search

ZimmermanGroup / pyGSM

Thermal and photochemical reaction path optimization and discovery
MIT License
54 stars 26 forks source link

Support different Hamiltonian for xTB level of theory #28

Open awvwgk opened 3 years ago

awvwgk commented 3 years ago

Seems like GFN2-xTB is hardcoded as method for the xtb interface:

https://github.com/ZimmermanGroup/pyGSM/blob/f1526ee73f8796a0871720857613789886e60886/pygsm/level_of_theories/xtb_lot.py#L41

Being able to select the method can be useful in case GFN2-xTB is not suitable for the system of interest (hydrogen transfers). Also, a possibility to communicate charge and spin to xtb or set a higher electronic temperature would be nice here.

ms860309 commented 3 years ago

Seems like GFN2-xTB is hardcoded as method for the xtb interface:

https://github.com/ZimmermanGroup/pyGSM/blob/f1526ee73f8796a0871720857613789886e60886/pygsm/level_of_theories/xtb_lot.py#L41

Being able to select the method can be useful in case GFN2-xTB is not suitable for the system of interest (hydrogen transfers). Also, a possibility to communicate charge and spin to xtb or set a higher electronic temperature would be nice here.

Hi, So, I add some args to support different xtb hamiltonian, accuracy and electronic temperature. Default value of the xtb hamiltonian is "GFN2-xTB", accuracy is 1.0 and electronic temperature is 300.0K. Is it reasonable for using these default values?

Jianyi

awvwgk commented 3 years ago

Those defaults match the standalone xtb and the Python API default, this should be a reasonable choice.