Figure 1

Main idea of this experiment. A drop of fluid (OMCTS) containing nonadsorbing fluorescent dye (coumarin 153) was placed between crossed cylinders of mica, and load was applied causing formation of a flattened contact ($\approx 50\mu \mathrm{m}$ diameter) larger than the laser spot ($\approx 0.5\mu \mathrm{m}$ diameter). The load was 2–3 MPa; this is the force, squeezing the solid surfaces together, divided by contact area. A focused femtosecond laser caused two-photon excitation. The laser power was $\approx 1\mathrm{m}\mathrm{W}$ at the sample and data were collected at 5–50 kHz sampling frequency. Fluorescence intensity fluctuations resulted as dye molecules diffused into and out of the focus spot. At an average instant, the signal resulted from one dye molecule.Reuse & Permissions

Figure 2

Illustrative fluorescence intensity-intensity autocorrelation functions, $g_N(\tau )$, normalized to unity at short time and plotted against logarithmic time lag, $\tau$. To cleave mica sheets, the conventional method was employed (see text). The thickness of the liquid film was $3.0\pm 0.4\mathrm{n}\mathrm{m}$ relative to air, corresponding to $\approx 3–4$ molecular layers. The curves compare diffusion at rest in the unconfined bulk (circles; $D=180\mu {\mathrm{m}}^2/\mathrm{s}$) and at two radial positions (“$a$”) within the contact zone of radius $r\approx 25\mu \mathrm{m}$; $a/r=0.7$ (triangles) and $a/r=0.5$ (squares). The center is at $a=0$. These data are taken at rest (open symbols) and while sliding (filled symbols) at shear rates ${10}^4$ and ${10}^2{sec}^{-1}$, respectively. Sliding was performed at 1–256 Hz such that it was unidirectional for half the period then reversed direction, and so on repetitively; for the data shown here, it was 256 and 32 Hz, respectively. Lines through the data for the bulk are fits to a single diffusion process. As discussed in the text, diffusion during sliding was 2–5 times faster than at rest.Reuse & Permissions

Figure 3

Similar to Fig. 2 except using a different protocol to cleave mica sheets, the method of Frantz and Salmeron [16]. As in Fig. 2, the thickness of the liquid film was $3.0\pm 0.4\mathrm{n}\mathrm{m}$ relative to air. The radial position within the contact spot was $a/r=0.6$ (see legend of Fig. 2 to identify $a/r$). The data were acquired without sliding (circles) and during sliding at 256 Hz at peak shear rate $2\times {10}^4{s}^{-1}$ (squares). Panels (a) and (b) show results from independent experiments. As discussed in the text, these data are consistent with the conclusions from Fig. 2. Lines through the data show fits to the model of a single effective diffusion coefficient.Reuse & Permissions

Figure 4

Ratio ($D_{\mathrm{r}\mathrm{e}\mathrm{l}}$) of the effective diffusion coefficient during sliding (defined in text) to that at rest, plotted against peak shear rate. The data represent the average of more than 30 experiments, with error bars indicated. The inset shows the low shear part of this data.Reuse & Permissions