6.1 Background

Certain polymer blends and composites achieve low friction coefficients (μ < 0.2) and low wear rates (k < 10−6 mm3/Nm) without intentional external lubrication of the contact area [1,2,3]; as such, these materials are known as solid lubricants. Unfortunately, no general rules exist to guide the design of such materials, and solid lubricant materials discovery almost always involves substantial trial and error. Fillers, depending on their properties, can affect the tribology of polymers by preferentially supporting load [4], arresting subsurface cracks [5], and modifying interfacial shear strength [6] to name a few mechanisms. However, none of these functions explains the orders of magnitude effects some small changes in filler composition [2, 7], loading [2, 8], or environment [9, 10] are known to have on the wear rate of the system. Normally, the polymer produces wear debris during sliding against a hard metallic counterface, most commonly steel. These debris particles are dragged through the contact and are eventually ejected from the wear track in most cases. In special cases, the debris can adhere to the counterface to initiate the formation of a layer called the transfer film. Regardless of the materials used, low wear sliding of polymeric materials always leaves a thin and continuous transfer film on the steel counterpart. Likewise, high wear sliding of polymers always leaves a thick and patchy transfer film. The studies of polymer solid lubricants in the literature provide strong evidence that the properties of transfer films and the measured wear rates are related [11,12,13,14,15,16,17,18,19,20].

Briscoe [13] first recognized the relationship between transfer film properties and polymer wear rate and concluded that the primary role of the filler is to help adhere the debris to the counterface. He suggested that active fillers degrade the relatively inert polymer chains thereby creating new opportunities for bonding with the counterface. Bahadur and Gong [21] asserted that the role of a filler is to enhance counterface adhesion by having the active filler itself create the bond between the polymer debris and counterface. Both theories support the need for a tribochemically adhered transfer film that reduces wear by protecting the polymer from the hard and high energy counterface. Bahadur and Tabor [22] conducted a clever experiment to test this idea. They used low wear polymers to form high quality transfer films and then measured the wear rates of traditionally high wear polymers against presumably protective predeposited transfer films. Interestingly, these high quality predeposited films had no effect on the wear rate of the high wear polymers. On further inspection, they found that the predeposited transfer film was removed almost immediately during sliding against the new polymer pin. Based on this observation, they concluded that the high-quality transfer films must be a consequence of the small debris generated during low wear sliding and not the cause of low wear. This is supported by evidence from Ricklin [23], Tanaka and Kawakami [24], Blanchet and Kennedy [5], and Burris and Sawyer [2], all of whom concluded that effective fillers reduced the wear of polytetrafluoroethylene (PTFE) by disrupting subsurface damage and thereby reducing debris size.

PTFE-based polymer systems have been the subject of numerous studies of the relationship between tribochemistry, transfer film adhesion, and tribological performance. Gong et al. [25] and Blanchet et al. [26] used XPS (X-ray photoelectron spectroscopy) to study the effects of tribochemistry on transfer film adhesion and wear reduction of PTFE-based materials. Although they observed clear evidence of tribochemistry in the form of metal fluorides, these reaction products occurred in both low and high wear systems. Furthermore, because fluorine is monovalent, metal-fluorides represent the transfer of fluorine from the polymer to the counterface and not a bond between them. Both studies concluded that tribochemically induced transfer film adhesion was not the primary means of wear reduction in the PTFE system.

Although much of the recent literature suggests that tribochemistry is simply a consequence of bond-breaking during tribological contact, other studies have shown clear evidence that it can play a critical role in wear reduction. During studies of a particularly low wear alumina-PTFE system in environments of variable humidity, Krick et al. [9] found that the wear rate increased by orders of magnitude as the availability of gaseous water was systematically removed. This observation provides strong evidence that tribochemistry plays an important role in the wear reduction of PTFE and, potentially, polymers in general [9, 10, 27]. To date, however, the link between tribochemistry, transfer film adhesion, and wear reduction remains uncertain.

The way in which transfer films develop may have equally important implications for elucidating causation between transfer film qualities, adhesion, and wear rate. Based on XPS studies, Gong et al. [28] concluded that PTFE-based transfer films grow over time. They interpreted their result as evidence of a thickening process in which new layers deposit onto old layers; the role of the filler, they proposed, is to strengthen the bond at these internal interfaces within the transfer film. Blanchet et al. [26] found similar evidence of growth but concluded that transfer film growth was the result of new debris adhering to remaining areas of exposed counterface. Ye et al. [29] used in situ optical microscopy to study the development of a transfer film in real time. Their results showed that the transfer film developed as new debris adhered to remaining areas of exposed counterface. The results also showed that existing transfer films thickened with increased sliding. More importantly, they showed that the transfer film of a low wear PTFE-based system was very stable; many of the debris particles deposited early in the test could be identified at the end of the test. The results demonstrated that the transfer films of this very wear resistant polymer are themselves very wear resistant. Ye et al. [6] studied the wear rate of this transfer film directly and found it to be exceedingly resistant to removal with an effective wear rate orders of magnitude lower than that of the parent polymer. Interestingly, the wear rate increased by many orders of magnitude when the surface energy of the wear probe increased beyond a threshold value, thus demonstrating that a low wear condition simply reflects the presence of a single weak interface to accommodate sliding.

There is little doubt based on the literature that transfer films play one or more critical roles in governing the wear response of polymeric solid lubricants. Nonetheless, the specific way in which fillers, polymers, stresses, and environments govern the development and properties of the transfer films remain unclear. This chapter reviews the most recent advancements in this field that have helped to shape our understanding of the cause-effect relationships between materials design, transfer film properties, and tribological performance of these polymer systems.

6.2 Properties of Transfer Films

6.2.1 Topography

Transfer films and their properties are central to most hypothesized mechanisms of polymer wear reduction. The reason is well-illustrated by Fig. 6.1. Two 5 wt% alumina-PTFE nanocomposites were prepared identically and tested under the same tribological conditions (50 mm/s, 6.4 MPa). The only known differences were the size, morphology, and phase of the alumina filler particles. The Δ-Γ phase alumina-PTFE composite produced a moderate wear rate of 10−5 mm3/Nm, while the α phase alumina-PTFE composite produced a low wear rate of 10−7 mm3/Nm. There are marked visual differences between the transfer films produced by these materials. The transfer film of the moderate wear rate composite is thicker and patchier, while the transfer film of the low wear rate composite is thinner and more complete. It is logical to conclude that the 100X difference in wear rates is somehow related to the obvious visual differences between the transfer films formed by these similar composite materials.

Fig. 6.1
figure 1

Two types of transfer films of PTFE nanocomposites: (a) transfer film of a moderate wear system (K~ 10−5 mm3/Nm, 44 nm Δ: Γ alumina) and (b) transfer film of an ultra-low-wear system (K~10−7 mm3/Nm, 80 nm α alumina)

Full size image

There are numerous examples in the literature for which improved wear resistance of a filled polymer is accompanied by comparable visual improvements in the appearance of the transfer film [5, 14,15,16,17, 21, 22, 30,31,32,33]. Wang et al. [7] found that nanoscale ZrO2 significantly reduced the wear of polyetheretherketone (PEEK). Scanning electron microscopy revealed that the transfer film of unfilled PEEK was “thick, lumpy, and incoherent” while that of ZrO2-PEEK nanocomposite was “thin, uniform, and coherent.” They obtained similar results from a study of nanoscale SiC in PEEK [8] and described the transfer films as “thin, uniform, and tenacious” based on similar post-test visual observations of transfer films. The authors attributed the tribological benefits of nanofillers in PEEK to improvements in quality of the transfer films. Li et al. [9] found that the addition of nanoscale ZnO to polytetrafluoroethylene (PTFE) reduced its wear rate while improving the “uniformity and tenacity” of the transfer films. Sawyer et al. [10] described the transfer films of low wear nanoscale alumina reinforced PTFE as “well adhered, smooth, and continuous.” Bhimaraj et al. [11] found that alumina filled polyethylene terephthalate (PET) nanocomposites produced more “coherent and uniform” transfer films with increasing nanoparticle loading. McCook et al. [12] noted that more “uniform” transfer films accompanied improved friction and wear performance of their epoxy nanocomposites.

Although there is clearly a ubiquitous relationship between wear rate and transfer film quality, it is difficult to assess the state of the art in general due to a lack of standard definitions for the terms used to describe these properties. “Thinness” can be quantified and is the obvious place to begin defining transfer film quality. A traditional 1-D thickness measurement is illustrated in Fig. 6.2. In these measurements, a sharp stylus is lightly loaded and dragged across the width of the transfer film, while a displacement transducer tracks the deflection of the tip. The height of the transfer film was determined here using the bare counterface as a baseline for subtraction. The average height of the transfer film was defined as the difference between the average height within the wear track and the average height outside the wear track [27]. Blanchet and Kennedy [5] were among the first to use this method to determine the thickness of unfilled PTFE transfer films, which they reported to be in the range from 4 to 20 μm. Pitenis et al. [27] used this method to show that the transfer films of alumina-PTFE nanocomposites thickened with increased sliding distance and approached ~1 μm at steady state (Fig. 6.2). When they removed environmental moisture, the transfer films thickened and wear rates increased [10].

Fig. 6.2
figure 2

Stylus profilometry measurements of transfer film thickness. (Image reprinted from [27] with permission)

Full size image

Many transfer films are discontinuous or patchy. In these cases, the thickness from the method described above will depend on thickness and coverage (the output thickness will be half the transfer film thickness at 50% coverage). A different approach is necessary if the goal is to assess thickness directly. Burris and Sawyer [34] used optical profilometry to obtain the same type of surface profile information, but instead of quantifying the average thickness as described above, they quantified the maximum thickness, which depends only on heights of the transferred debris particles. McElwain [35] used mapping stylus profilometry to quantify maximum transfer film thickness for similar materials. Despite differences in the techniques used, their data suggest that the wear rate of this system is proportional to the cube of the maximum transfer film thickness (Fig. 6.3).

Fig. 6.3
figure 3

Observed correlation between transfer film thickness and wear rates of various PTFE-based systems. (Image reprinted from [19, 34] with permission)

Full size image

Laux and Schwartz [36] proposed a robust method to quantify average thickness that is independent of the coverage; their method is illustrated in Fig. 6.4. When the transfer film is statistically distinct from the roughness, two peaks will emerge in the probability density function of surface height; the lower elevation peak is associated with the counterface and the higher elevation peak is associated with the transfer film. Laux and Schwartz fit two Gaussian curves to the data set and used the distance between the fitted peaks to quantify the most probable thickness of the transfer film. Using this method, they found no strong correlation between wear and transfer films thickness for unfilled PEEK tested at varying conditions [36, 37]. The correlation between transfer film thickness and wear rate appears to be system specific.

Fig. 6.4
figure 4

Laux and Schwartz’s method of measuring transfer film thickness using the counterface’s surface height histogram: (a) scanning white light interferometry image of the wear track, (b) optical image of the wear track, and (c) probability density of the wear track height distribution. (Image reprinted from [36] with permission)

Full size image

Transfer film coverage is similarly intuitive as a potential driver of wear protection; many researchers have used adjectives like complete, uniform, and continuous to describe them [30, 31, 33, 38,39,40]. In theory, the film area fraction (X), which is the ratio of film coverage area to total area, is easy to quantify using optical microscopy, electron microscopy, or profilometry images. In some cases, by simply applying a threshold pixel value, transfer film images can be converted into binary maps from which area fraction is readily calculated, as illustrated by Fig. 6.5. Images size is also critical to the differentiation between the film and the counterface as shown by Ye et al. [41]. Bhimaraj et al. [42] studied the wear and transfer film coverage of polyethylene terephthalate (PET) nanocomposites with different filler loadings and although transfer film area fraction increased with increased filler loading, area fraction and wear were not well correlated. Laux and Schwartz [36] used the same method to quantify transfer film area fraction and studied the effect of sliding direction on the wear of PEEK. They found that the wear rate decreased by 60% with a 100% increase in the transfer film area fraction when the sliding direction changed from single direction to reciprocating. Similarly, Ye et al. [41] found that the transfer film area fraction of a PTFE nanocomposite increased with decreased wear rate throughout a single wear test.

Fig. 6.5
figure 5

Example of converting color images or height profiles to pixel representations of transfer film (black) and exposed counterface (white) to measure (a) area fraction and (b) free-space length. (Image reprinted from [41] with permission)

Full size image

Transfer film thickness and coverage likely play an important role in wear mitigation but neither factor alone appears to be a reliable predictor of wear rate. Laux and Schwartz [37] suggested that transfer film heterogeneity/discontinuity is likely to be more directly relatable to the wear rate. This effect of nonuniformity is consistent with observations from Blanchet et al. [26] that transfer films developed by gradually filling in remaining areas of exposed counterface.

Because covered areas are theoretically protected, future transfer and wear are more likely governed by the nature of the uncovered areas than by the nature of the covered areas; this is consistent with the transfer films in Fig. 6.1. The high quality transfer film is thinner and covers more area, but the visual impression of “uniformity” comes from the fact that the characteristic size of the “gaps”/transfer film-free space/areas of exposed counterface have been substantially reduced. Thus, we propose that heterogeneity should be defined by the characteristic size of the uncovered areas or the free-spaces and not by their total area or area fraction.

Ye et al. [41] proposed that the free-space length (Lf), or the size of the characteristic free-space, limits the size of the wear fragment and may therefore serve as a more reliable topographical predictor of wear rate. The method used to measure free-space length begins with a representative image of the surface, which is then converted into a black and white image; black represents transfer film and white free-space. Their custom Matlab script used Monte Carlo simulation to identify the smallest randomly placed square for which the most probable number of intersecting transfer film pixels is zero; this length is the free-space length. The method is illustrated in Fig. 6.5 using optical and stylus based measurements of the same region. In theory, decreased free-space lengths should produce smaller debris; smaller debris consume less volume while producing thinner and more stable transfer films. Ye et al. [41] studied the relationship between free-space length and wear rate using low wear alumina-PTFE during the transition from high wear during run-in to low wear at steady state. Optical observations showed that the free space decreased with sliding distance as new debris gradually filled in the gaps in the transfer film as described previously by Blanchet et al. [26]. Changes in the free-space length were accompanied by changes in the wear rate as shown in Fig. 6.6a. Wear rate is plotted versus free-space length in Fig. 6.6b; the results suggest that reduced free-space length reduces wear down to a minimum value below which the wear rate remains constant at a value associated with the polymer on a perfect transfer film.

Fig. 6.6
figure 6

(a) Transfer film free-space length and wear rate of an alpha-alumina PTFE nanocomposite during the course of a single wear test. (b) In-situ wear rate plotted against transfer film free-space length. (Image reprinted from [41] with permission)

Full size image

The relationship between free-space length and wear rate in Fig. 6.6a suggests that changes in free-space length caused changes in wear rate, particularly since increased wear rate at 1 km of sliding was preceded by an increase in free-space length. This fact combined with the ubiquitous observation that transfer films of all polymers become thinner and more continuous with reduced wear rates suggests that there may be some universal relationship between transfer film topography and polymer wear. Haidar et al. (manuscript under review by the journal Wear) tested the generality of the relationship between wear rate and transfer film topography using 10 different but representative polymers and polymer composites to determine the strength of correlation between wear rate and transfer film thickness, area fraction, and free-space length; the results are shown in Fig. 6.7. Although there is no reason to believe there should be a universal relationship between wear rate and transfer film properties, the data generally reflect the expected relationships; wear rate tended to decrease with increased coverage, decreased thickness, and decreased free-space length. The best fits to the data are shown as dashed lines. The variation between the data and the bet-fit model was used as an indicator of uncertainty when using the model fit to independently predict wear rate; the grey region represents the mean plus and minus one uncertainty. The uncertainty in the predicted wear rate is 8x based on the free-space length alone, 17x based on the area fraction alone, and 21x based on the thickness alone.

Fig. 6.7
figure 7

A case study of correlation comparison between three transfer film quantifiers and the composite pin’s wear rate. Materials studied by Haidar et al. (manuscript submitted to the journal Wear) include: PTFE, PEEK, PET, PPS, Epoxy, 5% wt. γ-Al2O3 + PTFE, 5% wt. α-Al2O3 + PTFE, 5% wt. γ-Al2O3 + PΕΕΚ, 5% wt. α-Al2O3 + PΕΕΚ and 32% wt. PEEK + PTFE

Full size image

The existing studies show that transfer film topography has limited utility as an independent predictor of wear rate. Nonetheless, despite enormous variations in the material properties of the polymers, their wear properties, and the characteristics of their transfer films, there was a remarkably strong correlation between wear rate and free-space length. Future efforts by other investigators to quantify the topographical qualities of transfer films and relate them to wear rates of polymers will clarify whether any general relationship exists between the topographical properties of transfer films and the wear of polymers.

6.2.2 Adhesion

Most hypotheses of polymer wear resistance are rooted in transfer film adhesion. Many have suggested the adhesion strength at the film-counterface interface determines the tenacity of the transfer film and the wear rate of the polymer. Bahadur and Tabor first showed evidence that transfer films are most likely adhered by mechanical engagement of the debris with the topographical features of the counterface [22]. While studying the effect of counterface finish on the transfer and wear of various PTFE composites, Burris and Sawyer [34] showed that while even extreme roughness and smoothness failed to affect transfer and wear of these materials at steady state, scratches oriented in the direction of sliding disrupted the formation of the transfer film and increased wear rates by orders of magnitude. They concluded that debris was simply pushed from the contact because the surface lacked the features necessary to entrap the debris. Harris et al. [43] systematically studied the wear of the same material system against surfaces with a unidirectional finish. The lowest wear rates and best transfer films were observed when the surface roughness was aligned against the sliding direction. Wear resistance and transfer film quality degraded as the orientation approached the sliding direction. Laux and Schwartz studied the effect of roughness orientation on the transfer of polyetheretherketone (PEEK) using unidirectional surface finishes with a circular wear path. They found that the transfer films were thickest and patchiest when the sliding direction approached the roughness orientation [36]. TEM studies of the transfer films of a low wear epoxy nanocomposite [44] showed evidence that the transfer film initiated in the scratches. In sum, these results suggest strongly that polymer transfer films initiate by some form of mechanical engagement with topographical features of the counterface.

In cases of random surface finish, recent studies have shown that the low wear alumina-PTFE system can only achieve low wear and quality transfer films if there is sufficient water available to fuel a particular tribochemical pathway [9]. During sliding in humid environments, chain scission of the polymer backbone led to the formation of carboxylates, which directly linked degraded polymer chains to the counterface and filler particles. [27, 45]. Thus, both mechanical and chemical factors are necessary but neither sufficient for the formation of quality transfer films.

Transfer films only protect the polymer from contacting the counterface if they persist. Many have used the fact that transfer films are thin and continuous during low wear as evidence that they are also “tenacious,” which implies persistence, longevity, and resistance to removal. However, Bahadur and Tabor showed that even high quality transfer films were quickly removed when the parent low wear polymer was replaced with a high wear polymer [22]. The preexisting transfer film had no wear reducing effect on the high wear polymer because the transfer film lacked tenacity. To date the relationship between wear rate, quality, persistence, and adhesive strength remains unresolved.

In 2013, Ye et al. [29] set out to determine the extent to which the transfer films of low wear alumina-PTFE persisted during sliding. Fig. 6.8 shows images of the transfer film after ~500 m of sliding. This point in the experiment is particularly interesting because it shows how the steady state transfer film initiates. At 485 m, the transfer film consists of a sparse collection of micron and even submicron debris fragments. After 20 m of additional sliding, it is clear that many of the fragments remain. As sliding continues, these debris fragments grow, merge, and homogenize into larger domains that can easily be identified after tens of thousands of additional sliding cycles. This study provides direct insight into the formation mechanisms of a specific transfer film and demonstrates, despite prior evidence to the contrary, that persistence is likely a necessary component of ultra-low-wear sliding of polymeric solid lubricants [40, 46, 47].

Fig. 6.8
figure 8

Images of transfer film development in an ultra-low-wear alumina–PTFE nanocomposite: (a) images illustrating the evolution of the steady state transfer film as a function of distance slid; (b) wear volume as a function of distance slid. (Image reprinted from [29] with permission)

Full size image

Persistence is almost certainly a function of the adhesive strength between the film or debris fragments and the counterface. Unfortunately, transfer film adhesion has proven difficult to quantify. The most common means to quantify coating adhesion is the scratch test. Typically, the normal force is ramped as an indenter slides across the coating surface; the load at which coating failure occurs provides an indirect measure of adhesion strength [48]. A more direct method is the peel test [49, 50]. Schwartz and Bahadur [33] bonded a copper tab to predeposited alumina-PPS nanocomposite transfer films and measured the peel strength as a function of alumina loading. They found that transfer film adhesion strength increased with increased filler loading (up to 2%) and counterface roughness. The results were consistent with debris size regulation and improved engagement of smaller debris fragments with the rough surface. In a subsequent paper, Bahadur and Sunkara [16] measured the wear rates and the transfer film adhesion strength of PPS nanocomposites of different filler types and loadings. Although the wear rates of the composites were both higher and lower than that of the unfilled polymer, their values correlated strongly with the adhesive strength of the transfer film (Fig. 6.9).

Fig. 6.9
figure 9

Measured transfer film counterface bond strength versus composite’s wear rates. (Image reprinted from [16] with permission)

Full size image

The scratch and peel tests have important practical limitations. Neither test provides a useful measure of adhesion strength when the interfacial shear strength exceeds the shear strength of the film [48]. Furthermore, free-spaces from heterogeneity confound the measurements by providing direct bond sites between the adhesive and the counterface in peel tests and by nucleating failure in scratch tests. Lastly, the measurements are confounded by potential contributions from the addition of a second contact interface (adhesive-transfer film and indenter-transfer film). Agrawal and Raj [51, 52] solved these issues with the thin-film stretch test illustrated in Fig. 6.10. The film and counterface are loaded in tension until cracks form and stabilize or until the counterface fails. Mathematically, they showed that the ratio of the crack spacing to coating thickness is proportional to the normalized adhesive strength, which is defined as the ratio of adhesive shear strength to cohesive shear strength. This method requires no additional interface and provides spatial specificity for statistical evaluation. One potentially significant downside of the method is that it only measures the ratio of adhesive strength to cohesive strength; measurements of absolute adhesive strength require knowledge of the absolute cohesive strength of the transfer film. The other downside of the method is that the substrate must have a larger ultimate strain than the transfer film.

Fig. 6.10
figure 10

Thin-film stretch method for measuring transfer film adhesion strength by (a) pulling a sample with a predeposited transfer film in tension along the sliding direction in the native wear tests until transverse cracks within the film initiated and (b) observing the transverse cracks, which have an average crack spacing of λ. This value was found to be inversely proportional to the adhesion shear strength at the film-substrate interface. (Image reprinted from [41] with permission)

Full size image

Although the normalized adhesive strength is often viewed as a limitation of the stretch method, it is likely a better predictor of transfer film persistence than adhesive strength alone. One could argue that delamination is favored if the cohesive strength significantly exceeds the adhesive strength. Delamination is an undesirable failure route because it leaves the surface completely unprotected. Cohesive failure would be expected when the adhesive strength exceeds the cohesive strength. Cohesive failure leaves the counterface protected and is therefore the preferred failure mode. For this reason, the normalized adhesive strength is likely a more valuable metric of coating adhesion than the adhesive strength.

Ye et al. [6] used the thin-film stretch method to study the evolution of the normalized adhesive strength of the low wear alumina-PTFE transfer film from run-in to steady state; the results are shown in Fig. 6.11a. In the beginning of the test, wear rates were high and the cohesive strength of the transfer film exceeded its adhesive strength, which suggests that delamination was the favored failure mode. In situ studies of the thick flaky transfer films of this material during run-in have shown that they were removed and replenished on each pass of the pin [29]; this fact implies that transfer films failed to protect the polymer during the run-in. This is supported by recent experiments, which showed that removing the transfer film by replacing the counterface with a fresh counterface had no detectable effect on the wear volume produced by the pin (Fig. 6.12) [53]. In this case, visual improvements in the transfer film during the run-in period were caused by reduced wear not the cause of reduced wear [22].

Fig. 6.11
figure 11

(a) Normalized transfer film adhesion strength plotted against the sliding distance of an ultra-low wear PTFE nanocomposite. (b) Composite wear rate plotted against the normalized film adhesion strength. (Image reprinted from [6] with permission)

Full size image
Fig. 6.12
figure 12

Wear volume plotted against sliding distance for a test in which the developing transfer film was removed after cycle 43 and replaced with a fresh counterface. The results show that the transfer film and wear rate develop as they would in an uninterrupted test, which suggests that the transfer film develops in response to the developing surface properties of the pin. (Image reprinted from [53] with permission)

Full size image

The transition to low wear was accompanied by the first observations of adherent debris (20 m of sliding in Fig. 6.8). Unfortunately, the fragments were too small during this transition period to produce detectable cracking patterns [29]. However, as Fig. 6.8 demonstrates, these debris fragments merged and homogenized with increased sliding to create larger domains with detectable cracking patterns. The first possible measurement in the steady state sliding period revealed clear evidence that the bond between the transfer film and the counterface significantly exceeded the strength of the film itself, which suggests that cohesive failure was favored over delamination. This transition in normalized adhesive strength accompanies improved topographical quality of the transfer film (Fig. 6.6), the onset of transfer film persistence (Fig. 6.8), and dramatically reduce wear rate (Fig. 6.11), all of which is consistent with the expected transition from delamination to cohesive failure. The wear rate decreased as the normalized adhesive strength increased with sliding distance, which suggests that improved bonding of the transfer film contributes as a cause of wear reduction.

Another way to measure the transfer film persistence is through a direct transfer film wear rate measurement. From a simple control-volume analysis, this value represents the minimum possible wear rate of the solid lubricant system at steady state as noted by Blanchet and Kennedy [5]. Wang et al. [54,55,56] used bare steel spheres to quantify the wear resistance of predeposited transfer films. Standard gravimetric and volumetric wear measurements are difficult due to the fact that transfer films are thin and discontinuous. Because increased friction is expected when metal first contacts metal, the friction coefficient provides a useful indicator of coating failure. Wang et al. [55, 56] used this friction coefficient transition method to show that PTFE composite transfer films persisted 10X longer distances than those of unfilled PTFE. Using similar methods, Li et al. [57] showed that the wear lives of their transfer films were sensitive to the load, speed, and the counterface roughness used in creating the transfer film. Urueña et al. [58] used the same friction-based steel ball-on-film configuration to study wear rates of the low-wear alumina–PTFE transfer films described previously as persistent and well adhered. They used mean film thickness measurements with reasonable geometric assumptions to estimate the worn volume and wear rate at failure. As expected, the wear rate of the transfer film decreased as the wear rate of the polymer decreased during the development of the transfer film (Fig. 6.13a). Surprisingly, however, the wear rates of the transfer films were orders of magnitude larger than the wear rate of the parent system. This result implies that the transfer film was worn immediately and played no role in wear reduction, which is unlikely based on the observations and reasoning discussed previously. The alternative is that the test itself had artificially increased the wear rate of the transfer film, which is more likely given the differences in the contacting materials and pressures.

Fig. 6.13
figure 13

Comparison of wear rates of an alumina-PTFE nanocomposite (K) and the corresponding transfer film (Kfilm) at predefined cycles of development in the native wear test: (a) Result from Urueña et al. [58] using a steel sphere to measure transfer film wear rates, (b) Result from Ye et al. [6] using a polyethylene sphere to measure transfer film wear rates. All error bars represent 95% confidence intervals. (Image reproduced from [6, 58] with permission)

Full size image

Ye et al. [6] tested the wear rates of the same transfer films using a high-density polyethylene sphere to provide a better mechanical surrogate to the original polymer pin. The transfer film wear rate and the original solid lubricant pin wear rate are plotted against the accumulated sliding distance used to generate the transfer film in Fig. 6.13b. During running-in, the transfer film wear rate is much larger than the pin’s wear rate; the transfer films were easily/immediately removed, which is consistent with the results of direct optical observations of the transfer film evolution [29]. After the transition (~7 m of accumulated sliding distance in this case), the transfer film wear rate decreased sharply and stayed well below the pin’s wear rate; this is consistent with the persistence observed from direct optical observations [29] and the transition from weak to strong normalized adhesion strength (Fig. 6.11).

Despite significant potential for differences in the source materials, preparation procedures, and testing conditions between these independent studies, the steady state wear behaviors of the polymer nanocomposites were indistinguishable (Fig. 6.13a, b). Clearly, the wear behavior of this material is robust and insensitive to these differences. The strong agreement in polymer wear rates makes the orders of magnitude disagreements in transfer film wear rates that much more remarkable.

Ye et al. [6] repeated the measurements with spherical probes of varying material to clarify the cause of the enormous differences in the wear response of the same transfer film to different probes. Ultimately, the wear rate of the ultra-low wear transfer film did not vary systematically with contact pressure, friction coefficient, or shear stress during sliding. It did vary systematically as a function of probe surface energy as shown in Fig. 6.14a. Both PTFE and HDPE promoted ultra-low wear behavior of the transfer films. When tested against polymer probes of higher surface energy, the wear rate increased by orders of magnitude. They proposed that low surface energy probes reduced wear by 5 orders of magnitude by setting up a preferred slip system for which pure interfacial sliding becomes favorable over adhesive wear. For this system, interfacial sliding became favorable when the surface energy of the probe was less ~30 mJ/m2, which is comparable to estimates of that of the film. These results reinforce the fact that the wear resistance of a material, particularly a transfer film, can depend as strongly on the properties of the countersurface as it does on the properties of the wear surface.

Fig. 6.14
figure 14

(a) Surface energies of the probes versus ultra-low-wear Al2O3/PTFE transfer film wear rates in microtribometry experiments. Error bars represent the 95% confidence interval; and (b) three-body wear model involving a pin (A), transfer film (B), and counterface (C) (Reprinted with permission from [6])

Full size image

6.2.3 Mechanical Properties

Mechanical properties are also believed to be important contributors to the stability of transfer films and the wear resistance of polymers. Gong et al. [28] first proposed that increased transfer film cohesive strength discourages transfer film failure at internal interfaces and thereby stabilizes the film. However, as illustrated previously, stronger transfer films may be more likely to delaminate without a corresponding increase in strength at the counterface. Mechanical properties can also provide useful information about filler accumulation and polymer degradation, both of which change mechanical properties like hardness, toughness, and modulus.

The primary challenge to measuring the mechanical properties of thin films is isolating the properties of the film from those of the much stiffer and harder substrate [59, 60]. Friedrich et al. [61] used micro-indentation and measured transfer film hardness of two fiber-reinforced PEEK composites with different fiber orientations (normal vs. parallel) relative to the sliding surface. They found that composites produced lower wear rates and harder transfer films when fibers were oriented normally to the sliding interface. Randall et al. [62, 63] studied the correlation between transfer film hardness and wear of several ceramic coatings. They found the most wear resistant coating had transfer films that were ~30% harder than either the sliding counterpart suggesting a possible metal alloying effect. Ye [64] measured transfer film hardness and modulus of the low-wear alumina-PTFE system using AFM-based nanoindentation and found that the transfer film hardens and stiffens during sliding. However, Chang et al. [65] showed that the indentation hardness of thin (200 nm) epoxy transfer films had been artificially increased by contributions from the substrate. When transfer films thin and harden, it becomes difficult to differentiate between hardening of the film material and the effect of the substrate [61,62,63].

The substrate effect may be avoided by indenting on the running film, which forms on the sliding surface of the bulk solid lubricant. McCook et al. [18] measured hardness and modulus for running films of low wear PTFE-epoxy composites and found that the hardness and modulus of the worn surface were uniformly reduced. The results suggested that the running film was PTFE rich, which likely contributed to the low measured values of friction coefficient and wear rate. Krick et al. [66] measured the mechanical properties of running films from a low wear alumina-PTFE nanocomposite as a function of sliding distance. The film hardened and stiffened with increased sliding distance and decreased wear rate (Fig. 6.15). They concluded that filler accumulation and PTFE degradation were both likely contributors. Their results indicate that the cohesive strength of the running film increased by >2X. If we assume the transfer film is identical to the mated running film, then the absolute adhesive strength of these transfer films increased by >10X from run-in to steady state based on the results from Fig. 6.11.

Fig. 6.15
figure 15

Average (a) hardness and (b) reduced modulus of worn and unworn PTFE and PTFE–Al2O3 nanocomposites, for indentation contact depths in the range of 100–150 nm. Error bars represent ±1 standard deviation from the mean. (Image reprinted from [66] with permission)

Full size image

6.2.4 Tribochemistry

Tribochemistry is an important factor governing the adhesion and mechanical properties of transfer films. Briscoe [13] first proposed that effective fillers reduce wear by promoting polymer degradation and thereby increasing the adhesive strength of the transfer film. Bahadur and Tabor [22] found that even high quality transfer films did not stick around long after the parent pin was replaced and concluded that transfer film adhesion must be primarily mechanical in nature. Studies on PTFE-based solid lubricants have shown that ultra-low wear sliding is consistently accompanied by small wear debris, thin transfer films, and brown discoloration of both sliding surfaces [7, 9, 10, 19, 27, 29, 45, 66, 67]. This discoloration indicates tribochemical degradation, which is particularly remarkable for PTFE given its unique resistance to chemical attack.

A recent study by Krick et al. [9] provided the most compelling evidence to date that a specific chemistry is required for ultra-low wear sliding of the low wear α phase alumina-PTFE nanocomposite system. As they removed water from the environment, the wear rate of the system increased by two orders of magnitude from 10−7 mm3/Nm to 10−5 mm3/Nm, the latter of which is more typical of PTFE composites. This increase in wear rate in dry environments was accompanied by the loss of the brown discoloration and a transition toward large flaky debris fragments. The inability of this low wear material to achieve low wear in dry environments has been reproduced independently by Pitenis et al. [10] and Khare et al. [53]. Since the environmental constituents have no obvious effects on the mechanical or structural properties of the composite, the most reasonable conclusion is that wear rates increased due to the loss of favorable tribochemistry at the sliding interface; thus, a specific tribochemistry appears necessary for ultra-low wear of this system.

XPS studies of the transfer films of this low wear alumina-PTFE system by Burris et al. [68] revealed an unexpected peak at 288 eV, which they attributed to a tribochemical degradation product. Computational modeling by Onodera et al. [69, 70] showed that end-chain carboxyl groups formed in humid environments and suggested that they help bond PTFE transfer films to metal counterfaces. Experimental infrared spectroscopic (IR) studies by Pitenis et al. [10] and Harris et al. [45] showed that the new tribochemical product was a metal chelate salt of perfluorinated carboxylic acid (Fig. 6.16). Harris et al. [45] proposed a realistic reaction route that involved mechanical rupture of PTFE chains and fibrils, reaction of chain ends with oxygen and then water to form carboxylic acid, and, finally, chelation of the acid at the chain end to metal atoms in the counterface and filler particles. The reaction products increase with sliding distance (Fig. 6.16), which helps explain why these transfer films become more persistent, better adhered, stronger, and more wear resistant with increased sliding distances.

Fig. 6.16
figure 16

Infrared reflectance results from the metal surface after one cycle of sliding (light grey line), 100 k (gray line) and 1 M (black line) cycles. (Reprinted with permission from [27, 45])

Full size image

6.3 Causes and Consequences: A Case Study with a Low Wear Composite

Despite the compelling evidence from Pitenis et al. [10] and Harris et al. [45] that tribochemistry is necessary to anchor transfer films, recent studies of the run-in period of this particular PTFE nanocomposite provides further insights into the timing of favorable tribochemistry and its role in low wear sliding of polymers. Ye et al. [29] showed that the transfer films were compositionally and mechanically identical to the unworn composite and Pitenis et al. [10] found a comparable lack of carboxylates during run-in. Furthermore, Khare et al. [53] showed that, unlike at steady state, the run-in and transition behaviors of this material are completely independent of environment. Thus, the transition to low wear during run-in appears to occur without the benefit of favorable tribochemistry. Ye et al. [29] showed that the wear rate and the transfer film of these low wear alumina-PTFE composites were indistinguishable from those of unfilled PTFE initially, which suggests that the crack arresting role of the filler is negligible initially. With each pass, however, they observed reduced wear rate and thinner transfer films of characteristically smaller debris. Although the transfer films were immediately dislodged throughout run-in, Khare et al. [53] showed that the transfer film had no effect on the wear rate of the polymer during run-in; they did this by interrupting the test and replacing the transfer film with a fresh steel counterface. If anything, the wear rate of the pin was lower when the transfer film was removed from the contact, which demonstrates that the perceived improvements in transfer film quality during the run-in period are consequences of an increasingly wear resistant running surface and not the cause of progressively reduced wear rates. The result suggests that filler accumulation and mechanical conditioning of the running surface (e.g., filler accumulation, polymer orientation, polymer crazing) are responsible for the transition from large debris and high wear to small debris and low wear.

During the run-in phase, wear continues to decrease until effectively stopping at a critical point Ye et al. [29] defined as the transition. The system wear rate is well below 10−6 mm3/Nm and Ye et al. [29] report the first compositional evidence of polymer degradation at this point. Interestingly, Khare et al. [53] observed indistinguishable run-in and transition features when they removed ambient moisture to preclude favorable tribochemistry. This is not entirely surprising given that the run-in is independent of tribochemistry [10, 29] and suggests that low wear of this system does not require favorable tribochemistry, at least initially. These results suggest that mechanical debris size regulation is the first step toward low wear in this particular system. These debris fragments can become so small that they automatically adhere to produce stable low energy surfaces. Mechanical stability must precede tribochemical stability because the products of improbable chemical events take time to accumulate [10]. As low wear sliding continues, adhered debris fragments grow by scavenging material from the pin (Fig. 6.8). As debris fragments grow, adhesion energy becomes less competitive with elastic energy until they destabilized at a critical thickness Ye et al. [6] estimates to be on the order of ~1 μm for PTFE. However, if the humidity is sufficient, tribochemical reactions begin to link the polymer chemically to the counterface, thus increasing the critical thickness for instability as the film grows. This, we propose, explains why run-in is independent of environment, why low wear was achievable but unstable in dry nitrogen [53], and why the system can sustain low wear indefinitely in humid environments [9]. The cross-linking effect of tribochemistry is consistent with the substantial increases in adhesion strength [6, 29, 58], hardness [64, 66], and modulus [64, 66] for this system from run-in to steady state.

6.4 Summary

This chapter reviews the latest methods to quantify the topographical, adhesive, mechanical, and chemical properties of polymeric transfer films. Topographical features of interest, including thickness, area fraction, and free-space length, have been quantified primarily with profilometry and microscopy. Both thickness and area fraction tend to correlate strongly with wear rate for specific systems but less so across systems. Based on limited studies, the free-space length appears to be the best of the three for predicting wear rates in general. It is interesting to note that the free-space length is most closely related to the visual cues that have motivated use of the terms “uniform,” “coherent,” and “continuous” [31, 38, 71,72,73,74]. Nonetheless, despite consistent evidence that transfer film properties correlate strongly with wear rate, no universal function is likely to emerge.

Although peel and scratch tests are typical for qualifying coating adhesion, such measurements are confounded by the discontinuous nature of transfer films. The tensile cracking test eliminates the need for a second interface and is not affected by discontinuity in the transfer film. Additionally, it provides a direct measure of adhesion strength per unit cohesion strength; since adhesive failure is far more detrimental to tribopolymer wear than cohesive failure, this property is perhaps the most direct measure of transfer film adhesion performance. Lastly, although a wear test appears to be the most direct measure of tenacity, we have shown that the measured wear rates of transfer films can vary by many orders of magnitude depending on the surface properties of the probe material used; caution is necessary when interpreting the results of transfer film wear tests in an absolute sense for this reason.

Hardness and modulus are the most common mechanical metrics of transfer film performance and both are typically quantified with micro and nano-indentation. Because transfer films are thin and soft, the results are confounded by effects from the counterface, which effectively converts an inherently quantitative measurement into a qualitative assessment. Some have quantified the mechanical properties of the running film on the soft polymer surface as a way to address this issue of transfer film characterization. However, caution must be exercised when using running films as surrogates for transfer films because there is some evidence that the films are compositionally distinct despite developing at the same tribological interface.

Finally, we have used a particularly well-studied alumina-PTFE system as a case study for wear reduction. Based on the methods described throughout, its transfer films are thin, well-covered, persistent, well-adhered, chemically distinct, and mechanically stronger than the unworn composite. These changes are strongly associated with a favorable tribochemical reaction; polymer chains break mechanically during sliding, react with gaseous oxygen and water from the environment to form carboxylic acid, then chelate to metal surfaces to strengthen the film and bond it to the counterface. The end result is a stable and protective film that minimizes the wear rate of the system.

The mechanisms of wear mitigation in this system, particularly the tribochemical aspects, are likely unique given that the four order of magnitude wear reducing effect of fillers appears to be limited to fluoropolymers [2, 67]. Nonetheless, it is interesting that transfer film uniformity, thickness, coverage, bond strengthening, chemical degradation, and mechanical strengthening are recurring themes in the tribopolymer literature. Although the exact mechanisms are likely system-specific, the quantitative methods described here may be used broadly to expose common trends and elucidate general causation relationships. Ultimately, knowledge of these causation relationships will be critical to materials discovery by design rather than the more traditional trial and error-based approaches.