search

usnistgov / COSMOSAC

A Benchmark Implementation of COSMO-SAC
MIT License
51 stars 29 forks source link

Issues with getting cosmo files from DMOL3 #10

Open drdrjacobs opened 3 years ago

drdrjacobs commented 3 years ago

It looks like this COSMO-SAC implementation is based on using cosmo files from DMOL3. This raises two issues:

  1. DMOL3 is a commercial package and so this severely limits the usefulness of this implementation. Is there any way to use an open source electronic structure package to generate compatible cosmo files?

  2. Even if I wanted to purchase DMOL3, the product seems to have disappeared. It looks like it used to be a product of Accelrys which is now known as Biovia. However, I can't really find any info about it on their website. It seems to have just disappeared from the internet?

ianhbell commented 3 years ago
  1. Yup, this is an issue, but it is/was what we had to work with. Please see the limitations on the .cosmo files we provided
  2. Indeed, but you would then need to reparameterize the model; it is rather sensitive to the precise form of the sigma profiles. These folks have made progress along the lines of what you mention: https://github.com/lvpp/sigma