The measurements of the dewetting line are shown in Figure 11. As in Figure 9, the data are plotted vs. T - Ts. The green data of Figure 11(a) was data taken at X3 = 0.395 (this data is offset vertically for clarity). At this concentration Ts > Tw so no dewetting transition was found. The red and blue data of Figure 11 are from four measurements with Ts < Tw. The 3He concentrations used for these measurements were X3 = 0.197 (a), 0.152 (b), 0.110 (c), and 0.0905 (d). These concentrations were measured directly, as the helium was mixed, and the measured bulk separation temperatures agreed with the data of Ref. [32] to within 2 mK. The frequency of the microbalance was allowed to settle at each temperature before the data was taken, as described above for the prewetting and wetting transition measurements.
For each of the
dewetting transition measurements, data taken while cooling is plotted
in blue in Figure 11. Starting at the highest temperatures for each concentration,
the frequency shift increased slightly as the temperature was lowered and
the concentrated layer on top of the dilute films began to thicken. At
a point close to but still distinctly above the bulk phase separation temperature,
there was a sharp break in the slope of the cooling curves. For each measurement
this point is marked in Figure 11 with blue arrows. At this point the films
switched from thick wetting bi-layer films to relatively thin non-wetting
concentrated films. Since the microbalance is sensitive primarily to the
3He content of the films, the frequency shift can either increase
or decrease at this point because a concentrated film could contain more
or less 3He than a slightly thicker bi-layer film. As the experiment
was further cooled the frequency shift increased rapidly for all four concentrations
as the thickness of the concentrated films diverged to their gravitationally
limited value as bulk phase separation was approached. For T < Ts
the substrate was again wet, this time by a concentrated film, and the
frequency remained fairly constant.
The data taken while warming is plotted in red in Figure 11. As the temperature was raised above Ts, the frequency shift decreased as the concentrated liquid films became thinner, leaving the substrate non-wet. As the temperature was raised further the films became quite thin, and only at a temperature considerably warmer than Ts did the bi-layer films re-wet the substrate. As was the case for the wetting and prewetting transitions shown in Figures 8 and 9, the bi-layer films were nucleated at the edges of the cesium and so the transition temperatures measured while warming (marked with red arrows in Figure 11) are the triple point induced dewetting transition temperatures. The transitions measured while cooling (blue arrows) are spinodals.