[Pw_forum] Temperature continuously going down in cp.x run

Uli Aschauer aschauer at princeton.edu
Fri Nov 27 16:32:11 CET 2009


Dear Vega,

Your choice of dt=10 seems a bit large to me for dynamics. You will have to find a good value of dt combined with emass (for which you take the default value of 400) to conserve the energy in your system. To understand what this is all about I would suggest you to read at least the relevant section (2.4.4) herein: http://www.fz-juelich.de/nic-series/Volume1/marx.pdf

For the Nose thermostat you not only set the target temperature but also the frequency. The default may be completely out of the range of the frequencies you try to excite in your system. As Paolo suggested, you will almost certainly have to tune the Nose parameters to get effective thermostatting.

As for the amplitude of the randomization it does not really matter all that much. You will just spend a lot less computer time if you choose a value to give you something close to the temperature you want. You can randomize a few different amplitudes and then run dynamics without thermostat to see what temperature you converge to after a few hundred steps (given your dynamics is setup correctly). This allows you to make sort of a calibration of the temperature as a function of the randomization amplitude for your system.

Best regards,
-uli

_________________________________________________________
Uli Aschauer
Princeton University
Chemistry Department
Frick Laboratory, Room 213
Princeton, NJ 08544
Tel: +1-609-258-0116
Email: aschauer at princeton.edu
_________________________________________________________

On 27.11.2009, at 04:10, vega lew wrote:

> dear all,
> 
> I want to calculate certain heterogeneous reaction on a solid surface.
> First input the initial structure from previous pw.x relaxations.
> Second I fixed ions and let electron damping calculation run; 
> Then un-fixed the ions and let both electron  and ions damping calculations run.
> Third I gave a random motion to the atoms in the system with amprp(1)~(3) = 0.01 
> and then repeated second step to bring the system to the ground state.
> Fourth I added temperature (300K) to the system, using verlet schedule.
> 
> But the temperature continuously going down, from 1000K to 150K during the fisrt 3000 steps.
> I think there might be something wrong with my cp.x calcuation.
> Do you think the amprp(1)~(3) = 0.01 of random motion is too small?
> 
> Here is the input of my dynamics process,
>  &CONTROL                                                                                 
>                        title = 'CP' ,                                        
>                  calculation = 'cp' ,                                                  
>                 restart_mode = 'restart' ,                                           
>                       outdir = '/disk5/nanjing/vega/cp/300/' ,   
>                       wfcdir = '/tmp/' ,                                                  
>                   pseudo_dir = '/home/software/nanjing/vega/espresso-4.0.1/pseudo/' ,     
>                       prefix = 'cp_molecular' ,
>                        nstep = 99999 ,
>                           dt = 10,
>                       iprint = 5,
>                        isave = 150,                                                      
>  /                                                                                        
>  &SYSTEM                                                                                  
>                        ibrav = 8,                                                         
>                    celldm(1) =24.8624,
>                    celldm(2) = 0.8520,
>                    celldm(3) = 1.6964,
>                          nat = 147,                                                       
>                         ntyp = 3,
>                         nr1b = 20,
>                         nr2b = 20, 
>                         nr3b = 20,
>                         nbnd = 630,                                                         
>                    ecutwfc   =   30,
>                    ecutrho   =  300,
>  /                                                                              
>           
>  &ELECTRONS                          
>            electron_dynamics = 'verlet',                         
>  /                                                                             
>  &IONS                                                                                    
>                 ion_dynamics = 'verlet'
>              ion_temperature = 'nose',          
>                        tempw = 300,          
>  /
> ..................
> ..................
> 
> best wishes,
> 
> vega
> 
> -- 
> ==================================================================================
> Vega Lew ( weijia liu)
> Graduate student
> State Key Laboratory of Materials-oriented Chemical Engineering
> College of Chemistry and Chemical Engineering
> Nanjing University of Technology, 210009, Nanjing, Jiangsu, China
> ******************************************************************************************************************
> Email: vegalew at gmail.com
> Office: Room A705, Technical Innovation Building, Xinmofan Road 5#, Nanjing, Jiangsu, China  
> ****************************************************************************************************************** 
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://www.democritos.it/mailman/listinfo/pw_forum

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20091127/7f7dffdb/attachment.html>


More information about the users mailing list