2026-07-09 BSSE (counterpoise) sub-step

Add a BSSE sub-step to the ORCA plug-in that returns the counterpoise-corrected energy and gradient of a molecular complex, for use as machine-learned-force-field (MLFF) training data. It wraps the ORCA Compound script BSSEGradient.cmp (D. G. Liakos & F. Neese, 2024/2025).

This is the fast, ORCA-specific first step (two fragments, neutral singlet only, driven by a single ORCA Compound process). It was later generalized to an arbitrary number of fragments with independent per-fragment charge, backed by a new shared library rather than a second ORCA-specific special case – see the 2026-08-03 campaign, whose canonical design document lives in the sibling seamm_bsse repository. This Compound-script path is kept as that generalization’s permanent N = 2 regression reference, not deleted, even though run() no longer calls it.

Contents:

• BSSE sub-step – scope and design
  • Goal
  • What the Compound script computes
  • How it runs today (the raw script)
  • Design in SEAMM
  • Known limitations (scope boundaries of Phase 1)
  • Validation
  • Citations
  • Implementation task list
• Implementation notes
  • 2026-07-09 – built and validated