[CP2K:10764] Why doesn't the Cons Qty include the potential energy of shell (shell model) in NVE ensemble?

[zuohen... at gmail.com]

Dear Matthias,

Yes, that's very helpful. 
I will test the energy drift in NVE ensemble to determine the length of 
time step, before I use the core-shell model.
I actually found the OH- was broken when I use shell model for oxide with 1 
fs time step.
If I used 0.5fs time step, the energy plot would be much flatter even the 
temperature was up to 500K.

I think if the energy drift (per atom) in μeV/ps < 0.1, the tiny drift might be ignored.[1]
But I am not very certain whether it is correct for core-shell model.

Thank you very much~

[1]Steneteg P, Abrikosov I A, Weber V, et al. Wave function extended Lagrangian Born-Oppenheimer molecular dynamics. Physical Review B, 2010, 82(7): 075110.



Best wishes,
Zuoheng




在 2018年9月20日星期四 UTC+2下午4:21:36，Matthias Krack写道：
>
> Dear Zuoheng
>
>  
>
> Though the regression test inputs may serve as a kind of initial input 
> template for an application case, the settings for these tests are rather 
> sloppy allowing for a quick processing in order to check a specific program 
> feature or functionality. The core-shell model requires smaller time steps 
> than rigid-ion force fields because of the usually faster motion of the 
> light shell particles. At ambient conditions 0.5 fs might be a good guess, 
> but smaller and smaller time steps are required for increasing temperatures 
> until the break-down of the approach (“polarisation catastrophe”) at high 
> temperatures will occur. The shell particles should always stay close to 
> their core particles. Core-shell distances of more than 0.5A indicate that 
> the approach is going to break-down.
>
> There is the option to minimise (relax) the core-shell distances (
> MOTION/SHELL_OPT 
> <https://manual.cp2k.org/trunk/CP2K_INPUT/MOTION/SHELL_OPT.html> section) 
> if you specify for MAX_DISTANCE a value greater that zero and that value is 
> surpassed. That is useful for initialisation or structure optimisation 
> runs, but the energy conservation is affected when such relaxation steps 
> are randomly performed during an MD run. MAX_DISTANCE has only an effect if 
> the SHELL_OPT section is explicitly specified.
>
>  
>
> HTH
>
>  
>
> Matthias
>
>  
>
> *From:* cp... at googlegroups.com <javascript:> <cp... at googlegroups.com 
> <javascript:>> *On Behalf Of *zuoh... at gmail.com <javascript:>
> *Sent:* Donnerstag, 20. September 2018 10:10
> *To:* cp2k <cp... at googlegroups.com <javascript:>>
> *Subject:* Re: [CP2K:10764] Why doesn't the Cons Qty include the 
> potential energy of shell (shell model) in NVE ensemble?
>
>  
>
> Dear Matthias,
>
>  
>
> Thank you for your reply.
>
> Could I ask an another question? 
>
> I found core-shell model should use a smaller time step (0.1 - 0.2fs). 
> However the time step of some U2O simulation in cp2k test files is larger 
> (0.5fs even 1fs). 
>
> Is it because you use a heavier shell mass (MASS_FRACTION of O is 0.1) 
> [1] and use a MAX_DISTANCE = 0.5 angstrom?
>
> How do I determine the MAX_DISTANCE in my simulation of O2- and OH- ions 
> with shell model?
>
>  
>
> Thank you very much~
>
>  
>
> Best wishes,
>
> Zuoheng
>
>  
>
> [1] Devynck F, Iannuzzi M, Krack M. Frenkel pair recombinations in UO 2: 
> importance of explicit description of polarizability in core-shell 
> molecular dynamics simulations. Physical Review B, 2012, 85(18): 184103.
>
>
> 在 2018年9月19日星期三 UTC+2下午7:55:28，Matthias Krack写道：
>
> Dear Zuoheng
>
>  
>
> md_ener%epot contains the potential energy of all particles including the 
> shell particles.
>
>  
>
> Best regards
>
>  
>
> Matthias
>
>  
>
> *Von:* cp... at googlegroups.com <cp... at googlegroups.com> *Im Auftrag von *
> zuoh... at gmail.com
> *Gesendet:* Mittwoch, 19. September 2018 17:15
> *An:* cp2k <cp... at googlegroups.com>
> *Betreff:* [CP2K:10755] Why doesn't the Cons Qty include the potential 
> energy of shell (shell model) in NVE ensemble?
>
>  
>
> Dear CP2K developers,
>
> I am sorry to disturb you. I have some questions.
> Recently I want to analyze whether the total energy is converged with 
> core-shell model in NVE ensemble. I didn't use extra thermostat for shell.
> I don't know why the Cons Qty energy = 
> md_ener%ekin+md_ener%epot+md_ener%ekin_shell but not including the 
> potential of shell? Is the Cons Qty the total energy? In NVE ensemble, 
> which energy should be constant? ekin+epot+ekin_shell+epot_shell or 
> ekin+epot+ekin_shell?
>
> Thank you very much~
>
> Best wishes,
> Zuoheng
>
> PS:The whole code of get_econs_nve function is following:
>
> SUBROUTINE get_econs_nve(md_env, md_ener, para_env)
>       TYPE(md_environment_type), POINTER                 :: md_env
>       TYPE(md_ener_type), INTENT(inout)                  :: md_ener
>       TYPE(cp_para_env_type), POINTER                    :: para_env
>
>       CHARACTER(LEN=*), PARAMETER :: routineN = 'get_econs_nve', routineP 
> = moduleN//':'//routineN
>
>       TYPE(force_env_type), POINTER                      :: force_env
>       TYPE(thermostat_type), POINTER                     :: 
> thermostat_coeff, thermostat_shell
>
>       NULLIFY (force_env, thermostat_coeff, thermostat_shell)
>
>       CALL get_md_env(md_env, force_env=force_env, thermostat_coeff=
> thermostat_coeff, &
>                       thermostat_shell=thermostat_shell)
>       md_ener%constant = md_ener%ekin+md_ener%epot+md_ener%ekin_shell
>
>       CALL get_thermostat_energies(thermostat_shell, md_ener%
> thermostat_shell_pot, &
>                                    md_ener%thermostat_shell_kin, para_env)
>       md_ener%constant = md_ener%constant+md_ener%thermostat_shell_kin+
> md_ener%thermostat_shell_pot
>
>    END SUBROUTINE get_econs_nve
>
>  
>
> -- 
> You received this message because you are subscribed to the Google Groups 
> "cp2k" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to cp2k+... at googlegroups.com.
> To post to this group, send email to cp... at googlegroups.com.
> Visit this group at https://groups.google.com/group/cp2k.
> For more options, visit https://groups.google.com/d/optout.
>
> -- 
> You received this message because you are subscribed to the Google Groups 
> "cp2k" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to cp2k+... at googlegroups.com <javascript:>.
> To post to this group, send email to cp... at googlegroups.com <javascript:>.
> Visit this group at https://groups.google.com/group/cp2k.
> For more options, visit https://groups.google.com/d/optout.
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20180921/cd6a5520/attachment.htm>