search

lucpaoli / SAFT_ML

2 stars 0 forks source link

Investigate parameters predicting by ETH-Z "Language of Molecules" paper #13

Closed lucpaoli closed 10 months ago

lucpaoli commented 11 months ago

SI available in zotero, including parameters CSV. Also here: https://polybox.ethz.ch/index.php/s/j5CaVeeBU6Te2y8

MichaelGadaloff commented 11 months ago

Created dataframe of parameters, and amended with molecular weight (calculated from SMILES).

Methane has a non-zero mu (dipole) parameter, as do the very long-chain alkanes (makes little sense as none of these compounds are polar).

Plotted mepsilon and msigma^3 against alkane molecular weight to observe trend - should be linear (per SAFT-VR modelling of the phase equilibrium of long-chain n-alkanes, 1999). Trend observed not particularly linear.

Used FeOs in python (https://feos-org.github.io/feos/installation.html) to plot methane pressure-density envelope using ETH parameters with PC-SAFT and compare with NIST data. There are large discrepancies between the curves.

lucpaoli commented 10 months ago

Michael has done a really good breakdown. Further analysis shouldn't be necessary. Closing the issue.