next up previous contents index
Next: Ground state-excited state transition Up: Coupled-cluster calculations, CC Previous: Cubic response functions: *CCCR   Contents   Index


Calculation of excitation energies: *CCEXCI

In the *CCEXCI section the input that is specific for coupled cluster linear response calculation of electronic excitation energies is given. Coupled cluster linear response excitation energies are implemented for the iterative CC models CCS, CC2, CCSD, and CC3 for both singlet and triplet excited states. For singlet excited states the non-iterative models CC(2)(=CIS(D)) and CCSDR(3) are also available. For understanding the theoretical background for some aspects of the CC3 calculations consult also Ref.
citeChristiansen:JCP105,Hald:JCP113,Christiansen:JCP103,Hald:JCP115.

\fbox{
\parbox[h][\height][l]{12cm}{
\small
\noindent
{\bf Reference literature:...
...wblock {\em J.~Chem.~Phys.}, {\bf 113},\hspace{0.25em}7765, (2000).
\end{list}}}

.CC2PIC
Functions as CCSPIC but the picking is based on CC2 excitation energies.
.CCSDPI
Functions as CCSPIC but the picking is based on CCSD excitation energies.
.CCSPIC

READ(LUCMD,*) OMPCCS
Keyword for picking a state with a given CCS excitation energy. Optimize a number of states in CCS with different symmetries and this option will pick the one in closest correspondence with the given input excitation energy, and skip the states at higher energies and other symmetries in the following calculation. Useful for example in numerical Hessian calculations on excited states.

.MARGIN
Specifies the maximum allowed deviation of the actual excitation energy from the input excitation energy when using CCSPIC, CC2PIC and CCSDPI.

.NCCEXC

READ (LUCMD,*) (NCCEXCI(ISYM,1),ISYM=1,MSYM)

Give the number of states desired. For singlet states only, one single line is required with the number of excitation energies for each symmetry class (max. 8).
If also triplet states are desired an additional line is given in same format.

.NOSCOM
For CC3 calculations only: indicates that no self-consistent solution should be seeked in the partitioned CC3 algorithm.

.OMEINP

READ (LUCMD,*) (NOMINP(ISYM,1),ISYM=1,MSYM)
DO ISYM = 1, MSYM
DO IOM = 1, NOMINP(ISYM,1)
READ (LUCMD,*) IOMINP(IOM,ISYM,1),EOMINP(IOM,ISYM,1)
ENDDO
ENDDO

A way to provide an input omega for the partitioned CC3 algorithm or restrict the self-consistent solution to specific states. If OMEINP is not specified the program uses the best choice available to it at that moment based on previous levels of approximations (CCSD or even better CCSDR(3)) and calculates all states as given by NCCEXC. IOMPINP is 1 for the lowest excited state of a given symmetry, 2 for the second lowest etc.
By giving an 0.0 input excitation energy (as EOMINP) the program takes the best previous approximation found in this run - otherwise the user can specify a qualified guess (perhaps from a previous calculation which is now restarted).

.THREXC

The threshold for the solution of the excitation energies and corresponding response eigenvectors. The threshold is the norm of the residual for the eigenvalue equation. (Default: 1.0D-04).

.TOLSC
For CC3 calculations only: Set tolerance for excitation energies for obtaining a self-consistent solution to the partitioned CC3 algorithm. Tolerance refers to the eigenvalue itself in the self-consistency iterations of the default solver for CC3. Not used in DIIS solver (see .R3DIIS). (Default: 1.0D-04).

.R3DIIS
Use DIIS solver for CC3. This solver only scales linearly with the number of excited states in comparison to the default solver which scales quadratically. However this solver might fail in cases with states dominated by double or higher excited determinants. (Default: OFF).


next up previous contents index
Next: Ground state-excited state transition Up: Coupled-cluster calculations, CC Previous: Cubic response functions: *CCCR   Contents   Index
Dalton Manual - Release 1.2.1