Fixing a bond in DFT

[Wei]

Dear Pankaj, Ari and Marco,

I'd also like to fix a bond during a MD (DFT) run (not the absolute atomic 
positions). I tried to specify CP2K_INPUT / MOTION / CONSTRAINT / 
COLLECTIVE and CP2K_INPUT / FORCE_EVAL / SUBSYS / COLVAR by following an 
expample from regtest /Fist/regtest-12/water_3_dist_2.inp.

However, the calculation always crashes with an error:

 *
 *** 12:16:30 ERRORL2 in topology_input:read_constraints_section processor  
***
 ***      0  err=-300  condition FAILED at line 561                         
***
 *

 ===== Routine Calling Stack =====

            3 topology_control
            2 quickstep_create_force_env
            1 CP2K
 CP2K| condition FAILED at line 561
 CP2K| Abnormal program termination, stopped by process number 0


Could you please give me some hint on this issue? I put my input file 
below. Thanks in advance!

Best Regards

Wei

input 

&FORCE_EVAL
  METHOD QUICKSTEP
  &DFT
    BASIS_SET_FILE_NAME ./GST.qbs
    POTENTIAL_FILE_NAME ./GTH_POTENTIALS
    &MGRID
      CUTOFF 300
    &END MGRID
    &QS
      EPS_DEFAULT 1.0e-12
      EXTRAPOLATION ASPC
      EXTRAPOLATION_ORDER 0
      MAP_CONSISTENT
    &END QS
    &SCF
      EPS_SCF 1.0e-7
      MAX_SCF 40
      MAX_SCF_HIST 2
      &OT ON
        MINIMIZER DIIS
        PRECONDITIONER FULL_KINETIC
        ENERGY_GAP 0.002
        STEPSIZE 0.0875
      &END OT
      &OUTER_SCF ON
        MAX_SCF 800
        EPS_SCF 1.0e-7
      &END OUTER_SCF
      SCF_GUESS RANDOM
#     SCF_GUESS HISTORY_RESTART
    &END SCF
    &XC
      &XC_FUNCTIONAL PBE
      &END XC_FUNCTIONAL
    &END XC
  &END DFT
  &SUBSYS
    &CELL
      ABC   18.62463  18.62463  18.62463
    &END CELL
    &COORD
 Ge 10.0 10.0 10.0  bond
 Te 13.0 10.0 10.0  bond
    &END COORD
    &COLVAR
      &DISTANCE
        ATOMS 1 2
      &END DISTANCE
    &END COLVAR
    &KIND Ge
      BASIS_SET DZVP-MOLOPT-SR-GTH-q4
      POTENTIAL GTH-PBE-q4
    &END KIND
    &KIND Te
      BASIS_SET DZVP-MOLOPT-SR-GTH-q6
      POTENTIAL GTH-PBE-q6
    &END KIND
  &END SUBSYS
&END FORCE_EVAL
&GLOBAL
  PREFERRED_FFT_LIBRARY FFTW
  PRINT_LEVEL LOW
  PROJECT GeTe.md
  RUN_TYPE MD
&END GLOBAL

&MOTION
  &MD
    ENSEMBLE LANGEVIN
    STEPS    20
    TEMPERATURE  1100
    TIMESTEP 2.0
    &LANGEVIN
      GAMMA 4.0E-3
      NOISY_GAMMA 3.5E-5
    &END LANGEVIN
  &END MD
  &CONSTRAINT
    &COLLECTIVE
      MOLNAME  bond
      COLVAR 1
    &END COLLECTIVE
    &COLLECTIVE
      MOLNAME  bond
      COLVAR 2
    &END COLLECTIVE
  &END CONSTRAINT
&END MOTION

On Friday, May 30, 2014 2:18:07 PM UTC+2, Pankaj Mishra wrote:
>
> Dear Ari and Marco,
>
> Thank you so much. It was helpful.
>
> with regards--
>
> Pankaj
>
> On Saturday, 24 May 2014 13:00:05 UTC+2, Ari Paavo Seitsonen wrote:
>>
>> Dear Pankaj Mishra,
>>
>>   You can do it during molecular dynamics using section
>>
>> CP2K_INPUT / MOTION / CONSTRAINT / COLLECTIVE
>>
>> and specifying it in
>>
>> CP2K_INPUT / FORCE_EVAL / SUBSYS / COLVAR
>>
>>   During geometry optimisation ("GEO_OPT") it is not possible to 
>> constraint a bond exactly, but one can work with RESTRAIN and make the 
>> condition be _almost_ fulfilled.
>>
>>   The FIXED_ATOMS works only if the bond to be fixed is along one 
>> Cartesian axis and the specific coordinate along the direction of the bond 
>> can be fixed; this is usually not possible, but it depends on your system.
>>
>>     Greetings from Zurich,
>>
>>        apsi
>>
>> Le vendredi 23 mai 2014 11:17:44 UTC+2, Pankaj Mishra a écrit :
>>>
>>> Dear All,
>>>
>>> I am wondering whether it is possible to fix  a particular bond distance 
>>> in molecule while using quickstep. 
>>> For example: I have a H2O molecule and i want to fix 1 OH bond length to 
>>> some particular value. Then treat the molecule with DFT (rather than MM 
>>> with force field). 
>>>
>>> Please suggest if you have some idea.
>>>
>>> Thank you so much
>>>
>>>
>>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20141113/1f81fce8/attachment.htm>