Question

[maaren]

I was wondering if you are familiar with this paper:

Two-phase thermodynamic model for computing entropies of liquids reanalyzed
J. Chem. Phys. 147, 194505 (2017); https://doi.org/10.1063/1.5001798

They have identified three issues with the orginal 2PT model.

[mcaroba]

Thanks for bringing this paper to my attention, I wasn't aware of it (I haven't actively worked on the code too much in the past 1.5 years). I just had a quick look, I'll try to have a more detailed read in the next few days.

I had already noticed some of these issues. The problem with too large gas DoS was quite obvious. I am dealing with it since the first version of the code by limiting the gas DoS to be at maximum the total DoS, whenever the code identifies the issue. The discussion about the different formulas for excess entropy can be found in McQuarrie's book, as one of the excersices, if I remember correctly.You can see the discussion in these notes I wrote for a course at my university: https://mycourses.aalto.fi/mod/resource/view.php?id=335027 (pages 29 and 30). I would need to check the implementation and original argumentation to see whether this is a problem (and how big a problem it is).

The issue with the fluidicity is actually something I've given some thought to. The current definition, and also the proposed solution with a positive exponent, cannot be right. This is because the asymptotic behavior of the gas-like DoFs as one approaches the diffusivity of an equivalent hard sphere gas at zero pressure must be horizontal.

One thing that I found very useful is to normalize the DoS to the expected number of degrees of freedom. This improves convergence by up to one order of magnitude. This discussion can be found in our 2016 JCP paper: https://aip.scitation.org/doi/figure/10.1063/1.4973001 (discussion surrounding fig. 3).