Page 1 of 1

Dynamical Matrix calc. gives different ZPE than when using IBRION = 5

Posted: Thu Jul 05, 2018 9:04 pm
by gmlp
Hi,

I have done a frequency analysis. 1 using IBRION = 3 and the scripts for the creation of the dynamical matrix (specifically to calculate harmonic vibrational frequencies) and 2. Using IBRION 5 from VASP. I see that in both cases the codes use the displacement method to create the matrix. And after calculating the zero point energy and the vibrational entropy contribution I get different results. I am calculating 3 molecules of HCO3 adsorbed on a 3x3 slab.

I see that with IBRION= 5 I do not get the same amount of real frequencies as in IBRION = 3. Could you please help me figure out why is that?



The following are the frequencies and additional info using the IBRION = 3
DOF info:
Using 1 central atom
Central Coordinates: 0.4976434771192025 0.1383167840154998 0.5630636305468041
Central atom 1: 0.4976434771192025 0.1383167840154998 0.5630636305468041
----------------------------------------------------------------------
15 atoms were found within a radius of 4.875 of atom 64,
leading to 45 degrees of freedom selected.

38.085979 cm^{-1} ... 1
30.570475 cm^{-1} ... 1
19.711584 cm^{-1} ... 1
11.437344 cm^{-1} ... 1
8.964256 cm^{-1} ... 1
6.373125 cm^{-1} ... 1
5.757739 cm^{-1} ... 1
4.892928 cm^{-1} ... 1
3.588654 cm^{-1} ... 1
2.448143 cm^{-1} ... 1
1.238374 cm^{-1} ... 1
0.760153 cm^{-1} ... 1
0.624412 cm^{-1} ... 1
0.502125 cm^{-1} ... 1
0.406576 cm^{-1} ... 1
0.316430 cm^{-1} ... 1
52.770962 cm^{-1} ... 0
93.979432 cm^{-1} ... 0
107.686412 cm^{-1} ... 0
122.065795 cm^{-1} ... 0
141.598913 cm^{-1} ... 0
221.158248 cm^{-1} ... 0
238.700994 cm^{-1} ... 0
243.650008 cm^{-1} ... 0
474.840483 cm^{-1} ... 0
499.484473 cm^{-1} ... 0
506.768252 cm^{-1} ... 0
531.472803 cm^{-1} ... 0
615.540117 cm^{-1} ... 0
646.138546 cm^{-1} ... 0
657.121232 cm^{-1} ... 0
693.577552 cm^{-1} ... 0
741.469525 cm^{-1} ... 0
974.408719 cm^{-1} ... 0
1007.450491 cm^{-1} ... 0
1052.949364 cm^{-1} ... 0
1200.522127 cm^{-1} ... 0
1271.413712 cm^{-1} ... 0
1356.066712 cm^{-1} ... 0
1452.036431 cm^{-1} ... 0
1527.581567 cm^{-1} ... 0
3568.446239 cm^{-1} ... 0
3671.846038 cm^{-1} ... 0
ZPE and TS where calculated with the positive values of the frequencies
Temperature 298.15 K
ZPE Real freqs 1.467399 eV
Svib Real 0.001 eV/K
TS Real 0.358 eV

The following are the frequencies obtained using IBRION = 5
all of them real.
3679.283044 cm-1
3678.540514 cm-1
3678.173699 cm-1
1533.412588 cm-1
1533.078522 cm-1
1518.535168 cm-1
1405.916829 cm-1
1343.308088 cm-1
1343.143696 cm-1
1210.716903 cm-1
1195.624801 cm-1
1195.367639 cm-1
1012.358529 cm-1
1007.246599 cm-1
1007.168414 cm-1
742.951239 cm-1
742.862939 cm-1
739.602760 cm-1
660.717357 cm-1
657.624578 cm-1
657.553530 cm-1
617.273156 cm-1
614.602743 cm-1
614.531587 cm-1
532.536816 cm-1
531.849184 cm-1
511.149273 cm-1
253.004982 cm-1
232.319112 cm-1
231.936557 cm-1
218.140541 cm-1
217.811696 cm-1
217.457239 cm-1
146.244238 cm-1
146.123083 cm-1
130.769788 cm-1
125.278204 cm-1
106.337706 cm-1
105.447779 cm-1
85.259257 cm-1
83.755986 cm-1
83.546512 cm-1
48.830509 cm-1
48.367460 cm-1
43.653843 cm-1
Temperature 298.15 K
ZPE 2.262 eV
Svib 0.00265 eV/K
TS 0.7895 eV

Again the question is, why could be the reason to get different number of real frequencies with the two methods? Where could I be making a mistake? It seems that with IBRION = 5 I get almost 2 times the frequencies I get with IBRION = 3.

Thanks,
Laura

NB: IBRION = 5 is expensive for some of my calculations and sometimes not even 5 days is enough to complete 57 displacements ( in some of my systems) and I am left with incomplete vibrational analysis.
I would like to be able to use IBRION = 3 as I can continue the frequency calculation by generating the degrees of freedom that where not calculated before.

Re: Dynamical Matrix calc. gives different ZPE than when using IBRION = 5

Posted: Fri Jul 06, 2018 8:49 pm
by graeme
The first calculation does not make sense in terms of the number of negative modes. I would need to see that calculation to see what might be going wrong. The typical problems are related to the displacement size, which should be on the order of 0.01 - 0.001 Ang, accurate forces with ediff=1e-8, and making sure that symmetry is turned off with ISYM=0. Then, check that the calculation completed so that all displacements and forces were done and that you don't have zeros in the Hessian matrix.