iOpenShell » Q-Chem specific questions » Excitation energy is negative when using EOM-EE-CCSD?

Excitation energy is negative when using EOM-EE-CCSD?

Page: 1

Author Post
Member
Registered: Jan 2018
Posts: 2
Dear all,

I want to calculate the singlet and triplet excited states of CH3I using EOM-EE-CCSD method. The results are so weird that I cannot explain.
Here is my input.

$molecule
0 1
H 0.7515158257 -0.6022789095 -1.8859686281
C 0.0000275280 -0.0000480643 -1.3297166588
H 0.1457682315 0.9524366931 -1.8856497472
H -0.8975405428 -0.3496077564 -1.8858774957
I -0.0000005683 0.0000001805 0.1707858230
$end

$rem
INPUT_BOHR FALSE
METHOD eom-ccsd
BASIS def2-tzvp
EE_SINGLETS [2]
EE_TRIPLETS [2]

CC_SYMMETRY FALSE

CC_MEMORY 2500
MEM_TOTAL 4000

$end

I used the geometry I got from Molpro, and the same basis set. Firstly, the Hartree-Fock energy I got from Molpro is -334 a.u., but q-chem returned me -2700 a.u., which I don't understand. Secondly, when I calculate the excited states, the excitation energy is negative for the first singlet and triplet excited state.

Does anyone have any sense what might be the reason for those two strange problems?

Thanks!
Administrator
Registered: Oct 2017
Posts: 5
This clearly indicates that SCF has converged to a wrong solution. I run this calculation and confirm that the singlet excitation energy is negative. This is a good test to check if SCF converges to the global minimum, but EOM-EE is an overshot for this purpose - CIS should be enough. It looks like the default guess is wrong. From CIS calculation the leading excitation amplitude to the lowest singlet corresponds to HOMO-LUMO excitation. Therefore, you can read orbitals as a guess and re-populate them. I run it with MOM_START=1, but it did not work. Also, start from the triplet state is worth to try.

Too negative energy may indicate about a problem with ECP. Could you attach the output from Molpro?

Page: 1

iOpenShell » Q-Chem specific questions » Excitation energy is negative when using EOM-EE-CCSD?