CASSCF phosphorescence

If you think you found a bug this is where you can report it
Post Reply
frj
Posts: 11
Joined: 09 May 2014, 08:37
First name(s): Frank
Last name(s): Jensen
Affiliation: Aarhus University
Country: Denmark

CASSCF phosphorescence

Post by frj » 18 May 2014, 19:29

I have some 'unusual' results for the transition moments in a phosphorescence calculation, that hints at a Dalton problem.
The LiH tests supplied with Dalton check out OK.
The actual system is MgH+ with an uncontracted aug-cc-pVDZ run in C2v (z-axis along the molecular axis). The problem is basis set independent.
Transition moments are labelled as <dipole,SO>. By symmetry, the <x,y> moment should be the same as <y,x> except for a sign. Similarly, the <z,x> should match <z,y>, and <x,z> should match <y,z> in magnitude. This checks out at the HF level, both with SO and with the MNF-SO approximation. It is also OK with a full-CASSCF wavefunction and the SO-MNF, but not with the 'real' SO integrals.
Besides the symmetry problem, the magnitude of the transition moments also depends on which and how many ROOTS that are specified in the *QUADRA section. Specifying e.g. 0 1 0 0 gives different results than 1 1 1 1, and also different if requesting higher roots, like 2 2 2 0, for the same transition of the same symmetry.
The differences are not just numerics. The <x,y> moment is 0.000656 while the <y,x> is 0.009559. The <x,z> is 0.462241 while the <y,z> is 0.117397. The corresponding MNF-SO values are 0.000010 and 0.000463. The HF results suggest that the MNF approximation underestimates the value by roughly a factor of 10.
Excitation energies are OK.

.mol and .dal files are given below

Frank

.DAL:
**DALTON INPUT
.RUN RESPONSE
**INTEGRALS
.SPIN-ORBIT
**WAVE FUNCTIONS
.HF
.MCSCF
*SCF INPUT
.THRESH
1.0d-9
.DOUBLY OCCUPIED
4 1 1 0
*CONFIGURATION INPUT
.SYMMETRY
1
.SPIN MUL
1
.INACTIVE
1 0 0 0
.ELECTRONS
10
.CAS SPACE
5 2 2 0
*OPTIMIZATION
.STATE
1
**RESPON
*QUADRA
.PHOSPHORESCENCE
.ROOTS
1 1 1 1
**END OF



.MOL:
INTGRL
basis set exponent optimization
CO
2 0 1 1.0d-15
1. 1 2 1 1
H 0.00000 0.000000 0.000000
F 5 0
13.0100000
1.9620000
0.4446000
0.1220000
0.0297400
F 2 0
0.7270000
0.1410000
12. 1 3 1 1 1
Mg 0.00000 0.000000 2.000000
F 13 0
47390.0000000
7108.0000000
1618.0000000
458.4000000
149.3000000
53.5900000
20.7000000
8.3840000
2.5420000
0.8787000
0.1077000
0.0399900
0.0148800
F 9 0
179.9000000
42.1400000
13.1300000
4.6280000
1.6700000
0.5857000
0.1311000
0.0411200
0.0093500
F 2 0
0.1870000
0.0595000

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

Re: CASSCF phosphorescence

Post by olav » 20 May 2014, 06:16

Hi Frank,
Please provide the mol file as an attachmend instead - some of the formatting is lost in the forum. And one of your output files as well.

Cheers,
Olav

frj
Posts: 11
Joined: 09 May 2014, 08:37
First name(s): Frank
Last name(s): Jensen
Affiliation: Aarhus University
Country: Denmark

Re: CASSCF phosphorescence

Post by frj » 20 May 2014, 07:09

Files attached.
Attachments
test14.out
(123.5 KiB) Downloaded 369 times
test14.mol
(1.46 KiB) Downloaded 384 times
test14.dal
(318 Bytes) Downloaded 433 times

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

Re: CASSCF phosphorescence

Post by olav » 21 May 2014, 18:11

Frank,
The problem is really that the program should abort in response because triplet response requires a determinant representation of the wave function. CSFs do not work.
When I add *OPTIMIZATION.DETERMINANTS I observe that the inconsistencies you describe disappear, both regarding x-y symmetry and number of roots requested.

Regards,
Olav

frj
Posts: 11
Joined: 09 May 2014, 08:37
First name(s): Frank
Last name(s): Jensen
Affiliation: Aarhus University
Country: Denmark

Re: CASSCF phosphorescence

Post by frj » 22 May 2014, 06:50

Olav,
Yep, that solves the problem.
Thanks for the help.
Might be a good idea to check for .DETERMINANTS in the code and bomb if not. The MNF-SO results looked reasonable, but actually also change when using .DETERMINANTS

Frank

Post Reply

Who is online

Users browsing this forum: No registered users and 3 guests