This study demonstrated robust vibration-activated lubricity; to our knowledge, this is the first such demonstration. Subjecting microscale tribological contacts directly to QCM oscillations reduced friction coefficients to ~ 0.01–0.05 regardless of load, sliding speed, probe size, substrate material, instrument, and cantilever stiffness. This is in contrast to prior observations, which report vibration-activated lubricity as being limited to relatively narrow sets of experimental conditions. Specifically, previous studies show the loss of vibration-activated lubricity when the oscillation frequency deviated more than several-fold from the resonant frequency of the instrument or when the oscillation speed was less than the sliding speed. For example, Riedo et al. observed drastically reduced friction but only when the oscillation frequency was comparable to the resonant frequency of the instrument and only at relatively slow sliding speeds (at 8 µm/s but not at 150 µm/s) [8]. Similarly, Jeon et al. found that friction increased by at an order of magnitude or more when the oscillation frequency was much less than or greater than the resonant frequency of their system (~ 20 kHz) [30]. We detected no such limitations from instrument resonance or sliding speed. We observed vibration-activated lubricity at sliding speeds between 5 and 1,000 µm/s using a constant 5 MHz oscillation frequency with two instruments whose resonant frequencies were 10 to 10,000-fold smaller (~ 525 kHz and ~ 200 Hz). This is in direct contrast to previous studies, which consistently show the disruption of vibration-activated lubricity at very high oscillation frequency (relative to instrument resonance) [8, 9, 30].
We attribute these unusual findings to two key features of our experiments. The first is that oscillations of prescribed frequency were applied directly to the substrate. In most other studies, the oscillator had to be coupled mechanically to the sample through an intervening spring/mass system. In our case, the substrate oscillated at the prescribed QCM frequency without any requirement for mechanical coupling to the instrument. Additionally, because the oscillation speeds were likely between 10 and 100 mm/s, we were able to maintain vibration-activated lubricity up to 1 mm/s, the maximum sliding speed of either instrument. In effect, this approach decoupled the vibration-activated lubricity phenomenon from the instrument and other experimental variables.
As our sharp probe experiments (Fig. 3A and 3B) demonstrate, direct oscillation of the substrate doesn’t guarantee robust vibration-activated lubricity. Jeon et al. made similar observations during direct oscillation experiments with sharp AFM probes [30]. Like us, they found no friction reducing effect from substrate oscillations when the oscillation frequency was much larger than the resonant frequency of the cantilever. This observation, that robust vibration-activated lubricity breaks down below some critical length scale conflicts with the hypothesis that it is more difficult to achieve in multi-asperity contacts compared to single-asperity contacts [38].
The key difference, we propose, is the relative ease with which nanoscale and microscale probes can track the oscillating substrate. Consider our sharp AFM probe, which had a radius of ~ 8 nm and an estimated contact stiffness of ~ 100 N/m. Assuming 5 MHz oscillations on a 2 mm long track, the inertial force (tip mass times peak acceleration ~ 1 fN) and spring force (100 nN) were below the friction force (~ 1µN = 0.2·5 µN). Thus, we expect friction to have prevented the interfacial slip necessary to cause vibration-activated lubricity. Jeon et al. [30] restored vibration-activated lubricity by oscillating the substrate near cantilever resonance; this, we propose, activates large tip displacements that break the frictional contact. This would help explain why nanoscale examples of vibration-activated lubricity have, thus far, been limited to oscillation frequencies in the vicinity of resonance [8, 9, 30].
By contrast, consider the 50 µm diameter probe. In this case, the inertial force of the oscillating probe alone would be ~ 500 µN (negligible spring force). Assuming a 0.2 friction coefficient, we would expect the probe to slip at normal loads below 2.5 mN and stick to the oscillating substrate at greater loads; remarkably (and likely coincidentally), this anticipated transition is equal to the observed transition (Fig. 4B). Increasing the probe diameter to 100 µm increases the expected critical load to 18 mN, well beyond our maximum load of 5 mN. The critical load for the 1.5 mm diameter probe is expected to approach 6 N. Multi-asperity contacts may impede vibration-activated lubricity when mechanical coupling is needed; this study shows that multi-asperity contacts promote vibration-activated lubricity when the need for mechanical coupling is eliminated.
By eliminating the dependence on mechanical coupling, we were also able to overcome the limitations on sliding speed previously observed. In the study by Riedo et al. [8], vibration-activated lubricity vanished when they increased sliding speeds from 8 µm/s to 150 µm/s. In that case, the maximum speed was limited by the oscillation speed, which was limited by the resonance characteristics of the instrument. Here, we had no such limitations from instrument resonant frequency, which was as low as 200 Hz. In this case, we were limited by the maximum sliding speed of the instrument, but beyond that we were only limited by the speed of the oscillator, which was likely between 10 and 100 mm/s.
Finally, our results provide direct experimental evidence that normal ‘hopping’ was not a significant contributor to the friction-reducing effect of oscillation. The best evidence prior to this study was interferometry data showing minimal interfacial dilatation during oscillation [23]. We were able to maintain interfacial tension and a small but positive friction force during QCM oscillation. Thus, while it is clear that hopping does have a significant friction reducing effect, it is also true that purely lateral oscillations can provide similar effects without any need for changes in normal force. Thus, vibration-activated lubricity can occur due to normal hopping or due to purely lateral oscillations that reduce or overcome the barriers to slip.
-
Closing Remarks