10886.

■

REFERENCES

(1) Geoghegan, M.; Krausch, G. Prog. Polym. Sci. 2003, 28, 261−302.

1801

DOI: 10.1021/ma502591x Macromolecules 2015, 48, 1795−1803

Article

Macromolecules (2) Richardson, H.; Carelli, C.; Keddie, J. L.; Sferrazza, M. Eur. Phys. J. E 2003, 12, 437−440. (3) Reiter, G.; Hamieh, M.; Damman, P.; Sclavons, S.; Gabriele, S.; Vilmin, T.; Raphael, E. Nat. Mater. 2005, 4, 754−758. (4) Damman, P.; Gabriele, S.; Coppee, S.; Desprez, S.; Villers, D.; Vilmin, T.; Raphael, E.; Hamieh, M.; Al Akhrass, S.; Reiter, G. Phys. Rev. Lett. 2007, 99, 036101. (5) Thomas, K. R.; Chenneviere, A.; Reiter, G.; Steiner, U. Phys. Rev. E 2011, 83, 021804. (6) Barbero, D. R.; Steiner, U. Phys. Rev. Lett. 2009, 102, 248303. (7) Doi, M.; Edwards, S. F. The Theory of Polymer Dynamics; Oxford Science: Oxford, 1986. (8) Chung, J. Y.; Chastek, T. Q.; Fasolka, M. J.; Ro, H. W.; Stafford, C. M. ACS Nano 2009, 3, 844−852. (9) Fleer, G. J.; Cohen Stuart, M. A.; Scheutjens, J. M. H. M.; Cosgrove, T.; Vincent, B. Polymers at Interfaces; Chapman and Hall: London, 1993. (10) Guiselin, O. Europhys. Lett. 1992, 17, 225−230. (11) Durning, C. J.; O’Shaughnessy, B.; Sawhney, U. Macromolecules 1999, 32, 6772−6781. (12) Napolitano, S.; Wübbenhorst, M. J. Phys. Chem. B 2007, 111, 9197−9199. (13) Napolitano, S.; Lupascu, V.; Wubbenhorst, M. Macromolecules 2008, 41, 1061−1063. (14) Fujii, Y.; Yang, Z. H.; Leach, J.; Atarashi, H.; Tanaka, K.; Tsui, O. K. C. Macromolecules 2009, 42, 7418−7422. (15) Napolitano, S.; Pilleri, A.; Rolla, P.; Wübbenhorst, M. ACS Nano 2010, 4, 841−848. (16) Napolitano, S.; Wübbenhorst, M. Nat. Commun. 2011, 2, 260. (17) Koga, T.; Jiang, N.; Gin, P.; Endoh, M. K.; Narayanan, S.; Lurio, L. B.; Sinha, S. K. Phys. Rev. Lett. 2011, 107, 225901. (18) Gin, P.; Jiang, N.; Liang, C.; Taniguchi, T.; Akgun, B.; Satija, S. K.; Endoh, M. K.; Koga, T. Phys. Rev. Lett. 2012, 109, 265501. (19) Asada, M.; Jiang, N.; Sendogdular, L.; Gin, P.; Wang, Y.; Endoh, M. K.; Koga, T.; Fukuto, M.; Schultz, D.; Lee, M.; Li, X.; Wang, J.; Kikuchi, M.; Takahara, A. Macromolecules 2012, 45, 7098−7106. (20) Jiang, N.; Shang, J.; Di, X.; Endoh, M. K.; Koga, T. Macromolecules 2014, 47, 2682−2689. (21) Asada, M.; Jiang, N.; Sendogdular, L.; Sokolov, J.; Endoh, M. K.; Koga, T.; Fukuto, M.; Yang, L.; Akgun, B.; Dimitriou, M.; Satija, S. Soft Matter 2014, 10, 6392−6403. (22) Inoue, R.; Kawashima, K.; Matsui, K.; Kanaya, T.; Nishida, K.; Matsuba, G.; Hino, M. Phys. Rev. E 2011, 83, 021801. (23) Inoue, R.; Nakamura, M.; Matsui, K.; Kanaya, T.; Nishida, K.; Hino, M. Phys. Rev. E 2013, 88, 032601. (24) Zheng, X.; Sauer, B. B.; Vanalsten, J. G.; Schwarz, S. A.; Rafailovich, M. H.; Sokolov, J.; Rubinstein, M. Phys. Rev. Lett. 1995, 74, 407−410. (25) Zheng, X.; Rafailovich, M. H.; Sokolov, J.; Strzhemechny, Y.; Schwarz, S. A.; Sauer, B. B.; Rubinstein, M. Phys. Rev. Lett. 1997, 79, 241−244. (26) Lin, E. K.; Wu, W. I.; Satija, S. K. Macromolecules 1997, 30, 7224−7231. (27) Lin, E. K.; Kolb, R.; Satija, S. K.; Wu, W.-l. Macromolecules 1999, 32, 3753−3757. (28) Fakhraai, Z.; Forrest, J. A. Science 2008, 319, 600−604. (29) Ilton, M.; Qi, D.; Forrest, J. A. Macromolecules 2009, 42, 6851− 6854. (30) Yang, Z.; Fujii, Y.; Lee, F. K.; Lam, C.-H.; Tsui, O. Science 2010, 328, 1676−1679. (31) Koga, T.; Li, C.; Endoh, M. K.; Koo, L.; Rafailovich, M. H.; Narayanan, S.; Lee, D. R.; Lurio, L.; Sinha, S. K. Phys. Rev. Lett. 2010, 104, 066101. (32) Wallace, W. E.; Zanten, J. H. c.; Wu, W. L. Phys. Rev. E 1995, 52, R3329−R3332. (33) van Zanten, J. H.; Wallace, W. E.; Wu, W. L. Phys. Rev. E 1996, 53, R2053−2056. (34) Sinturel, C.; Vayer, M.; Morris, M.; Hillmyer, M. A. Macromolecules 2013, 46, 5399−5415.

(35) Koga, T.; Seo, Y. S.; Shin, K.; Zhang, Y.; Rafailovich, M. H.; Sokolov, J. C.; Chu, B.; Peiffeer, D.; Satija, S. K. Macromolecules 2003, 36, 5236−5343. (36) Koga, T.; Seo, Y. S.; Zhang, Y.; Sin, K.; Kusano, K.; Nishikawa, K.; Rafailovich, M. H.; Sokolov, J. C.; Chu, B.; Peiffer, D.; Occhiogrosso, R.; Satija, S. K. Phys. Rev. Lett. 2002, 89, 125506. (37) Koga, T.; Ji, Y.; Seo, Y. S.; Rafailovich, M. H.; Sokolov, J. C.; Satija, S. K. J. Polym. Sci., Part B: Polym. Phys. 2004, 42, 3282−3289. (38) Li, Y.; park, E.; Lim, K.; Johnston, K. P.; Green, P. F. J. Polym. Sci., Part B: Polym. Phys. 2007, 45, 1313−1324. (39) Chebil, M. S.; Vignaud, G.; Grohens, Y.; Konovalov, O.; Sanyal, M. K.; Beuvier, T.; Gibaud, A. Macromolecules 2012, 45, 6611−6617. (40) Gupta, R. R.; Lavery, K. A.; Francis, T. J.; Webster, J. R. P.; Smith, G. S.; Russell, T. P.; Watkins, J. J. Macromolecules 2002, 36, 346−352. (41) Koga, T.; Seo, Y. S.; Hu, X.; Kwanwoo, S.; Zhang, Y.; Rafailovich, M. H.; Sokolov, J. C.; Chu, B.; Satija, S. K. Europhys. Lett. 2002, 60, 559−565. (42) Wissinger, R. G.; Paulaitis, M. E. J. Polym. Sci., Polym. Phys. Ed. 1987, 25, 2497−2510. (43) Goel, S. K.; Beckman, E. J. Polymer 1993, 34, 1410−1417. (44) McHugh, M. A.; Krukonis, V. Supercritical Fluids Extraction Principles and Practice; Butterworth-Heinemann: Boston, MA, 1994. (45) Condo, P. D.; Paul, D. R.; Johnston, K. P. Macromolecules 1994, 27, 365−371. (46) Nishikawa, K.; Tanaka, I.; Amemiya, Y. J. Phys. Chem. 1996, 100, 418−421. (47) Sirard, S. M.; J, Z. K.; Sanchez, L. C.; Green, P. F.; Johnston, K. P. Macromolecules 2002, 35, 1928−1935. (48) Li, X. X.; Vogt, B. D. Polymer 2009, 50, 4182−4188. (49) Koga, T.; Gin, P.; Yamaguchi, H.; Endoh, M.; Sendogdular, L.; Kobayashi, M.; Takahara, A.; Akgun, B.; Satija, S. K.; Sumi, T. Polymer 2011, 52, 4331−4336. (50) Mendoza-Galvan, A.; Trejo-Cruz, C.; Solis-Canto, O.; LunaBarcenas, G. J. Supercrit. Fluids 2012, 64, 25−31. (51) Frank, B.; Gast, A. P.; Russell, T. P.; Brown, H. R.; Hawker, C. J. Macromolecules 1996, 29, 6531−6534. (52) Shin, K.; Hu, X.; Zheng, X.; Rafailovich, M. H.; Sokolov, J.; Zaitsev, V.; Schwarz, S. A. Macromolecules 2001, 34, 4993−4998. (53) Huang, F. H.; Li, M. H.; Lee, L. L.; Starling, K. E.; Chung, F. T. H. J. Chem. Eng. Jpn. 1985, 18, 490−496. (54) Russell, T. P. Mater. Sci. Rep. 1990, 5, 171−271. (55) Fleming, G. K.; Koros, W. J. Macromolecules 1986, 19, 2285− 2291. (56) Seeck, O. H.; Kaendler, I. D.; Tolan, M.; Shin, K.; Rafailovich, M. H.; Sokolov, J.; Kolb, R. Appl. Phys. Lett. 2000, 76, 2713−2715. (57) Koga, T.; Seo, Y. S.; Jerome, J.; Ge, S.; Rafailovich, M. H.; Sokolov, J. C.; Chu, B.; Seeck, O. H.; Tolan, M.; Kolb, R. Appl. Phys. Lett. 2003, 83, 4309−4311. (58) Li, X. X.; Vogt, B. D. Macromolecules 2008, 41, 9306−9311. (59) Li, X. X.; Vogt, B. D. J. Supercrit. Fluids 2009, 51, 256−263. (60) Jia, X. Q.; McCarthy, T. J. Langmuir 2002, 18, 683−687. (61) Scheutjens, J. M. H. M.; Fleer, G. J. J. Phys. Chem. 1980, 84, 178−190. (62) Zajac, R.; Chakrabarti, A. Phys. Rev. E 1995, 52, 6536−6549. (63) O’Shaughnessy, B.; Vavylonis, D. Phys. Rev. Lett. 2003, 90, 056103. (64) Jones, R. L.; Kumar, S. K.; Ho, D. L.; Briber, R. M.; Russell, T. P. Nature 1999, 400, 146−149. (65) Lin, E. K.; Wu, W.; Satija, S. K. Macromolecules 1997, 30, 7224− 7231. (66) Arita, H.; Mitamura, K.; Kobayashi, M.; Yamada, N. L.; Jinnai, H.; Takahara, A. Polym. J. 2013, 45, 117−123. (67) Nemoto, N.; Kishine, M.; Inoue, T.; Osaki, K. Macromolecules 1990, 23, 659−664. (68) Nemoto, N.; Kishine, M.; Inoue, T.; Osaki, K. Macromolecules 1991, 24, 1648−1654. (69) Rouse, P. E. J. Chem. Phys. 1953, 21, 1272−1278. (70) Lai, P. Y. Phys. Rev. E 1994, 49, 5420−5430. 1802

DOI: 10.1021/ma502591x Macromolecules 2015, 48, 1795−1803

Article

Macromolecules (71) Milchev, A.; Binder, K. Macromolecules 1996, 29, 343−354. (72) Desai, T. G.; Keblinski, P.; Kumar, S. K.; Granick, S. Phys. Rev. Lett. 2007, 98, 218301. (73) Maier, B.; Radler, J. O. Phys. Rev. Lett. 1999, 82, 1911−1914. (74) Manias, E.; Chen, H.; Krishnamoorti, R.; Genzer, J.; Kramer, E. J.; Giannelis, E. P. Macromolecules 2000, 33, 7955−7966. (75) Thompson, R. L.; McDonald, M. T.; Lenthall, J. T.; Hutchings, L. R. Macromolecules 2005, 38, 4339−4344. (76) Daoulas, K. C.; Harmandaris, V. A.; Mavrantzas, V. G. Macromolecules 2005, 38, 5780−5795. (77) Huang, Z. Y.; Ji, H. N.; Mays, J. W.; Dadmun, M. D. Macromolecules 2008, 41, 1009−1018. (78) Patton, D.; Knoll, W.; Advincula, R. C. Macromol. Chem. Phys. 2011, 212, 485−497. (79) Dadmun, M. Macromolecules 1996, 29, 3868−3874. (80) Eastwood, E. A.; Dadmun, M. D. Polymer 2002, 43, 6707−6717. (81) Eastwood, E. A.; Dadmun, M. D. Macromolecules 2002, 35, 5069−5077.

1803

DOI: 10.1021/ma502591x Macromolecules 2015, 48, 1795−1803