visualise the transition behavior in molpro

[Bin Han]

Dear all,

I want to visualise the d-f transition of a molecule centering with Ce3+ atom, so I take  single Ce3+ atom as an example to visualise. I have tried TDDFT and CASPT2 methods to calculate the transition behavior of Ce. However, as we know some excited states may not happen d-f transition, so I want to check whenther an excited state happens d-f transition. One commonly used method is to use natural bond orbital (NBO), but I don't know how to use it in molpro. So I choose to check the ground state density and the excited state density, but I still do not know how to do this. 

In TDDFT calculation, I find some problems in calculating the excited state of Ce3+ ,

 

Starting iteration 30

 State     Energy   Energy Change  norm(resL)  norm(resR)  nvec  orth  tot  conv
 Warning: left and right eigenvectors not orthonormal:   5.8229633904939327E-008
   1            NaN          NaN         NaN         NaN    21   193   199    F
 Warning: left and right eigenvectors not orthonormal:   8.3871009781799582E-008
   2            NaN          NaN         NaN         NaN    21   193   199    F
 Warning: left and right eigenvectors not orthonormal:   7.3206037039079628E-008
   3            NaN          NaN         NaN         NaN    21   193   199    F
   4     0.01017564   0.00000062    0.37D-05    0.39D-02    21   193   199    F
   5     0.01073153  -0.00000042    0.11D-04    0.52D-02    21   193   199    F
   6     0.01816445  -0.00000012    0.38D-05    0.41D-02    21   193   199    F
   7     0.21433549  -0.00000001    0.32D-05    0.26D-02    21   193   199    F
   8     0.21670081  -0.00000000    0.19D-05    0.94D-03    21   193   199    F
   9     0.21689404  -0.00000001    0.18D-05    0.15D-02    21   193   199    F
  10     0.22566058  -0.00000004    0.37D-05    0.12D-02    21   193   199    F
  11     0.22574483  -0.00000003    0.36D-05    0.22D-02    21   193   199    F
  12     0.34461772   0.00000000    0.33D-05    0.12D-02    21   193   199    F

 Starting iteration 31

 State     Energy   Energy Change  norm(resL)  norm(resR)  nvec  orth  tot  conv
   1     0.00277679          NaN         NaN         NaN    16   193   199    F
   2     0.01030182          NaN    0.20D-01    0.15D-01    16   193   199    F
   3     0.01131705          NaN    0.68D-01    0.34D-01    16   193   199    F
   4     0.01436524   0.00418961    0.15D+00    0.17D+00    16   193   199    F
   5     0.01862961   0.00789808    0.68D-01    0.52D-01    16   193   199    F
   6     0.21433609   0.19617164    0.15D-02    0.16D-02    16   193   199    F
   7     0.21670084   0.00236534    0.35D-03    0.66D-03    16   193   199    F
   8     0.21689414   0.00019333    0.78D-03    0.47D-03    16   193   199    F
   9     0.22566071   0.00876667    0.43D-03    0.97D-03    16   193   199    F
  10     0.22574533   0.00008475    0.13D-02    0.93D-03    16   193   199    F
  11     0.34461776   0.11887293    0.43D-03    0.86D-03    16   193   199    F
  12     0.62836140   0.28374367    0.14D+00    0.15D+00    16   193   199    F

 Starting iteration 32

 State     Energy   Energy Change  norm(resL)  norm(resR)  nvec  orth  tot  conv
   1            NaN          NaN         NaN         NaN    11   199   199    F
   2            NaN          NaN         NaN         NaN    11   199   199    F
   3            NaN          NaN         NaN         NaN    11   199   199    F
   4            NaN          NaN         NaN         NaN    11   199   199    F
   5            NaN          NaN         NaN         NaN    11   199   199    F
   6            NaN          NaN         NaN         NaN    11   199   199    F
   7            NaN          NaN         NaN         NaN    11   199   199    F
   8            NaN          NaN         NaN         NaN    11   199   199    F
   9            NaN          NaN         NaN         NaN    11   199   199    F
  10            NaN          NaN         NaN         NaN    11   199   199    F
  11            NaN          NaN         NaN         NaN    11   199   199    F
  12            NaN          NaN         NaN         NaN    11   199   199    F

 Convergence reached. Total number of Hv calculations: 598

 ///////////////////////////////////
 Excitation energies for symmetry 1:
 ///////////////////////////////////

     n      eig               oscill.stren.          eig (eV)
     1             NaN               NaN                    NaN
 ? Error
 ? NaN detected by ieee_is_nan
 ? The problem occurs in is_nan

 GLOBAL ERROR fehler on processor   0

I do not konw how to deal with this, and I have refer to the manual of molpro to visualise the transition,

https://www.molpro.net/manual/doku.php?id=time-dependent_density_functional_theory#analysing_the_spectrum

but I find the setting do not work at all (I have take Cu atom as example).

Can someone give me some suggestions? Thanks a lot.

Best regards,

Bin

[andreas...@gmail.com]

Dear Bin,

the first problem appears to be a singlet instabilty issue. You can check this with

{tddft,...; stabil}

printing the lowest eigenvalues of the orbital rotation hessian.

(Note that after a recent update Molpro now stops with a corresponding

error message and no longer proceeds with negative eigenvalues).

It turns out that this can not easily be corrected for the system. I tried with

{uks,b3lyp,so_sci,so_sci_thresh=-0.001; wf,27,1,1}

which uses a second order optimisation. You can observe that this converges

to a different state that is still, however, not stable wrt to orbital rotations if you

check with the tddft program afterwards. (Note that the so optimisation for KS

in Molpro is currently in an experimental state, so use only for such difficult

cases). I do not know how it can be solved for this particular case.

I checked other functionals and, so far, only observed LDA to be working

using

{uks,lda,so_sci,so_sci_thresh=-0.001; wf,27,1,1}

{tddft,states=[-12.1]}

which finds a very small excitation energy of about 0.3 eV for the lowest

excitation.

The second problem is just because the visualisation options currently have only

been implemented for closed-shell tddft calculations. I have added a note in the

manual for this.

Best wishes,

Andreas