DFTMD for 32H2O cannot conserved

[Nico Holmberg]

Hi,

In NVE simulations, total energy is the conserved quantity not the Fermi 
energy (=the energy of the HOMO). You check if the total energy is 
conserved by looking at the *ener file from your MD simulation. To perform 
constant Fermi-energy simulations, you would need to run grand canonical 
DFT (µVT ensemble) with a variable number of electrons. This not possible 
with the standard DFT module in CP2K, but afaik the linear scaling DFT is 
capable of running such calculations but I have no experience with the 
module. You can find grand canonical DFT support in some other codes, e.g. 
Quantum Espresso (see e.g. this paper 
<http://dx.doi.org/10.1021/acs.jctc.6b01232>).

BR,

Nico

keskiviikko 7. helmikuuta 2018 10.53.49 UTC+2 王锋 kirjoitti:
>
> Hello, I am running NVE ensemble for 32H2O, appendix is my input file. 
> This is a plot of E(fermi)-t, it's not conserved. So could someone tell me 
> the reason?
>
>
> <https://lh3.googleusercontent.com/-44I1YyJUVl4/Wnq9_3WhRUI/AAAAAAAAAB8/XZFZDwCga9g8DF1_mjgvFRYkFhy5v-jVwCLcBGAs/s1600/020AFADCC367CB78DD0DED3058992A83.png>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20180207/29ef553b/attachment.htm>