Unexpected Tribological Synergy in Polymer Blend Coatings

Sep 25, 2017 - Friction coefficients were determined for each sample using the reversal method described by Burris and Sawyer.(52) As was typical of t...
8 downloads 13 Views 4MB Size
Download PDF
Research Article www.acsami.org

Unexpected Tribological Synergy in Polymer Blend Coatings: Leveraging Phase Separation to Isolate Domain Size Effects and Reduce Friction Jillian A. Emerson,† Nikolay T. Garabedian,‡ Axel C. Moore,§ David L. Burris,*,‡ Eric M. Furst,*,† and Thomas H. Epps, III*,†,∥ †

Department Department § Department ∥ Department ‡

of of of of

Chemical & Biomolecular Engineering, University of Delaware, Newark, Delaware 19716, United States Mechanical Engineering, University of Delaware, Newark, Delaware 19716, United States Biomedical Engineering, University of Delaware, Newark, Delaware 19716, United States Materials Science & Engineering, University of Delaware, Newark, Delaware 19716, United States

S Supporting Information *

ABSTRACT: We employed a systematic processing approach to control phase separation in polymer blend thin films and significantly reduce dynamic friction coefficients (μ)s. We leveraged this modulation of phase separation to generate composite surfaces with dynamic friction coefficients that were substantially lower than expected on the basis of simple mixing rules, and in several cases, these friction coefficients were lower than those of both pure components. Using a model polyisoprene [PI]/polystyrene [PS] composite system, a minimum μ was found in films with PS mass fractions between 0.60 and 0.80 (μblend = 0.11 ± 0.03); that value was significantly lower than the friction coefficient of PS (μPS = 0.52 ± 0.01) or PI (μPI = 1.3 ± 0.09) homopolymers and was comparable to the friction coefficient of poly(tetrafluoroethylene) [PTFE] (μPTFE = 0.09 ± 0.01) measured under similar conditions. Additionally, through experiments in which the domain size was systematically varied at constant composition (through an annealing process), we demonstrated that μ decreased with decreasing characteristic domain size. Thus, the tribological synergy between PS and PI domains (discrete size, physical domain isolation, and overall film composition) was shown to play an integral role in the friction and wear of these PS/PI composites. Overall, our results suggest that even high friction polymers can be used to create low friction polymer blends by following appropriate design rules and demonstrate that engineering microstructure is critical for controlling the friction and adhesion properties of composite films for tribologically relevant coatings. KEYWORDS: tribology, polymer blends, phase separation, polymer composite, dynamic friction coefficient, domain size

■

INTRODUCTION The properties of polymer blend films can be varied through careful selection of constituents, composition, and processing methods. Such composites are promising materials for a wide range of industrial applications and demonstrate improvements in mechanical, thermal, optical, and electrical properties1,2 in comparison to analogues made from their homopolymer constituents making polymer blend films candidates for organic optoelectronic devices, pressure sensitive adhesives, batteries, radiation shielding, and anti-reflection coatings.3−9 In addition to typical design considerations such as composition- and constituent-based behavior, the structure and size scale of the phase separated morphology are equally important to the resulting material properties.10−16 For blended tribo-materials that are based on polymers such as polyamides,17,18 polyethylene,17,19 poly(tetrafluoroethylene) [PTFE],20−24 and poly(ether ether ketone) [PEEK],20−22,24 studies suggest that the polymer blend constituents behave synergistically in the composite, with friction coefficients, wear rates, or both in the blended state that are lower than those of © XXXX American Chemical Society

either of the pure homopolymers. In one notable example that motivates the current work, Burris and Sawyer reported that the PEEK/PTFE blend with the lowest friction and wear had a dynamic friction coefficient (μ) and wear rate well below those of either constituent,22 which agrees qualitatively with the results from Bijwe et al.20 However, a similar study by Briscoe and co-workers revealed no such minimum μPEEK/PTFE with increasing PEEK loading (from 0 to 100 wt %).21 Burris and Sawyer hypothesized that the remarkable performance of their low friction and low wear PEEK/PTFE blend was a result of the unusually small-sized cryo-ground PEEK resin that affected domain size and continuity in the polymer blend.22 Nonetheless, to our knowledge, no studies have directly evaluated such structural effects on the tribological properties of polymer blends. Received: July 12, 2017 Accepted: September 11, 2017

A

DOI: 10.1021/acsami.7b10170 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Fisher Scientific) were triple rinsed with toluene before and after cleaning with ultraviolet-ozone (Jelight, Model 342). Film Preparation and Annealing. Solutions for the PS, PI, and PS/PI blend casting were prepared in toluene at a total polymer content of 2.5 wt %. For all solutions, the composition was determined gravimetrically. The mass fraction of PS, xPS, was given by mPS/[mPS + mPI], for which mPS and mPI were the masses of PS and PI, respectively. The solutions were blade cast using a custom flow coating apparatus48 with added convection to produce uniform films. Toluene, which is PS selective, was chosen for its compatibility with both polymers and its relatively short evaporation time, which further promoted film uniformity during casting. Select films were annealed in solvent vapor to induce changes in the domain size without altering the overall composition of the film. These films were placed in a sealed chamber with a solvent reservoir containing ∼5 mL of o-xylene to create a solvent-rich vapor environment that led to swelling of the films. o-Xylene was selected as the annealing solvent because it evaporated more slowly than toluene (expanding control over the annealing time) and was nearly neutral for PS and PI. Because of these properties, o-xylene provided tunability in the degree of phase separation in the PS/PI blend, imparting mobility to PS while reducing the PI surface wetting layer. The films were exposed to the solvent for various times to allow chain rearrangement and generate larger domains. Following exposure, the chamber was unsealed to rapidly quench the annealed film microstructure. Blend Film Characterization. The morphology of the polymer blend films was determined via optical microscopy (OM) and atomic force microscopy (AFM). Bright field optical micrographs were captured using a Nikon Eclipse LV100 microscope with a Prior OptiScan II control box and a Nikon DS-Fi1 camera connected to a Nikon Digital Sight controller in reflection mode. For these PS/PI blend films, the surface morphology was the same as the through-film morphology (see also Figure S1). Domain sizes were characterized by the chord length, πA/p. The area, A, and perimeter, p, of the droplet phase were measured using ImageJ’s “Analyze Particle” routine, which measured objects in either binary or thresholded images.49 First, the image was scanned until the edge of a droplet domain was located. Second, the domain was outlined and measured. Third, the selected domain was made invisible to the program to prevent double counting. Finally, once all of the domains were located, measured, and counted, the results were summarized. AFM also was used to study the phase separated structures. Atomic force micrographs were captured on a Veeco Dimension 3100 V operating in tapping mode. The film morphologies were imaged using silicon probes (Tap 150G, Budget Sensors) with a force constant of 5 N/m and resonant frequencies between 120 kHz and 180 kHz. Film thicknesses were measured with a spectral reflectometer (Filmetrics, F20-UV) and confirmed through AFM scratch testing. All films were ∼1 μm in thickness. Dynamic Friction Coefficient and Adhesion Measurements. Dynamic friction coefficients were measured using a custom linear reciprocating tribometer, illustrated schematically in Figure 1, operated in a sphere-on-flat geometry. The 6.35 mm diameter sphere (probe) was made of either borosilicate glass (∼80 nm RMS roughness) or high-density polyethylene (HDPE, ∼800 nm RMS roughness) to test for contact area effects. Against glass, the contact pressure at a given load increased with film stiffness and, thus, PS content; against HDPE, the contact pressure was largely set by HDPE and, thus, approximately independent of PS content. A linear reciprocating stage was used to move the films beneath the probe affixed to a flexible cantilever. Each sample was tested at velocities of 0.5, 4.6, and 7.8 mm/s with a 2 mm path length for 100 cycles at a constant load of 5 mN. Low sliding speeds were selected to reduce any effects from frictional heating.50,51 The normal and friction forces were measured using two independent capacitive displacement sensors to measure deflections of the cantilever beam in the normal and frictional directions. Friction coefficients were determined for each sample using the reversal method described by Burris and Sawyer.52 As was typical of these materials, the friction coefficients varied significantly during the run-in

Many polymer blends are immiscible and phase separate into two classes of structures: (1) dispersed [droplet] phases and (2) co-continuous phases.25 The route by which phase separation occurs is strongly influenced by the composition and experimental time scales associated with accessing the phase separated regime, and the overall phase separation size scale (domain size) in films largely is controlled by the kinetics of the phase separation process along with the thermodynamic incompatibility of the materials.25−29 Co-continuous structures, which promote mass, charge, or thermal transport through connected paths in the material,11 can be produced by targeting a composition near the thermodynamic critical blend concentration, through which it is possible to trap the spinodal decomposition phase evolution to maintain the targeted structure.1,2 Droplet structures with small domain sizes, which impact the toughness and stiffness of the polymer blend, can be produced by kinetically trapping a more asymmetric mixture of the blend constituents and accessing the nucleation and growth regime.30−38 As a result of the final film casting and the requirements for kinetic arrest to achieve targeted domain morphologies, processing strongly influences the blend morphology.6,9−16,25−28,30−38 The trapping of the morphology can be achieved by kinetic arrest of one or both of the polymers, due to polymer vitrification or crystallization as a result of solvent loss during casting.30−42 Composite domain sizes can be tuned further by annealing approaches such as solvent annealing and thermal annealing.43,44 During annealing, the polymer chains have increased mobility, which commonly leads to a growth in domain size as a result of coalescence of the phase separated regions, typically via a coarsening or an Ostwald ripening process.45 In all, the final film microstructure is governed by the thermodynamics and kinetics of phase separation along with the processing of the material.6,9−16,25−28,30−38 In this work, we leverage the controlled phase separation of polystyrene [PS] and polyisoprene [PI], both of which are relatively high friction polymers, to tune the tribological properties of the polymer blends; the resulting friction coefficients obtained were well below those of either constituent, approaching values of ∼0.1, which is considered a benchmark for low friction.46 We systematically demonstrate that these friction coefficients can be tuned by changing the relative composition of the two polymers in the solution formulation or by modulating the domain size through postcasting processing methods (e.g., solvent vapor annealing). The changes in μ were correlated to the film microstructure as well as adhesion between the blend films and the test probes. Overall, this work provides an enhanced understanding of structure−tribology relationships toward producing new multicomponent, functional coatings.

■

EXPERIMENTAL SECTION

Materials. Polystyrene [Mn = 892 kg/mol, Đ = 1.04] was purchased from Scientific Polymer Products. Polyisoprene [Mn = 541 kg/mol, Đ = 1.13, 95% 1,4 content] was synthesized by anionic polymerization in cyclohexane using sec-butyllithium as an initiator.47 PS/PI was selected as a model system as its phase behavior is well understood, and the phase separation can be controlled readily, allowing the interplay between morphology and friction coefficient to be studied. Toluene (certified ACS) and o-xylene (puriss. p.a.) were purchased from Fischer Scientific and Sigma-Aldrich, respectively. The purchased polymer and solvents were used as received. Prior to use, glass microscope slides (25 mm × 75 mm, part # 12-550-A3 from B

DOI: 10.1021/acsami.7b10170 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 1. Schematic of the custom linear reciprocating tribometer used in this study. The spherical probe (6.35 mm diameter) was either glass or HDPE, and the probe was affixed to a flexible stainless-steel cantilever. The sample was attached to the linear reciprocating stage. Then, the probe was loaded at a set normal force (FN) using the piezo positioning stage. The linear reciprocating stage moved the sample at a set velocity and for a predetermined number of cycles under the probe. The FN and friction force (FF) were measured by two capacitance probes arranged perpendicular to each other during the experiment.

Figure 2. Effect of PS mass fraction (xPS) on μ measured with a glass probe at 3 different speeds: 7.8 mm/s (green triangles, ▲), 4.6 mm/s (blue squares, ■), and 0.5 mm/s (red diamonds, ⧫). At all speeds, PI had the highest μ, and the friction coefficient generally decreased with increasing PS content. For 0.60 ≤ xPS ≤ 0.80, μ was at a minimum and lower than either homopolymer. The gray solid line (region) represents the friction coefficient of bulk PTFE, which ranged from 0.07 ± 0.01 at 0.5 mm/s to 0.10 ± 0.01 at 7.8 mm/s (see also Figure S4 in the Supporting Information). The error bars represent the propagated error from the fluctuations in steady-state friction coefficient on a single film in combination with the variation in steady-state friction coefficients from measurements on different films. The colored lines are to guide the eye.

period and stabilized at steady state.52 The μ of any one sample was obtained using data from steady-state sliding; these values were averaged over multiple samples (N = 2 for glass, N = 4 for HDPE) at the same conditions. Note: additional HDPE tests were run as a result of the increased variability in the HDPE f riction coeff icient. The temporal responses of various samples are provided in the Supporting Information Figure S2. Adhesion was studied by measuring the pull-off force with the same instrument and probes. The probe was placed into contact with the surface at loads that varied from 0.3 to 15.8 mN, left in contact for a dwell time of 0, 4, 16, or 80 s to provide the opportunity for contact aging, and retracted from the surface at a rate of 1.2, 12.0, or 23.7 μm/s to probe for contributions from viscoelasticity.53−55 Contact aging tests were performed both under purely static contact (leaving the probe in stationary contact with the film) and after sliding contact (rastering the probe across the film surface at 4.6 mm/s for 5, 20, 50, or 100 cycles before aging). The pull-off force was proportional to the normal force and independent of other variables (aging time, sliding/ static, approach/retraction rate). Therefore, the adhesion coefficient (α), which was determined from the slope of pull-off force versus normal force, was used to quantify the adhesion at each condition of interest.

(0.11 ± 0.03) near xPS = 0.70 for a 0.5 mm/s probe velocity was comparable to those exhibited by bulk PTFE (0.07−0.11) during in-house testing at the same experimental conditions; it also was similar to values reported in the literature for PTFE analyzed under similar conditions.50 The μ increased at xPS ≥ 0.75 and became largely independent of composition and speed at xPS ≥ 0.85. The wear track data in Figure 3 provided information about the structure of the film, the mechanics of contact, and the wear of the sample. As demonstrated in these images, changes in composition affected which of the constituents acted as the primary phase (see also Figure S1), the size scale of the microstructure, and the effective modulus of the polymer blend. Interestingly, one of the films that produced the lowest friction coefficient (xPS = 0.70) also showed negligible visual evidence of wear; in fact, the wear track of this film was difficult to detect due to the absence of defects, debris, and other wear-induced features that typically provided optical contrast between the unworn and worn regions of the film. However, of particular interest, the wear track width was insensitive to composition from 0.20 ≤ xPS ≤ 0.75, falling between 70 and 130 μm. At higher PS content, the wear track width decreased significantly to 20−40 μm (see Figure S5 for a plot of the wear track widths versus blend composition). Because the contact radius scales inversely with the effective contact modulus to the one-third power,56−58 these wear track widths suggested that the transition from lower friction to higher friction (0.75 ≤ xPS ≤ 0.85) was accompanied by an order of magnitude increase in effective contact modulus. Against glass, the effective contact modulus (i.e., of both surfaces), and changes thereof, appear to have been dominated by the mechanical properties of the softer polymer blend films. Additionally, minimal friction and wear of this system

■

RESULTS The effect of PS/PI blend composition on friction and wear properties was studied through a series of tribometry experiments using glass as the counterbody. For pure PS films, μPS (xPS = 1) was largely insensitive to velocity and had values between 0.46 and 0.52 (Figure 2); these values matched well with literature reports for PS probed with steel (0.40 ≤ μPS ≤ 0.60).50 For pure PI films, μPI (xPS = 0) was much higher, ranging from 1.3 to 2.1 and increased monotonically with increasing speed. These values were reasonably consistent with those from Grosch, who found that rubber friction on smooth surfaces (e.g., glass) exhibited a maximum μ between 2.5 and 3.51 The μ values for polymer blends probed by glass beads are shown in Figure 2, as a function of PS mass fraction for sliding speeds of 0.5, 4.6, and 7.8 mm/s. The friction coefficients decreased with increased PS content up to xPS = ∼ 0.75, and a minimum in μ was noted. The lowest friction coefficient value C

DOI: 10.1021/acsami.7b10170 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 3. Optical micrographs of wear tracks on as-cast blend films following tribometry with the glass probe. The white lines indicate the visual location of the wear tracks in the blend films (see Figure S5 for a plot of wear track widths). Increasing speed increases the width of the wear track, with the higher speeds also having a more visible wear track. Additionally, higher xPS films have smaller wear tracks. The wear tracks are difficult to see in the xPS = 0.70 and xPS = 0.75 samples. The area-weighted chord length for each of the films is listed below the image for reference (see also Figure S6). The scale bars represent 50 μm in all images.

accompanied the apparent transition in effective contact modulus that occurred at xPS ∼ 0.80 (see also Figure S5). Thus, a second set of experiments was performed to decouple the contact area and pressure from the mechanical properties of the polymer blends, wherein a soft (relative to the films) HDPE (E of ∼0.4 GPa) sphere was used as the contacting countersurface. In these experiments, the effective contact modulus was set by the softer HDPE probe and, thus, largely independent of changes in the mechanical properties of the film. The effect of the counterface material on the tribological response of these films is illustrated in Figure 4. The friction coefficient of PI homopolymer was significantly lower and less speed-dependent against HDPE than against glass; a similar trend was noted in the xPS = 0.25 specimen. At xPS > 0.25, the frictional response of the polymer blends did not depend strongly on counterface material. These results suggest that the unexpected minimum in friction coefficient was attributable to the nature of the polymer blend itself and largely independent of the contact area and other extrinsic system-dependent factors, such as effective contact modulus and work of adhesion between the two surfaces.56−59 Pull-off measurements were conducted to investigate the effect of adhesion on friction in these systems. In a perfectly elastic contact, the pull-off force is directly proportional to the probe radius and the work of adhesion between the mating materials.58 For PS/PI blends, we found that pull-off forces increased in proportion to the applied load, which motivated the use of an adhesion coefficient.46 No systematic effects from contact aging or retraction rate were detected; therefore, variable aging and rate experiments were treated as repeated tests, as described in the Experimental Section. See Figure S7 for more information. The film composition and probe material had significant effects on the adhesion coefficient, α, and those results are summarized in Figure 5a. Against glass, PS produced 2 orders of magnitude smaller pull-off forces than PI. This lack of

Figure 4. Effect of composition (xPS) on μ measured with an HDPE probe (orange) for three different speeds: 7.8 mm/s (▲), 4.6 mm/s (■), and 0.5 mm/s (◆). The data points represent the average of time-averaged μ from multiple replicates. As with the glass probe (data shown in teal), there was a low μ region measured with the HDPE probe around xPS ≈ 0.75. The error bars represent the standard deviation of the friction coefficient values from the replicates, which are larger than the error propagated from the uncertainties in the timeaveraged friction coefficients. The shaded regions indicate the approximate range of friction coefficients measured with each probe.

adhesion by stiff materials is a well-known phenomenon and is attributable to the fact that they lack the compliance needed to conform to the rough and stiff counterbody;58 in other words, the “real” areas of contacts are far smaller than “nominal” areas of contact when the surfaces are stiff and rough. Because the surface energies of PS and PI were comparable, differences in their α’s against glass likely were closely related to differences in the real contact area, which depends on the composite topography and modulus, rather than differences in the work of adhesion. D

DOI: 10.1021/acsami.7b10170 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

Figure 5. (a) Adhesion coefficient of PS (dark gray) and PI (light gray) homopolymer films with glass and HDPE probes. The data represent the average adhesion coefficient from all approach/retraction speeds with the standard deviation shown by the error bars. The adhesion coefficients (ratio of pull-off force to normal force) between PS and the probes were significantly smaller than the adhesion coefficients between PI and the probes. (b) Comparison of the average dynamic friction coefficient (◆) and average adhesion coefficient (Δ) as a function of composition when probed using HDPE. For the adhesion coefficient, each data point represents the average adhesion coefficient from all retraction experiments on a given film, as described in the Experimental Section. The friction coefficient is the average of the rate-dependent friction coefficients against HDPE, and the error bars represent the standard deviation in the friction coefficient data.

this work is the first study to isolate the effect of microstructure from the effect of composition on the tribological properties of a polymer blend.

Replacing the glass probe with an HDPE probe, which had approximately 1/10th of the surface energy and 1/100th the modulus of glass, reduced the work of adhesion and the effective contact modulus of each material pair. Against HDPE, PI produced significantly less adhesion, which suggested that the adhesion of PI was dominated by the work of adhesion of the material pair. Conversely, PS produced greater adhesion against HDPE, which reinforced the hypothesis that the adhesion of PS was far more sensitive to the real area of contact than it was to the work of adhesion. Nonetheless, the pull-off forces produced by PS were more than an order of magnitude smaller than those produced by PI against both probe materials. The average adhesion coefficient against HDPE is plotted as a function of PS mass fraction in Figure 5b. α exhibited the same minimum at xPS = 0.75 as the friction coefficient (Figure 4); in fact, the xPS = 0.75 film produced negligible adhesion despite the larger contact area imposed by the soft HDPE probe. Additionally, the friction coefficient (Figure 4) and adhesion coefficient (Figure 5) against HDPE followed the same qualitative trend with PS content, which suggested that adhesion was the primary mechanism of frictional dissipation. The fact that adhesive friction coefficients were significantly greater than the corresponding adhesion coefficients could be attributed to two well-known effects: (1) stored elastic energy helped break up adhesive contacts during pull-off and (2) initially small areas of real contact grew under the influence of shear in materials that deformed plastically.46,60 The addition of PS to PI changed the composition of the film as well as the size scale of microstructural features, both of which possibly affected contact mechanics, real areas of contact, work of adhesion, friction, and wear. The lowest friction system (xPS = 0.75) was annealed to coarsen domains and enable the probing of microstructural size scale effects at a constant composition. As illustrated in Figure 6, the size scale of the domains within the polymer blend increased with annealing time. More importantly, the friction coefficient increased with domain size; an ∼100% increase in friction accompanied an ∼150% increase in microstructure size scale. To our knowledge,

■

DISCUSSION There are several examples in the tribology literature in which the friction coefficient or wear rate of a polymer blend or composite was lower than that of either constituent.22,61−63 Because the effect could not be attributed to rules of mixtures, these minima were attributed to material synergies that have proven difficult to isolate or test in traditional material systems. Furthermore, these blends involve materials whose tribological properties were expected to interact in synergistic ways; for example, PTFE and HDPE are thought to reduce the friction via reduced interfacial shear strength, whereas PEEK and other high strength engineering polymers are thought to reinforce the composite and reduce the real contact area.22 The model polymer blend used in this study had no obvious tribological synergy. In comparison to PTFE, the soft PI phase had high surface energy and, thus, high friction (μ > 0.80). The dynamic friction coefficient of PS was on the order of 0.5, which is typical of polymers used as reinforcing phases in polymer blends.46 Despite the relatively high dynamic friction coefficients of both homopolymer constituents, the xPS = 0.70 blend film exhibited low dynamic friction coefficients (μ < 0.11) rivaling those of PTFE during sliding under similar conditions (Figure 2). Most previous studies have attributed comparable tribological minima to the preferential drawing of low shear strength running films of the soft constituent (e.g., PTFE) over the relatively harder composite surface; the end outcome is reduced friction from a combination of reduced shear strength over a reduced total area of real contact.62,64,65 However, our results are inconsistent with this hypothesis. First, pull-off measurements on the unworn polymer blend revealed negligible adhesion at the composition with the lowest friction and wear despite the absence of a pre-existing running film. Second, subjecting the film to sliding, to intentionally create a running film, had no detectable effect on the adhesion coefficient. Third, E

DOI: 10.1021/acsami.7b10170 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

cannot be attributed directly to the effect of composition on roughness or modulus. Burris and Sawyer hypothesized that the tribological performance of their low friction and low wear PEEK/PTFE blend was due to the unusually small size of the cryo-ground PEEK resin and its effect on the domain size and continuity in the polymer blend;22 however, the PEEK/PTFE composite system precluded any direct testing of this hypothesis. Here, we used film post-processing (solvent vapor annealing) to systematically vary the size scale of the microstructural domains while holding composition constant. The results in Figure 6 demonstrated that friction coefficient decreased significantly with decreasing domain size scale, especially at the higher probe velocities. To our knowledge, this study is the first to show that reducing domain size scale in composite films has an independent friction-reducing effect. Our results also provide some insights into how purely structural characteristics of the polymer blend might have been useful in overcoming the relatively poor frictional attributes of the constituent polymers. The data suggest that the friction behavior was adhesive in nature. Adhesion in our systems depended on the work of adhesion (surface energies and chemical compatibilities) and real area of contact. The relatively low adhesion exhibited by PS can be attributed to its stiffness and inability to conform to the counterbody; conversely, compliant polymers like PI are known to produce high adhesion because they are able to conform to even rough surfaces. The minimum friction blend also exhibited negligible adhesion. This result, we propose, is due to preferential load support from the stiff phase,69−71 which discourages snap-in by the soft phase. We expect this effect to be enhanced as the length scale of the soft phase decreases. Nonetheless, even stiff polymers exhibiting negligible pull-off forces, such as PS, produce high friction coefficients that are typically in the range of 0.3−0.6.69 This discrepancy between adhesive pull-off forces and adhesive friction coefficients is attributed primarily to (1) the breakup of adhesive contacts by stored elastic energy during pull-off; and (2) the shear-induced growth and coalescence of initially small contact junctions during sliding.46 We believe the phase separated microstructure of the optimal [minimum friction] polymer blend mitigated friction by (1) discouraging snap-in of the soft phase during contact; and (2) physically impeding the growth and coalescence of contact junctions during sliding, suggesting that wear and its frictional consequences can be arrested by physically discretizing the contact area via microstructural manipulation. In summary, physically isolating the contact areas by manipulating the composite microstructure via polymer phase separation created a structural synergy that produced unexpectedly soft films given the high loading of the stiff component, negligible adhesion due to preferential load support by the stiff phase, and low friction due to the inhibition of unmitigated junction growth and coalescence. At loadings below xPS = 0.75, friction increased in proportion to PI content due, presumably, to increased load support from, and snap-in by, the PI domains. At loadings above xPS = 0.75, the PI domains likely were too sparse or small to interfere with the growth of the PS contact junctions, which led to PS-like frictional properties. At compositions near the friction minimum, the friction increased proportionally to the size scale of the phase-separated microstructure. Thus, this study demonstrates that tribological properties of these polymer blends and, likely, those of other polymer blends, depend just

Figure 6. Optical micrographs for PS/PI blend films at a constant xPS = 0.75. Films were annealed for (a) 0 h [as-cast], (b) 4 h, (c) 8 h, (d) 10 h, (e) 14 h, (f) 16 h, (g) 24 h, (h) 25 h, and (i) 48 h; the domain size increased with annealing time. Low intensity in (a) is the result of PI wetting layer formed on as-cast film. The brightness and contrast of the optical micrographs are enhanced for visual clarity. The scale bars represent 25 μm. (j) Effect of chord length on μ of the annealed films in (a)−(i) measured with a glass probe for 3 different speeds: 7.8 mm/ s (green triangles, ▲), 4.6 mm/s (blue squares, ■), and 0.5 mm/s (red diamonds, ⧫). μ at 7.8 and 4.6 mm/s increased with increasing chord length. μ remained nearly constant at a probe speed of 0.5 mm/s. The error bars represent the propagated error from the fluctuations in steady-state friction coefficient in combination with the variation in steady-state friction coefficient from measurements on different films.

unlike PTFE and HDPE, whose mechanical properties may promote the formation of a running film,66−68 there is no reason to believe either constituent used in this study is a natural running film former. In this case, the transfer of either constituent to either surface (e.g., polymer blend components to glass or HDPE to the blend film) does not explain the low friction coefficients found near xPS = 0.75. Changes in composition caused corresponding changes in effective modulus (Figure S5) and roughness (Figure S1), both of which are known to affect real areas of contact and adhesive forces. We explored these potential effects by varying the counterface from glass, which is stiffer and smoother (E ∼ 100 GPa, ∼80 nm RMS roughness) than the films, to HDPE, which is softer and rougher (E ∼ 0.4 GPa, ∼800 nm RMS roughness) than the films (E ∼ 1 GPa, Ra ∼ 300 nm). Although this change significantly affected the pull-off forces, it had a minimal effect on friction coefficients of the blend films in the vicinity of the optimal (friction minimum) composition. Thus, the unexpected tribological synergy demonstrated in this work F

DOI: 10.1021/acsami.7b10170 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces

University of Delaware for financial support. J.A.E. and E.M.F. also acknowledge NSF CBET-1235955 and NASA Grant No. NNX10AE44G for funding. N.G. and D.L.B. acknowledge NSF CMMI-1434435 for financial support. A.C.M. acknowledges the University of Delaware for financial support. The authors also acknowledge Dr. S. Mastroianni for synthesis of the PI homopolymer. We thank the UD W. M. Keck Microscopy Facility for use of the AFM. Use of the AFM in this work was supported, in part, by the Delaware COBRE program, with a grant from the National Institute of General Medical Sciences − NIGMS (5 P30 GM110758-02) from the National Institutes of Health.

as strongly on the nature of the phase-separated microstructure as they do on the composition of the polymer blend and the tribofilms that typically form.62,64,72

■

CONCLUSIONS We leveraged the phase separation in a model PS/PI blend system for rational design of the friction and adhesion properties in coatings. A minimum in μ, which was lower than the dynamic friction coefficients of either the PS or PI homopolymers, was found between 0.60 ≤ xPS ≤ 0.80 using both glass and HDPE probes. This low μ resulted from synergies generated upon the formation and tuning of the PS/ PI microstructure in the blend films and was not predicted on the basis of standard mixing rules. The low μ in our model polymer blend was comparable to the dynamic friction coefficient of a standard high-performance tribo-material, PTFE. The results provide direct evidence that the nature of friction reduction was linked to an effect of the phase-separated microstructure (as opposed to more traditional effects of lubricating running and transfer films) that has not, to our knowledge, been described previously. Our results suggest that tribologically relevant materials can be engineered from tribologically poor constituents by manipulating the domain size and composition of the polymer blend. Lastly, this study demonstrates that compositional/microstructural control and a working understanding of the appropriate microstructure− friction coefficient relationship are needed to engineer the next generation of high performance solid lubricants.

■

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.7b10170. Additional dynamic friction coefficient traces for HDPE and glass probes. Dynamic friction coefficient traces and OM images of wear of a PI film with glass. Dynamic friction coefficient trace of PTFE with glass and HDPE probes. OM and AFM images of blend film morphologies. Plot of wear track width as a function of composition. Plot of mass fraction vs domain size. Plot of adhesion coefficient as a function of dwell time for blend films (PDF)

■

■

ABBREVIATIONS PS=polystyrene PI=polyisoprene PTFE=polytetrafluoroethylene PEEK=poly(ether ether ketone) HDPE=high-density polyethylene FN=normal force (load) FF=friction force xPS=mass fraction PS with respect to total polymer μ=dynamic friction coefficient μPS=dynamic friction coefficient of PS homopolymer μPI=dynamic friction coefficient of PI homopolymer Fp‑o=pull-off force Mn=number-average molecular weight Đ=dispersity mPS=mass polystyrene mPI=mass polyisoprene A=area p=perimeter AFM=atomic force microscopy OM=optical microscopy N=number of samples α=adhesion coefficient REFERENCES

(1) Paul, D. R. Polymer Blends; Elsevier Science: New York, 2012. (2) Utracki, L. A. Polymer Blends Handbook; Springer: Dordrecht, Netherlands, 2003. (3) Fabiano, S.; Himmelberger, S.; Drees, M.; Chen, Z.; Altamimi, R. M.; Salleo, A.; Loi, M. A.; Facchetti, A. C. Charge Transport Orthogonality in All-Polymer Blend Transistors, Diodes, and Solar Cells. Adv. Energy Mater. 2014, 4, 1301409. (4) Yu, G.; Nishino, H.; Heeger, A. J.; Chen, T. A.; Rieke, R. D. Enhanced Electroluminescence from Semiconducting Polymer Blends. Synth. Met. 1995, 72, 249−252. (5) Javier, A. E.; Patel, S. N.; Hallinan, D. T.; Srinivasan, V.; Balsara, N. P. Simultaneous Electronic and Ionic Conduction in a Block Copolymer: Application in Lithium Battery Electrodes. Angew. Chem., Int. Ed. 2011, 50, 9848−9851. (6) Müller-Buschbaum, P.; Ittner, T.; Maurer, E.; Körstgens, V.; Petry, W. Pressure-Sensitive Adhesive Blend Films for Low-Tack Applications. Macromol. Mater. Eng. 2007, 292, 825−834. (7) Nunes-Pereira, J.; Costa, C. M.; Lanceros-Mendez, S. Polymer Composites and Blends for Battery Separators: State of the Art, Challenges and Future Trends. J. Power Sources 2015, 281, 378−398. (8) Colaneri, N. F.; Schacklette, L. W. EMI Shielding Measurements of Conductive Polymer Blends. IEEE Trans. Instrum. Meas. 1992, 41, 291−297. (9) Walheim, S.; Schäffer, E.; Mlynek, J.; Steiner, U. NanophaseSeparated Polymer Films as High-Performance Antireflection Coatings. Science 1999, 283, 520−522.

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected] (T.H.E.). *E-mail: [email protected] (E.M.F.). *E-mail: [email protected] (D.L.B.). ORCID

Jillian A. Emerson: 0000-0001-7536-1941 Thomas H. Epps III: 0000-0002-2513-0966 Author Contributions

The manuscript was written through contributions of all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS J.A.E. and T.H.E. acknowledge the Thomas and Kipp Gutshall Professorship in Chemical and Biomolecular Engineering at the G

DOI: 10.1021/acsami.7b10170 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces (10) Hussain, F.; Hojjati, M.; Okamoto, M.; Gorga, R. E. Review Article: Polymer-Matrix Nanocomposites, Processing, Manufacturing, and Application: An Overview. J. Compos. Mater. 2006, 40, 1511− 1575. (11) Macosko, C. W. Morphology Development and Control in Immiscible Polymer Blends. Macromol. Symp. 2000, 149, 171−184. (12) Nicho, M. E.; García-Escobar, C. H.; Arenas, M. C.; AltuzarCoello, P.; Cruz-Silva, R.; Güizado-Rodríguez, M. Influence of P3HT Concentration on Morphological, Optical and Electrical Properties of P3HT/PS and P3HT/PMMA Binary Blends. Mater. Sci. Eng., B 2011, 176, 1393−1400. (13) Ruderer, M. A.; Wang, C.; Schaible, E.; Hexemer, A.; Xu, T.; Müller-Buschbaum, P. Morphology and Optical Properties of P3HT: MEH-CN-PPV Blend Films. Macromolecules 2013, 46, 4491−4501. (14) Moons, E. Conjugated Polymer Blends: Linking Film Morphology to Performance of Light Emitting Diodes and Photodiodes. J. Phys.: Condens. Matter 2002, 14, 12235−12260. (15) Qiu, L.; Lim, J. A.; Wang, X.; Lee, W. H.; Hwang, M.; Cho, K. Versatile Use of Vertical-Phase-Separation-Induced Bilayer Structures in Organic Thin-Film Transistors. Adv. Mater. 2008, 20, 1141−1145. (16) Riscanu, D.; Favis, B. D.; Feng, C.; Matsuura, T. Thin-Film Membranes Derived from Co-Continuous Polymer Blends: Preparation and Performance. Polymer 2004, 45, 5597−5609. (17) Chen, Z.-B.; Li, T.-S.; Yang, Y.-L.; Zhang, Y.; Lai, S.-Q. The Effect of Phase Structure on the Tribological Properties of PA66/ HDPE Blends. Macromol. Mater. Eng. 2004, 289, 662−671. (18) Chen, Z.; Liu, X.; Li, T.; Lü, R. Mechanical and Tribological Properties of PA66/PPS Blend. II. Filled with PTFE. J. Appl. Polym. Sci. 2006, 101, 969−977. (19) Brostow, W.; Datashvili, T.; Huang, B. Tribological Properties of Blends of Melamine-Formaldehyde Resin with Low Density Polyethylene. Polym. Eng. Sci. 2008, 48, 292−296. (20) Bijwe, J.; Sen, S.; Ghosh, A. Influence of PTFE Content in PEEK−PTFE Blends on Mechanical Properties and Tribo-Performance in Various Wear Modes. Wear 2005, 258, 1536−1542. (21) Briscoe, B. J.; Yao, L. H.; Stolarski, T. A. The Friction and Wear of Poly(tetrafluoroethylene)-Poly(etheretherketone) Composites: An Initial Appraisal of the Optimum Composition. Wear 1986, 108, 357− 374. (22) Burris, D. L.; Sawyer, W. G. A Low Friction and Ultra Low Wear Rate PEEK/PTFE Composite. Wear 2006, 261, 410−418. (23) Stuart, B. H. Tribological Studies of Poly(ether ether ketone) Blends. Tribol. Int. 1998, 31, 647−651. (24) Lu, Z. P.; Friedrich, K. On Sliding Friction and Wear of PEEK and Its Composites. Wear 1995, 181−183, 624−631. (25) Bates, F. S. Polymer-Polymer Phase Behavior. Science 1991, 251, 898−905. (26) Rubinstein, M.; Colby, R. H. Polymer Physics; Oxford University Press: New York, 2003. (27) Bansil, R. Phase Separation in Polymer Solutions and Gels. J. Phys. IV 1993, 03, C1-225−C1−235. (28) Geoghegan, M.; Krausch, G. Wetting at Polymer Surfaces and Interfaces. Prog. Polym. Sci. 2003, 28, 261−302. (29) Emerson, J. A.; Toolan, D. T. W.; Howse, J. R.; Furst, E. M.; Epps, T. H., III Determination of Solvent−Polymer and Polymer− Polymer Flory−Huggins Interaction Parameters for Poly(3-hexylthiophene) via Solvent Vapor Swelling. Macromolecules 2013, 46, 6533− 6540. (30) Wang, P.; Koberstein, J. T. Morphology of Immiscible Polymer Blend Thin Films Prepared by Spin-Coating. Macromolecules 2004, 37, 5671−5681. (31) Wei, J. H.; Coffey, D. C.; Ginger, D. S. Nucleating Pattern Formation in Spin-Coated Polymer Blend Films with Nanoscale Surface Templates. J. Phys. Chem. B 2006, 110, 24324−24330. (32) Yamamura, M.; Nishio, T.; Kajiwara, T.; Adachi, K. Effect of Stepwise Change of Drying Rate on Microstructure Evolution in Polymer Films. Drying Technol. 2001, 19, 1397−1410.

(33) Sprenger, M.; Walheim, S.; Budkowski, A.; Steiner, U. Hierarchic Structure Formation in Binary and Ternary Polymer Blends. Interface Sci. 2003, 11, 225−235. (34) Walheim, S.; Boltau, M.; Mlynek, J.; Krausch, G.; Steiner, U. Structure Formation via Polymer Demixing in Spin-Cast Films. Macromolecules 1997, 30, 4995−5003. (35) Dalnoki-Veress, K.; Forrest, J. A.; Stevens, J. R.; Dutcher, J. R. Phase Separation Morphology of Spin-Coated Polymer Blend Thin Films. Phys. A 1997, 239, 87−94. (36) Jukes, P. C.; Heriot, S. Y.; Sharp, J. S.; Jones, R. A. L. TimeResolved Light Scattering Studies of Phase Separation in Thin Film Semiconducting Polymer Blends During Spin-Coating. Macromolecules 2005, 38, 2030−2032. (37) Müller-Buschbaum, P.; O’Neill, S. A.; Affrossman, S.; Stamm, M. Phase Separation and Dewetting of Weakly Incompatible Polymer Blend Films. Macromolecules 1998, 31, 5003−5009. (38) Arias, A. C.; Corcoran, N.; Banach, M.; Friend, R. H.; MacKenzie, J. D.; Huck, W. T. S. Vertically Segregated Polymer-Blend Photovoltaic Thin-Film Structures through Surface-Mediated Solution Processing. Appl. Phys. Lett. 2002, 80, 1695−1697. (39) Wang, H.; Shimizu, K.; Kim, H.; Hobbie, E. K.; Wang, Z.-G.; Han, C. C. Competing Growth Kinetics in Simultaneously Crystallizing and Phase-Separating Polymer Blends. J. Chem. Phys. 2002, 116, 7311−7315. (40) Shi, W.; Han, C. C. Dynamic Competition between Crystallization and Phase Separation at the Growth Interface of a PMMA/PEO Blend. Macromolecules 2012, 45, 336−346. (41) Shi, W.; Yang, J.; Zhang, Y.; Luo, J.; Liang, Y.; Han, C. C. Lamellar Orientation Inversion under Dynamic Interplay between Crystallization and Phase Separation. Macromolecules 2012, 45, 941− 950. (42) Wang, H.; Shimizu, K.; Hobbie, E. K.; Wang, Z.-G.; Meredith, J. C.; Karim, A.; Amis, E. J.; Hsiao, B. S.; Hsieh, E. T.; Han, C. C. Phase Diagram of a Nearly Isorefractive Polyolefin Blend. Macromolecules 2002, 35, 1072−1078. (43) Harris, M.; Appel, G.; Ade, H. Surface Morphology of Annealed Polystyrene and Poly(methyl methacrylate) Thin Film Blends and Bilayers. Macromolecules 2003, 36, 3307−3314. (44) Albert, J. N. L.; Epps, T. H., III Self-Assembly of Block Copolymer Thin Films. Mater. Today 2010, 13, 24−33. (45) Fortelný, I.; Ž ivný, A.; JŮ za, J. Coarsening of the Phase Structure in Immiscible Polymer Blends. Coalescence or Ostwald Ripening? J. Polym. Sci., Part B: Polym. Phys. 1999, 37, 181−187. (46) Rabinowicz, E. Friction and Wear of Materials; Wiley: New York, 1995. (47) Bailey, T. S.; Pham, H. D.; Bates, F. S. Morphological Behavior Bridging the Symmetric AB and ABC States in the Poly(styrene-bisoprene-b-ethylene oxide) Triblock Copolymer System. Macromolecules 2001, 34, 6994−7008. (48) Stafford, C. M.; Roskov, K. E.; Epps, T. H., III; Fasolka, M. J. Generating Thickness Gradients of Thin Polymer Films via Flow Coating. Rev. Sci. Instrum. 2006, 77, 023908. (49) Ferreira, T.; Rashband, W. ImageJ User Guide - IJ 1.46; 2010− 2012. http://rsbweb.nih.gov/ij/docs/user-guide.pdf. (50) McLaren, K. G.; Tabor, D. Visco-Elastic Properties and the Friction of Solids: Friction of Polymers: Influence of Speed and Temperature. Nature 1963, 197, 856−858. (51) Grosch, K. A. Visco-Elastic Properties and the Friction of Solids: Relation Between the Friction and Visco-Elastic Properties of Rubber. Nature 1963, 197, 858−859. (52) Burris, D. L.; Sawyer, W. G. Addressing Practical Challenges of Low Friction Coefficient Measurements. Tribol. Lett. 2009, 35, 17−23. (53) Tiwari, A.; Dorogin, L.; Bennett, A. I.; Schulze, K. D.; Sawyer, W. G.; Tahir, M.; Heinrich, G.; Persson, B. N. J. The Effect of Surface Roughness and Viscoelasticity on Rubber Adhesion. Soft Matter 2017, 13, 3602−3621. (54) Hammerschmidt, J. A.; Gladfelter, W. L.; Haugstad, G. Probing Polymer Viscoelastic Relaxations with Temperature-Controlled Friction Force Microscopy. Macromolecules 1999, 32, 3360−3367. H

DOI: 10.1021/acsami.7b10170 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX

Research Article

ACS Applied Materials & Interfaces (55) Grosch, K. A. The Relation Between the Friction and ViscoElastic Properties of Rubber. Proc. R. Soc. London, Ser. A 1963, 274, 21−39. (56) Hertz, H. Miscellaneous Papers; Macmillan and Co.: London, 1896. (57) Hertz, H. Ueber die Berührung Fester Elastischer Körper. J. Reine Angew. Math. 1882, 92, 156−171. (58) Johnson, K. L.; Kendall, K.; Roberts, A. D. Surface Energy and the Contact of Elastic Solids. Proc. R. Soc. London, Ser. A 1971, 324, 301. (59) Brostow, W.; Lobland, H. E. H. Materials: Introduction and Applications; John Wiley & Sons, Inc.: Hoboken, NJ, 2016. (60) Sikorski, M. E. The Adhesion of Metals and Factors that Influence It. Wear 1964, 7, 144−162. (61) Liu, X.; Li, T.; Tian, N.; Liu, W. Tribological Properties of PTFE-Filled PMIA. J. Appl. Polym. Sci. 1999, 74, 747−751. (62) Friedrich, K.; Lu, Z.; Hager, A. M. Recent Advances in Polymer Composites’ Tribology. Wear 1995, 190, 139−144. (63) Palabiyik, M.; Bahadur, S. Mechanical and Tribological Properties of Polyamide 6 and High Density Polyethylene Polyblends with and without Compatibilizer. Wear 2000, 246, 149−158. (64) Burris, D. L.; Boesl, B.; Bourne, G. R.; Sawyer, W. G. Polymeric Nanocomposites for Tribological Applications. Macromol. Mater. Eng. 2007, 292, 387−402. (65) McCook, N. L.; Burris, D. L.; Bourne, G. R.; Steffens, J.; Hanrahan, J. R.; Sawyer, W. G. Wear Resistant Solid Lubricant Coating Made from PTFE and Epoxy. Tribol. Lett. 2005, 18, 119−124. (66) Makinson, K. R.; Tabor, D. The Friction and Transfer of Polytetrafluoroethylene. Proc. R. Soc. London, Ser. A 1964, 281, 49−61. (67) McCook, N. L.; Boesl, B.; Burris, D. L.; Sawyer, W. G. Epoxy, ZnO, and PTFE Nanocomposite: Friction and Wear Optimization. Tribol. Lett. 2006, 22, 253−257. (68) Wang, A. A Unified Theory of Wear for Ultra-High Molecular Weight Polyethylene in Multi-Directional Sliding. Wear 2001, 248, 38−47. (69) Lancaster, J. The Effect of Carbon Fibre Reinforcement on the Friction and Wear of Polymers. J. Phys. D: Appl. Phys. 1968, 1, 549. (70) Tanaka, K.; Kawakami, S. Effect of Various Fillers on the Friction and Wear of Polytetrafluoroethylene-Based Composites. Wear 1982, 79, 221−234. (71) Blanchet, T. A. A Model for Polymer Composite Wear Behavior Including Preferential Load Support and Surface Accumulation of Filler Particulates. Tribol. Trans. 1995, 38, 821−828. (72) Ye, J.; Burris, D. L.; Xie, T. A Review of Transfer Films and Their Role in Ultra-Low-Wear Sliding of Polymers. Lubricants 2016, 4, 4.

I

DOI: 10.1021/acsami.7b10170 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX