[CPMD-list] question on the QUENCH[IONS, ELECTRONS, CELL, BO] keyword

[Eduard Schreiner]

Hey Hua,


On Wednesday 04 October 2006 17:05, huawj wrote:
> Dear All，
> 
> I tried the keyword 	
>          QUENCH BO ELECTRON   
> for a (31H2O +HCl) system at 300K(which started from a equilibrated configuration using classical MD and relaxed for 1 ps using CPMD),
> but get confused about the following questions:  
> 
> (1)  From the cpmd-manual, I can read the illustrations for this keyword:
> "The velocities of the ions, wavefunctions or the cell are set to _zero_ at the beginning of the run.
> With the option BO the wavefunctions are converged at the beginning of the MD run.  "
> 
> But why here the resulting EKINC of the first step is not _zero_ , but a positive value instead ?
> 
>        NFI    EKINC   TEMPP           EKS      ECLASSIC          EHAM         DIS    TCPU
>          1  0.00040   302.0    -490.88033    -490.74549    -490.74509   0.486E-04    8.31
>          2  0.00362   285.0    -490.87575    -490.74849    -490.74487   0.189E-03    8.34
>          3  0.01012   300.0    -490.87335    -490.73940    -490.72928   0.412E-03    8.32
>          4  0.01866   280.9    -490.87342    -490.74801    -490.72935   0.727E-03    8.33
>          5  0.02698   300.0    -490.87535    -490.74140    -490.71442   0.112E-02    8.34
>            ......
>         48  0.02502   315.3    -490.88021    -490.73946    -490.71443   0.959E-01    9.44
>         49  0.02513   316.6    -490.88090    -490.73955    -490.71443   0.995E-01    9.39
>         50  0.02529   317.8    -490.88163    -490.73972    -490.71443   0.103E+00    9.45

The EKINC is not zero, because the system is (at this time) not on the BO surface anymore since
one integration step is already performed.

> 
> 
> 
> (2) Why the "quench to the BO surface" step should not take more than 50 steps ?
> 

I understand it more as a rule of thumb. 
In some cases it takes much more than 50 steps to converge the wavefunction, depending on the chosen
convergence criteria.
(see also subsection "9.3 Wavefunction Convergence" in the "Hints and Tricks" section of the manual)

> (3) In the case the running takes more than 50 steps, I adjusted the convergence parameter (CONVERGENCE ORBITALS 2.0E-5), but nothing changed in the output file. Even no statement about the convergence creteria can be found in the output file. Why can this happen?
> 
> (4) Suppose I have converged the wavefunctions with the creteria  2.0E-5 in less than 50 steps, is there anything else should I do before
> I start a simulation of trajectory using a Nose-Hoover chain thermostat? Should I reoptimize the wavefunction to get a much higher convergence creteria, e.g. 1.0E-6 ?  

I think that is not neccessary. But I would first do an MD with 

NOSE IONS MASSIVE

before using the chain to ensure proper equilibration.

For the usage of the Nose-Hoover chain thermostat check also 
"9.6.1 Choosing the Nos´e-Hoover chain thermostat parameters"
in the manual.

Apart from this, is your time step not too large?


> 
> Thank you very much. 
> 
> ==input file for quenching the electrons to the BO surface======
> 
> &CPMD
> MOLECULAR DYNAMICS CP
> RESTART WAVEFUNCTION  COORDINATES VELOCITIES
> QUENCH BO ELECTRONS
> TRAJECTORY XYZ
> TEMPERATURE
>  300.0
> TEMPCONTROL IONS
>  300.0 30.0
> TIMESTEP
> 7.0
> MAXSTEP
> 50
> EMASS
> 1000.0
> MEMORY BIG
> &END
> 
> 
> 
> 
> 
> 
> 
> 　　
> Best regards, 				
> 
> Weijie Hua
> Institute of Theoretical and Computational Chemistry,
> Key Laboratory of Mesoscopic Chemistry,
> School of Chemistry and Chemical Engineering,
> Nanjing University,Nanjing 210093,
> P.R.China
> Email:huawj at itcc.nju.edu.cn
> 2006-10-04
> _______________________________________________
> CPMD-list mailing list
> CPMD-list at cpmd.org
> http://cpmd.org/mailman/listinfo/cpmd-list
> 

-- 

--

=============================================================================
Eduard Schreiner                     e-mail: eduard.schreiner at theochem.rub.de
Lehrstuhl fuer Theoretische Chemie          Phone: ++49 (0)234/32-22121
Ruhr-Universitaet Bochum - NC 03/52         Fax:   ++49 (0)234/32-14045
D-44780 Bochum                              http://www.theochem.rub.de
=============================================================================