Mulliken charges for MP2 ground state from RICC2 module

[acapobia]

I'm sorry if this matter was partially addressed in:

http://www.turbo-forum.com/index.php?topic=185.0  (Topic: Charge analysis on excited state with RICC2)
http://www.turbo-forum.com/index.php?topic=180.0  (Topic: calculating charges on excited state)
http://www.turbo-forum.com/index.php?topic=285.0  (Topic: -proper crashes)
        
but something remains unclear to me. I have big molecules (say 50 atoms) which I studied at ri-mp2 level using the very powerful parallel ricc2 module. Now I need Mulliken charges and I wish to extract them from the current density (instead of computing it again via the rimp2 module and then running rimp2 -proper) which, according to TM manual, is stored in the file mp2-gsdn-1a-000-total.cao.

So, following the manual and the suggestions from the above links, after mp2 calculation via ricc2 was done,
I added the following keys in control

$anadens
calc mulliken.MP2 from
mp2-gsdn-1a-000-total.cao
$pop

and then, after switching to serial mode
I tried both
ricc2 -fanal > output.fanal
and
ricc2 -proper> output.proper
but my outputs  were:
--------------------------------------------------------
  New density "mulliken.MP2" is
  calculated from the following factors and densities:
--------------------------------------------------------
     1    1.000000"mp2-gsdn-1a-000-total.cao"
--------------------------------------------------------

Now analysing density file "mulliken.MP2"


Population analysis disabled for this density ... skipped

Finally my questions.  Is it possible to get Mulliken charges from ground states MP2 densities obtained by ricc2 computations and, if so, what is the right procedure ?

Thanks in advance

Amedeo

[Arnim]

Hi,

it is by far more easier than that to get these results. 

If know, that you want to look at the Mulliken charges, already before you
start your geometry optimizations, simply add $pop in the control file.
Then start your optimization using ricc2 (serial or parallel) and you
will find the analysis in the job.last file.

Cheers,
Arnim

[acapobia]

I'm very thankful for your quick and kind reply, however I do not need to optimize the geometries of my molecules, nevertheless I have already computed relaxed dipole moments adding
$response
  semicano
  fop relaxed  operators=xdiplen,ydiplen,zdiplen
in my control file, so there is a density available to further 'manipulation';  I also tried ricc2 > ricc2.out with the $pop key set in control but I didn't get any charge.
Now, without starting any not-needed (to me) optimization, how can I get Mulliken charges from available MP2 density?

Thanks again.

Amedeo

[Arnim]

This should do the trick:

kdg end
kdg ricc2
echo "\$ricc2" >> control
echo "  mp2" >> control
echo "  geoopt model=mp2" >>control
echo "\$pop" >>control
echo "\$end" >>control
ricc2 | tee ricc2.out

[acapobia]

OK, perhaps I was unclear in my previous post and anyway I wish to  thank Arnim once again. I didn't look for a script which could automatize some procedure, but I want to know if density needed for  first order one-electron properties calculations (via the $response key) is enough to make TM to compute charges. Strictly speaking I computed energies and relaxed dipoles of a lot of molecules 50-60 atoms each and it would be very annoying to run once again the same rimp2 computations (via ricc2) setting geoopt in control, just to get Mulliken charges.

However I begin to suspect this will be indeed the case. In fact I discovered the following topic  http://www.turbo-forum.com/index.php?topic=221.0, in which Christof addresses the whole question.

I apologize if I duplicated a topic, but really I missed it.

Thank you all.

Amedeo