## error in polarizability calculation at imaginary frequencies

(It is not necessary to upload input files, they can be found in the output file.)

greifsw
Posts: 11
Joined: 28 May 2016, 21:05
First name(s): Amit
Last name(s): Sharma
Affiliation: Wright State University
Country: United States

### error in polarizability calculation at imaginary frequencies

Hi
I am not experienced with Dalton so excuse me if the question is trivial. As a TEST calculation I am trying to reproduce numbers from the following paper:
Atomic Data and Nuclear Data Tables 96 (2010) 323–331
https://www.sciencedirect.com/science/a ... onent-id18

Table 3 of the paper gives dynamic polarizability for He atom over 50 imaginary frequencies. When I try to run this calculation I get the following error. Looking at the output the dynamic polarizability values for smaller frequency values see ok but for higher frequency values the numbers are completely wrong. Can anyone please point out what is wrong with my calculation.

Here is the warning/error.

832: *** INFO, negative eigenvalues in reduced matrix. Symmetry 2, triplet F Dimension of reduced matrix = 6
839: *** INFO, negative eigenvalues in reduced matrix. Symmetry 2, triplet F Dimension of reduced matrix = 6
846: *** INFO, negative eigenvalues in reduced matrix. Symmetry 2, triplet F Dimension of reduced matrix = 6
853: *** INFO, negative eigenvalues in reduced matrix. Symmetry 2, triplet F Dimension of reduced matrix = 6
860: *** INFO, negative eigenvalues in reduced matrix. Symmetry 2, triplet F Dimension of reduced matrix = 6
867: *** INFO, negative eigenvalues in reduced matrix. Symmetry 2, triplet F Dimension of reduced matrix = 6
874: *** INFO, negative eigenvalues in reduced matrix. Symmetry 2, triplet F Dimension of reduced matrix = 6
881: *** INFO, negative eigenvalues in reduced matrix. Symmetry 2, triplet F Dimension of reduced matrix = 6
888: *** INFO, negative eigenvalues in reduced matrix. Symmetry 2, triplet F Dimension of reduced matrix = 6
895: *** INFO, negative eigenvalues in reduced matrix. Symmetry 2, triplet F Dimension of reduced matrix = 6

Thanks
Amit
PS attached are Dalton files.
Attachments
he.mol
pol.dal

magnus
Posts: 461
Joined: 27 Jun 2013, 16:32
First name(s): Jógvan Magnus
Middle name(s): Haugaard
Last name(s): Olsen
Affiliation: Hylleraas Centre, UiT The Arctic University of Norway
Country: Norway

### Re: error in polarizability calculation at imaginary frequencies

Not certain, but I don't think you can get what you want through **PROPERTIES but you can do it through **RESPONSE as follows (see 29.1.11 in the Dalton manual for more details):

Code: Select all

``````**DALTON INPUT
.RUN RESPONSE
**WAVE FUNCTIONS
.HF
**RESPONSE
*ABSORP
.ALPHA
.IMAG F
.FREQUE
51
0.0
0.00178065
...``````

hjaaj
Posts: 317
Joined: 27 Jun 2013, 18:44
First name(s): Hans Jørgen
Middle name(s): Aagaard
Last name(s): Jensen
Affiliation: Universith of Southern Denmark
Country: Denmark

### Re: error in polarizability calculation at imaginary frequencies

An additional comment: what you asked for (and got) was dynamic polarizabilities at 50 real frequencies (not imaginary frequencies). When the specified frequency is greater than excitation energies you get negative eigenvalues (and possibly negative polarizabilites), and that is what the INFO statements tell you.

greifsw
Posts: 11
Joined: 28 May 2016, 21:05
First name(s): Amit
Last name(s): Sharma
Affiliation: Wright State University
Country: United States

### Re: error in polarizability calculation at imaginary frequencies

Thank you!!!
That works, the dispersion coeff. calculated with above input comes out to be 1.38 (HF method), and, with DFT/B3LYP the value is 1.61. Reported value is 1.46 calculated with the relativistic CI method. Next I have to try this with CI. Will this approach also work for excited states calculated with other methods - CASSCF/MRCI. The Dalton manual seem to say that RESPONSE module is generalized.

I am mainly interested in calculating dispersion coeff. for excited states. I plan to adopt this approach for excited state. Do you see any flaw in this approach?

magnus
Posts: 461
Joined: 27 Jun 2013, 16:32
First name(s): Jógvan Magnus
Middle name(s): Haugaard
Last name(s): Olsen
Affiliation: Hylleraas Centre, UiT The Arctic University of Norway
Country: Norway

### Re: error in polarizability calculation at imaginary frequencies

Yes, the RESPONSE module is general except for coupled cluster which had its own response implementation. So it should work but may require additional input also in the response section.

greifsw
Posts: 11
Joined: 28 May 2016, 21:05
First name(s): Amit
Last name(s): Sharma
Affiliation: Wright State University
Country: United States

### Re: error in polarizability calculation at imaginary frequencies

As a followup, I have setup calculation to compute RESPONSE for Li atom ground state at MCSCF level of theory. If this is successful then I would try excited states. I run into the negative eigenvalue and negative polarizabilites error for this calculation (output attached). I looked for additional input under RESPONSE but not sure what else to add. Any suggestions?
Attachments
pol_li.out