Learn how easy it is to sync an existing GitHub or Google Code repo to a SourceForge project! See Demo

Close

#174 accept GenNBO 5 output

closed
nobody
5
2010-03-25
2010-03-23
No

Accept GenNBO 5 output as a source of Natural Bond Orbital data.
At the moment NBO output appearing in electronic structure system (e.g. Gaussian) output is accepted (provided $NBO AONBO=P $END is given as input to the ESS). Running NBO analysis as GenNBO (standalone rather than embedded in an ESS) can be quite advantageous, for the parameters of the NBO analysis can be changed without rerunning the ESS job which can take hundreds of times longer to run than GenNBO. It would be nice if JMol accepted GenNBO output for NBO visualization. This is not a large change, because the GenNBO output format is the same as for the NBO output appearing in the middle of ESS output.

Secondly, this is handy because the GenNBO output derived from any ESS output can be accepted. No modifications are necessary to make use of a new/different ESS as input to GenNBO provided that the ESS outputs the GenNBO input .47 format file format (Jmol doesn't need to know that format). ESSs that don't have NBO 5 as standard (e.g. Gaussian 03+09, ADF) do support .47 format output so GenNBO 5 can be used.

It may be possible to take GenNBO output and insert it into ESS output so that it can read by Jmol, but it would be nice if the GenNBO output (typically suffixed .nbo) could be read as is.

Discussion

  • Bob Hanson
    Bob Hanson
    2010-03-23

    sounds great to me. Send us some output examples of various types.

     
  • Bob Hanson
    Bob Hanson
    2010-03-23

    note -- we'll need:

    -atom symbols or atomic numbers
    -coordinates in Bohr or Angstrom
    -atomic basis function coefficients

    are those in the NBO output?

     
  • Hmm, won't upload a 3.5MB file - send it to you directly?

     
  • Yes, everything you need is in the GenNBO output.

    There is a 256KB limit for upload? Unzipped, a sample output file is 15MB.

     
  • zipped GenNBO 5 output file

     
    Attachments
  • Zipped small GenNBO 5 output file attached.

     
  • Bob Hanson
    Bob Hanson
    2010-03-23

    The issue is that the NBO output does not list the AO basis functions. So Jmol has no way to know how to build the orbitals. And also I think it doesn't have atom positions.

    But I think I have a solution. I added a bit of code that allows multiple NBO sections and basically starts over with a new set of orbitals if a second NBO output section is found. So all you should have to do is concatentate the NBO output onto the end of the original ESS output, and you are good to go. This should work with Gaussian, GAMESS, Jaguar, QChem, NWChem, and PSI, because they all access that same code.

    What do you think?

     
  • Bob Hanson
    Bob Hanson
    2010-03-23

    Make that Gaussian, GAMESS, Jaguar, QChem, and PSI

     
  • Gaussian 03 and 09 are supplied with NBO 3.1 installed in the ESS. My license for NBO 5 is a single research group license, and can't be installed in the ESSs because they are site wide. If NBO 5 output were appended to the end of ESS output which contained NBO 3.1 output, your code would have to figure out which version of NBO was in use at what part of the output and switch to the right formatting and positioning. The text of headings used for locating data sections could change between NBO 3.1 and NBO 5. I don't think you should deal with NBO 3.1 because it is not for serious use, but it is what is used by default in at least Gaussian 03 and 09.

     
  • Bob Hanson
    Bob Hanson
    2010-03-24

    Shouldn't be a problem with the version of Jmol I just uploaded. Just append the NBO output to the Gaussian output. Doesn't matter that two versions of NBO output are in there. Jmol will take the last it sees.

     
  • Bob Hanson
    Bob Hanson
    2010-03-25

    • status: open --> closed