search

grimme-lab / xtb

Semiempirical Extended Tight-Binding Program Package
https://xtb-docs.readthedocs.io/
GNU Lesser General Public License v3.0
595 stars 148 forks source link

Dielectric constants for benzene, toluene #571

Open cneiss opened 2 years ago

cneiss commented 2 years ago

Dear all,

the dielectric constants implemented in xtb of benzene and toluene seem strange to me: both are set to 7.0 which is rather different from the literature values (benzene: 2.274, toluene: 2.38). Is this a reason to worry or even a bug in the program? I am using GFN2 and xtb version 6.4.1 (afa7bdf).

Best, Christian

awvwgk commented 2 years ago

Marcel @MtoLStoN, could you have a look? I recall that for GBSA the benzene/toluene parameters were copied from chloroform, but we redid them properly for ALPB, at least in GFN1-xTB(ALPB) this seems to be the case:

https://github.com/grimme-lab/xtb/blob/863a58a499e0a7768a6793e44079ecbbeaf00925/include/param_alpb_toluene.fh#L1-L2

However, for GFN2-xTB(ALPB) there is still a dielectric constant of 7

https://github.com/grimme-lab/xtb/blob/863a58a499e0a7768a6793e44079ecbbeaf00925/include/param_alpb_toluene.fh#L67-L68

MtoLStoN commented 2 years ago

I recall that for GBSA the benzene/toluene parameters were copied from chloroform, but we redid them properly for ALPB, at least in GFN1-xTB(ALPB) this seems to be the case:

Interesting, this also seems to be the case for GFN-FF (ALPB) https://github.com/grimme-lab/xtb/blob/863a58a499e0a7768a6793e44079ecbbeaf00925/include/param_alpb_toluene.fh#L133-L134 I probably readjusted the parameters for GFN2-xTB(ALPB) after some optimization problems and used the GBSA parameters as templates. I will check this.

steinmig commented 2 years ago

@MtoLStoN @awvwgk Is there any update on this? Is it fine to use toluene with GBSA?

MtoLStoN commented 2 years ago

There is no update yet, but the parameters are still optimized to reproduce the particular solvent (even with the wrong dielectric constant). Changing the dielectric constant will likely require a full reparameterization. This is certainly something I should look at, but given the accuracy range of the ALPB method, it is not something where I would expect a serious impact. So this is not very high on my to-do list at the moment.

GBSA and ALPB will still give reasonable results for toluene and benzene.