Unexpected decomposition of molecules during AIMD

[冬tlyer]

Dear users,

I let three moleclues(H2SO4,organic,H2O)  running  AIMD in my computer. But after running the Aperiodic molecular system  for a while. I took out the track-file to display in VMD. And i find  the  the organic  unexpectedly decomposed(as show below).

 So the question is how to change my input file or add some other keyword for solve this question.  In other words, what caused my system to decompose.The input-file and the molecules inages as attach is present below.

thanks!

tlyer

[HB H]

Hi 冬，

1). Talk to your superviser if you are a student~

2). The reason that the organic unexpectedly decomposed is probabaly that it wants to decompose at current conditions (i.e. NVT, 300K, etc).

3). You can prevent it from happening by constraining or fixing bond length in MD calculations, if you konw what you are doing. 

Good luck!

Haibei

冬tlyer <tsd19...@gmail.com> 于2020年10月6日周二 上午9:16写道：

--
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+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/7b089edd-05de-43dc-aa21-c3321f486255n%40googlegroups.com.

[冬tlyer]

Hello haibei,

 

Thanks for you reply.

I can get your meaning "constrain bood length".  But i try another input file in my system, the decomposition did not occur again. And i complare the twe file  finding that the some differenence is the QS section , the XC_GRID  and the MG_GRID.  So in my mind ,maybe these sections get the AIMD to different output. However, i don‘t know what section solve the situation of decomposition.

As for fix bood length , I feel it will cause errors in the calculation results . But this is only my feelling.

The new input-file and the old input-file as attach is present below.

[HB H]

Hi, 

It seems to me you have a convergence problem. So you have to first choose a reasonable cutoff value and the basis set, and the functional as well, before MD calculations.

Plz read the follow instructions:

https://www.cp2k.org/howto:converging_cutoff 

https://www.cp2k.org/exercises:2018_uzh_cmest:faq

You get a lot of work to do!

Good luck~

Haibei

冬tlyer <tsd19...@gmail.com> 于2020年10月6日周二 上午11:42写道：

--

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+uns...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CAECQVt8VggY4hO0FjdegEOruAZYgk32kqt8RnHvgWFO6UiLEOw%40mail.gmail.com.