Lowest Eigenvalues of the unoccupied subspace spin did not converge in ot_eigensolver -- UV intensities

Pablo M. De Biase pablod... at gmail.com
Wed Nov 10 23:10:14 CET 2010


Hello,

1) I am trying to compute the uv-spectrum of a coumarin (using TDDFPT) but I
have found the following warning that makes me doubt about my results:


  Lowest Eigenvalues of the unoccupied subspace spin  1
 -----------------------------------------------------
 WARNING : did not converge in ot_eigensolver
 number of iterations  3000  exceeded maximum
 current gradient / target gradient 0.00007879514188995091  /  0.00001
      -0.11296455     -0.05366109     -0.04693513      0.01409883
       0.02124071


I have increased to 10 times the default number of iterations
(using MAX_ITER_LUMO 3000) but still not converging. I don't understand how
important is this for the computation of the excited states, why is not
converging, and what other keywords can I use to make it converge.

2) Anyway, I got this result (that I do not trust because of the warning):


  excited state :   1            2.115067

                   48 -> 49             0.993        0.987
                   47 -> 49             0.115        1.000

  excited state :   2            3.128501

                   48 -> 49             0.095        0.009
                   47 -> 49             0.774        0.608
                   46 -> 49             0.605        0.976
                   46 -> 51             0.052        0.979

  excited state :   3            3.653467

                   48 -> 49             0.059        0.004
                   47 -> 49             0.578        0.338
                   47 -> 51             0.112        0.350
                   46 -> 49             0.761        0.931
                   46 -> 51             0.083        0.938
                   46 -> 52             0.055        0.941
                   45 -> 49             0.086        0.951
                   44 -> 51             0.056        0.955


I would like to know which is the UV intensity of each excited state which I
can use for constructing the UV-Spectrum.

Thanks to everybody in advance,

Pablo

PS: I am attaching the input, output and restart files inside a tar/bzip2
file.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20101110/024bb483/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: cou152-spectra.tar.bz2
Type: application/x-bzip2
Size: 778180 bytes
Desc: not available
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20101110/024bb483/attachment.bin>


More information about the CP2K-user mailing list