The problem of basis set selection for molecular dynamics of Mg2+ aqueous solutions

xyfeng

unread,

Apr 30, 2024, 5:32:20 AM4/30/24

to cp2k

Dear all, I am trying to use cp2k for the molecular dynamics of Mg2+ solution systems, but the calculated forces of Mg are too large and not very good.I suspect that it's because the basis set I chose (DVZP-MOLOPT-SR-GTH) doesn't describe larger systems well.

So, Is there any recommended basis sets suitable for magnesium ion solution system that can balance accuracy and cost?

Marcella Iannuzzi

unread,

Apr 30, 2024, 8:08:09 AM4/30/24

to cp2k

Hi ..

The provided information are very limited, but, if you are using GPW,

it is possible that the PW cutoff is too low for Mg.

Regards

Marcella

xyfeng

unread,

Apr 30, 2024, 10:05:03 AM4/30/24

to cp2k

Thanks for your response! This is my input file. The force calculated is about twice larger than that of vasp results(PAW-PBE)

input (3).inp

Marcella Iannuzzi

unread,

Apr 30, 2024, 10:19:01 AM4/30/24

to cp2k

Hi

The basis set for Mg that you are using contains an exponent of 30, for which the cutoff of 400 Ry is much too small.

In this case it is probably better to switch to GAPW.

Anyway I would suggest to use PP and BS in the UZH libraries, which are currently the best performing MOLOPT settings.