How to calculate phosphorescence and spin-orbit coupling

Find answers or ask questions regarding Dalton calculations.
Please upload an output file showing the problem, if applicable.
(It is not necessary to upload input files, they can be found in the output file.)

Post Reply
blueyangliu
Posts: 5
Joined: 06 Apr 2015, 14:56
First name(s): Bao
Last name(s): Yang
Affiliation: CCZU
Country: China

How to calculate phosphorescence and spin-orbit coupling

Post by blueyangliu » 22 Apr 2015, 06:52

Dear All,
I want to calculate the phosphorescence and SOC of transition-metal Pt complexes. listed below are my input and output files, I am not sure it is right or wrong? Could anyone give me some advice?
**DALTON INPUT
.RUN RESPONSE
**INTEGRALS
.SPIN-ORBIT
**WAVE FUNCTION
.DFT
B3LYP
**RESPONSE
*QUADRATIC
.PHOSPHORESCENCE
.ROOTS
1
**END OF DALTON INPUT

the part of output file
Phosphorescence - length gauge:
Oscillator strength (/2PI) (H_SO) 3.848929E-09
Dipole strength [a.u.] (H_SO) 2.115217E-06
Dipole strength E-40 [esu**2 cm**2] 1.366530E-01
Total transition rate (H_SO) 7.617659E-02 s-1
Total phosphorescence lifetime (H_SO) 1.312739E+01 s

the lifetime is much longer than experiment result, and what is the "Total transition rate"? how to analyze the spin-orbit coupling and the spin-orbit constant?

olav
Posts: 143
Joined: 28 Aug 2013, 06:20
First name(s): Olav
Last name(s): Vahtras

Re: How to calculate phosphorescence and spin-orbit coupling

Post by olav » 22 Apr 2015, 07:14

The input is correct. DFT is not known for its predictive power when it comes to transition-metal compounds, in particular for non-standard properties like this.

Please look up "transition rate" in any textbook on quantum mechanics.

If you by SOC mean the singlet-triplet transition matrix element, it is a separate type of calculation in Dalton, see e.g. the rsp_lrso test case in the distribution.

Regards,
Olav

You

blueyangliu
Posts: 5
Joined: 06 Apr 2015, 14:56
First name(s): Bao
Last name(s): Yang
Affiliation: CCZU
Country: China

Re: How to calculate phosphorescence and spin-orbit coupling

Post by blueyangliu » 22 Apr 2015, 15:18

Hi Olav, Thanks a lot for your reply.
Actually, I want to calculate the singlet-triplet transition matrix element. So I checked the rsp_lrso test and searched the keywords in Dalton2015 manual, My molecule has 298 electrons and 858 orbitals, I think it is a big molecule for soc calculation, so how to set the ".INACTIVE SPACE" and ".CAS SPACE" if we take account of calculation time and accuracy? Thanks in advance


**DALTON INPUT
.RUN RESPONSE
.NEWTRA
**INTEGRALS
.SPIN-ORBIT
**WAVE FUNCTIONS
.MCSCF
*CONFIGURATION INPUT
.SYMMETRY
1
.SPIN MULTIPLICITY
1
.INACTIVE SPACE

.ELECTRONS
2
.CAS SPACE

*CI VECTOR
.PLUS COMBINATIONS
*OPTIMIZATION
.DETERMINANTS
**RESPONSE
.TRPFLG
*LINEAR
.SINGLE RESIDUE
.SPIN-ORBIT
**END OF DALTON INPUT

taylor
Posts: 545
Joined: 15 Oct 2013, 05:37
First name(s): Peter
Middle name(s): Robert
Last name(s): Taylor
Affiliation: Tianjin University
Country: China

Re: How to calculate phosphorescence and spin-orbit coupling

Post by taylor » 22 Apr 2015, 15:48

I am going to make some suggestions here that I hope you will take as constructive thinking, rather than criticism. But it seems to me that launching immediately into trying to calculate phosphorescence lifetimes for a Pt complex (you do not give any information as to the structure of the molecule) that your current calculation involves almost 900 basis functions and 300 electrons, and deciding that because DFT/B3LYP gives unsatisfactory results you will switch to CASSCF is somewhat akin to someone who decides to learn to drive, and having had a negative experience trying to learn by driving a 40-ton 18-wheeler truck, thinks that switching to a Formula 1 race car is a better alternative. As just one issue to consider, an MCSCF calculation will not run integral-direct, and you will therefore have to cope with disk storage of integrals. If most of the integrals are nonzero (again, without knowing your molecule it is impossible to say) you will need over 1TB of disk for the integrals and labels, plus additional disk needed to sort and transform integrals. Furthermore, this method is not parallelized and the computer time will likely be nontrivial.

My thinking is that you would be much better advised to find a model Pt-complex that exhibits at least some of the features of your target molecule, but which is considerably smaller. Replacing larger ligands with "stubs", such as replacing, say, triphenylphosphines with PH$_3$ groups, or an S-bound cysteine with an SH group, should at least provide a start. I stress this because CASSCF and MCSCF more generally is very far from a black-box approach! It is often necessary to look at expanding or shrinking the active space to get a balanced (I say "balanced" here, not necessarily "accurate"!) description of different electronic states. In addition, model complexes can often be constructed that have higher symmetries than C$_1$, which not only reduces the computation time but makes the identification of particular excited states much easier.

Once you have investigated such a model at a higher level, like CASSCF, you can then compare it to DFT approaches. If you find one that agrees reasonably well with your model system CASSCF calculation, it is at least plausible, although by no means guaranteed, that this agreement would hold for your target molecule. If you find no DFT approach agrees with the higher-level calculation for your model system, the odds are, unfortunately, that no DFT approach will be acceptable for your target system.

This general approach --- find a model system that is at least some sort of approximation to your desired system but which is amenable to a range of high-level treatments, and use this model to calibrate more approximate methods, is something that many of us, including myself, have been advocating for many years. Indeed, if you can get hold of a copy of the lecture notes from the biennial European Summerschool in Quantum Chemistry, you will find a detailed article of mine on how to use this philosophy. The notes are now old and omit many methods one might use today, but the philosophy, I submit, is still sound...

Best regards
Pete

Bernd S.
Posts: 17
Joined: 19 Feb 2015, 13:31
First name(s): Bernd
Last name(s): Schimmelpfennig
Affiliation: KIT
Country: Germany

Re: How to calculate phosphorescence and spin-orbit coupling

Post by Bernd S. » 22 Apr 2015, 16:06

Peter has summed up already a lot of important points but I dare to add one more:

For a Pt complex you need to include relativity in some form and in case you go for the cheap way by using a pseudo-potential the spin-orbit integrals provided by DALTON will not be generated with the correct Hamiltonian. Even with an all-electron basis set and scalar relativity you should be careful about the correct SO-operator. This is independent of using MCSCF or DFT.

Best,


Bernd

taylor
Posts: 545
Joined: 15 Oct 2013, 05:37
First name(s): Peter
Middle name(s): Robert
Last name(s): Taylor
Affiliation: Tianjin University
Country: China

Re: How to calculate phosphorescence and spin-orbit coupling

Post by taylor » 22 Apr 2015, 16:10

Thanks very much to Bernd for the additional information: I was myself a bit worried about this aspect but felt it was too far outside my expertise to be commenting on. As you see, there is very much to consider in doing these calculations!

Best regards
Pete

blueyangliu
Posts: 5
Joined: 06 Apr 2015, 14:56
First name(s): Bao
Last name(s): Yang
Affiliation: CCZU
Country: China

Re: How to calculate phosphorescence and spin-orbit coupling

Post by blueyangliu » 23 Apr 2015, 02:15

Thank you all, Pete and Bernd. What about ADF in calculating spin-orbit of Pt complexes ?

Bernd S.
Posts: 17
Joined: 19 Feb 2015, 13:31
First name(s): Bernd
Last name(s): Schimmelpfennig
Affiliation: KIT
Country: Germany

Re: How to calculate phosphorescence and spin-orbit coupling

Post by Bernd S. » 23 Apr 2015, 08:46

ADF should be able to do it with TD-DFT including SO-coupling either from the beginning or as a perturbation.

Post Reply

Who is online

Users browsing this forum: No registered users and 2 guests