Excited State of Lithium Atom

[ZHIBO LIU]

Hello Molpro users, I am very new to molpro and a part of my study needs to compute the energy for each excited state of Lithium. In this case I want the electron in 2s orbital to be excited to 2p,3s,3p,3d,4s,4p,4d,4f. The result I got sort of matches the value reported by experiments, however when I give "wf,3,1,1" and call for 8 states in casscf I think it returns me the energy for 2s (ground), 3s, then 2 degenerate 3d, 4s, and after that the energy difference jumps directly from ~4.65eV to ~6.63eV which seems like a 5d orbital I guess? I think I failed to get 4d and 4f (wf input card is changed for f orbital in another run). Is it the case that 4d and 4f being skipped? If so what could be the cause and solution for this? Thank you in advance!

The input file I used looks like:

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
***,test1



gprint,orbital!,civector



geometry = {Li}

basis=aug-cc-pCV5Z
 
WF_1=3
WF_2=1
WF_3=1
N_state=8


uhf

{casscf

WF,WF_1,WF_2,WF_3
occ,9,5,5,2,5,2,2,1 

!closed,1.1

expec2,lxx,lyy,lzz

state,N_state

!NATORB,state=1.1
!NATORB,state=2.1
!NATORB,state=3.1
!NATORB,state=4.1
!NATORB,state=5.1
!NATORB,state=6.1
!NATORB,state=7.1
!NATORB,state=8.1
!NATORB,state=9.1
!NATORB,state=10.1

}

my_eng_cas=energy(:)

{mrci
occ,9,5,5,2,5,2,2,1
WF,WF_1,WF_2,WF_3
STATE,N_state
}


my_eng_mrci=energy(:)
my_mrcid=energd0(:)



table,my_eng_cas,my_eng_mrci,my_mrcid

---

%%%%%%%%%%%%%%%%%%%%%%%%%%

[Alexander MITRUSHCHENKOV]

Hi Liu,

I guess you would need much bigger basis (especially diffuse d/f orbitals) to describe correctly the Rydberg states. Please also note that for Li (3 electrons) the excited states are not simply one orbital excitations and starting from some energy you will have a mixture.

Best wishes,

Alexander

--
You received this message because you are subscribed to the Google Groups "molpro-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to molpro-user...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/molpro-user/a1d71b7a-debe-498f-85f3-1db4e7407068n%40googlegroups.com.

-- 
Dr. Alexander Mitrushchenkov, IGR
Laboratoire de Modélisation et Simulation Multi Echelle
UMR 8208 CNRS
Univ Gustave Eiffel
5 Bd Descartes
77454 Marne la Vallée, Cedex 2, France

Phone:    +33(0)160957316
Fax:      +33(0)160957320
e-mail:   Alexander.Mi...@univ-eiffel.fr

[Jacky LIEVIN]

Dear Liu,

As pointed out by Alexander, standard basis sets like aug-cc-pCV5Z are not optimized for the description of the Rydberg states.

The following paper and references therein should help you.

Launoy, T., Loreau, J., Dochain, A., Liévin, J., Vaeck, N., & Urbain, X. (2019). Mutual Neutralization in Li+-D- Collisions: A Combined Experimental and Theoretical Study. The Astrophysical journal, 883(1), 85. 

Best wishes,

Jacky

To view this discussion on the web, visit https://groups.google.com/d/msgid/molpro-user/65aed58b-5f00-c4d8-0064-1aa65f901325%40univ-eiffel.fr.

[ZHIBO LIU]

Thank you very much Mr. Mitrushchenkov and Mr. Lievin for the suggestion, I tried the basis by Gim&Lee (table 1) that you cited in the paper mentioned. Now the energies turn out much more reasonable and closer! But please forgive me for not understanding the technical details in your paper, the ACV5Z+G basis that you mentioned is obtained from a package called AUTOSTRUCTURE, after some searching I found there are limited information about this package as well as another package called SUPERSTRUCTURE, is this an external package that is apart from Molpro? 

Also thank you really very much for repling me Mr. Lievin, I was actually choosing the Lithium atom as the starting case becasue your paper was the only one that I found that reads very detailed especially for me - a beginner. And indeed, the final goal for me is to obtain the nonadiabatic coupling that will be, the same, used in a Landau-Zener equation for evaluating the transistion probability. If you don't mind, could you please also give me some suggestions about the specification of active space when scanning for the PEC of the 1Σ+ state Li~H? If I understood it right from what molpro manual suggests, the "occ" is the input line that controls the active orbitals, their example of the use of "occ" is done with O2 molecule which I think is not very hard to follow since it's written as the MO list, but when it comes to the case of Li~H, I tend to consider that the two 1s electrons are not a part of the MOs, so would the "occ" in this case to be the combined AOs (the two 1s electron in Li) and MOs (the two valance electrons for Li and H)? I am not quite sure what would be the "occ" that discibe each excited state under 1Σ+. I noticed that it is common to give a larger "occ" input then the "occ" that coresponds to the expected electron configuration by AO or MO theory, so would it desirable if to use a large "occ"  (if disregard the computational load since it is diatomic) for a more acurate result when not sure about the correct "occ".

Thank you very much!

Best Regards,

[Jacky LIEVIN]

Hi,

AUTOSTRUCTURE is not interfaced with Molpro. It is an independent program performing atomic calculations in basis sets of Slater type orbitals. For each atomic state it provides linear combinations of Slater functions describing each atomic orbital, which can the be fitted to gaussian type functions by least square fit. These gaussian functions can be added to the valence ACV5Z basis set to allow the description of the Rydberg states.

I don't have experience in using AUTOSTRUCTURE. In our work on Li+ + D- (and on HeH+ see Loreau 2010 in the LiD paper) we didn't performed these calculations by ourselves, but asked to colleagues who are familiar with AUTOSTRUCTURE to run these calculations.

Please note that an alternative solution for describing Rydberg states is to add to the Molpro basis input even tempered gaussian basis sets describing s, p, d, f… atomic symmetries, whose parameters are to be optimized to match with atomic results. An example is given in Laruelle et al. J. Phys. Chem. A113 13210-13220 (2009) and Boyé-Péronne at al. J. Chem. Phys. 141, 174317 (2014) for the Rydberg states of acetylene and vinylidene.

As for the question about « occ », the rule is to include in the active space all orbitals which can have a significant weight in the calculated states.

In your input for the Li atom, occ,8,5,5,2,5,2,2,1 seems OK for describing the 1s, 2s, 3s, 4s, 2p, 3p, 4p, 3d, 4d and 4f orbitals correlated to the symmetries of the D2h subgroup used by Molpro.

However, for describing the ground state and the doublet states resulting from the excitation of the 2s orbital to 2p,3s,3p,3d,4s,4p,4d,4f, you should include in the state averaged CASSCF calculation all the degenerated components of the corresponding electronic states (also correlated to D2h):

wf,3,1,1;state,7;

wf,3,2,1;state,4;

wf,3,3,1;state,4;

wf,3,4,1;state,2;

wf,3,5,1;state,4;

wf,3,6,1;state,2;

wf,3,7,1;state,2;

wf,3,8,1;state,1;

In your run limited to the Ag symmetry only (wf,3,1,1;state,8), it is normal that it only returns the energy for 2s (ground), 3s, then 2 degenerate 3d and 4s. P and F states belong to anti-symmetric u species.

In diatomic calculations like LiH, you should do exactly the same, but correlate the atomic orbitals of interest of Li and H to the C infinity v subgroup symmetry.

Best wishes,

Jacky

To view this discussion on the web, visit https://groups.google.com/d/msgid/molpro-user/f75d65d3-6c8d-4038-a5c4-853801395b6dn%40googlegroups.com.