The temperature I get is different from what I set when do massive and global thermostat.

[汤羽]

Dears cp2k's users,

      I'm doing a vibration analysis exercise.When I did massive and global thermostat, the temperature I get is different from what I set. Could you tell me what causes this error? I would appreciate you generous help. The input and output file are listed as follow.

[Julian Heske]

Hello,

in order to run a MD calculation at a specific tempertare, you have to equilibrate the system first to reach your target temperature. The time you need to equilibrate your system is dependent on your initial configuration. Since you are starting with a structure which has a very high potential energy (-157,5 vs. ~ -168 a. u. of the others, see out.txt), this energy remains as kinetic energy in the system and the thermostat will need a long time to remove the energy from the system. You can first do a geometry optimization ( CP2K_INPUT / MOTION / GEO_OPT ), which leads to an optimized structure. Then start your MD run with your obtained structure as initial configuration. However, you should then still take a look at the temperature and not use the equilibration steps for your analysis.

Kind regards,

Julian

[汤羽]

Hello Julian,

     Thanks for your suggestion. I will do geometry optimization immediately. However, I'm a little confused about the last sentence  in your reply, whether I should use the structure gotten by geometry optimization to run massive and global thermostat firstly and than run others processing? I hope i have expressed my question clearly. 

      Tangyu.

Julian Heske <jjh...@gmail.com> 于2021年3月26日周五 下午11:03写道：

--
You received this message because you are subscribed to a topic in the Google Groups "cp2k" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/cp2k/MnB_Ejspnpw/unsubscribe.
To unsubscribe from this group and all its topics, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/ac99e428-7df6-4e03-8b6e-cbf7c16009dbn%40googlegroups.com.

[汤羽]

Hello,

    Thanks for your reply. I will change my input files according to your suggestion and read the paper carefully. Thanks a lot.

Travis <polla...@gmail.com> 于2021年3月27日周六 下午1:44写道：

Hi,

Yes, use the configuration from geometry optimization to start MD. Though in your original post, the thermostat coupling constant is much greater than the length of the trajectory you ran, whether massive or global - it's not doing anything here. The velocities are randomized and scaled to the target temperature by default as well so you should expect differences in the initial temperature fluctuations. After running for some time and reaching equilibrium, the average temperature from both thermostat settings should be equal. The massive thermostat controls every degree of freedom so it should get to the target temperature in fewer MD steps, making it ideal for equilibration. You should exercise caution in using MASSIVE for a given thermostat and coupling constant when you are interested in dynamic properties during data collection (aka production) runs. See, J. Chem. Theory Comput. 2013, 9, 2887−2899.

-T

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/a95c6cd7-9007-48ae-be3a-0167f5dbd763n%40googlegroups.com.