Robert W. Carpick, PhD, University of Pennsylvania
William P. King, PhD, University of Illinois (Urbana-Champaign)
Our research has continued to focus on real-time, nanoscale measurements of friction forces using AFM probes with an integrated solid-state heater. The work builds on the complementary expertise of our two research groups in a new direction in the field of nanotechnology. Thermal AFM probes are having a substantial impact in nanotechnology, for example, on tip-based nanomanufacturing as recently highlighted by Wei et al. (Science, 2010, 328, 5984, 1373-1276), who used thermal probes to locally reduce graphene oxide to graphene. Nanotribology has received extensive attention because of its vital importance for the performance and reliability of nanoscale devices where friction and adhesion are a serious threat to their lifetime (see for example Bhaskaran et al. in Nature Nanotechnology, 2010, 5, 3, 181-185), as well as being such a common physical force that is yet to be well understood at both macroscopic and nanoscopic length scales.
We used custom thermal probes to measure nanoscale friction as a function of temperature (from room temperature to ~250 degrees Celsius) between silicon tips and both silicon and silicon oxide samples. The ability to vary temperature adds a new degree of freedom to such nanoscale friction experiments, which originates from the rapid heating of the cantilever tip (within microseconds) compared to heating and cooling the entire sample (as done in all previous nanotribology literature). Our approach not only increases throughput but, more importantly, allows us to study the process of capillary condensation and its influence on friction forces in real time.
First, we quantitatively investigated the nanotribological behavior of single asperity contacts between the silicon tip and a silicon sample (both with a native oxide) in an ambient environment. In this case, we estimate the tip temperature could be raised to approximately 120 degrees Celsius. We observed a strong dependence whereby friction increased significantly with temperature, and decreased with sliding velocity. This is in contrast to cryogenic measurements where the opposite trends were reported. The behavior is reversible, indicating that the effects are not due to damage or modification of the tip or sample. The temperature and velocity dependences are suppressed and in some cases slightly reversed in a dry nitrogen environment, demonstrating that the effects seen in ambient arise from capillary condensation of water at the tip-substrate interface. As the formation of capillaries has been shown to be a thermally activated process, increasing the tip temperature helped to form the bridges resulting in higher friction forces. As the sliding velocity was increased it became more difficult for the capillary to follow the tip, thus friction decreased with velocity. Above a certain critical velocity, the tip broke free from the capillary bridge completely and friction plateaued or even showed a slight increase with velocity, consistent with the classical Tomlinson model.
Due to the thermal conductivity of silicon, the temperature at the tip-sample contact was estimated to be limited to approximately 120 degrees C. Therefore, in a second set of experiments, we chose silicon oxide as a substrate material. Its lower thermal conductivity prevented the tip's heat from being transported as rapidly into the bulk of the sample, enabling the contact to reach a substantially higher steady-state contact temperature of approximately 250 degrees C.
As with the Si sample, in ambient atmosphere and up to a tip temperature of 90 degrees C, a thermally-activated capillary bridge formed between tip and substrate, increasing friction and adhesion. For higher temperatures, water was driven away from the contact, reducing adhesion and friction forces. Thus, there was no capillary component left, and both friction and adhesion reach a plateau at significantly lower respective values than at room temperature. The capillary's ability to follow the sliding tip is limited, producing a decrease in friction with speed followed by a plateau for tip temperatures below 90 degrees C. For higher heater temperatures, the behavior resembles that of a classical solid-solid contact; friction increases with sliding speed. The overall friction force is much lower at higher temperatures.
The behavior at various fixed temperatures and during real-time temperature variation depends on temperature and speed. With the appropriate choice of speed, it is possible to obtain a decrease or increase in friction in real-time by increasing the temperature. This behavior is suppressed in dry nitrogen where friction always decreases upon heating, independent of sliding speed.
These results demonstrate the fascinating potential of thermal AFM tips to probe phenomena at the nanoscale and also show the ability to control nanoscale friction, adhesion, and capillarity in-situ and in real time. As well, these results have wide impact, as nanoscale water condensation and capillary formation is of major interest in the fields of scanning probe microscopy, nanotribology, and tip-based nanomanufacturing, as well as in the study of confined fluids.
The results have been published in one conference paper and two journal papers.