frequency calculations

[mhyman]

Hello,

I've been performing frequency calculations on oxygen-oxygen stretches. I've performed frequency calculations of O2 on Pt(111) to compare with Eichler and Hafner's work PRL V79 p.4481 (1997), which compare very favorably to experiment. I've used a slightly different calculation setup -

3 frozen substrate layers
1/9 ML O2 coverage
PAW potentials
5x5x1 Monkhorst Pack k-point mesh
ENCUT = 400
POTIM = 0.04

For both the magnetic (top-bridge-top) and nonmagnetic (top-fcc-bridge) states, my adsorption energies and geometries compare very well to Eichler and Hafner's values. And the O-O stretching frequency for the magnetic state is very close (within 3 cm-1). However, the O-O stretch for the nonmagnetic state is 826cm-1, compared to 690 cm-1 calculated by Eichler/Hafner. The optimized geometry was converged with forces on the oxygen atoms are < 0.005 eV/A. Can anyone offer suggestions as to what is going on?

Thanks,
Matt

[alex]

How do your values compare to experiment? That's always the perfect test ...
Was the coverage the same as in Hafner/Eichler's work?
Have you used decent pseudopotentials for oxygen? O-O bond comes with r=124pm, so O_s is too bad in forces, but might give reasonable structures ...

Hth
Alex

[mhyman]

That's the thing. Structures and energies agree very well. It's just the frequency. The frequencies calculated by Eichler and Hafner compare very favorably to experiment. The question is: why would the frequency be so far off when everything else is correct?