4.4 Counterpoise and basis-set superposition error
These are unexecuted teaching inputs and starting models. Original diagrams are schematics, not calculated results. Validate version-specific syntax, licensed or authorized data, numerical convergence and the scientific model before using this workflow.
4.4.1 Model, units and provenance
Keep basis/potential files and executable versions traceable. Grid controls use Ry; common energy/force outputs use hartree and hartree/bohr. Read the unit in each output heading.
Shared inputs, conventions and evidence
4.4.2 Unexecuted inputs and explicit deltas
Use the accompanying instructions to identify the parent calculation and placement of every delta; a snippet is not automatically a standalone input. Preserve all blank-line and file-provenance requirements.
4.4.2.1 Input block 1
&KIND Oghost
ELEMENT O
BASIS_SET DZVP-MOLOPT-GTH
GHOST T
&END KIND
&KIND Hghost
ELEMENT H
BASIS_SET DZVP-MOLOPT-GTH
GHOST T
&END KIND
4.4.3 Worked investigation
Intuition. In an atom-centered basis, each fragment can borrow the partner's basis functions when the dimer is calculated. The separated monomer calculation normally lacks those functions, creating an unbalanced comparison. Counterpoise evaluates fragments in the full dimer basis using ghost centers: basis functions remain but nuclei/electrons of the ghost partner are absent. This corrects a basis imbalance, not a deficient functional.
Original delta and prerequisites. Use one fixed water dimer from case 15. Perform three electronic calculations in the same cell/geometry and with the same full orbital basis: AB; A with B ghosted; B with A ghosted. In SUBSYS, distinguish real and ghost labels in COORD, then add matching kinds such as:
See input block 1 above.
Do not attach a real nuclear potential to the ghost kinds. The real fragment retains its normal KIND definitions. Verify the output electron count; a neutral water fragment should still have eight valence electrons with the chosen GTH potentials even though six basis centers are present. This explicit ghost workflow avoids assuming automatic BSSE fragment/configuration settings.
Workflow. 1. Label monomer atom indices before editing any KIND. 2. Run AB at frozen dimer geometry. 3. Ghost B and compute EA^AB; then ghost A and compute EB^AB, keeping the fragment charge/spin correct. 4. Form counterpoise interaction Eint,CP=EAB^AB−EA^AB−EB^AB. 5. Compute uncorrected monomer energies in their own bases and compare the interaction values. 6. Repeat at larger basis and at one longer separation. Treat empirical dispersion separately and consistently: ensure ghost centers do not contribute fictitious pair corrections, or calculate the electronic CP terms without D3 and add the physical real-atom interaction correction once.
Interpretation and checks. Ghost basis functions can lower a fragment energy by increasing flexibility. The difference measures basis borrowing for that geometry and basis, not the total basis-set incompleteness. Counterpoise may change a minimum position, so a CP single point at an uncorrected optimized geometry is not a fully CP-optimized structure. For periodic solids, fragment definitions and ghost lattices are more subtle; do not extrapolate this isolated dimer recipe blindly.
Pitfalls, exercise and diagram. Accidentally retaining a partner's electrons computes a different species. Accidentally removing its basis defeats the correction. Choose a larger basis and determine whether CP and uncorrected curves move toward each other; convergence is empirical, not guaranteed by a label.
4.4.4 Related calculations
- 4.3 Noncovalent interaction and dispersion choices
- 5.1 A neutral slab and its electrostatic boundary