CIS Transition-Density Cube Files

PYSEQM can export real-space CIS transition densities as Gaussian cube files for MNDO-type methods: MNDO, AM1, and PM3. These files help visualize the spatial character of a transition from the ground state to an excited state.

Generating cube files

Use scripts/cis_transition_density_cubes.py to run a CIS calculation for one XYZ geometry and write cubes for its leading excited states. For example, this calculates ten AM1/CIS states and writes transition-density cubes for the first three:

python scripts/cis_transition_density_cubes.py molecule.xyz \
    --method AM1 \
    --n-states 10 \
    --cube-states 3 \
    --out-dir cis_transition_density

Supported --method values are MNDO, AM1, and PM3. The default is AM1.

The output directory contains:

• cis_energies.csv: excitation energies and oscillator strengths for all requested CIS states.

• state_01_transition_density.cube, etc.: signed transition-density volumes for the requested leading states.

The grid spacing defaults to 0.20 Angstrom and the grid extends 3.0 Angstrom past the outermost atoms. Adjust these values when needed:

python scripts/cis_transition_density_cubes.py molecule.xyz \
    --method PM3 \
    --spacing 0.15 \
    --padding 4.0

--chunk-size controls how many grid points are evaluated at once. Reduce it if cube generation uses too much memory.

Basis and orthogonalization

The exporter evaluates PYSEQM’s normalized Slater-type valence basis: one 1s function per hydrogen and [ns, npx, npy, npz] functions per heavy atom. The CIS amplitudes are first transformed from PYSEQM’s orthonormal NDDO AO representation into this non-orthogonal Slater basis using symmetric Lowdin orthogonalization:

\[R_{\mathrm{STO}} = S^{-1/2} R_{\mathrm{NDDO}} S^{-1/2}.\]

This transformation is necessary before evaluating the real-space density.