Molcas Forum

David

Member

Registered: 2017-05-16

Posts: 85

[SOLVED] CASPT2 energies in vacuo differ a lot from that in crystal

Hi Molcas developers,

In a CASPT2 single point calculation, I included backgroud charge in ESPF module to simulate aggregate state in crystal, however, results seem very unreasonable, S0-S1 energy gap are 27.5, 16.5, 21.7 kcal/mol at the CASSCF, CASPT2, MS-CASPT2 level, respectively. As a comparison, I do the same computation in vacuo ( background charge isn't considered),  S0-S1 energy gap  are 93.7, 61.3, 61.5 kcal/mol at the CASSCF, CASPT2, MS-CASPT2. I have also tested this job with TDDFT method by G09, the difference in energy is about 2 kcal/mol within and without of  background charges. Can you help me fix this problem? Thank you!

&GATEWAY
 coord 
 s1.Opt.xyz

 basis 
 ano-rcc-vdzp

 group 
 nosym 

 ricd
END OF INPUT 

&SEWARD
END OF INPUT 

&ESPF
 multipoleorder
 0

 external
23506
 -2.420537 -10.684757  -2.068428  -0.370507   0.0000000E+00  0.0000000E+00  0.0000000E+00 angstrom
 -2.260866 -10.621717   2.636676  -0.132357   0.0000000E+00  0.0000000E+00  0.0000000E+00 angstrom
 -7.044123  -7.619631  -3.088281   0.110441   0.0000000E+00  0.0000000E+00  0.0000000E+00 angstrom
 -2.591088  -8.088226  -6.976354   0.116782   0.0000000E+00  0.0000000E+00  0.0000000E+00 angstrom
 -0.084204  -7.069786  -3.593425  -0.400579   0.0000000E+00  0.0000000E+00  0.0000000E+00 angstrom
....
....
....
END OF INPUT 

&RASSCF
 nactel
 12 0 0

 charge
 0

 ras2
 10

 symmetry
 1

 spin
 1

 ciroot
 8 8 1

 lumorb
END OF INPUT

&CASPT2
 multi
 2 1 2

 ipea
 0.0

 imag
 0.2
END OF INPUT

Ignacio

Administrator

From: Uppsala

Registered: 2015-11-03

Posts: 1,085

Re: [SOLVED] CASPT2 energies in vacuo differ a lot from that in crystal

Could it be that you believe you are giving the charge coordinates in angstrom, but you are actually giving them in atomic units? According to the documentation, if you want them in angstrom, you have write that next to the total number, not next to each charge. (Or just convert your coordinates to bohr.)

David

Member

Registered: 2017-05-16

Posts: 85

Re: [SOLVED] CASPT2 energies in vacuo differ a lot from that in crystal

Thank you very much, Ignacio! Based on your advice, I added "angstrom" in number n and final results turn normal. I have ever used background charge in bohr unit, so this time I directly replace bohr with angstrom without further consideration, sorry for that. Again, Thank you for your help!

Could it be that you believe you are giving the charge coordinates in angstrom, but you are actually giving them in atomic units? According to the documentation, if you want them in angstrom, you have write that next to the total number, not next to each charge. (Or just convert your coordinates to bohr.)

Last edited by David (2020-04-19 03:30:38)