From the Molecular Hamiltonian to Predictions

Introduction

Computational chemistry starts with a physical model before it starts with software. For an isolated collection of electrons and nuclei in a common nonrelativistic approximation, the Hamiltonian lists the energy contributions that the model includes: motion of electrons and nuclei, attraction between unlike charges, and repulsion between like charges. Solving the associated Schrödinger equation yields wavefunctions and energies from which properties can be estimated. The exact many-particle solution is generally too difficult, so practical methods approximate either the Hamiltonian, the wavefunction or the way it is solved. Understanding the terms makes those choices visible.

Core explanation

Write the full model schematically as Ĥ = T̂ e + T̂ N + V̂ eN + V̂ ee + V̂ NN . T̂ e represents electron kinetic energy, and T̂ N represents nuclear kinetic energy. V̂ eN is negative because electrons and positively charged nuclei attract. V̂ ee and V̂ NN are positive because like charges repel. The many-particle stationary Schrödinger equation is ĤΨ = EΨ, where Ψ depends on electron and nuclear coordinates and E is an allowed energy for the model. This five-term organization is given in a university quantum-chemistry text and in current primary molecular-model research.

In atomic units, the electron mass, elementary charge magnitude, reduced Planck constant and Coulomb constant are each one by convention. Then the nonrelativistic terms take an uncluttered form. For electrons indexed i and nuclei indexed A with charges Z A and masses M A in electron-mass units, T̂ e = −½Σ i∇² i and T̂ N = −Σ A[1/(2M A)]∇² A. The electron–nucleus attraction is V̂ eN = −Σ iΣ A Z A/r iA, electron–electron repulsion is V̂ ee = Σ {i<j}1/r ij, and nuclear–nuclear repulsion is V̂ NN = Σ {A<B}Z AZ B/R AB. Distances are in bohr and energies in hartree in this form. The index restriction i<j or A<B prevents counting a pair twice; including i=j would introduce meaningless self-interaction in this elementary particle list.

The electron–nucleus term couples electronic and nuclear coordinates. Electrons respond to where nuclei are, and nuclear forces respond to the electronic distribution. That coupling is one reason the full equation is difficult. A common first step is to hold nuclear positions R fixed while solving an electronic problem. Then T̂ N is absent from that fixed-geometry electronic equation , while V NN is a constant for each geometry and must still be included in the total fixed-geometry energy. The following page develops the Born–Oppenheimer reasoning and its limits. A molecular-structure lesson from New York University uses the hydrogen molecular ion to show how an electronic energy plus nuclear repulsion becomes a potential-energy curve.

An energy eigenvalue is not directly a reaction free energy, a bond strength or a spectrum. To estimate a reaction energy, calculate comparable reactant and product states with balanced composition, charge and spin, then subtract energies using the same method and reference convention. To estimate a room-temperature equilibrium quantity, add nuclear vibration, rotation, translation, solvation and statistical populations as appropriate. To predict a measured spectrum, calculate transition energies and intensities under a suitable excited-state model. The Hamiltonian identifies what physics is in principle present; the chosen approximation and the target observable determine what prediction is actually justified.

The nonrelativistic Coulomb Hamiltonian also has limits. Heavy elements may need relativistic terms for accurate electronic structure. Molecules in an external magnetic field, under light, on a surface or in solvent require additional environment or field descriptions. Spin-dependent interactions may matter for spectroscopy or magnetic properties. Finite basis sets approximate wavefunctions and introduce numerical incompleteness even when the Hamiltonian terms are conceptually correct. It is therefore useful to distinguish model error from solution error : missing physics differs from a poor approximation to the model one intended to solve.

The sign of each potential term gives useful checks, but no single term determines bonding. Bringing nuclei together increases their mutual repulsion, yet it can also make favorable electron–nucleus interactions possible. Electron kinetic energy and electron–electron repulsion change as the wavefunction reorganizes. A stable bond reflects the minimum of the total energy curve, not simply “attraction wins” as an isolated statement. Separating terms aids interpretation only when their coordinated change is considered.

Step-by-step reasoning

1. List all electrons and nuclei and state the modeled charge, spin and environment. 2. Write electron and nuclear kinetic terms and three pairwise Coulomb interaction families. 3. Check attraction is negative, repulsions positive and every distinct pair counted once. 4. If nuclei are held fixed, remove nuclear kinetic motion from the electronic equation but retain nuclear repulsion in total energy. 5. Solve or approximate the resulting equation at each relevant geometry and check convergence. 6. Convert energies into the requested observable with consistent states, thermal or environmental corrections and uncertainty.

Visual explanation

Draw two positive nuclei and two negative electrons. Use blue arrows between unlike pairs for attractive V eN and red arrows between like pairs for repulsive V ee and V NN. Around each particle draw a small motion arrow for kinetic terms. Then show a series of nuclear separations R with a total fixed-geometry energy curve; mark a minimum. Annotate that the curve includes changing electronic energy and the positive nuclear-repulsion term at every R. This prevents mistaking a single Coulomb term for the whole binding energy.

Real-world analogy

A financial balance sheet lists income and several kinds of cost. The total result is not one line item; changing a project can alter several lines at once. A molecular Hamiltonian likewise separates energy contributions whose total determines predicted states. The analogy is limited because quantum kinetic and Coulomb terms are operators acting on a shared wavefunction, not independent bills that can be optimized one by one.

Real-world example

A chemist models H₂⁺, which has two protons and one electron. Because there is only one electron, the electron–electron repulsion term is absent. At fixed internuclear distance R, the remaining electronic operator includes electron kinetic energy and attraction to both protons, while the proton–proton repulsion +1/R is added to the total fixed-geometry energy in atomic units. Calculating this energy for several R values gives a potential-energy curve. Its minimum predicts a preferred separation in the fixed-nuclei model; vibrational motion can then be considered on that curve.

Why?

Why is nuclear–nuclear repulsion retained when the nuclei are treated as fixed? “Fixed” means their positions are parameters during the electronic calculation, not that the positive charges cease interacting. At one chosen R, their repulsion is a constant with respect to electronic coordinates, but it changes when comparing different R values. Omitting it would severely distort a bond-energy curve and could falsely favor arbitrarily close nuclei.

Common misconception

“A molecule binds because electron–nucleus attraction is negative.” Negative attraction is necessary in this model, but electron kinetic energy and both repulsions also change with geometry. Another mistake is counting every electron pair twice by summing both i,j and j,i without a factor of one-half. A third is equating one fixed-geometry electronic energy difference with an experimental standard Gibbs energy; thermal, ensemble and environment terms may be substantial.

Worked example

For H₂⁺ in atomic units, take two protons separated by R = 2.0 bohr. The nuclear–nuclear repulsion term is Z AZ B/R = 1/2.0 = +0.50 hartree. At R = 4.0 bohr it is +0.25 hartree. The repulsion therefore falls by 0.25 hartree on separation. This does not prove the total energy falls by 0.25 hartree, because the electron's kinetic and attraction energies change as its wavefunction adjusts at the new geometry. If separate electronic calculations gave, purely as an illustrative input, electron contributions of −1.10 and −0.70 hartree at the two geometries, total fixed-geometry energies would be −0.60 and −0.45 hartree. In that invented data the shorter geometry is lower in total energy despite its larger nuclear repulsion. The example shows why all terms must be combined; the illustrative electronic numbers are not asserted physical benchmarks for H₂⁺.

Quick check

1. Which term is absent for a molecular ion containing only one electron? Answer: Electron–electron repulsion V ee is absent because no electron pair exists. 2. Why can the nuclear-repulsion energy be treated as constant in one fixed-geometry electronic calculation but not across several geometries? Answer: Nuclear positions are fixed within one calculation, but their separations and hence V NN change between geometries.

Exam focus

Name the five Hamiltonian terms and their physical signs. In atomic units, use pair sums without double counting. Distinguish the full electron-plus-nucleus problem from the fixed-nuclei electronic problem. Include nuclear repulsion when constructing a potential-energy curve. Explain why an electronic-energy result needs further steps to compare with a measured free energy or spectrum. The H₂⁺ one-electron example is a good check of which terms vanish.

Advanced insight

The exact separation between energy components depends on the chosen model and coordinate conventions, but predicted observables come from the complete, consistently handled Hamiltonian. Electron correlation, the failure of simple mean-field methods for some states, is not an omitted Coulomb term: it arises from how well the interacting many-electron equation is solved. In contrast, a missing solvent or relativistic effect changes the modeled Hamiltonian or environment itself. That distinction guides method development and error budgets throughout computational chemistry.

Summary

The basic nonrelativistic molecular Hamiltonian contains electron and nuclear kinetic energies, electron–nucleus attraction, electron–electron repulsion and nuclear–nuclear repulsion. Its Schrödinger equation connects these modeled interactions to energies and wavefunctions. Fixed-nuclei calculations simplify the problem but retain nuclear repulsion in geometry comparisons. Useful chemical predictions require a suitable solution method and careful conversion of model energies into the actual observable.

Practice questions

1. How many electron–electron pair terms are present for three electrons before considering symmetry or spin? Answer: Three distinct pairs: (1,2), (1,3) and (2,3). 2. What is V NN for two unit-charge nuclei 5 bohr apart in atomic units? Answer: V NN = 1/5 = +0.20 hartree. 3. If nuclei are fixed for an electronic calculation, is their kinetic-energy operator included in the electronic equation? Answer: No. Nuclear motion is set aside in that fixed-geometry electronic problem, though nuclear repulsion remains in the total energy. 4. Why does a predicted electronic-energy difference not automatically equal a solution-phase equilibrium free-energy difference? Answer: The latter also depends on thermal motion, entropy, solvation, standard states and possible conformer populations.