search

ramess101 / MBAR_GCMC

3 stars 0 forks source link

Pressures #5

Closed ramess101 closed 5 years ago

ramess101 commented 6 years ago

@msoroush @jpotoff

So far, I have only presented MBAR-GCMC results for coexistence densities. This is because, even with the standard histogram reweighting approach, I am not able to get reliable pressures. I tried to replicate the method reported in your recent Alkyne publication, namely:

image

Specifically, for the hexane example discussed previously, I computed ln(Xi) for a range of mu values with a fixed temperature (set equal to the highest simulated value of 510 K). I then plotted ln(Xi) as a function of N (obtained either from histogram reweighting or MBAR, they are essentially the same):

image

As expected, the slope is almost exactly 1 in this low-density regime. The intercept is used to compute an absolute pressure. However, my pressures are extremely different than those reported for hexane:

image

image

image

image

Note that the deviation between my log(P) values and those of Potoff appears to increase significantly with respect to decreasing temperature. However, on an absolute scale, the deviation in pressure only increases slightly with increasing temperature from 0.7 to 1.4 bar. This suggests that the intercept ln(Xi) vs N is the main reason why my absolute pressures are off, although this will not completely resolve the temperature trend observed for the absolute pressure deviation.

Do you see a fundamental mistake in how I am computing the pressures? Does my intercept plot look similar to what you typically see?

Also, I noticed that your method appears to be slightly different than that reported by Panagiotopoulos in Monte Carlo methods for phase equilibria of fluids

image

From this statement, it appears that they actually performed an additional low density simulation where the system is assumed to behave like an ideal gas. Do you also perform an additional low density simulation?

ramess101 commented 6 years ago

Hi Richard,

We use the lowest temperature gas phase simulation to calculate C constant. In our hist file, it is usually hist3a.dat. You can give it a try to see if it solves your problem.

Sincerely, Mohammad

ramess101 commented 6 years ago

@msoroush @jpotoff

Thanks for the insight. I tried that as well but got a similar result. I will continue to work on this and let you know if I have any additional questions.

ramess101 commented 6 years ago

Hi Richard,

I think you will get your answer by looking at this directory. Look at patch and phase programs.

https://github.com/GOMC-WSU/Workshop/tree/master/GCMC/argon/patch/program Sincerely, Mohammad