VSCF PES for a linear molecule


Just Got Here
Hi,

I have run a Vibrational Self Consistent Field calculation on a CO2 molecule with NWChem. The resulting potential energy surface stored in the co2.coup file (I chose a pair mode coupling) surprised me, because it used only 3 normal modes of the molecule. That is, the PES looked like this:

ModeA ModeB ProgressionAlongMode Energy .....

1 2
1 2
...
1 3
1 3
...
2 3
2 3
...
End of file

But the CO2 molecule has 4 normal modes! There is one symmetric stretch, one asymmetric stretch and two bending modes (call one bend_X and the other bend_Y). bend_X and bend_Y (mode 3 and mode 4) are degenerate. But the coupling of modes 3 and 4 should still be included in the PES, because progression along both mode 3 and 4 at the same time corresponds to a nuclear configuration which is not explored by the PES NWChem has output. To clarify:

Energy(Q1=0,Q2=0,Q3,Q4) is not equal to Energy(Q1=0,Q2=0,Q3=0,Q4)

where the Qn here signify progression along the nth normal mode (Qn =0 means at equilibrium for that mode)

I suspect this is a problem with my input file geometry specification:

"start
restart
title co2

geometry noautoz units angstrom
C 0.00000000 0.00000000 0.00000000
O 0.00000000 0.00000000 1.18792006
O 0.00000000 0.00000000 -1.18792006
end

basis
C library sto-3g
O library sto-3g
end

vscf
coupling pair
ngrid 25
iexcite 1
end

task scf vscf"

Any ideas why the PES output is lacking a normal mode and how I could get NWChem to output that information?

Thanks,
Andrew

Forum Vet
Bug to be fixed in next release
You found a bug, as it is not picking up a keyword that says the molecule is linear. This will be fixed in the upcoming release. In the mean time, you can get it to work by putting:

set vib:linear .true.

just before the "task scf vscf" line.

Bert

Just Got Here
I've set the calculation running again and we'll see how it goes. I probably could've worked that out, but it would've taken a long time. Thanks heaps.



Forum >> NWChem's corner >> General Topics