Basis-Set Superposition Error

Introduction

A weak complex can appear more stable than it really is because its fragments gain access to one another's basis functions. In a finite atom-centered basis, the calculation on AB uses functions centered on both A and B, whereas separate calculations on A and B normally use only their own functions. Part of the apparent energy lowering on forming AB can therefore reflect extra mathematical flexibility rather than a physical interaction. This basis-set superposition error, or BSSE, is especially important to examine when the true binding energy is small.

Core explanation

Let E AB^(AB) denote the electronic energy of complex AB in its full combined basis. Let E A^(A) and E B^(B) denote the monomer energies calculated in their own bases at the geometries they have inside the complex. An uncorrected fixed-geometry interaction energy is ΔE raw = E AB^(AB) − E A^(A) − E B^(B). Negative values indicate stabilization under this sign convention. The problem is that the complex can use more functions than either monomer was allowed in the comparison. A basis function centered on B may help describe A's electron density when the fragments are close, even if the two do not have a real attraction of that size.

The Boys–Bernardi counterpoise idea evaluates each monomer in the full AB basis. In the A calculation, B's basis functions remain as ghost functions, but B's nuclei and electrons are removed; the reverse applies to B. Write these monomer energies E A^(AB) and E B^(AB). The counterpoise interaction energy becomes ΔE CP = E AB^(AB) − E A^(AB) − E B^(AB). Now all three energy terms use the same set of basis-function locations. IUPAC defines counterpoise correction as a method for correcting basis-set superposition error. This does not change the physical Hamiltonian of A into AB: ghost centers supply mathematical functions only.

For a variational calculation with consistent settings, giving an isolated monomer additional functions can only lower or retain its variational energy. Thus E A^(AB) is commonly no higher than E A^(A), and similarly for B. Replacing the higher monomer energies in the subtraction usually makes ΔE CP less negative than ΔE raw: the uncorrected interaction looked too attractive. The direction of this diagnostic should be derived from the chosen sign convention rather than memorized without context. Correlated methods, different optimization choices or numerical noise add practical complications, so inspect the actual terms.

The fixed-geometry condition matters. An interaction energy compares AB with fragments frozen at their complex geometries. A binding or dissociation energy comparing relaxed isolated fragments also includes deformation of each monomer. If A changes shape substantially on binding, that deformation cost is a real physical term and should not be mislabeled BSSE. Counterpoise calculations must use identical fragment coordinates and method settings to isolate basis borrowing. Charge, spin and basis placement need to be defined for each fragment; assigning electrons differently changes the chemical question.

BSSE generally shrinks as an appropriate atom-centered basis approaches completeness, but a counterpoise-corrected finite-basis result is not automatically closer to the exact interaction energy. Ordinary basis incompleteness can push in the opposite direction, and the two errors can partly cancel before correction. A correction may be large when the basis is too small for the question, signaling that a larger basis or a systematic sequence is needed. Research on counterpoise methods compares molecular and atomic estimates across increasingly large bases, illustrating why BSSE should be assessed within a convergence study.

Diffuse functions can reduce some representation limits for weak complexes, yet they can also make fragment-basis borrowing more extensive at particular separations or trigger near-linear dependence. The magnitude of a CP correction is therefore not a universal quality score for a basis. Report both raw and CP interaction energies when the distinction matters, and show how the target value behaves as the basis is enlarged. Temperature, solvent, zero-point motion and intermolecular geometry sampling are separate from this electronic basis artifact.

Step-by-step reasoning

1. Fix the geometry of complex AB and define fragments A and B, their charges and spin states. 2. Compute the complex energy in the full AB basis at the chosen electronic method. 3. Compute each frozen-geometry monomer in its own basis for the raw interaction energy. 4. Recompute each monomer with the partner's basis functions present as ghost centers. 5. Form both ΔE raw and ΔE CP with the same sign convention and compare them. 6. Repeat with a better basis if the correction is material to the chemical conclusion.

Visual explanation

Draw two fragments with basis-function clouds around their nuclei. In the separate-monomer pictures, A and B each have only their own clouds. In the complex picture, overlapping clouds supply a larger combined mathematical space. Then draw the ghost calculation of A: B's nucleus and electrons disappear, but dotted basis-function clouds remain at B's positions. This shows how counterpoise makes the monomer comparison use the same basis locations as the complex without pretending that B physically remains.

Real-world analogy

Imagine comparing two runners when the paired event gives them a longer, better-surfaced track but the solo event uses short rough tracks. A faster paired result could partly reflect the better track rather than teamwork. Counterpoise gives each solo runner the same track for the comparison. The analogy is limited: basis functions are mathematical representational freedom, and a quantum calculation's variational behavior is not ordinary athletic performance.

Real-world example

A researcher predicts a weak hydrogen-bonded dimer to have a small negative electronic interaction energy. The counterpoise calculation makes the interaction substantially less negative, showing that finite-basis borrowing contributed to the apparent binding. The researcher repeats the calculation with a larger polarized and diffuse basis and checks how raw and corrected values converge. Only then are zero-point and thermal corrections considered for comparison with a measured equilibrium. A single raw energy would have hidden an uncertainty comparable with the claimed effect.

Why?

Why do ghost functions change a monomer energy when there is no ghost nucleus? The monomer's orbitals are represented as linear combinations of available basis functions. A function placed at the partner's former position can help describe the monomer's tail or polarization even without an attractive nucleus there. The Hamiltonian remains that of the monomer, but its finite mathematical approximation improves. That improvement measures the unfair basis advantage that the monomer lacked in the raw comparison.

Common misconception

“Counterpoise adds a ghost atom to the chemistry.” It adds only basis functions, with no ghost nuclear charge or electrons. Another misconception is equating the CP correction to the entire physical interaction. It estimates a finite-basis artifact. A third is applying it after allowing the fragment geometries to relax and then calling the result a fixed-geometry interaction energy. Geometry deformation must be treated separately. Finally, a tiny CP correction does not prove the electronic method or solvent model is accurate.

Worked example

Use invented energies in hartree at one fixed dimer geometry: E AB^(AB) = −200.015, E A^(A) = −100.000 and E B^(B) = −100.000. The raw interaction is −0.015 hartree, about −39.4 kJ mol⁻¹. With ghost functions, suppose E A^(AB) = −100.004 and E B^(AB) = −100.003. Then ΔE CP = −200.015 − (−100.004) − (−100.003) = −0.008 hartree, about −21.0 kJ mol⁻¹. The 0.007-hartree difference, about 18.4 kJ mol⁻¹, is the counterpoise diagnostic in this example. The complex is still predicted attractive under both calculations, but its apparent strength is strongly basis-sensitive. These numbers are pedagogical, not a measured dimer.

Quick check

1. What remains at a ghost center in a counterpoise monomer calculation? Answer: The partner's basis functions remain, while its nuclei and electrons are omitted. 2. Why is raw complex-minus-monomer energy often too negative in a small atom-centered basis? Answer: The complex can borrow partner-centered functions, gaining variational flexibility that the separate monomers lack.

Exam focus

Write the raw and counterpoise formulas with superscripts indicating which basis each energy uses. State the fragment geometry and sign convention. Explain why ghost functions do not represent physical ghost atoms. Distinguish interaction from deformation or dissociation energy. Describe counterpoise as a basis-error diagnostic and compare it across larger bases before making a quantitative binding claim.

Advanced insight

An interaction curve requires counterpoise treatment at each separation if BSSE varies with distance. A correction calculated only at one minimum does not automatically fix the whole curve or its vibrational properties. For strongly covalent reactions, defining meaningful fragments and consistent electron states can be more difficult than for weak dimers. Methods using delocalized plane-wave representations face a different basis-comparison problem from small atom-centered sets. The general lesson is to align the representational spaces of compared states when extracting a small difference from large total energies.

Summary

Basis-set superposition error occurs when a finite-basis complex gains representational flexibility from partner-centered functions that isolated fragments lack. Counterpoise recomputes frozen fragments in the full complex basis using ghost centers, producing a more balanced comparison. It is most informative alongside basis convergence and a clear separation of electronic interaction, geometry deformation and environmental effects.

Practice questions

1. In a CP calculation of A, does B's nucleus remain at its complex position? Answer: No. Only B-centered basis functions remain as ghost functions. 2. If ΔE raw = −12 kJ mol⁻¹ and ΔE CP = −7 kJ mol⁻¹, what was the CP change under this sign convention? Answer: The result became 5 kJ mol⁻¹ less attractive, indicating a 5 kJ mol⁻¹ basis-borrowing contribution in this diagnostic. 3. Why must fragment geometries be fixed when calculating an interaction energy at a chosen complex geometry? Answer: Relaxing fragments adds real deformation-energy differences and confounds the basis comparison. 4. Does a CP-corrected value automatically equal the complete-basis interaction energy? Answer: No. Other basis incompleteness and electronic-method errors remain, so a convergence study is still needed.