Screwer

[Beverly]

Hello,
  I often encounter at least one imaginary frequency and have seen the 'screwer' tool mentioned a few times. I would really appreciate it if someone could give me some guidance as how to use it.

  Thanks
  Beverly

[uwe]

Hi Beverly,

screwer is an interactive script. Just run it in the directory where your input is. You will be asked along which mode you want to distort the structure - choose the one with the imaginary frequency. After that, you will find a new section in the control file: $newcoord

To use the new coordinates, you have to copy and paste them from $newcoord to $coord (usually in the coord file).

Another possibility is to run TmoleX and read in the control file of your job. Open the molecular viewer and click on the button that looks like a Laplace operator. This will open a new window where you can choose the mode with the imaginary frequency. Enter a higher temperature (if needed) and use the slider to move the structure along the chosen mode. Save the coordinates and either save all files or run the optimization directly from TmoleX.

Regards,

Uwe

[Beverly]

Hi,
 Thanks Uwe that made a things a lot clearer. I am wondering though if this will also be alright to use in cosmoprep as i am calculating in solvent. My imaginary frequency is about 6 cm-1, do i just re-run define from the same directory as my original when I have copied my coordinates or must i begin in a new directory?

Thank you
Regards
Beverly

[uwe]

Hi,

you can either reuse your old input and just replace the coordinates, or start from scratch with just the coordinates. Reusing your old orbitals can speed up convergence of the first SCF step, but if the change in geometry is too large, the opposite effect can occur. So I would recommend to start from scratch.

cosmoprep and COSMO work as before since the new coordinates have the same order (and the same number) of atoms. It is only the molecular orbitals that are different when moving the structure along a vibrational mode.

Regards,

Uwe

[Beverly]

Hello again,
  When I used screwer for the first time the run seemed to fail with a message saying that it may be neccessary to increase scfinterlimit or to check occupation.
 When selecting eht with my new coords in define i got a message saying that my HOMO-LUMO gap was less than 0.05 au, could you please explaina little bit about these messages. I'm very new to the program.

many thanks
Beverly

[jbaltrus]

Hi Beverly,

screwer is an interactive script. Just run it in the directory where your input is. You will be asked along which mode you want to distort the structure - choose the one with the imaginary frequency. After that, you will find a new section in the control file: $newcoord

To use the new coordinates, you have to copy and paste them from $newcoord to $coord (usually in the coord file).

Another possibility is to run TmoleX and read in the control file of your job. Open the molecular viewer and click on the button that looks like a Laplace operator. This will open a new window where you can choose the mode with the imaginary frequency. Enter a higher temperature (if needed) and use the slider to move the structure along the chosen mode. Save the coordinates and either save all files or run the optimization directly from TmoleX.

Regards,

Uwe

Uwe,

how does it work? Do I just paste $newcoords into the coord file and continue? But there is bunch of redundant coordinate definition from the old coords, do I have to recalculate redundants or after submitting system will do that automatically?

[uwe]

Hi,

yes, copy and paste of the coordinates is sufficient.

The internal coordinates that are stored in the coord file just define the topology, the values of the internal coordinates which are also in the coord file are ignored (except when you have fixed internal coordinates). Distorting the structure along a vibrational mode usually does not change the topology, unless you have bond breaking or something like that. The Turbomole modules statpt and relax read in the cartesian coordinates, the definition of the internal coordinates and then generate the values of the internal coordinates themselves.

If you run into trouble after replacing the coordinates because of a bad set of internal coords, jobex will automatically generate new internal redundant coordinates and continue the calculation with them. If that fails too, optimizations will continue using cartesian coordinates (unless you have frozen internal coordinates - in that case jobex will stop with an error message).

If you know that the distorted coordinates will cause problems, just shortly call define, enter the geometry menu, enter 'ired' there, and continue to the end by hitting <enter> until you get to the last, the general, menu - where you can simply quit with q or *.

Regards,

Uwe