Page 1 of 1

Question about calculated vibrational frequecies

Posted: Fri Mar 05, 2010 6:55 pm
by jianmin
Dear Prof. Graeme,

I have calculated the vibrational frequencies of propanoic acid (CH3CH2COOH) adsorbed onto Pd(111). Since the molecule has 11 atoms, I got 33 modes as follows:
33.514594 cm^{-1} ... 1
15.126440 cm^{-1} ... 1
37.403031 cm^{-1} ... 0
64.535768 cm^{-1} ... 0
73.538973 cm^{-1} ... 0
94.334503 cm^{-1} ... 0
132.603136 cm^{-1} ... 0
230.998630 cm^{-1} ... 0
289.930830 cm^{-1} ... 0
427.942812 cm^{-1} ... 0
465.571976 cm^{-1} ... 0
553.286629 cm^{-1} ... 0
617.179544 cm^{-1} ... 0
793.115648 cm^{-1} ... 0
825.255346 cm^{-1} ... 0
993.992770 cm^{-1} ... 0
1062.385355 cm^{-1} ... 0
1070.608972 cm^{-1} ... 0
1115.927258 cm^{-1} ... 0
1206.613829 cm^{-1} ... 0
1251.469340 cm^{-1} ... 0
1330.988036 cm^{-1} ... 0
1381.933614 cm^{-1} ... 0
1410.439249 cm^{-1} ... 0
1453.801138 cm^{-1} ... 0
1459.014032 cm^{-1} ... 0
1722.589432 cm^{-1} ... 0
2947.790830 cm^{-1} ... 0
2999.201999 cm^{-1} ... 0
3050.128255 cm^{-1} ... 0
3070.807761 cm^{-1} ... 0
3077.268803 cm^{-1} ... 0
3656.764091 cm^{-1} ... 0

My questions are:
1. Are these results worth trusting?
2. There are 2 imaginary frequencies, could I just omit these 2 imaginary ones?

Thanks.

Re: Question about calculated vibrational frequecies

Posted: Sat Mar 06, 2010 4:05 am
by graeme
If the molecule is really adsorbed to the surface there should not be any negative modes. If it is weakly physisorbed, you can ignore the lowest 6 modes - 3 rotation and 3 translation.

Make sure that the modes are converged with respect to the accuracy of the forces (ediff) and the displacement size. Also, include some surface Pt atoms if the molecule is chemisorbed.

Re: Question about calculated vibrational frequecies

Posted: Mon Mar 08, 2010 7:06 pm
by jianmin
Prof. Graeme, thank you so much for all the tips. I will follow your suggestions.
I use EDIFF=1.0E-7 and displacement=0.01. I assume these values are accurate enough for dyn_matr calculations.

Re: Question about calculated vibrational frequecies

Posted: Thu Apr 29, 2010 9:48 pm
by jianmin
I still got some imaginary vibrational frequencies for molecules which strongly bind to Pd(111) surface, such as CO.
One thing I haven't tried is to include some Pd atoms into the dyn_matr calculation. Could you tell me which Pd atoms I should include? The Pd atoms binding to the molecules?