| We have simultaneously measured the frequency shift and the interaction-induced dissipated energy per cycle above a maximum in a topographic image of KBr(001). The measurements were performed at room temperature in UHV using the 2nd flexural mode of a silicon cantilever (1039369 Hz) at 11 constant amplitudes /A /between 12.8 and 0.51 nm. The dissipative energy rises smoothly above the noise level, then almost saturates as the closest approach distance extends below the attractive force minimum. Whereas the conservative force vs. distance curves extracted from the frequency shift were smooth and coincided apart from noise, the dissipative energy decreased smoothly by only a factor of 2 and became less noisy as the amplitude was reduced from 12.8 to 0.51 nm. This behavior, and the magnitude of the dissipative energy suggest that the dominant cause of dissipation is a force hysteresis loop narrower than peak-to-peak amplitude, but significantly smeared by thermal activation. However, some velocity-dependent damping with a coefficient decaying with increasing distance is also present. Using an amplitude of 285 pm, one thousand approach-retraction curves were recorded, separated by intervals for atom-tracking (Nanonis: AT4). In most cases, one of three distinct dissipative energy vs. distance curves are obtained, whereas the conservative force curves are almost the same. This is consistent with the higher sensitivity of the dissipative energy to the tip apex configuration. In a few cases we observed jumps between curves corresponding to two of those long-lived configurations, accompanied by significant deviations between approach and retraction. The tracked 3D motion of the sample relative to the tip at a frequency shift of -150 Hz exhibits strong correlations between more frequent jumps (strongly reduced by averaging over each tracking interval). This suggests that tip changes occur over a range of time scales. Further implications of those results will be discussed. |
|