Need help in running AIMD simulations on Ta2 dimer supported on C2N monolayer

[DAR MANZOOR]

Dear CP2K developers,

I am new to CP2K and am trying to run abinitio MD simulation on a Ta2 dimer supported on C2N monolayer. However, as i run the calculation the structure completely breaks and looks too weird. I am attaching the input and output files here. I will be grateful if someone can look into the files and help me troubleshoot the issue. Thanks in advance.

--

Regards & best wishes
Dr. Manzoor A. Dar,

Assistant Professor,

Dept. of Chemistry,

IUST Awantipora.

[DAR MANZOOR]

Dear Users,

I will be extremely grateful if anyone can take some time to look into the above problem.

Best wishes,

Manzoor

[Rizwan Nabi]

Dear Manzoor, 

Please check If the charges are correct. From Your Coordinate file, the valency seems unsatisfied.

I can not see any double bonds in the structure. 

Regards

Rizwan

[DAR MANZOOR]

Dear Rizwan,

Thanks for your reply. I am using an overall charge of zero. I have run the same system in VASP and it relaxes smoothly. However here I faced the same problem even when I tried to relax the structure. The structure just breaks away whether I run the geometry optimization or MD.

Best wishes,

Manzoor

--
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/2ae6eb9e-43b0-4c6f-8b5a-417f7e343749n%40googlegroups.com.

[Rizwan Nabi]

Dear Manzoor,

Can you just add H atoms on all the C atoms and Re-run the calculations. I guess the problem is with the Carbon valency. Please let me know if this works?

Regards

Rizwan

-------

Dr Rizwan Nabi
Research Associate "European Research Council Fellow"
Department of Chemistry, The University of Manchester
Oxford Road, Manchester, M13 9PL, United Kingdom
t: +44 (0)742 447 4940
e: rizwa...@manchester.ac.uk

w: http://nfchilton.com/people.html

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CAKNVsq43v%2BuBTPumOkLO%3D23cpVwkfNU08-vrvPTTvhhoz2KMQg%40mail.gmail.com.

[DAR MANZOOR]

Dear Dr. Rizwan,

I don't think there is a problem with the structure of C2N monolayer. Its structure is well documented in literature. However, I feel I may be making some mistake in terms of the format of the xyz file and cell parameters in the input file. If you know how generate the xyz coordinates for VASP CONTCAR in the CP2K format, that might help me troubleshoot this problem.

Manzoor

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CAFme2NQPnoweyGbbCzwgQ9KuQS%2BtdxW3oJdmZh8%3DGDu1dJf6WQ%40mail.gmail.com.

[Rizwan Nabi]

Dear Manzoor,

Please send me your CONTCAR, I will Send you XYZ format.

Regards

Rizwan

-------

Dr Rizwan Nabi
Research Associate "European Research Council Fellow"
Department of Chemistry, The University of Manchester
Oxford Road, Manchester, M13 9PL, United Kingdom
t: +44 (0)742 447 4940
e: rizwa...@manchester.ac.uk

w: http://nfchilton.com/people.html

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CAKNVsq6wTECgsUoViVLCDumvUw9qw5a_c%2BreHBtCweZna7uMZg%40mail.gmail.com.

[Fabian Ducry]

Dear Manzoor,

a very simple way is to use ASE (a python package). If you have it installed (e.g. "pip install ase" or if you use conda "conda install -c conda-forge ase") just run the following two lines in python:

from ase import io

io.read('CONTCAR').write('C2NTa2.xyz')

This will convert the CONTCAR to C2NTa2.xyz. The second line of c2n.xyz contains Lattice="xx xy xz yx yy yz zx zy zz" which are the 9 numbers that define the lattice. You can use these to set A xx xy xz B yx yy yz and C zx zy zz in the &CELL section of the input file. If you explicitly set A, B,and C then ANGLE is not needed anymore.

Cheers,

Fabian

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CAFme2NT_Sx5DwfJfH_OopyQgN1njCvTDupgpgSbMJ9XVmoPktg%40mail.gmail.com.

[DAR MANZOOR]

Dear Fabian,

Thanks for your suggestions. I will try it out your suggestions and get back to you on this. Thanks again.

Best wishes,

Manzoor

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/3145f5ad-35f6-e185-6512-6dc1cca83377%40gmail.com.

[Lucas Lodeiro]

Hi Manzoor,

As I see, the xyz file is well defined, the length bonds are good, but, when you check the periodicity in the first vector, there are overlapped atoms (VESTA always shows replicated atoms, I fails to transform from VASP coordinates to XYZ due to put more atoms. Also, the periodicity in the second vector is weird, if you see it, the replicas do not match very well, and the distance between them is so long.

Also, on the input there are some definitions which are not the better... I suggest:

1- Increase the CUTOFF. You are using a 13 electron Ta pseudo, maybe you need higher plane waves for it.

2- EPS_DEFAULT 1.0E-12 is better for the EPS_SCF defined.

3- Adding Grimme's correction could be a good idea.

4- Change from DIAGONALIZATION to OT scheme. OT is faster and robust for gapped materials. In case that your system does not have a band gap (metallic), you have to use DIAGONALIZATION + SMEARING + ADDED_MOS.

Regards

To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/CAKNVsq637_dFdA9W0gpG_gjnx%2BBjP2gN3h8J1254PGHFkRb1pA%40mail.gmail.com.