[CPMD-list] Energy conservation continued... (fwd)

[Nick Cooper]

Dear all,

Please disregard the message I posted below, as I have found that mu
is in units of energy.time^2 by dimensional analysis and the sigma_i is
just the EKINC in the output file produced by CPMD.

Sorry for wasting your time, though I'm sure posting the question helped
me to find a solution!

Best Regards,

Nick Cooper

---------- Forwarded message ----------
Date: Thu, 23 Oct 2003 10:55:30 +1000 (EST)
From: Nick Cooper <ncooper at physics.usyd.edu.au>
To: cpmd-list at cpmd.org
Subject: Energy conservation continued...

Firstly, thank you very much to those who replied to my initial question
regarding achieving energy conservation at elevated temperatures.

However, having increased the order of the Suzuki integrator and
decomposition ratio using

NOSE PARAMETERS
4 4 4 6 125 4

a constant downwards drift of the "conserved" quantity was observed.

Given this, I suspect that I have set the fictitious kinetic energy for
the electrons (the first number in the NOSE ELECTRONS keyword of CPMD)
too small. Having read "Adiabaticity in first-principles molecular dynamics"
by Blochl and Parrinello: PRB 45, 9413 (1992), I came upon their expression
for E_kin,ad (eqn 6 in the paper), which provides an estimate for the
transition point between two regimes. However, I have a couple of questions
regarding this expession:
1) I assume mu (EMASS in CPMD) is in electron mass units and M proton
mass units?
2) How do I determine the value of the of what's contained in the sigma_i?

Sorry for asking such a basic question but I haven't got anyone to ask
at present, so I was hoping someone here could enlighten me.

Thank you for your time.

Regards,

Nick Cooper