Page 1 of 1

DynamicMatrix_CH4

Posted: Sat May 23, 2020 11:10 am
by microomran
Dear Dr.Graeme

I am trying to use dynamic matrix to get vibratioonal frequencies for some simple molecules in box.
I started with methane relaxation at PREC= High ,then followed by dynamic matrix calculations (files attached).However,the imaginary frequencies persist even after increasing the accuracy and box size (In fact when I increase the box size the number of imaginary frequency increase from 3 to 8 which was a bit strange):

57.266885 cm^{-1} ... 1
3.407140 cm^{-1} ... 1
2.811486 cm^{-1} ... 1
1.643336 cm^{-1} ... 1
1.425706 cm^{-1} ... 1
0.739629 cm^{-1} ... 1
0.149094 cm^{-1} ... 1
0.000025 cm^{-1} ... 1
5.644886 cm^{-1} ... 0
1281.384962 cm^{-1} ... 0
1282.428479 cm^{-1} ... 0
1501.560956 cm^{-1} ... 0
2950.633396 cm^{-1} ... 0
3076.440907 cm^{-1} ... 0
3078.811713 cm^{-1} ... 0


Your kind advice is appreciated!

Regards,
Omran

Re: DynamicMatrix_CH4

Posted: Sat May 23, 2020 5:17 pm
by graeme
The problem has to do with ISYM. In your calculation, vasp is detecting symmetry so that when you make a small displacement, it is zeroing the components of the force which break symmetry.

Additionally, you have a tab instead of a space after your ISYM variable, so that even if you think you are setting it to zero, it was not. So uncomment the ISYM = 0 line and remove the tab. When I did that, I got the following frequencies:

79.630518 cm^{-1} ... 1
57.346688 cm^{-1} ... 1
4.912432 cm^{-1} ... 1
2.808476 cm^{-1} ... 1
5.931526 cm^{-1} ... 0
11.098396 cm^{-1} ... 0
1281.221222 cm^{-1} ... 0
1281.608603 cm^{-1} ... 0
1282.352277 cm^{-1} ... 0
1500.636448 cm^{-1} ... 0
1501.644785 cm^{-1} ... 0
2951.172226 cm^{-1} ... 0
3073.998233 cm^{-1} ... 0
3077.337065 cm^{-1} ... 0
3078.916644 cm^{-1} ... 0

where the first 6 are all near-zero and correspond to rotation and translation.

Re: DynamicMatrix_CH4

Posted: Sat May 23, 2020 10:18 pm
by microomran
Thank you...I have some doubts

How I can calculate ZPE/Svib in presence of imaginary frequencies to reproduce calculations as attached? dymmatrix.pl will ignore them?

In the case of of surface_adsorbate and clean surface calculations I managed to get all real frequency but never get same with molecules such as CH4 or H2O.

I saw you had earlier comment about that here but it was not clear how problem was solved: https://www.vasp.at/forum/viewtopic.php?t=138

Re: DynamicMatrix_CH4

Posted: Sun May 24, 2020 5:45 am
by graeme
Wait a sec. I just nicely and rapidly solved the problem that you posted and now you immediately post a different problem?!?

No, take a minute to think about your calculation. With the frequencies that I sent, there are 6 zero-ish modes associated with translation and rotation. The others are vibrations for which zpe can be calculated. Take some time to think about this and please try to post questions which require expert advice.

Re: DynamicMatrix_CH4

Posted: Sun May 24, 2020 1:10 pm
by microomran
Thank you.Sorry,I just misread the numbers.I appreciate your kind help.