Some examples of the namelist input used in DIRAC10.
Calculate the CCSD(T) energy and MP2 first order properties. The full calculation is given as input example 1.energy+dipole in the test directory:
&RELCCSD TIMING=T, DOENER=T, DOFOPR=T &END
&CCFOPR DOMP2G=T &END
A simple example is the oxygen molecule in which the 1s electrons are kept frozen. The active space then consists of 12 electrons, of which there are 10 closed shell electrons. The closed shell electrons go in the nonbonding 2s σg and σu orbitals and the bonding 2p σg and πu orbitals, while the 2 open shell electrons are distributed over the two antibonding 2p πg orbitals. Let us suppose that we want to take the MS=1 state as our reference and assign alpha spin to both our open shell electrons. In the double group (D2h*) we do not distinguish between σ and π, so we need to add all alpha electrons in gerade orbitals (2s σg + 2p σg + 2p πg = 4), beta electrons in gerade orbitals (2s σg+ 2p σg = 2), alpha electrons in ungerade orbitals (2s σu + 2p πupi_u = 3) and beta electrons in ungerade orbitals (2s σu + 2p πu = 3). This gives the following occupation:
NELEC=4,2,3,3
Note that this determinant does not represent the exact ground state of the oxygen molecule as this triplet is split by a few wave numbers due to spin-spin and (second order) spin-orbit interactions. The lowest state is the Omega=0 component that cannot be represented by a single determinant. This state can be calculated using the Fock space method (see under &CCFSPC).
Warning
check link to &CCFSPC
Coming back to the example given above, the oxygen molecule, we now show how this is done with the keywords given in the reference manual. We again want to take the M_S=1 state as our reference. The irreps of Dinfh* are ordered as 1/2,-1/2,3/2,-3/2,.... so we need to consider the omega value (giving the projection on the molecular axis of both spin and orbital angular momentum) of the occupied oribitals. The sigma orbitals go in the irreps 1/2 and -1/2 while the pi-orbitals span the four irreps (1/2,-1/2,3/2,-3/2). Putting an alpha electron in a sigma orbital will give an omega value of 1/2, while putting it in a pi-orbital can either give -1/2 (when put in the orbitals with orbital momentum -1) or 3/2 (when put in the +1 orbital). Similarly the beta electrons go in irreps -1/2 (for the sigma), -3/2 and 1/2 (for the pi). This makes the input for our example :
NELEC\_F1=2,3,1,0 NELEC\_F2=2,2,1,1
An example of the use of the Fock space method concerns the calculation of the lowest three states of molecular oxygen. These are obtained by distributing two electrons in the degenerate πg orbitals. The X ground state is a triplet that is split by second-order spin-orbit coupling and spin-spin interactions into a lowest MS=0 component and two higher MS=±2 components. This case is also discussed under the RELCCSD input as an example of the use of the NELEC keyword. In the Fock space approach we first perform a calculation on O2+2 and calculate the three X states, the degenerate open shell singlet a and the singlet b state in one single Fock space run. The FSCCSD input reads then
&RELCCSD NELEC=2,2,3,3, DOFSPC=T, DOENER=F &END
&CCFSPC MAXIT=100, NACTP=2,2,0,0, DOEA2=T, &END
This example can be found as test #52 REFERENCE_IT!!! in the DIRAC test set.
we give an example for the high spin state of the nitrogen atom.... ADD ME !