[CP2K-user] [CP2K:17693] Re: SGCPMD: problem with equilibration?

[Onofrio Tau]

Dear Alexandra,
I used to do ENSEMBLE NVE in &MD for the BOMD run. Probably 50 K of the 
TEMP_TOL is too strict, at least for Pt. As you can see from your last 
plot, the Temperature of the Pt region frequently drops to your lower 
limit, 280 K, even after 4000 fs (not good). This is why you get a Pt 
temperature drop in the SGCPMD run, because of a not good previous 
equilibration.
Regards.

Onofrio

Il giorno giovedì 15 settembre 2022 alle 13:34:39 UTC+2 a.celi... at gmail.com 
ha scritto:

> Dear Onofrio,
>
> thanks a lot for your fast response! Indeed, I did a NVE simulation with 
> Langevin  (as suggested here:https://www.cp2k.org/howto:md). Then I used 
> the geometry and wfn for the SGCPMD. 
> As you, I use TEMP_TOL, but I set it to the value of 50 for each region. 
> Do you think that it makes the difference?
>
>  This are the parameters that I used for the NVE:
>    &MD
>      ENSEMBLE  LANGEVIN
>      STEPS  6000
>      TIMESTEP     5.0E-01
>      TEMPERATURE     3.3000000000000006E+02
>      TEMP_KIND  T
>      SCALE_TEMP_KIND T
>      &LANGEVIN
>        GAMMA     1.0000000000000000E-03
>        NOISY_GAMMA     0
>      &END LANGEVIN
>      &THERMAL_REGION
>        DO_LANGEVIN_DEFAULT  F
>        &DEFINE_REGION
>      DO_LANGEVIN F
>          LIST  145..546
>          TEMPERATURE     3.3000000000000006E+02
>          NOISY_GAMMA_REGION     2.1999999999999998E-04
>          TEMP_TOL 50
>        &END DEFINE_REGION
>        &DEFINE_REGION
>          DO_LANGEVIN F
>          LIST  73..144
>          TEMPERATURE     3.3000000000000006E+02
>          NOISY_GAMMA_REGION     4.9999999999999996E-05
>          TEMP_TOL 50
>        &END DEFINE_REGION
>      &END THERMAL_REGION
>    &END MD
> The NOISY_GAMMA is not considered in the simulation. Is there something 
> not so good with the set of parameters I used?
>
> [image: NVE_00_TempLANGEVIN.png]
>
> I appreciate a lot your feedback!
> Best regards,
> Aexandra
> El jueves, 15 de septiembre de 2022 a las 12:53:45 UTC+2, 
> onofr... at gmail.com escribió:
>
>> Dear Alexandra,
>> have you done an equilibration step before starting the SGCPMD run?
>> As I can see from your plots it seems that it is not equilibrated yet. I 
>> suggest you to make a BOMD run first for better equilibration, then 
>> switching to sgcpmd, as follows:
>> 1. BOMD as NVE ensemble + TEMP_TOL 100
>> 2. switch to SGCPMD using RESTART.wfn of the previous equilibration step 
>> as a wfn guess
>> It works for me.
>> Hope it helps
>>
>> Regards
>>
>> Onofrio
>>
>> Il giorno giovedì 15 settembre 2022 alle 12:04:35 UTC+2 
>> a.celi... at gmail.com ha scritto:
>>
>>> Dear Cp2k users,
>>> I have carried on a second-generation Car-Parrinello (SGCP) MD 
>>> simulation for Pt(111)/water interface. Unit cell contains Pt: 6x6x4 (2 
>>> frozen layers), water: 134 molecules.
>>>
>>> I have performed a similar workflow as Onofrio ("SGCPMD: number of 
>>> steps per SCF run increase suddenly after initial 400 MD steps"). But I 
>>> have used the NOISY_GAMMA as in the publication *"First-Principles 
>>> Simulations of an Aqueous CO/Pt(111) Interface" 
>>> doi:10.1021/acs.jpcc.8b05933, *Pt: 5E-05 1/fs and for water 2.2E-04 
>>> 1/fs. Temperature is set to 330K. 
>>>
>>> Unfortunately, in my case along the SGMD the temperature of Pt region 
>>> drops to roughly 180 K, the temperature of the hole system is about 328K. I 
>>> expected that the temperature of Pt region would be closer to the water 
>>> after a equilibration time of 4ps. Does this mean that probably my system 
>>> is not well-equilibrated? Is this going to affect the dynamics and energy 
>>> of the system? (PS: the density profile seems ok, double peak on oxygen, 
>>> well-known for Pt(111))
>>>
>>> Many thanks in advance!
>>> Alexandra
>>>
>>> PS: I also observed the increase of the SCF steps after a restart!
>>>
>>> TOP: red dashed line: average temperature of system
>>> BOTTOM: temperature for each region
>>>
>>> [image: SGMD_00MD_Temp.png][image: SGMD_00MD_TempLAN.png]
>>>
>>

-- 
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+unsubscribe at googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/a81210ec-82e5-4249-9558-d16838ee3f4bn%40googlegroups.com.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20220915/0a8ffd22/attachment-0001.htm>