Proximity to Graphene Dramatically Alters Polymer Dynamics

Article Cite This: Macromolecules XXXX, XXX, XXX−XXX

pubs.acs.org/Macromolecules

Proximity to Graphene Dramatically Alters Polymer Dynamics Feipeng Yang,† Dillon Presto,† Yanbo Pan,∥ Kewei Liu,† Leyao Zhou,† Suresh Narayanan,‡ Yu Zhu,† Zhenmeng Peng,∥ Mark D. Soucek,§ Mesfin Tsige,† and Mark D. Foster*,† †

Department of Polymer Science, ∥Department of Chemical and Biomolecular Engineering, and §Department of Polymer Engineering, The University of Akron, Akron, Ohio 44325, United States ‡ Advanced Photon Source, Argonne National Laboratory, Lemont, Illinois 60439, United States

Downloaded via GUILFORD COLG on July 22, 2019 at 12:21:13 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

S Supporting Information *

ABSTRACT: Proximity to graphene has an extraordinary effect on the dynamics of entangled chains, as evidenced by strong slowing of thermally stimulated fluctuations at the surface of a thin supported film of polystyrene (PS) when a single layer of graphene is placed between the supporting substrate and the PS film. This slowing is due to a layer of highly viscous polymer next to the substrate being thicker for graphene (75 nm or 7.5Rg) than for silicon (13 nm or 1.3Rg). Molecular dynamics simulations of unentangled PS chain melts on silicon and on multiple layers of graphene show that the work of adhesion with graphene is roughly 4 times larger than that with silicon. In addition, PS chains adjacent to silicon readily form loops, with phenyl rings pointing predominantly toward the substrate. Chains adjacent to graphene more readily form trains, with the phenyl rings predominantly parallel to the graphene.

■

to 1Rg, no surface fluctuations were observed.9 This confinement was attributed to the physisorption of PS segments to the substrate, and inclusion of an elasticity modulus in the HCT was vital in order to represent the variation in relaxation time with the scattering vector9 for thicknesses less than 4Rg. Foster and co-workers studied melt films of custom-synthesized polystyrenes of different topologies such as linear,17 starbranched,21 end-branched star,17 and cyclic22 and illustrated the influence of molecular architectures on the thickness h* at which confinement effects appear.21 The differences in the value of h* for different architectures were attributed to the variations with the chain architecture in the thickness of the strongly adsorbed layer, which forms at the substrate upon annealing of the melt film.23 Foster et al. also demonstrated that the substrate surface chemistry can have a remarkable effect on the strongly adsorbed layer thickness and thus alter the surface fluctuations.24 They modified the surface chemistry of a silicon wafer by coating it with a very thin (4 nm) plasma-polymerized maleic anhydride (ppMA) coating, quantified the surface fluctuations, and observed a change in the strongly adsorbed layer thickness using an approach proposed by Guiselin.25 In 1992, Guiselin proposed that one could investigate the formation and structure of a layer of irreversibly, strongly adsorbed molecules at the melt/substrate interface by rinsing away chains not so strongly adsorbed. Guiselin et al.,25 Fujii et al.,26 and Napolitano et al.27,28 all used this approach to study various attributes of such strongly adsorbed layers. It was

INTRODUCTION The surface fluctuations of entangled melt films constitute a significant scientific and technological phenomenon with applications in wetting, adhesion,1,2 and tribology. Thermally stimulated fluctuations depend on the mobility of the entangled chains not only at the surface but also at larger depths in the film.3 Liquid surface fluctuations have been studied using X-ray reflectivity,4 off-specular scattering,4−8 and X-ray photon correlation spectroscopy (XPCS).9−18 The XPCS results can often be summarized or discussed by focusing on a plot of relaxation time normalized by film thickness (τ/h) as a function of the dimensionless in-plane scattering vector q∥h because a hydrodynamic continuum theory (HCT) by Jäckle46 predicts a universal behavior in such a plot.9,19,20 That is, the curves should overlap when the relaxation data at a temperature much higher than the bulk glass-transition temperature (Tg,bulk) for films with different thicknesses are plotted together if the hydrodynamics in the film can be described by a single bulk viscosity value through the whole film. However, when the thickness of a film of linear entangled PS melt chains is less than 4Rg, where Rg is the unperturbed chain radius of gyration, a confinement effect manifests itself and the surface fluctuations are slower than predicted by the HCT theory. For example, static off-specular X-ray scattering data6 show that the surface fluctuations for films of 90000 g/mol linear PS are suppressed. The authors suggested that this was due to strong van der Waals interactions with the Si substrate. Dynamics measurements with XPCS later showed that, when the thickness of an entangled melt film of somewhat larger M (123000 g/mol) decreased to about 4Rg, the universal behavior was lost; for such a film with thickness comparable © XXXX American Chemical Society

Received: February 13, 2019 Revised: May 26, 2019

A

DOI: 10.1021/acs.macromol.9b00317 Macromolecules XXXX, XXX, XXX−XXX

Article

Macromolecules

solution (HCl/H2O = 1:10 by volume), ultrasonically washed with spectroscopic grade acetone and reagent grade ethanol, and dried with high-purity (99.999%) nitrogen (N2). After this cleaning process, the copper foil was placed into the quartz tube of a homemade chemical vapor deposition system including a tube furnace (Lindberg/Blue Model TF55030A), gas flow control system, and a roughing vacuum pump. Keeping the Cu foil in the cold zone, the furnace was heated to 1040 °C under a flow of argon (Ar) and hydrogen (H2) at a pressure of 524 mtorr and at flow rates of 100 and 60 sccm, respectively. The Cu foil was moved into the hot zone and was annealed for 20 min. Then a flow of methane (CH4) was started at a rate of 6 sccm for 7 min under a pressure of 537 mtorr. The methane gas flow was stopped after the reaction. The Cu foil was moved to the cold zone using a magnetic rod and was cooled down to room temperature under the original flow rates of the Ar/H2 mixture. Graphene Transfer to Silicon Wafer. The graphene that formed on the Cu surface was transferred to a silicon wafer (El-Cat Inc., 0.7 mm × 16 mm × 20 mm) that was cleaned using piranha solution (H2O2/H2SO4 = 3:7 by volume) at 100 °C for 30 min and treated with dilute hydrofluoric acid (HF) to etch away the native oxide. (Warning: Piranha solution presents an explosion danger and should be handled with extreme care; it is a strong oxidant and reacts violently with organic materials. All work should be performed in a fume hood. Wear proper protective equipment.) A thermal release (TR) tape (Nitto Americas Inc., CA) was applied to the graphene-coated Cu foil with caution, and 6.25 N/cm2 compression force was applied for 10 min. The graphene on the opposite side of the Cu foil was removed by contact with oxygen plasma for 5 min. The Cu foil was removed from the graphene layer by etching in (NH4)2S2O8 aqueous solution (1 mol/L) for 12 h, rinsing with deionized (DI) water 10 times, and blow drying with high-purity N2. The graphene side of the graphene− TR tape assembly was adhered to the HF-treated silicon wafer at room temperature using a compression force of 12.5 N/cm2 for 10 min. Finally, the TR tape was released from the graphene by heating the wafer for a few seconds on a hot plate at 110 °C, leaving the graphene layer on the silicon wafer. The contaminants on the transferred graphene surface were removed by annealing the Graphene/Si with a gas flow of Ar and H2 (200 and 60 sccm, respectively) at 400 °C for 4 h. Raman Spectroscopy. Raman spectroscopy is widely utilized in the characterization of both the quality and the number of layers of graphene.38,39 Raman spectroscopy was done with a 532 nm laser Horiba LabRam HR micro-Raman spectrometer using a 50× objective, 400 μm pinhole, and 100 μm slit size. The spectra were collected from 10 spots and averaged in order to minimize the impact of local variations. An optical image corresponding to the Raman spectrum was obtained with an Olympus BX41 camera on the Raman spectrometer. Polystyrene Thin Film Preparation. Linear PS (Mw = 131000 g/mol, PDI = 1.05, synthesized by living anionic polymerization) was obtained from Polymer Source Inc., and was used as received. Toluene (99.5%, for spectroscopy ACS) was purchased from Acros Organics Inc. Two concentrations of PS solutions were made: 0.75 (PS 0.0401 g, toluene 5.3110 g) and 1.5 wt % (PS 0.0800 g, toluene 5.2541 g). Films of linear PS were spun-cast from their toluene solutions at a rate of 2000 rpm onto two types of substrates: piranhacleaned, HF-treated silicon wafers and silicon substrates covered with a single layer of graphene. All spin-cast films were annealed at 150 °C in high vacuum (approximately 1 × 10−7 Pa) for 12 h before XPCS measurements. X-ray Photon Correlation Spectroscopy (XPCS) and X-ray Reflectivity (XR). XPCS measurements were performed at beamline 8-ID-I at the APS, Argonne National Laboratory, using a reflection geometry, and the data were analyzed using a procedure described elsewhere.14,40 A partially coherent monochromatic X-ray beam (E = 7.35 keV, 20 × 20 μm2) was used, with the incident angle (θ = 0.14°) below the PS critical angle (0.16°). The X-ray beam penetration in the film was limited to 9 nm in depth, so scattering from the interface between the film and the substrate can be neglected and scattering from density fluctuations within the film can be neglected due to the

shown that the strongly adsorbed layer at the substrate interface is influenced by the surface energy of the substrate and changes in this layer cause changes at larger depths in the film or at its surface through long-range perturbations.28 We would anticipate that changing or tailoring the behavior of the strongly adsorbed layer is not only a matter of adjusting the surface energy of the substrate but could also involve more specific characteristics of that substrate surface, such as opportunities for the epitaxial arrangement of polymer segments on the substrate or opportunities for particular interactions between the functionality of the polymer and functionality on the substrate surface. Here, we show that modification of the substrate surface with a 2D material, namely, graphene, offers extraordinary opportunities to tune the surface fluctuations of thin films or the interactions of the chains in that film with the substrate. Such tailoring of melt dynamics near a solid substrate could have applications, for example, in designing highly effective corrosion protection systems for high value applications.29 Because of graphene’s exceptional properties such as impermeability, thermodynamic stability, transparency, and flexibility, there has been extensive interest in it since its discovery.30,31 Intensive efforts were made in the fabrication of large pieces of graphene of high quality,32 both for academic and industrial purposes. The potential application of graphene in the areas of composites, membranes, energy, coatings, biomedical drug delivery, sensors, and electronics is the driving rationale behind a wide variety of graphene-based research studies.33 For example, thin coatings of graphene were shown to substantially slow the short-term corrosion of metals.34 However, to gain advantage against corrosion in the longer term, it will probably be necessary to combine such ultrathin graphene coatings with conventional polymeric coatings.29 The fact that graphene is atomically thin suggests that it could be used to significantly impact the properties of a thin film adjacent to it while requiring very little material volume to achieve this impact; thus, a key objective is to understand the interactions at the interface between graphene and an adjacent polymer thin film. Molecular dynamics (MD) simulations by Rissanou and Harmandaris35 showed the effect of the number of graphene layers on the dynamical properties of a graphenesheet-supported film of PS 10-mer chains. Jiang et al.36 demonstrated that the most significant change in segmental relaxation times between PS on graphite (a close cousin of graphene) and PS on α-quartz occurs within 3 nm from the substrate. No experimental investigations for the graphene− polymer interfacial interaction have been reported so far, most probably because this is a “buried” interface and not easily probed. This problem can be addressed using XPCS37 to probe the surface fluctuations at a thin melt film surface that are a short distance from this interface and using the Guiselin25 approach to provide an artifact, the structure of which is intimately linked with that of the strongly adsorbed layer present at the melt/solid interface at the temperature of interest. Both methods provide information pertinent to the dynamics of polymer chain segments at the interface between graphene and the polymer thin film adjacent to it.

■

EXPERIMENTAL SECTION

Graphene Growth on Copper (Cu) Foil. A copper (Cu) foil (Sigma-Aldrich 349178, thickness 0.25 mm, 99.98% trace metal basis) was immersed in a solution of 1 mol/L (NH4)2S2O8 (1.14 g (NH4)2S2O8, 50 mL H2O) for 10 min, washed with dilute HCl B


Article

Macromolecules dominating contrast at the vacuum/melt interface. The intensity was recorded using a two-dimensional CCD camera located 4027 nm downstream of the sample.14 Full frame mode was used to capture the off-specular scattering, meaning the entire detector surface was used for every frame. The range of the in-plane scattering vector q∥ covered was 10−4 to 10−3 Å−1. The range of relaxation times accessible with the setup was 10−1 to 103 s. The X-ray beam intensity was about 5 × 109 photons/s, with coherence lengths of 140 and 7 μm in the vertical and horizontal directions, respectively. Each XPCS measurement was started after 20 min of thermal equilibration. The fluctuations in the coherent speckle pattern arise from the thermally stimulated surface fluctuations of the polymer film. The dynamics of these surface fluctuations were captured by the normalized intensity-intensity autocorrelation function (g2)

within a 2 ns additional run to match the lateral dimensions of the substrates. The periodic boundary condition in the z direction was then removed to generate a freestanding PS film. The film, which was about 10 nm in thickness, was placed 5 Å above the substrates. The PS/substrate systems were then cooled down from 600 to 560 K at a rate of 20 K/10 ns in the NVT ensemble. The simulations were continued to run in the NVT ensemble at 560 K for an additional 23 ns for the PS/silicon system and for an additional 36 ns for the PS/ graphene system. During the NVT production run, the position of each atom was recorded every 2 ps for data analysis.

■

RESULTS AND DISCUSSION Graphene Characterization. An optical image of the surface of a graphene monolayer on silicon and the corresponding Raman spectrum are shown in Figure 1. The

⟨I(q , t ′)I(q , t ′ + t )⟩ g2(q , t ′) =

⟨I(q , t ′)⟩2

(1)

where I(q∥, t′) is the scattering intensity at q∥ at time t′ and the angular brackets denote ensemble averages for the delay time t. A single-exponential decay function g2 = 1 + β exp(− 2t/τ) was used in fitting the g2 data, consistent with those data resulting from overdamped capillary waves,40 with β being the coherent contrast and τ being the relaxation time for equilibrium surface height fluctuations. None of the correlation functions discussed here displayed a “stretched” exponential shape. XR was performed before and after every XPCS measurement to acquire the thickness of the film and also to provide secondary evidence for X-ray beam radiation damage. Primarily, comparisons between correlation functions derived from the frames of the first and second halves of the run were made in order to exclude frames showing evidence of beam damage. Strongly Adsorbed Layer Measurement with X-ray Reflectivity (XR). After XPCS measurements, the linear PS films were rinsed with toluene (Acros Organics Inc., 99.5%, for spectroscopy ACS) by immersing them in toluene for 10 min, blow drying with high-purity N2, and then repeating this procedure four times before drying the film in high vacuum (approximately 1 × 10−7 Pa) at room temperature overnight. XR was measured using a Rigaku SmartLab Xray diffractometer with a sealed tube source (40 kV, 44 mA, λ = 1.54 Å) and parallel beam optic. A layered model was used in fitting the XR data to obtain values of thickness and scattering length density of each layer, as well as the microroughness of each interface between neighboring layers or media. Nanoindentation Measurement. The hardnesses and elastic moduli of the silicon wafer and various samples of PS on silicon or PS on supported graphene were estimated by analyzing the load− displacement curves garnered with a standardized loading and unloading process with a Hysitron Premier nanoindentation system.41−43 A Berkovich diamond tip was used, and the unloading curve was analyzed using the approach coded in the TriboScan software provided with the instrument. Each reported value was obtained as an average of results from three locations on the sample forming the vertices of an isosceles right triangle, with the sides of equal length being 200 μm long. Molecular Dynamics Simulation. All-atom molecular dynamics simulations using the optimized potential for liquid simulations-allatom (OPLSAA) force field were performed on a PS film containing 81 60-mer chains and adsorbed on a graphite or silicon substrate. The lateral dimensions of the substrates were 10.824 nm × 10.226 nm for graphite and 10.752 nm × 9.984 nm for silicon. The graphite substrate was five layers thick (1.31 nm), and the silicon substrate was 1.22 nm thick; the atoms of both substrates were frozen during the simulation. All the MD simulations were carried out using the LargeScale Atomic/Molecular Massively Parallel Simulator (LAMMPS) simulation package. An integration time step of 1 fs was used for all simulations with a Lennard-Jones cutoff radius of 1.2 nm. The initial polystyrene system was first equilibrated at 600 K for 10 ns with periodic boundary conditions in all directions. After NPT equilibration of the PS system, the dimensions in the x and y directions were adjusted using the LAMMPS deform command

Figure 1. (a) Reflection mode optical microscopy image of the surface of the transferred graphene on the silicon wafer. The spots are dust particles on the sample surface. (b) Raman spectrum of the graphene layer on silicon. The large values for the IG/ID and I2D/IG intensity ratios indicate a graphene sheet with low defect density.

Raman spectrum provides a chemical fingerprint for the graphene, averaged over the beam footprint. The three peaks seen are characteristic of graphene: the G mode at approximately 1580 cm−1, the D mode at approximately 1350 cm−1, and the 2D mode at approximately 2670 cm−1.44,45 The G mode, D mode, and 2D mode are a doubly degenerate zone center E2g mode, a ring breathing mode that becomes activated near defects, and a second-order mode of the D mode that does not require defects for activation, respectively.46 Information regarding the graphene quality and the number of layers can be derived from an analysis of the positions, line shapes, and relative intensities of these modes.47−49 It is well accepted that the ratio of the 2D peak C


Article

Macromolecules intensity to G peak intensity (I2D/IG) and the average full width at half-maximum (FWHM) of the 2D peak are key parameters reflecting the charge impurities on the graphene surface50−52 from which the grain size and the number of layers can be estimated.53 The grain size of graphene can be calculated from ID/IG using the equation54 ID/IG = C(λ)/La, where La (nm) is the grain size, λ (nm) is the wavelength of incident light, and C(λ) is equal to 2.4 × 10−10 λ.4 As the number of graphene layers increases, FWHM increases and I2D/IG decreases.55,56 When I2D/IG from the Raman spectrum is larger than unity, this is taken as evidence that only a single layer of graphene is present. In Figure 1, the intensity ratio of the G peak to the D peak is very large (∼20), indicating a large grain size or low defect density.57 The 2D peak intensity (∼7200) is slightly more than twice that of the G peak (∼3500), and the FWHM of the 2D peak is approximately 26 cm−1, indicating that the transferred film is a monolayer of graphene.52,58 X-ray Photon Correlation Spectroscopy (XPCS). The surface fluctuations of three kinds of samples were measured. Each sample was designated using a name that contains first the thickness, specified as a multiple of Rg, then “PS” for polystyrene, and finally “/Graphene” or “/Si” to denote the substrate in contact with the melt. Sample thicknesses of 8Rg (approximately 80 nm) and 3Rg (approximately 30 nm) were studied so that the designations for the four samples were “8Rg PS/Graphene”, “3Rg PS/Graphene”, “8Rg PS/Si”, and “3Rg PS/Si”. With an incident angle θ of 0.14° with respect to the surface, the penetration depth into PS is limited to 9 nm and the intensity of the scattering from the fluctuations of the film surface is much larger than the intensity of scattering due to density fluctuations inside the film. Thus, the XPCS data capture only the surface fluctuation dynamics. A typical 2D false color plot of the scattering intensity in the detector plane is shown in Figure 2a where the horizontal axis corresponds to the real space direction perpendicular to the sample surface and the vertical axis corresponds to the real space direction perpendicular to the scattering plane; the distance along the axis is denoted by the pixel number. This scattering pattern is also referred to as a speckle pattern. Each time that the speckle pattern is captured, the resulting data set, corresponding to a specific delay time since the beginning of the measurement, is referred to as a “frame”. One XPCS data set is composed of frames numbered 1 to N. A comparison of a plot resulting from an average of the intensities in each pixel over the first N/2 frames with a plot resulting from an average in each pixel over the second N/2 frames provides a quick qualitative indication of the degree to which the surface has fluctuated over the time scale of the experiment. The difference in surface fluctuation behavior between the case of a PS film on graphene and the case of the PS film on silicon is immediately evident from the raw XPCS data. Consider a pixel-by-pixel comparison of average intensities along the line shown in two false color intensity plots on the right side of Figure 2a. Since the scattering measured here is coherent scattering, when the intensity in a given pixel averaged over the first half of the run differs outside the statistical uncertainty from the intensity in that pixel averaged over the second half of the run, then one may infer that some real space feature of the surface relaxed over the time elapsed from the middle of the first half to the middle of the second half of the run. For a given set of experimental conditions, if few pixels show statistically significant changes in intensity,

Figure 2. (a) Upper left: an example speckle pattern from the CCD detector. Lower left: schematic of the grouping of intensity patterns from the first half of a run and the second half of a run for averaging. Right side: examples of the averaged intensity patterns for a region of interest on the detector (shown with the red rectangle) from the first and second halves of the run for 3Rg PS/Graphene at 220 °C. Both averages still have the character of speckle patterns. That is, there are jumps in intensity between neighboring pixels, making the pixels very evident, meaning the sample is not moving on the time scale over which the averaging was done. Lines mark the pixels for which intensities are plotted in panels (b) and (c). Plots of intensity in each pixel along the line shown in panel (a) for the first 336 s (open symbols, blue) and second 336 s (closed symbols, red) of a run for (b) 8Rg PS/Graphene at 170 °C and (c) 3Rg PS/Graphene at 220 °C. On the time scale of 672 s, the 8Rg sample surface changes some, but the 3Rg sample surface does not relax appreciably.

then the surface is not moving on that time scale. Shown in Figure 2b are the plots of averaged intensities along one row of pixels from the first 160 frames, that is, the first 336 s, and the second 160 frames, that is, the second 336 s, of an experiment. This is for the 8Rg thick sample on graphene at 170 °C. For some pixels, the intensities from the average of the first half and average of the second half are quite similar. For other pixels, the two averages differ significantly. Thus, the surface is relaxing on the time scale of 336 s. Figure 2c shows such a comparison for the 3Rg thick sample on graphene at a much higher temperature of 220 °C. Despite the sample being at a temperature 50 °C higher, the average intensities for the two sets of frames are the same within the counting uncertainty for nearly all the pixels. Thus, the surface of this thinner film is not fluctuating on this time scale. The g2 intensity-intensity autocorrelation functions, examples of which are shown in Figure 3a, were fit using the XPCS Analysis Program 3.0 (Copyright 2008) by Jiang and Sprung.59 Using the same data mask for all data ensured that the analysis reflected always the same length scales of motion, resulting in D


Article

Macromolecules

defining a lower bound. If the relaxation time were less than 336 s for at least some values of q, we should be able to see at least the first part of the relaxation in the g2 function. Since we cannot, we can say that the value of τ for 3Rg PS/Graphene, even at 220 °C, lies above 336 s for all values of q∥ studied here. This result is represented in Figure 3b by a dot-dashed line at a high value of τ for 3Rg PS/Graphene at 220 °C. In order to quantify measurable relaxations, a thicker sample of 8Rg PS on graphene was measured. At 170 °C, the surface fluctuations are measurable for this 8Rg film. However, even though the sample on graphene is much thicker than the 3Rg PS film on Si, its surface fluctuates much more slowly than does the surface of 3Rg PS on Si. The fluctuations of the film on graphene appear to be about 3 orders of magnitude slower than the surface fluctuations of the film of the same thickness on Si. This is a dramatic difference that results from the attractive PS/Graphene interaction being much stronger than the attractive PS/Si interaction. The stronger attraction leads to a more profound slowing down of the PS chains next to graphene and the formation of a thicker strongly adsorbed layer on graphene. According to the HCT, for polymer films all at the same temperature but with different thicknesses, the plots of τ/h versus q∥h must collapse onto a single universal curve. The deviation from this universal behavior seen in Figure 3b is evidence of confinement effects or failure of some other assumptions of the theory. We find that it is possible to recover the universal behavior if, instead of plotting τ/h versus q∥h, we plot τ/heff versus q∥heff using an effective layer thickness heff, which is conceived as a layer in which the viscosity is equal to the bulk value. The remainder of the layer is imagined to have dynamics so strongly slowed so as to effectively behave as though having an extremely high viscosity and therefore, for our purposes, immobile, functioning essentially as part of the substrate. We know that certainly there must be a gradient in mobility through the film, but this simple model approximating the gradient with a step function in viscosity is sufficient to recover the universal behavior with respect to film thickness, as shown in Figure 4 for the case of 170 °C. In order to determine what value of immobile layer thickness we need to achieve the overlap, we require from the literature24,60 values of surface tension and bulk viscosity for 131000 g/mol linear PS at different temperatures. These are given in Table 1. The

Figure 3. (a) g2 functions for PS films for wave vector 0.55 × 10−2 nm−1. The solid lines denote the best fits to the single exponential decay. The data for 8Rg PS/Graphene at 170 °C do not contain a complete relaxation, but data for somewhat higher values of q do. (b) τ/h vs q∥h for 3Rg PS and 8Rg PS thin films on silicon and graphene at temperatures as labeled. The dot-dashed curve for 3Rg PS/Graphene 220 °C indicates that the relaxation is outside the measurable range, even at 220 °C, and reflects the shape that the curve would have if consistent with the HCT. The dashed curves shown in the other data correspond to fits with the HCT using a single layer and an effective viscosity, which may or may not equal the experimentally measured bulk viscosity.

more consistent comparisons among different measurements. Fitting of the g2 data with the equation g2 = 1 + β exp(− 2t/τ), with β being the contrast, yielded values of the relaxation time τ. Plots of τ/h as a function of q∥h are shown for samples 8Rg PS/Graphene and 3Rg PS/Si in Figure 3b. Consider first the 3Rg PS/Si sample data at 170 °C. The normalized relaxation time decreases with a power-law dependence of about −3.8 as the dimensionless scattering vector increases, which is what one expects for overdamped surface fluctuations. When the thickness of the PS remains 3Rg but the substrate is changed from Si to graphene, the speckle patterns do not evolve at all over the measurement times used, even for temperatures as high as 220 °C. That is, the surface for the 3Rg PS/Graphene appears solid-like on the length and time scales probed. The corresponding g2 functions are horizontal lines with values above the baseline, and it is not possible to define a value for the relaxation time beyond

Figure 4. τ/heff versus q∥heff at 170 °C for PS thin films on silicon and graphene assuming for each an effective thickness equal to the actual thickness minus the thickness of a layer of very slow dynamics having the thickness shown in the legend (e.g., 13 nm for PS films on Si). E


Article

Macromolecules Table 1. Surface Tension and Bulk Viscosity of 131 kg/mol Linear PS at Various Temperatures T (°C)

γlinearb (mN/m)

150 160 170 180

31.3 30.6 29.9 29.3

ηbulk (Pa s) 3.4 9.7 3.1 1.2

× × × ×

106 105 105 105

Reprinted with permission from ACS Macro Lett. 2017, 6, 915−919. Copyright 2017. American Chemical Society. bUncertainty approximately ±5%.

highly viscous layer at 170 °C for 3Rg PS/Si and 8Rg PS/Si is taken to be 13 nm (approximately 1.3Rg) thick, while the highly viscous layer for 8Rg PS/Graphene must be taken to be 75 nm (approximately 7.5Rg) thick, meaning that heff for 8Rg PS/Graphene is only 5 nm, while for 3Rg PS/Si and 8Rg PS/Si, the values of heff are 17 and 67 nm, respectively. At 200 °C, the thicknesses required for the highly viscous layer drop to 11 and 64 nm on silicon and graphene, respectively. Determining these differences in highly viscous layer thicknesses needed to reconcile the data for the two types of substrates with the hydrodynamic continuum theory is another means of characterizing the strong interaction between the graphene layer and chains adsorbed to it. Experimental evidence of the structure of the strongly adsorbed part of the melt film was obtained using X-ray reflectivity (XR) and ellipsometry measurements after the rest of the film was rinsed away using the protocol of Koga et al.23 The XR curves for the strongly adsorbed layers are shown in Figure 5a. The layer that was left on the substrate after rinsing and drying was a collapsed version of the strongly adsorbed layer that was present in the melt.61 That is, in the melt state, there would have been portions of chains from the adsorbed layer that extended up into the melt adjacent to the adsorbed layer, either as “loops” or as “tails”. These parts of the chains could be extending away from the densest part of the layer when the rinsed sample is still in the solvent, but upon the removal from the solvent, the tails and loops would have collapsed onto the rest of the adsorbed layer since air is a poor solvent for the chains. Fitting the XR data with a model providing a discretized depth profile of the scattering length density (SLD) provides a description of the morphology of the strongly adsorbed layer. Both the fits and corresponding SLD profiles are shown in Figure 5a. Nominal thicknesses of these collapsed versions of the strongly adsorbed layers on silicon and graphene were calculated as the distance from the middle of the interface with air to the middle of the interface with the substrate. Those thicknesses were 3.4 ± 0.1 and 8.2 ± 0.2 nm for PS/Si and PS/Graphene, respectively. These values of thickness are also consistent with values from ellipsometry measurements obtained assuming a model of a PS layer on Si or PS layer on Si-supported graphene (Figure S1). The value of thickness on the silicon substrate is consistent with results from our earlier work24 and with results from Koga et al.23 Further details of the differences between the structures of the adsorbed layers on the two substrates can be gleaned from the SLD profiles. The SLD profile for the strongly adsorbed layer on Si after rinsing appears to describe a PS layer that has an SLD value in its densest part (the plateau) that is 9.7 × 10−6 Å−2, which is about 1% larger than the value of 9.59 × 10−6 Å−2 for high-molecular-weight PS. Then there is an interface with the air with an apparent width corresponding to a locally sharp

Figure 5. (a) X-ray reflectivity curves for 8Rg PS/Si and 8Rg PS/ Graphene samples after rinsing with the fits determined by nonlinear least square regression using a multilayer structural model and (inset) the corresponding SLD profiles. (b) Cartoons for the strongly adsorbed PS layers on silicon and graphene substrates.

interface with an rms roughness of about 8 Å, to a local gradual transition from this heightened density next to the substrate to low density next to air, or some mixture of interface local roughness and interface diffuseness. (From XR data alone, it is not possible to say whether this breadth is due to diffuseness, roughness, or both.) The SLD profile for the layer adsorbed to graphene shows a densest part with an SLD of 11.0 × 10−6 Å−2, which is 14% larger than the bulk SLD, which is a densification higher than that reported for PMMA adsorption onto SiOx62 (5%) but consistent in magnitude with a report for PS next to Si.63 However, the densification suggested by the SLD profile on graphene is over a larger depth than that reported in ref 63 for PS next to Si. Also, the transition in SLD from the PS layer to air for the layer on graphene is about 4.5 times broader than that seen here for PS on Si. Presumably, this interface width reflects how chains that do extend up into the melt but are tethered to the substrate directly by adsorption or indirectly by entanglement collapse when the sample is pulled out of the solvent. The thickness found here for the strongly adsorbed layer on Si is comparable to what has been reported in the literature26,64−66 for similar molecular weights, temperatures, and annealing times. Since scattering results provide a global, statistical picture of this intriguing layer next to the substrate and its change with the nature of the substrate, we have also performed MD simulations to gain insight into what sorts of individual chain conformations might be associated with these differences in overall adsorbed layer morphology. The simulation corresponds to the state of the chains on the substrate before washing. The simulations were performed on a 60-mer F


Article

Macromolecules (approximately 6k) PS film adsorbed on graphite and silicon substrates using the optimized potential for liquid simulationsall-atom (OPLS-AA) force field (see details in the Experimental Section). Snapshots of selected PS chains near a silicon surface and a graphite (five layers of graphene) surface are shown in Figure 6. For the PS chains next to the silicon,

Figure 6. Results from MD simulations showing how representative PS chains from a ∼10 nm thick melt film arrange themselves near (a,b) a silicon surface and (c,d) a graphite surface. In panels (a) and (c), the conformations of a few representative PS chains with parts adsorbed to the two surfaces are shown. Only the backbone atoms are shown, and each chain is shown in a different color for clarity. The cyan chain from panel (a) and the orange chain from panel (c) are shown with their phenyl rings in snapshots in panels (b) and (d), respectively. At the graphite surface, the chains have longer trains on the surface, while loops dominate the configurations of chains adjacent to the silicon surface. (b,d) The orientations of the phenyl ring vectors for the rings adjacent to the substrate also differ strikingly with the substrate type.

Figure 7. (a) Time-autocorrelation functions ⟨P2(t)⟩ of the phenyl ring vector adjacent to graphite (solid curves) and silicon (dashed curves) substrates at T = 560 K. The simulation box was divided in the z direction, that is, in the direction perpendicular to the substrate surface, into slabs of 0.5 nm thickness, and the z-dependent time autocorrelations are shown with different colors for the seven slabs closest to the substrates ranging (in the order of red, green, blue, violet, yellow, orange, and black) from red for phenyl rings within 0− 0.5 nm from the substrate surface to black for phenyl rings within 3− 3.5 nm from the substrate surface. The black curve is taken as representing the bulk behavior in the PS film. The distance of the slab above the substrate increases in the direction of the black arrow. After a distance of 2.5 nm, the bulk behaviors are recovered. The solid red and green curves indicate that the PS/Graphite case needs more than the 36 ns simulation data used here in order for the two autocorrelation curves to decay to zero. (b) Distribution of phenyl ring orientation, specified in terms of cosine of the angle between the phenyl ring vector and the surface normal, in the four slabs closest to the substrates with the same assignment of color to distances from the substrate as given in panel (a).

loops are formed and the vectors joining the backbone carbon atom in contact with the phenyl ring to the carbon atom in the para position of that ring (“phenyl ring vector”) are predominantly pointing toward the substrate. For the PS chains next to graphite, trains connected to few or no loops are formed. The phenyl ring vectors close to the graphite substrate tend to lie nearly parallel to the graphite surface with the plane of the phenyl ring perpendicular to the substrate surface and the −CH2 groups (not shown) pointing toward the substrate. Figure S2 shows the atomic mass density profiles as a function of distance from the two substrates. The chain conformations dominated by trains on the graphite surface resulted in an approximately 100% increase in local density relative to the bulk near the surface, while the chain conformations next to silicon, which were dominated by loops, increased the local density by about 10% relative to the bulk. The significant differences in the conformations and packing of the PS chains near the two types of substrates are the result of the huge difference in interaction between the PS chains and the substrates. The PS film/graphite substrate interaction was found to be about 5 times stronger than the PS film/silicon substrate interaction. This large difference in interaction energy should cause the dynamics of the chains adjacent to the two types of substrates to differ significantly as well. The segmental dynamics of the chains were investigated as a function of the distance from the substrate through calculations of the autocorrelation function of the phenyl ring vector. The results for both substrates are shown in Figure 7a as a semilogarithmic plot of the ensemble average of the

second rank order parameter P2, which is defined as ⟨P2(t )⟩ = 1 2 (3⟨cos2 θ(t )⟩ − 1), where θ(t) is the angle of rotation of the phenyl ring vector with respect to the surface normal at time t. The PS film was divided into slabs of 0.5 nm thickness in the direction perpendicular to the substrate surface, and P2 for each slab was computed and is represented by different colors in the figure for the seven slabs closest to the substrates. The P2 results indicate that there is a strongly adsorbed region of about 2.5 nm thickness (five slabs thick) on graphene and a relatively weakly adsorbed region of about 1 nm thickness (two slabs thick) on silicon. Note that the simulation results are for a film of PS 60-mers with no entanglement effects while the experimental results are for PS ∼1310-mers manifesting entanglement effects. The distributions of the phenyl ring orientation (in terms of cos θ) in different slabs of the film are also given in Figure 7b. G


Article

Macromolecules It was shown in earlier work from our group24 that a silicon substrate surface can also be modified so that the strongly adsorbed layer thickness on that substrate is much less than that in silicon. Foster et al.24 considered a coating of 3Rg PS on a silicon wafer covered with a 5 nm thick plasma-polymerized maleic anhydride (ppMA) film. They explained the XPCS data for that sample using a two-layer model in which the highly viscous layer was only 15 nm (1.5Rg) thick, and no layer at all was found by XR on the substrate after rinsing. Although the two-layer model with a highly viscous layer is an approximation, it is sufficient to illustrate clearly that the influence on the film hydrodynamics for PS on a substrate extends over a distance much bigger than the strongly adsorbed layer thickness left after rinsing for graphene, silicon, and the plasma-polymerized MA coating and that the distance over which this influence extends is by far the greatest next to graphene. How the distance tails of tethered chains penetrate into the film and differ between the cases of graphite and silicon substrates was also studied with the simulations. Figure 8 plots

the extent of the layer in which slowing down can be seen is amplified by entanglement. While the comparison of dynamics near the two substrates made above addresses differences in segmental dynamics, it is also possible from the simulation results to show that the dynamics on the atomic scale are quite different in the two cases. Figure 9a−c presents plots of atomic mean squared

Figure 8. Furthest extent of a given chain from the substrate vs closest approach of that chain for all chains in the simulation box having one part that came within 10 Å of the substrate surface. A larger fraction of the chains in the simulation have such a close approach in the case of graphite, and on average, these chains stretch farther from the substrate.

for each chain that comes within 10 Å distance of the substrate a single point with coordinates that correspond to the furthest distance any atom of the chain is from the substrate and the closest distance any atom of that chain comes to the substrate. All chains with an atom coming to within 10 Å of the substrate would be part of the highly viscous layer. For graphite, 32 chains have atoms close to the substrate and the farthest extension of a chain is 85 Å. For silicon, 23 chains have atoms close to the substrate and the farthest extension is 70 Å. This plot suggests that the chains that have some portion adsorbed to the substrate do extend out into the melt farther in the case of graphite than in the case of silicon, but the ratio of furthest extensions or even the ratio of mean extensions is considerably less than the ratio of roughly 5 between the apparent thicknesses of the two highly viscous layers seen experimentally for the two types of substrate. The simulations do not reflect the larger difference seen experimentally because the experimental system is highly entangled, while the simulation chains are much shorter and not entangled. We conjecture that

Figure 9. Atomic mean squared displacement as a function of time (a) in the direction parallel to the surface, (b) perpendicular to the surface, and (c) overall for atoms in polystyrene chains that are initially within 1 nm of the surface and separately for atoms that are between 5 and 6 nm of the surface for the two types of substrates.

displacement (MSD) as a function of time in the direction parallel to the surface, perpendicular to the surface, and overall for atoms in polystyrene that are initially within 1 nm of the surface and separately for atoms that are between 5 and 6 nm of the surface for the two types of substrates. The difference in MSD reached in 3 ns for the two substrates is much larger for PS chain atoms that are within 1 nm of the substrate and is more prominent for motion perpendicular to the surface (factor of approximately 5) than for motion parallel to the surface (factor of approximately 3.2). The motion of the atoms in the chains is decidedly slower next to graphite. H


Article

Macromolecules Nanoindentation. In the previous section, a large change in the surface fluctuation behavior of substrate-supported PS melt films was seen as a result of interposing a single monolayer of graphene between the silicon substrate and the PS melt. Is it possible that the PS−graphene interactions could likewise have a large effect on the behavior of the PS film once it has cooled through its glass-transition temperature and become a glass? Although the exact values of mechanical properties of thin films are challenging to extract from nanoindentation experiments of thin polymer films on hard substrates, results from an exploratory study are highly suggestive that the modulus of the film is substantially higher when on the graphene substrate than when on silicon. The force−displacement curves for the silicon wafer, 3Rg PS/Si, 3Rg PS/Graphene, 8Rg PS/Si, and 8Rg PS/Graphene are compared in Figure 10. Values of apparent elastic modulus and

value due to substrate effects, but at least in the case of AFM indentation, the factor by which the apparent value exceeds the actual value is determined by the ratio of contact area to film thickness (a/h) and Poisson’s ratio of the polymer.68 If we ignore adhesion, for constant indentation depth, the contact radius for the hemispherical end of the AFM indenter should be the same. If to a zeroth-order approximation the correction for the substrate effect with the Berkovich indenter varies analogously, the correction factor for the 3Rg thick films should be the same for the two substrates if Poisson’s ratio of the PS is the same on the two substrates. In this case, the apparent values of the modulus are both higher than the real values but should be higher by about the same factor and can be compared qualitatively. For both 3Rg and 8Rg PS films, the apparent elastic modulus and hardness values for PS on graphene were substantially larger than those on silicon. We conjecture that this is due to the stronger interactions between the PS and the graphene. For PS films on one type of substrate, when the thickness decreased from 8Rg to 3Rg, the values of apparent hardness and elastic modulus increased since the substrate effect became more important. For hardness, the apparent value on graphene was about 10 times larger than that on Si for both thicknesses. The similarity of the apparent modulus values for 3Rg and 8Rg films on graphene is puzzling since the modulus values on Si dropped by about a factor of 2 going from a thickness of 3Rg to a thickness of 8Rg. This difference should be studied further. In any case, these first nanoindentation results suggest that also the macroscopic properties of the glassy PS films formed from the melt state manifest a strong influence of the different strengths of interaction between the PS and substrate for Si and graphene.

■

Figure 10. Force−displacement curves for the silicon wafer, 3Rg PS/ Si, 3Rg PS/Graphene, 8Rg PS/Si, and 8Rg PS/Graphene using a displacement of 10 nm.

CONCLUSIONS XPCS measurements of surface fluctuations on thin melt films, reflectivity measurements of the strongly adsorbed layer in such films, and nanoindentation measurements of the glassy films formed upon cooling give evidence that a much stronger adsorption of PS chains to the substrate sharply alters the film’s dynamics and mechanical properties when a single layer of graphene is interposed between the Si substrate and the PS film. Slowing of the surface fluctuations suggests that the dynamical effects of the stronger adsorption can propagate up into the film over distances comparable to several times the unperturbed chain dimension, much farther than seen previously for entangled linear chains.10,24 The depth over which local viscosity of the film on graphene is enhanced is about 3 times the thickness of the collapsed strongly adsorbed layer, implying that both tethering at the graphene surface and extension of dynamically encumbered melt chains into the film are important for altering the melt film surface dynamics. MD simulation results for 6k chains are consistent with the picture of strongly favorable interactions at the graphene surface and changes in chain conformations and dynamics there, though the dynamics are probed on the much faster time scale of phenyl ring motions. The simulations find that the interaction of PS with graphite is much more favorable than that with silicon. This difference can be quantified using the work of adhesion Wa. Wa is the energy required to reversibly separate a unit area of interface between two materials to create a unit surface area of each material. It is calculated from the simulation as the difference between the energy of the polymer film on the substrate and the energies of the substrate and polymer film when infinitely

hardness of each sample derived from fitting the unloading curves ignoring the complication of the sample being a thin film on a much harder substrate67,68 are summarized in Table 2. The values found experimentally for the modulus and Table 2. Elastic Modulus and Hardness Values from Nanoindentation sample name silicon PS bulk 8Rg PS/Si 3Rg PS/Si 8Rg PS/Graphene 3Rg PS/Graphene

elastic modulus (GPa) 202 2.6 27.6 52 189 188

± ± ± ± ± ±

6 0.1 0.4 5 2 7

hardness (GPa) 21 0.09 0.44 1.1 4.7 12

± ± ± ± ± ±

1 0.02 0.05 0.2 0.2 1

hardness for the Si alone are in reasonable agreement with values in the literature.69 The hardness and elastic modulus values of bulk PS from the literature70 are quoted for comparison, while the rest of the apparent elastic modulus and hardness values in Table 2 are experimentally determined values. The maximum vertical displacement in each nanoindentation measurement was 10 nm; nonetheless, the apparent hardness and elastic modulus values were influenced by the substrate because deformation of the film at a given depth invokes processes deeper in the sample.43 The actual value of the modulus is known to be less than the apparent I


Article

Macromolecules

AC02-06-CH11357. We acknowledge the use of the Thin Film Deposition Chamber in the SOA Facility of the Institute of Polymer Science and Polymer Engineering and the assistance of Dr. Zhorro Nikolov.

far apart, normalized to the appropriate unit area. The values of Wa are 170 and 40 ergs/cm2 (or 0.17 and 0.04 J/m2) for the PS on graphite and silicon substrates, respectively. Also, next to graphite, long trains of segments are much more likely and the density of the melt is enhanced much more strongly than on the Si substrate. The orientation of the phenyl ring vector for rings close to the substrate also differs strongly for the two substrate types. For the graphite, the phenyl ring vector tends to lie nearly parallel to the substrate surface. For the Si, the phenyl ring vector points predominantly toward the substrate. Finally, it seems that these strong differences in the melt structure and dynamics correlate with marked differences in the properties of the films after cooling into the glassy state. The apparent moduli and hardness of the PS film are much larger when a single graphene layer is interposed between the Si and the PS film. Since graphene has already found wide use in various fields including electronics, sensors, supercapacitors, superconductors, and coatings,30,31 the results of this investigation could have broad implications. Properties of both melt and glassy thin films can be strongly altered by the imposition of a single graphene sheet between the film and substrate. Graphene layers on metals have already been shown to significantly reduce corrosion of metal in the short term.34 Perhaps when combined with polymeric coatings in which substantial densification occurs at the polymer/graphene interface, the failure of graphene to protect such metals on the longer term29 can be remedied and the hybrid coatings can demonstrate substantial synergy. In addition, the PS/graphene interface may serve as a useful extreme example for fundamental studies of how favorable interactions with the substrate affect the glasstransition temperature in thin polymer films.71−75

■

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.macromol.9b00317. Highly viscous layer thicknesses used for 3Rg PS/Si, 8Rg PS/Si, and 8Rg PS/Graphene samples at 170 and 200 °C; ellipsometry data and fits for 8Rg PS/Graphene and 8Rg PS/Si; PS segment density profiles as a function of distance from the substrate for silicon and graphite (PDF)

■

REFERENCES

(1) Tang, T.; Jagota, A.; Chaudhury, M. K.; Hui, C. Y. Thermal Fluctuations Limit the Adhesive Strength of Compliant Solids. J. Adhes. 2006, 82, 671−696. (2) Ghatak, A.; Chaudhury, M. K. Adhesion-Induced Instability Patterns in Thin Confined Elastic Film. Langmuir 2003, 19, 2621− 2631. (3) He, Q.; Narayanan, S.; Wu, D. T.; Foster, M. D. Confinement Effects with Molten Thin Cyclic Polystyrene Films. ACS Macro Lett. 2016, 5, 999−1003. (4) Tidswell, I. M.; Rabedeau, T. A.; Pershan, P. S.; Kosowsky, S. D. Complete Wetting of a Rough Surface: An x-Ray Study. Phys. Rev. Lett. 1991, 66, 2108−2111. (5) Tolan, M.; Vacca, G.; Wang, J.; Sinha, S. K.; Li, Z.; Rafailovich, M. H.; Sokolov, J.; Gibaud, A.; Lorenz, H.; Kotthaus, J. P. Thin Polymer Films on Rough Surfaces. Phys. B 1996, 221, 53−59. (6) Wang, J.; Tolan, M.; Seeck, O. H.; Sinha, S. K.; Bahr, O.; Rafailovich, M. H.; Sokolov, J. Surfaces of Strongly Confined Polymer Thin Films Studied by X-Ray Scattering. Phys. Rev. Lett. 1999, 83, 564−567. (7) Seo, Y. S.; Koga, T.; Sokolov, J.; Rafailovich, M. H.; Tolan, M.; Sinha, S. Deviations from Liquidlike Behavior in Molten Polymer Films at Interfaces. Phys. Rev. Lett. 2005, 94, 157802. (8) Seeck, O. H.; Müller-Buschbaum, P.; Tolan, M.; Press, W. Diffuse X-Ray Scattering in Specular Direction: Analysis of a Wetting Film. EPL 1995, 29, 699−704. (9) Jiang, Z.; Kim, H.; Jiao, X.; Lee, H.; Lee, Y. J.; Byun, Y.; Song, S.; Eom, D.; Li, C.; Rafailovich, M. H.; et al. Evidence for Viscoelastic Effects in Surface Capillary Waves of Molten Polymer Films. Phys. Rev. Lett. 2007, 98, 227801. (10) Jiang, Z.; Mukhopadhyay, M. K.; Song, S.; Narayanan, S.; Lurio, L. B.; Kim, H.; Sinha, S. K. Entanglement Effects in Capillary Waves on Liquid Polymer Films. Phys. Rev. Lett. 2008, 101, 246104. (11) Mukhopadhyay, M. K.; Jiao, X.; Lurio, L. B.; Jiang, Z.; Stark, J.; Sprung, M.; Narayanan, S.; Sandy, A. R.; Sinha, S. K. Thickness Induced Structural Changes in Polystyrene Films. Phys. Rev. Lett. 2008, 101, 115501. (12) Akgun, B.; Uǧur, G.; Jiang, Z.; Narayanan, S.; Song, S.; Lee, H.; Brittain, W. J.; Kim, H.; Sinha, S. K.; Foster, M. D. Surface Dynamics of “Dry” Homopolymer Brushes. Macromolecules 2009, 42, 737−741. (13) Koga, T.; Li, C.; Endoh, M. K.; Koo, J.; Rafailovich, M.; Narayanan, S.; Lee, D. R.; Lurio, L. B.; Sinha, S. K. Reduced Viscosity of the Free Surface in Entangled Polymer Melt Films. Phys. Rev. Lett. 2010, 104, No. 066101. (14) Wang, S. F.; Jiang, Z.; Narayanan, S.; Foster, M. D. Dynamics of Surface Fluctuations on Macrocyclic Melts. Macromolecules 2012, 45, 6210−6219. (15) Evans, C. M.; Narayanan, S.; Jiang, Z.; Torkelson, J. M. Modulus, Confinement, and Temperature Effects on Surface Capillary Wave Dynamics in Bilayer Polymer Films near the Glass Transition. Phys. Rev. Lett. 2012, 109, No. 038302. (16) Liu, B.; Narayanan, S.; Wu, D. T.; Foster, M. D. Polymer Film Surface Fluctuation Dynamics in the Limit of Very Dense Branching. Macromolecules 2013, 46, 3190−3197. (17) Wang, S. F.; Yang, S.; Lee, J.; Akgun, B.; Wu, D. T.; Foster, M. D. Anomalous Surface Relaxations of Branched-Polymer Melts. Phys. Rev. Lett. 2013, 111, No. 068303. (18) Zhao, J.; Sakellariou, G.; Green, P. F. Phase Behavior of Diblock Copolymer/Star-Shaped Polymer Thin Film Mixtures. Soft Matter 2016, 12, 3849−3853. (19) Jäckle, J. The Spectrum of Surface Waves on Viscoelastic Liquids of Arbitrary Depth. J. Phys.: Condens. Matter. 1998, 10, 7121− 7131.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Feipeng Yang: 0000-0002-5470-3241 Mark D. Soucek: 0000-0003-3865-4504 Mesfin Tsige: 0000-0002-7540-2050 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS This work was supported by funding from DoD through the Technical Corrosion Collaboration (FA7000-14-2-20016) and The University of Akron Research Foundation. Use of the Advanced Photon Source at Argonne National Laboratory was supported by the DOE’s Office of Science under contract DEJ


Article

Macromolecules (20) Sutton, M. A review of X-Ray Intensity Fluctuation Spectroscopy. C. R. Phys. 2008, 9, 657−667. (21) Zhang, F.; He, Q.; Zhou, Y.; Narayanan, S.; Wang, C.; Vogt, B. D.; Foster, M. D. Anomalous Confinement Slows Surface Fluctuations of Star Polymer Melt Films. ACS Macro Lett. 2018, 7, 834−839. (22) He, Q.; Wang, S. F.; Hu, R.; Akgun, B.; Tormey, C.; Peri, S.; Wu, D. T.; Foster, M. D. Evidence and Limits of Universal Topological Surface Segregation of Cyclic Polymers. Phys. Rev. Lett. 2017, 118, 167801. (23) Koga, T.; Jiang, N.; Gin, P.; Endoh, M. K.; Narayanan, S.; Lurio, L. B.; Sinha, S. K. Impact of an Irreversibly Adsorbed Layer on Local Viscosity of Nanoconfined Polymer Melts. Phys. Rev. Lett. 2011, 107, 225901. (24) Zhou, Y.; He, Q.; Zhang, F.; Yang, F.; Narayanan, S.; Yuan, G.; Dhinojwala, A.; Foster, M. D. Modifying Surface Fluctuations of Polymer Melt Films with Substrate Modification. ACS Macro Lett. 2017, 6, 915−919. (25) Guiselin, O. Irreversible Adsorption of a Concentrated Polymer Solution. EPL 1992, 17, 225−230. (26) Fujii, Y.; Yang, Z.; Leach, J.; Atarashi, H.; Tanaka, K.; Tsui, O. K. C. Affinity of Polystyrene Films to Hydrogen-Passivated Silicon and Its Relevance to theTg of the Films. Macromolecules 2009, 42, 7418−7422. (27) Napolitano, S.; Wübbenhorst, M. The Lifetime of the Deviations from Bulk Behaviour in Polymers Confined at the Nanoscale. Nat. Commun. 2011, 2, 260. (28) Napolitano, S.; Sferrazza, M. How Irreversible Adsorption Affects Interfacial Properties of Polymers. Adv. Colloid Interface Sci. 2017, 247, 172−177. (29) Schriver, M.; Regan, W.; Gannett, W. J.; Zaniewski, A. M.; Crommie, M. F.; Zettl, A. Graphene as a Long-Term Metal Oxidation Barrier: Worse than Nothing. ACS Nano 2013, 7, 5763−5768. (30) Novoselov, K. S.; Jiang, D.; Schedin, F.; Booth, T. J.; Khotkevich, V. V.; Morozov, S. V.; Geim, A. K. Two-Dimensional Atomic Crystals. Proc. Natl. Acad. Sci. 2005, 102, 10451−10453. (31) Geim, A. K.; Novoselov, K. S. The Rise of Graphene. Nat. Mater. 2007, 6, 183−191. (32) Wang, X.; Li, J.; Zhong, Q.; Zhong, Y.; Zhao, M. Wafer-Scale Synthesis and Transfer of Monolayer Graphene. Proc. IEEE Conf. Nanotechnol. 2013, 652−655. (33) Li, X.; Rui, M.; Song, J.; Shen, Z.; Zeng, H. Carbon and Graphene Quantum Dots for Optoelectronic and Energy Devices: A Review. Adv. Funct. Mater. 2015, 25, 4929−4947. (34) Prasai, D.; Tuberquia, J. C.; Harl, R. R.; Jennings, G. K.; Bolotin, K. I. Graphene: Corrosion-Inhibiting Coating. ACS Nano 2012, 6, 1102−1108. (35) Rissanou, A. N.; Harmandaris, V. Structural and Dynamical Properties of Polystyrene Thin Films Supported by Multiple Graphene Layers. Macromolecules 2015, 48, 2761−2772. (36) Jiang, N.; Sen, M.; Endoh, M. K.; Koga, T.; Langhammer, E.; Bjöörn, P.; Tsige, M. Thermal Properties and Segmental Dynamics of Polymer Melt Chains Adsorbed on Solid Surfaces. Langmuir 2018, 34, 4199−4209. (37) Seydel, T.; Madsen, A.; Tolan, M.; Grübel, G.; Press, W. Capillary Waves in Slow Motion. Phys. Rev. B 2001, 63, No. 073409. (38) Heller, E. J.; Yang, Y.; Kocia, L.; Chen, W.; Fang, S.; Borunda, M.; Kaxiras, E. Theory of Graphene Raman Scattering. ACS Nano 2016, 10, 2803−2818. (39) Graf, D.; Molitor, F.; Ensslin, K.; Stampfer, C.; Jungen, A.; Hierold, C.; Wirtz, L. Raman Imaging of Graphene. Solid State Commun. 2007, 143, 44−46. (40) Kim, H.; Rühm, A.; Lurio, L. B.; Basu, J. K.; Lal, J.; Lumma, D.; Mochrie, S. G. J.; Sinha, S. K. Surface Dynamics of Polymer Films. Phys. Rev. Lett. 2003, 90, No. 068302. (41) VanLandingham, M. R. Review of Instrumented Indentation. J. Res. Natl. Inst. Stand. Technol. 2003, 108, 249. (42) Briscoe, B. J.; Fiori, L.; Pelillo, E. Nano-Indentation of Polymeric Surfaces. J. Phys. D: Appl. Phys. 1998, 31, 2395−2405.

(43) VanLandingham, M. R.; Villarrubia, J. S.; Guthrie, W. F.; Meyers, G. F. Nanoindentation of Polymers: An Overview. Macromol. Symp. 2001, 167, 15−44. (44) Rümmeli, M. H.; Gorantla, S.; Bachmatiuk, A.; Phieler, J.; Geißler, N.; Ibrahim, I.; Pang, J.; Eckert, J. On the Role of Vapor Trapping for Chemical Vapor Deposition (CVD) Grown Graphene over Copper. Chem. Mater. 2013, 25, 4861−4866. (45) Zhao, P.; Kumamoto, A.; Kim, S.; Chen, X.; Hou, B.; Chiashi, S.; Einarsson, E.; Ikuhara, Y.; Maruyama, S. Self-Limiting Chemical Vapor Deposition Growth of Monolayer Graphene from Ethanol. J. Phys. Chem. C 2013, 117, 10755−10763. (46) Pang, J.; Bachmatiuk, A.; Fu, L.; Yan, C.; Zeng, M.; Wang, J.; Trzebicka, B.; Gemming, T.; Eckert, J.; Rummeli, M. H. Oxidation as A Means to Remove Surface Contaminants on Cu Foil Prior to Graphene Growth by Chemical Vapor Deposition. J. Phys. Chem. C 2015, 119, 13363−13368. (47) Pimenta, M. A.; Dresselhaus, G.; Dresselhaus, M. S.; Cançado, L. G.; Jorio, A.; Saito, R. Studying Disorder in Graphite-Based Systems by Raman Spectroscopy. Phys. Chem. Chem. Phys. 2007, 9, 1276−1290. (48) Kalbac, M.; Kong, J.; Dresselhaus, M. S. Raman Spectroscopy as a Tool to Address Individual Graphene Layers in Few-Layer Graphene. J. Phys. Chem. C 2012, 116, 19046−19050. (49) Cançado, L. G.; Reina, A.; Kong, J.; Dresselhaus, M. S. Geometrical Approach for the Study of G′ band in the Raman Spectrum of Monolayer Graphene, Bilayer Graphene, and Bulk Graphite. Phys. Rev. B 2008, 77, 245408. (50) Lee, J.-U.; Yoon, D.; Kim, H.; Lee, S. W.; Cheong, H. Thermal Conductivity of Suspended Pristine Graphene Measured by Raman Spectroscopy. Phys. Rev. B 2011, 83, No. 081419. (51) Ni, Z. H.; Yu, T.; Luo, Z. Q.; Wang, Y. Y.; Liu, L.; Wong, C. P.; Miao, J.; Huang, W.; Shen, Z. X. Probing Charged Impurities in Suspended Graphene Using Raman Spectroscopy. ACS Nano 2009, 3, 569−574. (52) Ferrari, A. C.; Meyer, J. C.; Scardaci, V.; Casiraghi, C.; Lazzeri, M.; Mauri, F.; Piscanec, S.; Jiang, D.; Novoselov, K. S.; Roth, S.; et al. Raman Spectrum of Graphene and Graphene Layers. Phys. Rev. Lett. 2006, 97, 187401. (53) Ferrari, A. C. Raman Spectroscopy of Graphene and Graphite: Disorder, Electron−Phonon Coupling, Doping and Nonadiabatic Effects. Solid State Commun. 2007, 143, 47−57. (54) Cançado, L. G.; Takai, K.; Enoki, T.; Endo, M.; Kim, Y. A.; Mizusaki, H.; Jorio, A.; Coelho, L. N.; Magalhães-Paniago, R.; Pimenta, M. A. General Equation for the Determination of the Crystallite Size La of Nanographite by Raman Spectroscopy. Appl. Phys. Lett. 2006, 88, 163106. (55) Reina, A.; Jia, X.; Ho, J.; Nezich, D.; Son, H.; Bulovic, V.; Dresselhaus, M. S.; Kong, J. Large Area, Few-Layer Graphene Films on Arbitrary Substrates by Chemical Vapor Deposition. Nano Lett. 2009, 9, 30−35. (56) Malard, L. M.; Pimenta, M. A.; Dresselhaus, G.; Dresselhaus, M. S. Raman Spectroscopy in Graphene. Phys. Rep. 2009, 473, 51−87. (57) Yan, Z.; Lin, J.; Peng, Z.; Sun, Z.; Zhu, Y.; Li, L.; Xiang, C.; Samuel, E. L.; Kittrell, C.; Tour, J. M. Toward the Synthesis of WaferScale Single-Crystal Graphene on Copper Foils. ACS Nano 2012, 6, 9110−9117. (58) Li, X.; Cai, W.; An, J.; Kim, S.; Nah, J.; Yang, D.; Piner, R.; Velamakanni, A.; Jung, I.; Tutuc, E.; et al. Large-Area Synthesis of High-Quality and Uniform Graphene Films on Copper Foils. Science 2009, 324, 1312−1314. (59) Sikorski, M.; Jiang, Z.; Sprung, M.; Narayanan, S.; Sandy, A. R.; Tieman, B. A Graphical User Interface for Real-Time Analysis of XPCS Using HPC. Nucl. Instrum. Methods Phys. Res., Sect. A 2011, 649, 234−236. (60) Plazek, D. J.; O’Rourke, V. M. Viscoelastic Behavior of Low Molecular Weight Polystyrene. J. Polym. Sci., Part A-2 1971, 9, 209− 243. (61) Carrillo, J.-M. Y.; Cheng, S.; Kumar, R.; Goswami, M.; Sokolov, A. P.; Sumpter, B. G. Untangling the Effects of Chain Rigidity on the K


Article

Macromolecules Structure and Dynamics of Strongly Adsorbed Polymer Melts. Macromolecules 2015, 48, 4207−4219. (62) Wu, W.; Majkrzak, C. F.; Satija, S. K.; Ankner, J. F.; Orts, W. J.; Satkowski, M.; Smith, S. D. The Density Profile at a Polymer/Solid Interface. Polymer 1992, 33, 5081−5084. (63) Jiang, N.; Cheung, J.; Guo, Y.; Endoh, M. K.; Koga, T.; Yuan, G.; Satija, S. K. Stability of Adsorbed Polystyrene Nanolayers on Silicon Substrates. Macromol. Chem. Phys. 2018, 219, 1700326. (64) Gin, P.; Jiang, N.; Liang, C.; Taniguchi, T.; Akgun, B.; Satija, S. K.; Endoh, M. K.; Koga, T. Revealed Architectures of Adsorbed Polymer Chains at Solid-Polymer Melt Interfaces. Phys. Rev. Lett. 2012, 109, 265501. (65) Housmans, C.; Sferrazza, M.; Napolitano, S. Kinetics of Irreversible Chain Adsorption. Macromolecules 2014, 47, 3390−3393. (66) Jiang, N.; Shang, J.; Di, X.; Endoh, M. K.; Koga, T. Formation Mechanism of High-Density, Flattened Polymer Nanolayers Adsorbed on Planar Solids. Macromolecules 2014, 47, 2682−2689. (67) Miyake, K.; Satomi, N.; Sasaki, S. Elastic Modulus of Polystyrene Film from near Surface to Bulk Measured by Nanoindentation Using Atomic Force Microscopy. Appl. Phys. Lett. 2006, 89, No. 031925. (68) Li, L.; Encarnacao, L. M.; Brown, K. A. Polymer Nanomechanics: Separating the Size Effect from the Substrate Effect in Nanoindentation. Appl. Phys. Lett. 2017, 110, No. 043105. (69) Bhushan, B.; Li, X. Micromechanical and Tribological Characterization of Doped Single-Crystal Silicon and Polysilicon Films for Microelectromechanical Systems Devices. J. Mater. Res. 1997, 12, 54−63. (70) Du, B.; Tsui, O. K. C.; Zhang, Q.; He, T. Study of Elastic Modulus and Yield Strength of Polymer Thin Films Using Atomic Force Microscopy. Langmuir 2001, 17, 3286−3291. (71) Kim, S.; Torkelson, J. M. Distribution of Glass Transition Temperatures in Free-Standing, Nanoconfined Polystyrene Films: A Test of de Gennes’ Sliding Motion Mechanism. Macromolecules 2011, 44, 4546−4553. (72) Bäumchen, O.; McGraw, J. D.; Forrest, J. A.; Dalnoki-Veress, K. Reduced Glass Transition Temperatures in Thin Polymer Films: Surface Effect or Artifact? Phys. Rev. Lett. 2012, 109, No. 055701. (73) Erber, M.; Tress, M.; Mapesa, E. U.; Serghei, A.; Eichhorn, K. J.; Voit, B.; Kremer, F. Glassy Dynamics and Glass Transition in Thin Polymer Layers of PMMA Deposited on Different Substrates. Macromolecules 2010, 43, 7729−7733. (74) Roth, C. B.; Dutcher, J. R. Glass Transition and Chain Mobility in Thin Polymer Films. J. Electroanal. Chem. 2005, 584, 13−22. (75) van Zanten, J. H.; Wallace, W. E.; Wu, W.-l. Effect of Strongly Favorable Substrate Interactions on the Thermal Properties of Ultrathin Polymer Films. Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top. 1996, 53, R2053−R2056.

■

NOTE ADDED AFTER ASAP PUBLICATION This paper was published ASAP on June 27, 2019. A typo was corrected in the Experimental Section. The revised paper was reposted on June 28, 2019.

L


Proximity to Graphene Dramatically Alters Polymer Dynamics

Recommend Documents