Can I calculate the internal energy from NPT simulations or does the thermostat complicate things?

[Ant]

I would be very grateful if someone could help me out with what is probably a newbie question.  I have been unable to find the answer in text books or online, however.  

I have existing NPT simulation results that I used to calculate density.  My question is, can I also use them to calculate the internal energy?  

I know that for an NPT simulation, the "conserved quantity" (column 6 of the .ener file) is not internal/total energy.  It is the energy of the system + the energy of the thermostat + the energy of the barostat.  However, is it a reasonable approach to calculate the internal energy by summing the potential and kinetic energies (columns 3 and 5, .ener file)?  Or does the added energy from the thermostat somehow mess things up?  

I found a paper where someone first ran CP2K NPT simulations and then used the cell volumes to run NVT simulations to get the internal energy (I assume because the fluctuations in V would be quite large for NPT and might lead to error in internal energy--my internal energy averages from my NPT simulations seem quite stable though).  So someone has used a non-NVE approach.  However, I have a niggling feeling that the thermostat might be doing something I don't understand.  Thank you.

[Lucas Lodeiro]

Hi!

I never do it using AIMD, but for classical dynamics, the internal energy (E or H) in well behaved NPT simulations gives you a very good approach. At least in my case, for classical NPT simulations, I could get values for gas solutions and other types of systems with an accuracy near chemical precision, respect to experimental data. I expect the same for AIMD, but how the AIMD are shorter (too short compared to classical) I do not know if the ergodicity is fulfilled, and the simulation is in a well behaved equilibrium. I guess the NPT is not the best ensemble to do that in AIMD due to the time coupling between the system and the barosthat is too long compared to simulation length... NVT ensemble could be better.

Regards!

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/83e65c46-ed6e-4816-a1fe-595abdca10a0o%40googlegroups.com.