Input Format¶

edf reads its directives from stdin. The first three records are positional and mandatory; everything after them is keyword-driven.

./edf-omp < input.edfinp > output.edfout

Comments: any line whose first non-blank character is # is ignored.

Terminator: END (case-insensitive) ends the input.

Record 1 — `ioverlap`¶

An integer on its own line that selects the source and format of the Atomic Overlap Matrix.

Value	AOM source
`0`	PROMOLDEN format or `filedat` is a fuzzy-partition keyword (see Record 2)
`1`	PROAIM format
`2`	TOPMOD (ELF basins)
`3`	TOPMOD (QTAIM basins)
`4`	AIMALL (QTAIM basins)
`-1`	Complex AOM — control passed to `edfx.f`
`-2`	Complex AOM — control passed to `edfcrit.f`
`-11`	Single-determinant fast path — control passed to `edfstd.f`

See AOM Formats for details on each format.

Record 2 — `filedat`¶

Either the name of the AOM file or one of the built-in fuzzy-partition keywords. Requires ioverlap = 0 when a keyword is used.

Value	Meaning
filename	Read AOM from this file
`mulliken`	Compute AOM using Mulliken population analysis
`lowdin`	Compute AOM using Löwdin population analysis
`mindef`	Compute AOM using Minimally Deformed Atoms (MinDef) of Fernández Rico et al.
`mindefrho`	MinDef with weights \(w_A = \rho_A/\rho\) (always gives \(S_{ii}^A > 0\))
`netrho`	Net-density weights (eliminates mixed-center cross terms)
`promrho`	Promolecular density weights: \(w_A = \rho_A^{\rm atom}/\rho\)
`heselmann`	Heselmann chemical-localization weights (JCTC 12, 2720, 2016)
`becke`	Becke fuzzy-cell partition

For mindefrho, netrho, and becke, optional integration parameters (NANG, NRAD, IRMESH, and for Becke also POW) may follow the keyword on the same line. See Integration Keywords.

Record 3 — `wfnfile`¶

The name of the .wfn wavefunction file. Must be in the GAMESS aimpac format.

Single-determinant: standard .wfn from GAMESS or Gaussian.
Multi-determinant (CASSCF): .wfn with determinant coefficients and active-space occupation numbers, written by a domestic version of GAMESS10.

The AOM integrals in filedat must correspond to the wavefunction in wfnfile.

Record 4 — `NGROUP ngroup`¶

The number of fragments into which the molecule is divided.

If ngroup ≤ 0, or NGROUP is absent, each atomic basin becomes its own independent fragment with full electron-count range (0 to total alpha/beta). In this case Records 5.i must be skipped.
Negative ngroup: the first |ngroup|−1 groups are specified in Records 5.i, and the last group automatically contains all remaining atoms.

A variant keyword NFRAG nfrag sets the fragment count used exclusively by the EOS/Lewis routines (nfnoeos.f, lewisoqs.f). If absent, NFRAG defaults to NGROUP.

Records 5.i — Fragment specification (i = 1 … NGROUP)¶

Two formats are accepted:

Format 1 (with electron count limits):

nfugrp  atom1  atom2  ...  minelecA  maxelecA  minelecB  maxelecB

Format 2 (full range assumed):

nfugrp  atom1  atom2  ...

Field	Meaning
`nfugrp`	Number of atoms in this group
`atom1 atom2 …`	Indices of the atoms belonging to this group (1-indexed)
`minelecA`, `maxelecA`	Min/max alpha electrons in this group
`minelecB`, `maxelecB`	Min/max beta electrons in this group

When Format 2 is used, the program sets minelec = 0 and maxelec = NALPHA (or NBETA).

Core electrons

The minelec/maxelec values include core electrons (as defined by COREMO). They are total electron counts, not valence-only counts.

Example — H₆ ring, 6 groups of 1 atom each:

Example — N₅⁺, 4 groups with atoms 2 and 4 together:

ngroup 4
1 1
1 3
1 5
2 2 4

Records 5bis.i — EOS/Lewis fragment specification¶

These are used only when NFRAG differs from NGROUP and EOS/Lewis analysis is requested. Format: nwithin atom1 atom2 … (similar to Records 5.i, Format 2, but for NFRAG fragments).

After the fragment specification, all remaining input is keyword-driven and optional (except END). See the Keywords section.

Input Format¶

Record 1 — ioverlap¶

Record 2 — filedat¶

Record 3 — wfnfile¶

Record 4 — NGROUP ngroup¶