Reading in hessian but with isotopic substitution

[mike]

Dear all,

I have the hessian from a frequency calculation using aoforce and would like to use this hessian to calculate the normal modes and IR spectrum of the same molecule but with some isotopic substitutions. I really want to place the mass of some solvation side groups (not present in my calculation) on the terminal hydrogen just to get a qualitative picture of how the modes may differ with the side groups present, but it is effectively the same as modeling a very heavy isotope of hydrogen.

I changed the mass of the h atoms concerned in the control file as follows:

h  37,42,71,76,114,119,148,153,189,194,223,228                                 \
   basis =h   6-31G**                                                          \
   mass  =190.33144000

I also set fa in the drv menu to true, activating spectroscopic analysis only, so turbomole would use the old hessian rather than calculating it again.
$drvopt
   cartesian  on
   basis      off
   global     off
   hessian    on
   dipole     on
   nuclear polarizability
   frequency analysis only

when I ran aoforce it took just a minute to diagonalise the new mass weighted hessian but when I had a look at the log file, there were no IR intensities for the new modes, all the |dDIP/dQ| and intensity (km/mol) values were zero for all modes. What's more, there were several warnings in the log file:

WARNING - symmetry breaking for
  pair ivec - jvec         1064        1069 -1.004777627758302E-005
.....
WARNING - SYMMETRY ANALYSIS FAILED

The isotopic substitutions do not break the symmetry, so I don't understand what is going on. And is it normal that the fa option does not calculate IR intensities for the normal modes, or is this related to the symmetry warnings?

Many thanks in advance for any help/suggestions,

Mike

[uwe]

Hi,

did you also try the $isosub keyword?

Regards,

Uwe

[mike]

Dear all

Thanks for the suggestion Uwe and yes I did indeed try the $isosub keyword, but I had the same problem, namely no IR intensities.

My molecule is quite large, so I thought I would do some testing on something smaller, namely benzene. The test comprised three steps:
1) geometry optimization
2) normal mode calculation using aoforce
3) frequency only (using drv menu in define) calculation using hessian obtained in step 2

When I performed these calculations using turbomole version 6.0, the IR intensities of all modes was zero in the frequency only output file, in contrast to the original aoforce output obtained in step 2 which had non-zero IR intensities. However when repeated all the steps using turbomole 6.1, the frequency only calculation yielded the same results as step 2, as it should do. Finally, when I took the hessian from a turbomole 6.0 aoforce run and performed frequency-only calculation using version 6.1, again all the modes had zero IR intensities. All calculations were b3-lyp/SVP, m5 grid, scfconv 9.

This is the same behaviour I observed for my larger molecule. I have the normal modes and hessian computed using version 6.0 (which took a considerable amount of time to complete) and tried performing a frequency only analysis both with and without isotopic substitution using version 6.0 and 6.1. In all cases all modes had zero IR intensity, so not very useful for plotting the IR spectrum.

I also noticed that the normalised atomic displacements of each mode were not the same in the frequency only output file and the original aoforce output file, even for version 6.1 were the frequency only IR intensities were correct.

Here are the atomic displacements of the first atom for the first few modes of benzene in the original aoforce output file followed by the equivalent lines in from the frequency only output. Note that the displacements of modes of e2u and e2g symmetry are different, while those of a2u, b1g remain the same.

  mode               7        8        9       10       11       12

     frequency         418.47   418.47   619.94   619.94   694.90   733.65
 
     symmetry            e2u      e2u      e2g      e2g      a2u      b1g
                 
        IR               NO       NO       NO       NO       YES      NO
|dDIP/dQ|   (a.u.)     0.0000   0.0000   0.0000   0.0000   0.0068   0.0000
intensity (km/mol)       0.00     0.00     0.00     0.00    81.19     0.00
intensity (  %   )       0.00     0.00     0.00     0.00   100.00     0.00
                 
       RAMAN             NO       NO       YES      YES      NO       NO
 
  1   c           x   0.00000  0.00000 -0.11337  0.30043  0.00000  0.00000
                  y   0.00000  0.00000 -0.11049 -0.19459  0.00000  0.00000
                  z   0.11782  0.20345  0.00000  0.00000  0.03414 -0.18346
 

    mode               7        8        9       10       11       12

     frequency         418.47   418.47   619.94   619.94   694.90   733.65

     symmetry            e2u      e2u      e2g      e2g      a2u      b1g

        IR               NO       NO       NO       NO       YES      NO
|dDIP/dQ|   (a.u.)     0.0000   0.0000   0.0000   0.0000   0.0068   0.0000
intensity (km/mol)       0.00     0.00     0.00     0.00    81.19     0.00
intensity (  %   )       0.00     0.00     0.00     0.00   100.00     0.00

       RAMAN             NO       NO       YES      YES      NO       NO

  1   c           x   0.00000  0.00000 -0.23166  0.22236  0.00000  0.00000
                  y   0.00000  0.00000  0.22236  0.02509  0.00000  0.00000
                  z  -0.13170  0.19475  0.00000  0.00000  0.03414 -0.18346

Any idea what could be going on here? My only idea was that the frequency only aoforce rerun may be using a different reference geometry which has been transformed in such a way that displacements of modes of some symmetries remain unchanged but others are changed. But the contents of the coord files are identical in both calculations.

As for my original problem, I think I will just have to recalculate the hessian of my large molecule using version 6.1, then I can use the new hessian to perform isotopic substitution and see the effect on the IR spectrum.

Regards,

Mike