In these set of runs, we require the amount of Gallium deposited is exact, in contrast to earlier runs where there was some variance in this amount. We also look at droplet statistics as a function of Gallium thickness.

Here is droplet width as computed by the autocorrelation function:

Droplet width computed via thresholding:

Note the droplet size predicted via autocorrelation function is larger than that predicted by thresholding, especially for thin Ga. Here is an example of the autocorrelation function for such a scenario:

Here is the final frame for this run.

We may also plot the number of droplets as a function of thickness:

The concave-up curves correspond to low temperature. For example, we may consider a thickness of 2.5 monolayers Ga and a flux of 0.4 monolayers per second and temperature at 350 K — a point on a concave-up curve — vs. 500 K — a point on a concave down. This data can also be plotted vs. temperature:

The various plots partition themselves according to Ga thickness. Consider the low temperature case T = 350 K, thickness = 1.5 monolayers Ga, deposition rate = 0.4 monolayers per second: here. There are indeed droplets that have formed, but this contradicts the above plot. This suggests our thresholding is too restrictive. Indeed if we lower the threshold, we obtain the following plot for number of droplets vs. temperature:

This changes, for example, the plot for droplet width:

Here we see the data for high temperature is unreliable. Here is an example for such parameters. Note there is nothing extraordinary about this — the thresholding is just too lenient in determining the number of droplets and cannot distinguish between rough, but flat film and droplets.

Further tweaking threshold yields the following plot for number of droplets vs. temperature:

Droplet width vs. Thickness:

### Like this:

Like Loading...

*Related*

## Leave a Reply