Molecular-orbital cubes

Export the signed amplitude of an orbital as a cube file. View it in a compatible molecular viewer. Check the source and grid normalization diagnostics before using the values quantitatively.

Status: independently checked orbital fields; each generated cube retains its own grid-convergence status.

For a real Gaussian AO basis, the signed field is

psi_i(r) = sum_mu C_mu,i chi_mu(r)

in bohr^-3/2. It is an orbital amplitude; its square is an orbital density. Overall MO phase has no physical significance. openWFN evaluates the source-normalized Cartesian or pure-spherical AO basis and applies recorded input-convention transforms where required.

Required data are isolated geometry, a supported Gaussian basis, and coefficients for the selected spin channel. Public orbital numbers are one-based. alpha uses the restricted channel on restricted inputs; requesting an unavailable beta channel fails explicitly. HOMO/LUMO selection follows the normalized occupations used by frontier analysis and HOMO selection remains usable when no virtual orbitals are present.

openwfn molecule.molden orbitals cube --mo homo --output homo.cube
openwfn molecule.molden orbitals cube --mo lumo --spin beta --output lumo-beta.cube --spacing 0.2 --padding 6
openwfn molecule.molden orbitals cube --mo 25 --output mo25.cube

Spacing and padding are in bohr, matching density commands. The molecular grid is bounded before allocation and AO evaluation uses bounded chunks. Cube comments record MO selection, spin, spacing, padding, and input hash. Structured results retain parser provenance, orbital energy and occupation, grid shape, raw ao_metric_norm = c^T S c, and squared_amplitude_integral.

from openwfn import load
result = load("molecule.molden").orbital_cube(
    "homo.cube", mo="homo", spacing_bohr=0.2
)
print(result.as_dict())

openWFN does not silently renormalize coefficients. If the source orbital norm or grid integral fails its tolerance, the result is partial and includes a warning. Refine spacing and padding for quantitative use. Writing a cube successfully does not mean its grid is converged.

Validation evidence

Analytic normalized-Gaussian tests establish sign, amplitude, and squared norm. Independent GBasis comparisons verify mapped Cartesian and pure-d AO evaluations. Cross-format tests compare equivalent FCHK, Molden, MWFN, WFN, and WFX representations with global MO phase handled explicitly.

The stable everyday-QC reference suite additionally compares HOMO/LUMO fields against PySCF AO evaluations for the documented molecular cases and records serialized cube-value errors in validation/everyday-qc/pyscf-report.json. Those field comparisons validate the orbital evaluator for the stated scope.

The cube result itself remains conditional on the requested grid. Some deliberately coarse release-corpus cube settings produce partial normalization diagnostics even though the field values agree with the independent reference. openWFN preserves that distinction rather than promoting every written file to a clean validated result.

validation/manifest.json and the everyday-QC validation page define the current evidence scope. The reference set does not establish arbitrary high-angular-momentum molecular coverage, ECP behavior, every program convention, or grid convergence for settings the user has not tested. Cube writing remains outside the read-only MCP interface.