Page 1 of 1

Lanczos excitation energies and oscillator strengths

Posted: 13 Oct 2015, 23:07
by Bruno
Hi,
I'm trying to compare the spectra of transition energy and oscillator strength for H2 molecule obtained with standard CCSD linear response and the spectrum of Lanczos algorithm.

I don't know why in my calculation on Lanczos algorithm, it doesn't show the spectrum of energy/oscillator strength (I'm asking for 10 B3u excited states). I don't know what is missing..
To make it clearer, I'm attaching my output file.
I would appreciate if anyone here could help me.

My input is:
**DALTON
.RUN ALL
**WAVE FUNCTIONS
.CC
*CC INPUT
.CC2
.NSYM
8
*CCLRLA
.OPERAT
XDIPLEN XDIPLEN
.CHAINL
10
.SUMRULES
**END OF DALTON INPUT

Basis file:
BASIS
q-aug-cc-pVQZ

Atomtypes=1
Charge=1.0 Atoms=2
H1 0.000000000 0.000000000 -0.700493131
H2 0.000000000 0.000000000 0.700493131

...
Best regards,
Bruno.

Re: Lanczos excitation energies and oscillator strengths

Posted: 14 Oct 2015, 10:28
by sonco
Dear Bruno,

that is my fault, the Lanczos code in the release was still very experimental and the manual is missing one important piece of information,
namely that is that two additional output files have to be saved when running Lanczos:

- the CCEXCIL file, which contains the computed oscillator strengths for each excitation energy
- the CCSPECX file, which contains the cross sections (and real and imaginary polarizability) over the specified grid of frequencies

As both are lengthy outputs, I had to separate them from the DALTON.OUT file, otherwise it will have become too long.

All the best
Sonia

Re: Lanczos excitation energies and oscillator strengths

Posted: 14 Oct 2015, 13:00
by Bruno
Dear Sonia,

That's ok.
I think I didn't get it. How do I save these files?

Best regards,
Bruno.

Re: Lanczos excitation energies and oscillator strengths

Posted: 14 Oct 2015, 13:21
by sonco
Dear Bruno,

the two files are in the scratch directory and you have to make sure you back them up. Do you use the dalton
script to run you calculations, or do you have a "hand-made" script?

VERY IMPORTANT: Please also note that you need to comment out .SUMRULES in the input.
Unfortunately in your current version of the code the .SUMRULES key prevents
the code from computing the oscillator strengths. Instead, add smthg like

.FREQ INTERVAL
0.1 0.2 0.01

I have modified the code to amend for that restriction, but it has not yet been ported into the release.

All the best
Sonia

Re: Lanczos excitation energies and oscillator strengths

Posted: 14 Oct 2015, 15:49
by Bruno
I am using the dalton scrip.


att,
Bruno

Re: Lanczos excitation energies and oscillator strengths

Posted: 14 Oct 2015, 16:25
by Bruno
It's ok, I added in the script to save these two files in the .tar archive

thanks!

att,
Bruno