[CP2K-user] [CP2K:10343] transition dipole moment

[Nuri Yazdani]

Hi Xiaoming,
You are right, the MOs are not normalised. However, and I am not 100% sure 
about this, I think that they shouldnt necessarily be normalized, since the 
basis functions (i.e. the gaussian orbital on each atom) are not orthogonal 
to one another. This is where the overlap matrix comes in..  
So, I think, if one wants to construct for example the electron density 
from the MOs and the basis functions, one should use the MO coefficients as 
is (i.e. do not renormalise the MOs yourself). 
I think, and again I am not totally confident about this, that if you print 
out the overlap matrix S, you can get the dipole moment (missing some 
prefactors, and for example the x component) by

A_ij,x = <mo_i| S x |mo_j>

The problem that I have is that I have already run really large MD 
calculations, and I only was printing out the MOs. I think the best I can 
do then is to use

A_ij,x ~ <mo_i| x |mo_j>

which from my understanding, ignores any contribution to A_ij,x from 
orbitals on two different atoms. For my system, A_ij,x will be dominated by 
s -> p transitions on one particular atomic species, so this approximation 
should be OK, but it may fail terribly for systems where the transition 
goes from an orbital on one atom type to an orbital on a different atom 
type..

On Wednesday, October 10, 2018 at 2:49:34 PM UTC+2, Xiaoming Wang wrote:
>
> Hi Nuri,
>
> Are you using the MOS? The mo coefficients are not normalized, how do you 
> solve this problem?
>
> Best,
> Xiaoming
>
> On Tuesday, October 9, 2018 at 7:31:19 AM UTC-4, Nuri Yazdani wrote:
>>
>> Hi Prof. Hutter,
>>
>> I am trying to do such calculations. I have printed out the MOS from my 
>> calculations to do this, however, I am quite confused by the spatial 
>> representation and ordering of the d-orbitals... 
>>
>> I am using the MOLOPT-DZVP basis set. For example, for Cs 
>> DZVP-MOLOPT-SR-GTH 1 2 0 2 6 3 2 1, I have 3x S, 6x P, and 5x D orbitals 
>> (this counting is also consistent with the length of my MOSs).
>>
>> The S and P orbitals I can construct from the basis set, and (if CGS are 
>> a linear combination of gaussians) the p orbitals are px = x*CGS, py = 
>> y*CGS, pz =z*CGS, however, what is the ordering and form of the d 
>> orbitals?? i.e. what is r^2 equal to in each of the 5 last lines of this 
>> page: https://www.cp2k.org/basis_sets? 
>>
>> Cheers,
>> Nuri
>>
>>
>>
>>
>> On Thursday, May 31, 2018 at 11:28:04 AM UTC+2, jgh wrote:
>>>
>>> Hi 
>>>
>>> this property is on our TO DO list. However, I cannot say when 
>>> it will become available. 
>>>
>>> If you want to calculate it yourself from the MO cube files 
>>> you need to use the Berry phase algorithm, meaning you need 
>>>
>>> mu = IMAG LOG <phi_a | exp(i*k*r) |phi_b> 
>>>
>>> However, this can be calculated much more efficiently from 
>>> the atomic basis functions. 
>>>
>>> regards 
>>>
>>> Juerg Hutter 
>>> -------------------------------------------------------------- 
>>> Juerg Hutter                         Phone : ++41 44 635 4491 
>>> Institut für Chemie C                FAX   : ++41 44 635 6838 
>>> Universität Zürich                   E-mail: hut... at chem.uzh.ch 
>>> Winterthurerstrasse 190 
>>> CH-8057 Zürich, Switzerland 
>>> --------------------------------------------------------------- 
>>>
>>> -----cp... at googlegroups.com wrote: ----- 
>>> To: cp2k <cp... at googlegroups.com> 
>>> From: Xiaoming Wang 
>>> Sent by: cp... at googlegroups.com 
>>> Date: 05/24/2018 05:07PM 
>>> Subject: [CP2K:10343] transition dipole moment 
>>>
>>> Hi, 
>>>
>>> Is it possible for CP2K to output the transition dipole moment <phi_A | 
>>> r | phi_B>  between two KS states, say HOMO to LUMO transition? 
>>> I tried to calculate the integral by printing the MO_CUBE files which 
>>> are the phi_A and phi_B and then do the integration. This is fine for 
>>> molecules or in other word non-periodic systems. But for periodic 
>>> systems, the result was not as expected, maybe due to the position 
>>> operator is ill-defined in that case. So how to evaluate the transition 
>>> dipole integral for periodic systems with the MO_CUBE files known? 
>>>
>>> Best, 
>>> Xiaoming     
>>>   -- 
>>>  You received this message because you are subscribed to the Google 
>>> Groups "cp2k" group. 
>>>  To unsubscribe from this group and stop receiving emails from it, send 
>>> an email to cp2k+... at googlegroups.com. 
>>>  To post to this group, send email to cp... at googlegroups.com. 
>>>  Visit this group at https://groups.google.com/group/cp2k. 
>>>  For more options, visit https://groups.google.com/d/optout. 
>>>   
>>>
>>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20181012/15c98a5d/attachment.htm>