[CPMD-list] What are classic energies for CP MD and PI MD?

[Lan-Feng Yuan (lyuan@Princeton.EDU)]

----- Original Message -----
From: Axel Kohlmeyer <akohlmey at vitae.cmm.upenn.edu>
Date: Wednesday, December 7, 2005 10:35 am
Subject: Re: [CPMD-list] What are classic energies for CP MD and PI MD?
To: "Lan-Feng Yuan (lyuan at Princeton.EDU)" <lyuan at Princeton.EDU>

> On Tue, 6 Dec 2005, Lan-Feng Yuan (lyuan at Princeton.EDU) wrote:
> 
> LY> > 
> 
> LY>   Yes, I am doing symmetry 0. 3N-3? Not 3N-6? Because 
> translation is
> 
> 3N-3 is what is used in DETDOF to compute the temperature.
> 
> LY> frozen, but rotation not? Anyway, it's still not quite right. For
> LY> example, for a H2O molecule at 150 K, I get:
> 
> 
> LY> 
> LY>                               MEAN VALUE       +/-  RMS DEVIATION
> LY>                                      <x>     [<x^2>-<x>^2]**(1/2)
> LY>  ELECTRON KINETIC ENERGY    0.382262E-04             0.289944E-04
> LY>  IONIC TEMPERATURE                145.36                   110.93
> LY>  DENSITY FUNCTIONAL ENERGY    -17.184344             0.132918E-03
> LY>  CLASSICAL ENERGY             -17.180351             0.289125E-04
> LY>  CONSERVED ENERGY             -17.180313             0.197829E-05
> LY> 
> 
> LY>   So ECLASSIC - EKS = 0.003993 Har = 0.1087 eV = 8.40 k_B*T 
> (k_B*T =
> LY> 0.01293 eV). I can't figure out how this 8.4 comes ...
> 
> hmm, i currently have no time to check this, but you may
> have to factor in the DOFs from the nose-hoover chains, 
> if you used them. 
> 
> axel.
> 
  Well, I certainly use Nose-Hoover chains, otherwise the temperature cannot be fixed around 150 K (although its fluctuation is huge because the system is so small). In my understanding, the instant temperature is simply a constant times the total nuclear kinetic energy, and the constant has to do degree of freedoms (3N-3 here). The tricky thing is what's ECLASSIC? Just sum of EKS and nuclear KE? Or thermostat KE also included? If thermostat energy is also there, then it has not much use in calculating other properties.

  Thanks! Best regards,

  Lan-Feng