Author Topic: EDD visualization  (Read 398 times)

valienpo

  • Newbie
  • *
  • Posts: 4
  • Karma: +0/-0
EDD visualization
« on: February 19, 2019, 04:43:11 am »
Hi,
I calculated ADC(2) excitation using Turbomole 7.2 and 7.3 and needed to visualize difference densities between excited and ground states. I confused by some inconsistency between my results and the recommendation of the Turbomole manual.
I used control with a "min setting"
$excitations
  irrep=a nexc=12
  spectrum states=all operators=diplen,qudlen
but didn't receive the needed files. 

1. Which minimum set of keywords should I add to control for getting files like
cc1td-<type>-<mult><irrep>-<number> ?
They should be something like
cc1td-cc2-gs-1a1-001
cc1td-adc2-xs-1a1-00*, where *=1,2,3,...
as it's described in "Default density analysis and visualization" for the canonical SCF orbitals.
if I want EDD for second singlet but I don't see such files to continue with
$anadens
calc my_favourite_diffden from
1d0 cc1td-adc2-xs-1a1-002
-1d0 cc1td-cc2-gs-1a1-001
$pointval

2. I tried to apply a recommendation from your forum that is a "max setting" for me
$excitations
 irrep=a multiplicity=1  nexc=8  npre=8  nstart=8
 spectrum states=all operators=diplen,qudlen
 exprop  states=all  operators=diplen,qudlen
 xgrad states=(a 1-3)
$response
  semicano
  fop relaxed  operators=diplen,qudlen
This calculation was more than twice longer in comparing with the min setting but returned me several interesting files
cc2-gsdn-1a-000-total.cao and 3 files of
cc2-xsdn-1a-00*-total.cao, where *=1,2,3 that corresponds to number from xgrad states=(a 1-3)
The job run CC2 instead of I required ADC(2) that is strange
$ricc2
adc(2)
Nevertheless, I used these files to try EDD according to  the $anadens described above.
The results were similar "Difference densities based on natural transition orbitals"
I could obtain such difference because CCRE0-1--1---* are created even with the min setting.

3. Just an extra question is for my common knowledge. Could you explain me the meanings of the following files:
adcp2-tm0f-1a-00* which appeared with my 1st min setting and even I tried to use them for EDD and obtained quite reasonable pictures;
cc2-tmf0-1a-00* which appeared with the max setting and has the main difference of tmf0 vs tm0f. {I required adc(2) but have got cc2}
cc2-xs1d-ur-1a-00*  which also appeared with the max setting. {the same calculation where I required adc(2) but have got cc2.}

4. The last note is about MOs. There is an inconsistency between the required numbering of states and returned data.
For instance, $pointval mo 33,34 could return 33a and 35a, though, as I checked, these 33a and 35a correspond to these numbers.

I apology for long detailed questions but I should describe all my attempt to present the most clear picture.
Thank you for any incoming answer.

 

valienpo

  • Newbie
  • *
  • Posts: 4
  • Karma: +0/-0
Re: EDD visualization
« Reply #1 on: February 22, 2019, 03:36:42 am »
Well, I see my question was too long.
The main question is how should I reorganize the control file in order to calculate difference densities based on canonical MOs?
The manual recommendations don't work because there are not necessary files after ADC(2) calculation of excited states.
EDD requires to have a file  cc1td-<type>-<mult><irrep>-<number>
I tried minimum
$excitations
  irrep=a nexc=12
  spectrum states=all operators=diplen,qudlen
then added
 xgrad states=(a 1-3)
that allowed me to get two new types of files
adcp2-xsdn-1a-00*-total.cao
adcp2-xs1d-ur-1a-00*.cao
but no file with *gs*
When I involved
 exprop  states=all  operators=diplen,qudlen
calculation switched the to CC2 level automatically but some *gs* file with appeared.
I have no other ideas and can't find a help.
Could anyone give me a fruitful advise?
Thank you!
 

Arnim

  • Developers
  • Sr. Member
  • *
  • Posts: 228
  • Karma: +0/-0
Re: EDD visualization
« Reply #2 on: February 26, 2019, 10:39:45 am »
Hi,

a minimal input would look like this.
For the relaxed excited state density add this: 
$ricc2
  adc(2)
  geoopt model=adc(2) state=(a 1)
$excitations
  irrep=a nexc=1
The file with the relaxed density is adcp2-xsdn-1a-001-total.cao.
And additionally for the ground state density this:
$response
  gradient
The relaxed density is then called mp2-gsdn-1a-000-total.cao.

Cheers,

Arnim



valienpo

  • Newbie
  • *
  • Posts: 4
  • Karma: +0/-0
Re: EDD visualization
« Reply #3 on: February 28, 2019, 04:06:24 am »
Thank you, Armin!
I will try.
If I want to get more than 1 EDD for excited states at the same time
should I write, for example state=(a 1-3)?
Which difference is between
geoopt model=adc(2) state=(a 1)
in the section $ricc2
and
 xgrad states=(a 1-3)
in the section $excitations ?
I started both to check but, anyway, your answer is interesting
Vladimir

valienpo

  • Newbie
  • *
  • Posts: 4
  • Karma: +0/-0
Re: EDD visualization
« Reply #4 on: February 28, 2019, 02:36:47 pm »
Armin,
I used your scheme and obtained density like
cc2-gsdn-1a-000-total.cao
but no density of excited states
Now I'm waiting results after
xgrad states=(a 1-3)
in the section $excitations