Comparison of MiPPE with other force fields

[ramess101]

@jpotoff @msoroush

JCED requires that we compare our results with those for other force fields.

Previously I had only included the TraPPE literature values as validation of our GCMC results for TraPPE. Now I include the Exp-6 model of Errington, LJ model of Mauricio (Vrabec), LJ+quadrupole model of Eckl (Hasse), and the AUA model of Bourasseau et al.

I use lines for GCMC results since these are so close together that it doesn't make sense to use data points.

Unfortunately, I cannot find the tabulated DeltaHv for Exp-6. From their figures, I know the values are pretty similar to MiPPE, but it would be nice to show that here.

From this figure it is clear that no other literature model is far superior to MiPPE.

[jpotoff]

Tabulated data for Errington's exp-6 cyclohexane model is here. From what I remember, it's pretty good. May be close to the Mie potential.

http://kea.princeton.edu/jerring/n-alkanes/new/index.html

[ramess101]

@jpotoff

Thanks, that was where I got the data. For cyclohexane it is here:

http://kea.princeton.edu/jerring/cyclic/ch/index.html

But for some reason they only have T, Psat, rhol, and rhov

[jpotoff]

@ramess101 nevermind, I see you have the densities and pressures already. Let me checj, we may have the data, or could run it if we don't. We recently added exp-6 potentials to GOMC.

[ramess101]

@jpotoff

No need to simulate it ourselves. This is really just to check the box for "comparisons with published MMS data"

[ramess101]

@jpotoff

From this figure we already know it is pretty good (diamonds are cyclohexane):

It would just be nice to quantify this in the figure

[msoroush]

It would just be nice to quantify this in the figure

I can do that using digiplot https://apps.automeris.io/wpd/

[ramess101]

@msoroush

Give it a try! I have used datathief for such purposes, but I am hesitant to purport that these are "their" data when there is some inherent uncertainty in pulling data from a figure. I guess I would be OK with it if we can get smooth looking deviations that are positive by a percent or so.

[msoroush]

T (K), dHV (kJ/mol) 341.01271202812876, 31.409450948572317 360.9903017657741, 30.13098411962444 380.73142459719486, 29.06363741741046 400.7090143348403, 27.785170588462588 420.68351300181604, 26.42231753023454 440.6502839921178, 24.848498898806078 460.36822379351656, 23.148255476990848 480.5621884780341, 21.11015803098799 500.4949576909703, 18.608090877477693 520.3983617325451, 15.304354545805815

[ramess101]

@msoroush

That was remarkably fast (datathief is much more tedious and time consuming). I will see what they look like now.

[msoroush]

@ramess101 Are we going to replace the MiPPE data with the smooth data that I am running (it should be ready by tomorrow morning)? For lambda 14, 18, 20 are you using predicted VLE from TraPPE simulation?

[ramess101]

@msoroush

Good questions.

Are we going to replace the MiPPE data with the smooth data that I am running (it should be ready by tomorrow morning)?

Yes, we are going to replace the predicted MiPPE data with the actual simulation data you are going to have.

For lambda 14, 18, 20 are you using predicted VLE from TraPPE simulation?

These are from direct GCMC simulations. The predicted VLE from TraPPE would look a lot worse than this (at least for 18 and 20 because of low number of effective snapshots).

[ramess101]

@msoroush @jpotoff

I will see what they look like now.

I think the values you pulled from that figure are pretty good. The percent deviation is slightly positive, smoothly varying, and just a few percent. Consistent with the figure. Thanks Mohammad!

[msoroush]

I think the values you pulled from that figure are pretty good. The percent deviation is slightly positive, smoothly varying, and just a few percent. Consistent with the figure. Thanks Mohammad!

No problem 😄

[ramess101]

@jpotoff @msoroush

I have now included the uncertainties from the literature data points except for Exp-6. Although I have the Exp-6 uncertainties (at least for rhol, Psat, and rhov) I am not sure if I should plot these uncertainties. Since I use a line to depict Exp-6 and our GCMC results, I should have a continuous line for the uncertainties. But this would make the plot very difficult to manage. Any suggestions?

We could just include a single representative error bar for each line. Or we could report the uncertainties in the caption. Or we could include a continuous uncertainty for MiPPE, TraPPE and Exp-6 but not the other iterations. That would not be too difficult to look at.

[jpotoff]

I suggest listing a representative or maximum uncertainty in the figure caption. The actual uncertainties for each point will be listed is supporting info with the tabulated data, so we are covered. The figure is getting busy and if they shrink it to single column it could be hard to read.

[ramess101]

True, we do report uncertainties in SI. So I think reporting it in the caption is better. As another idea, we can reference Figure 1 in the manuscript as this gives an estimate of uncertainty.

Rich

On Thursday, December 13, 2018, Jeffrey Potoff notifications@github.com wrote:

I suggest listing a representative or maximum uncertainty in the figure caption. The actual uncertainties for each point will be listed is supporting info with the tabulated data, so we are covered. The figure is getting busy and if they shrink it to single column it could be hard to read.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/ramess101/JCED_FOMMS_Manuscript/issues/15#issuecomment-447143872, or mute the thread https://github.com/notifications/unsubscribe-auth/AWUvhKZf_hwoboOvewy6prheg2s4Fdjdks5u4tTUgaJpZM4ZSayz .