[CPMD-list] question on thermal energy tranfer

[Tengfei Luo]

Dear All:
I have some question on the result of thermal energy equilibration. Experts, could you please look at my inputs to see if there is anything wrong?

This is what I have done:
1. I divided the 32 silicon atoms and 32 germanium atoms into to two group. Then, I set these two parts in different temperatures using NOSE CAFES, like the following:
 INPUT:
---------------------------------------------------------
&CPMD
  MOLECULAR DYNAMICS CP
  RESTART WAVEFUNCTION COORDINATES VELOCITIES RESTART
  TRAJECTORY XYZ
  NOSE IONS CAFES
    2
    1  32  100.0 1500
    33 64  50.0 900
  NOSE ELECTRONS 
    0.00012 7000 
  MAXSTEP
   10000
  SUBTRACT COMVEL ROTVEL
   3000
  TIMESTEP
   4.0
  STORE
   1000
&END

&DFT
  FUNCTIONAL LDA
&END

&SYSTEM
  ANGSTROM
  SYMMETRY
  6
  CELL
  11.316 1.0  1.0  0  0  0
  CUTOFF
  45.0
&END

&ATOMS
*SI_MT_LDA KLEINMAN-BYLANDER 
   LMAX=D
   32
   2.715500041283      5.431000082566      2.715500041283
`````
   9.504250144491      9.504250144491      1.357750020642

*Ge_SGS_LDA.psp  KLEINMAN-BYLANDER 
   LMAX=D
   32
  2.828999962063      5.657999924127      8.486999886190
`````
  9.901499867222      9.901499867222      7.072499905159
&END
-----------------------------------------------------------------------

2).After some thousands of steps, I then, erase all the thermostat, and let the two parts equilibrate by themselves:
INPUT:
-----------------------------------------------------------------------
&CPMD
  MOLECULAR DYNAMICS CP
  RESTART WAVEFUNCTION COORDINATES VELOCITIES RESTART
  TRAJECTORY XYZ
  MAXSTEP
   10000
  SUBTRACT COMVEL ROTVEL
   3000
  TIMESTEP
   4.0
  STORE
   2000
&END
``````
-----------------------------------------------------------------------

But I got a result very strange: the temperature of the two materials doesn't go a uniform temperature after the thermostates are erased no matter how many steps I run. I cannot explain this phenomena physically. Why the temperatures don't go to a same temperature, but there is indeed a little temperature drop of high temperature side and a little raise in the low temperature side. BTW, When I used silicon in both group and germanium in both group, the temperature get to a uniform temperature after some thoudands of steps. (I wrote a little code to process the trajectory file to evaluate the temperature of each part.) 

I would like to provide any further information if it's needed!

Thank you in advance!

Tengfei
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://cpmd.org/pipermail/cpmd-list/attachments/20070407/b85ce68f/attachment.html