Noncovalent Interaction Energies

Introduction

A weakly bound dimer may differ from its separated fragments by only a few kilojoules per mole. At that scale, a small basis-set artifact or a neglected entropy term can change whether the calculation predicts attraction at all. Noncovalent contacts include hydrogen bonding, permanent electrostatics, induction, exchange repulsion and dispersion. These contributions overlap and cooperate; a chemical label does not by itself give an interaction energy. The calculation needs an unambiguous energy definition, adequate electron correlation and a check for artificial stabilization caused by the finite basis.

Core explanation

For fragments A and B at a fixed complex geometry, define the electronic interaction energy as ΔEint = EAB(geometry AB) − EA(geometry A in AB) − EB(geometry B in AB). A negative value means the complex is electronically lower in energy than the two fragments frozen in their complex shapes. If the fragments are allowed to relax separately, the resulting difference is a binding or dissociation energy that includes deformation. A solution association free energy adds solvation, thermal motion, standard-state and entropy terms; it is not the same as either electronic quantity. State the geometry, method, basis and physical state alongside a reported number.

The main physical components have different origins. Electrostatic interaction reflects charge distributions, including permanent multipoles. Induction is the change in one fragment's density caused by the field of its partner. Dispersion comes from correlated electronic fluctuations and remains attractive even between neutral, nonpolar species. Exchange repulsion rises when occupied electronic regions overlap and the many-electron wavefunction must remain antisymmetric. A hydrogen bond can combine all of these; it is not a purely electrostatic line between two atom symbols. Symmetry-adapted perturbation theory offers a method-dependent decomposition, but the total energy difference is the primary quantity to compare across calculations. Primary XSAPT work discusses electrostatic, exchange, induction and dispersion treatment for noncovalent complexes.

A finite atom-centered basis causes basis-set superposition error (BSSE). In the complex, A can use basis functions centered on B, and B can borrow those on A. An isolated monomer calculation without its partner has fewer functions, so the raw complex-minus-monomers difference may appear too favorable. In the Boys–Bernardi counterpoise procedure, calculate each monomer at the complex geometry using the full dimer basis: partner nuclei and electrons are absent, but their basis functions remain as ghost centers. The corrected interaction energy is EAB(AB basis) − EA(AB basis) − EB(AB basis). This correction diagnoses basis borrowing; it does not create physical dispersion or repair every method error. Its importance generally shrinks as the basis approaches completeness, though convergence must be checked for the system at hand. Benchmark methodology reviewed by the researchers who assembled noncovalent datasets treats basis and correlation errors as distinct limits.

Method choice matters. Hartree–Fock lacks the ordinary long-range dispersion attraction because its independent-particle treatment omits correlated fluctuations. MP2 includes dispersion-like correlation but can overbind some π-stacked systems; a particular density functional may also need a compatible dispersion correction. High-level benchmark interaction energies for small suitable complexes often use coupled-cluster methods with basis extrapolation, while large systems require an affordable approximation validated against a relevant set. One should not trust an attractive small-basis result merely because it matches experiment: missing dispersion and excessive BSSE can accidentally cancel. That agreement may vanish when the geometry or chemical class changes.

Water and temperature complicate the comparison. A gas-phase hydrogen bond may be strong relative to separated gas molecules, but in water each partner can make favorable contacts with solvent. Association also reduces translational freedom. Thus a negative gas-phase electronic interaction energy does not guarantee a favorable aqueous binding free energy. If the goal is an equilibrium constant, calculate or estimate the entire free-energy change under consistent concentration and solvent conditions. If the goal is a force-field parameter, compare electronic energies across distances and orientations rather than calibrating against only a single minimum.

Step-by-step reasoning

1. Define fragments, charge and spin states; decide whether the target is a fixed-geometry interaction, relaxed binding energy or solution free energy. 2. Optimize or select the complex geometry and preserve its fragment coordinates for the fixed-geometry energy difference. 3. Compute complex and monomer energies with the same electronic method and compatible basis. 4. Evaluate monomers in the full complex basis using ghost functions to diagnose BSSE. 5. Check basis and correlation-method convergence against a suitable benchmark if accuracy demands it. 6. Add deformation, thermal, entropic and solvent terms only when the target observable requires them.

Visual explanation

Sketch an energy diagram with separated frozen fragments at zero, a lower complex minimum and a separate level for independently relaxed fragments. The vertical distance from frozen fragments to complex is ΔEint; the gap from relaxed fragments includes deformation. Beside this, draw A and B with overlapping basis-function circles. The circles centered on B remain as empty ghost centers during the counterpoise calculation for A. A second panel can show the same two molecules surrounded by water, emphasizing that solution binding includes displaced solvent.

Real-world analogy

Two people completing a task together may use tools borrowed from one another. If their individual performance is measured without those tools, the comparison unfairly exaggerates the benefit of pairing. Counterpoise gives each isolated fragment access to the same mathematical tools it had in the complex. The analogy describes basis consistency, not a literal transfer of electrons or the physical origin of binding.

Real-world example

A researcher compares several substituted aromatic dimers as models for molecular recognition. One geometry is stacked; another is T-shaped. Both may be stabilized by dispersion, but their electrostatic and repulsive terms differ. The researcher calculates a distance scan with a dispersion-capable method, checks counterpoise corrections and reports whether monomer geometries were frozen or optimized. A single interaction energy at one geometry would not establish the preferred arrangement in solution, because accessible orientations and solvent competition also affect the population.

Why?

Why does BSSE usually make the uncorrected interaction too negative? Adding a partner's basis functions lowers or leaves unchanged the variational energy available to each fragment in the complex basis. The isolated monomer energies, computed in their smaller bases, are artificially high relative to that flexible complex description. Subtracting them therefore exaggerates stabilization. Counterpoise puts monomer and complex calculations on the same basis footing. Other errors may still have either sign, so the final corrected energy is not automatically exact.

Common misconception

“A negative electronic interaction energy proves spontaneous association.” Entropy, solvent and standard-state effects may reverse the free-energy sign. “A hydrogen bond is only electrostatics” ignores polarization, dispersion and exchange. “Counterpoise is a dispersion correction” confuses a basis artifact with a physical electron-correlation contribution. “The strongest dimer at one geometry will dominate experimentally” overlooks conformer and orientation populations and competing environments.

Worked example

At a fixed geometry, suppose a hypothetical dimer has EAB − EA(own basis) − EB(own basis) = −18 kJ mol⁻¹. Recalculating each fragment in the full dimer basis raises the interaction-energy difference by 4 kJ mol⁻¹, giving a counterpoise-corrected ΔEint of −14 kJ mol⁻¹. The raw result overestimated electronic attraction by 4 kJ mol⁻¹ on this basis. If independently relaxed monomers lie 3 kJ mol⁻¹ below the frozen fragment sum, the relaxed-fragment binding energy becomes −11 kJ mol⁻¹ relative to those monomers. None of these three numbers is yet the aqueous association free energy; it would need solvent and entropy contributions. The illustrative numbers show why the chosen energy definition belongs beside the value.

Quick check

1. What geometry should the isolated fragments use in a fixed-geometry interaction energy? Answer: Their coordinates as found within the complex, so relaxation energy is excluded. 2. What does a ghost center supply in a counterpoise monomer calculation? Answer: The partner's basis functions without its nucleus or electrons.

Exam focus

Write the complex-minus-fragments energy expression with geometries specified. Explain why a negative result is electronic stabilization but not necessarily a favorable solution equilibrium. Distinguish counterpoise, basis convergence and dispersion treatment. In a numerical comparison, state whether fragment deformation and thermodynamic corrections have been included. Avoid assigning a contact to one physical component solely from its molecular drawing.

Advanced insight

Energy decomposition is useful for explanation but its individual components depend on the partitioning formalism, especially when charge transfer and polarization overlap. A distance scan can distinguish short-range exchange repulsion from longer-range attraction more clearly than a single minimum. Many-body environments introduce nonadditive induction and dispersion, so a pairwise dimer energy may not transfer directly to a crystal or biomolecular pocket. Benchmarks should cover relevant nonequilibrium separations, not only the lowest-energy structures.

Summary

Noncovalent interaction energies require a defined fragment reference and geometry. Electrostatics, induction, dispersion and exchange all contribute, often within a single hydrogen bond. Finite-basis borrowing can make binding appear too strong, and counterpoise provides a specific diagnostic. Method, basis, conformational, solvent and entropy effects must be assessed separately. Report what energy was calculated before interpreting its sign or comparing it with a measured equilibrium.

Practice questions

1. A raw interaction energy is −24 kJ mol⁻¹ and the counterpoise correction is +5 kJ mol⁻¹. What is the corrected result? Answer: −19 kJ mol⁻¹ at the stated geometry and method. 2. Why does Hartree–Fock alone often miss attraction between nonpolar fragments at long range? Answer: It omits the electron correlation responsible for ordinary London dispersion. 3. What extra contribution distinguishes a relaxed binding energy from a fixed-geometry interaction energy? Answer: The energy of deforming each isolated monomer from its own relaxed structure to its shape in the complex. 4. Can a negative gas-phase interaction energy be equated with a negative aqueous association free energy? Answer: No. Solvent competition, entropy and standard-state terms affect the aqueous free energy.