The Born-Oppenheimer approximation

Result

The Born-Oppenheimer approximation

T-074Home CU-301Threads quantum

Statement

Separating slow nuclei from fast electrons.

Why it matters

The full molecular Schrödinger equation, treating every nucleus and every electron on an equal footing, is intractable to solve exactly for essentially any molecule of chemical interest. The Born-Oppenheimer approximation is what makes molecular quantum chemistry practical at all: it separates the problem into an electronic part, solved at a fixed nuclear geometry, and a nuclear part, which then moves on the resulting potential energy surface — a separation so foundational that particle-in-a-box, hydrogen-atom-solution, huckel-theory, and lcao-molecular-orbitals all implicitly assume it, solving only the electronic problem at fixed nuclear positions without further comment.

The very concept of a molecule having a definite equilibrium geometry, a bond length, or a vibrational spectrum at all depends on this separation: without it, "nuclear positions" would not even be well-defined quantities to plot an energy against, since nuclei and electrons would be treated as one inseparable quantum system with no meaningful notion of a fixed molecular shape.

Hypotheses

Nuclei are far more massive than electrons (a proton outweighs an electron by a factor of roughly \(1800\)), so electrons respond to nuclear motion essentially instantaneously.This large mass ratio is the entire physical justification for the approximation: on the timescale over which nuclei move appreciably, electrons have already fully relaxed to whatever their instantaneous, lowest-energy configuration is for the nuclei's current positions, allowing electronic and nuclear motion to be decoupled. The total molecular wavefunction can be factorised as \(\Psi(\mathbf{r},\mathbf{R}) \approx \psi_{\text{elec}}(\mathbf{r};\mathbf{R})\,\chi_{\text{nuc}}(\mathbf{R})\).The semicolon notation is deliberate: the electronic wavefunction depends parametrically on the nuclear coordinates \(\mathbf{R}\) (solved anew for each fixed nuclear geometry) but not on nuclear momenta, while the nuclear wavefunction depends explicitly on \(\mathbf{R}\) as a genuine dynamical variable, moving on the potential surface the electronic calculation supplies. The approximation breaks down where two electronic states approach closely in energy as a function of nuclear geometry (an avoided crossing or a true conical intersection), since the assumption that electrons stay in one well-separated electronic state while nuclei move is precisely what fails there; such breakdowns are central to understanding photochemical reactions and non-radiative relaxation processes.

Proof

\hat{H} = \hat{T}_{\text{nuc}} + \hat{T}_{\text{elec}} + V(\mathbf{r},\mathbf{R})

The full molecular Hamiltonian contains nuclear kinetic energy, electronic kinetic energy, and a potential energy term depending on both electron and nuclear positions (nuclear-nuclear repulsion, electron-nuclear attraction, and electron-electron repulsion together); solving the full Schrödinger equation with this Hamiltonian exactly, for more than a few particles, is computationally intractable. A

\big[\hat{T}_{\text{elec}} + V(\mathbf{r},\mathbf{R})\big]\,\psi_{\text{elec}}(\mathbf{r};\mathbf{R}) = E_{\text{elec}}(\mathbf{R})\,\psi_{\text{elec}}(\mathbf{r};\mathbf{R})

Because \(\hat T_{\text{nuc}}\) is negligible on the electronic timescale (Hypotheses), the nuclear kinetic energy term is dropped from the electronic problem, and the electronic Schrödinger equation is solved treating the nuclear positions \(\mathbf{R}\) as fixed, externally specified parameters rather than dynamical variables. B

E_{\text{elec}}(\mathbf{R}) \equiv \text{potential energy surface}

Repeating Step 2's electronic calculation at many different fixed nuclear geometries \(\mathbf{R}\) traces out a function \(E_{\text{elec}}(\mathbf{R})\), the potential energy surface, giving the total electronic energy (including nuclear-nuclear repulsion) as a function of geometry alone. A

\big[\hat{T}_{\text{nuc}} + E_{\text{elec}}(\mathbf{R})\big]\,\chi_{\text{nuc}}(\mathbf{R}) = E\,\chi_{\text{nuc}}(\mathbf{R})

The nuclei are then treated as moving quantum mechanically not through the original, complicated multi-particle potential, but on the single, effective potential energy surface \(E_{\text{elec}}(\mathbf{R})\) supplied by the electronic calculation — the nuclear Schrödinger equation, whose solutions describe molecular vibration and rotation. B

\text{An equilibrium bond length/geometry corresponds to a minimum of } E_{\text{elec}}(\mathbf{R}).

Because a stable molecular geometry is, by definition, one at which nuclei experience no net force, and force is minus the gradient of potential energy, the equilibrium geometry is exactly the minimum (or a stationary point, more generally) of the potential energy surface constructed in Step 3. A

Result

\Psi(\mathbf{r},\mathbf{R}) \approx \psi_{\text{elec}}(\mathbf{r};\mathbf{R})\,\chi_{\text{nuc}}(\mathbf{R})

Reading. The intractable, fully coupled electron-nucleus problem is replaced by two much simpler, sequential problems: solve for the electrons at a fixed nuclear geometry to get a potential energy surface, then let the nuclei move on that surface.

Scope. Excellent for the great majority of ground-state chemistry, where electronic states remain well separated in energy across the geometries of interest; breaks down near electronic state crossings (Hypotheses), most relevant in photochemistry and certain reactive intermediates.

Corollaries & converses

variational-principle's approach to approximating the electronic ground-state energy, and hydrogen-atom-solution's exact electronic solution, are both explicitly solutions of the fixed-nuclei electronic problem of Step 2, not of the full coupled electron-nuclear problem.
The very existence of a molecular potential energy surface, and hence the entire concept of a reaction coordinate, transition state, or equilibrium bond length used throughout kinetics and structural chemistry, presupposes this separation; without it, "the energy as a function of geometry" would not be a meaningful, well-defined quantity at all.
Converse: observing genuinely anomalous vibrational or spectroscopic behaviour that cannot be explained by motion on a single, well-defined potential energy surface is itself evidence of Born-Oppenheimer breakdown (a nearby electronic state crossing), rather than an error in the vibrational analysis itself.

Fails without

Apply the approximation near an electronic state crossing or conical intersection: the assumption that nuclei remain on one well-separated electronic surface throughout their motion (Hypotheses, third point) fails, and non-adiabatic coupling allows population to transfer between electronic states — exactly the situation central to much of photochemistry.
Apply the approximation to a hypothetical system with comparable nuclear and electronic mass: without the large mass ratio (Hypotheses, first point), there is no justification for solving the electronic problem at a frozen nuclear geometry, since electrons would no longer respond much faster than the nuclei move.

Common errors

Treating the Born-Oppenheimer approximation as exact rather than as an approximation whose validity rests specifically on the large nuclear-to-electron mass ratio (Hypotheses).
Assuming a single potential energy surface adequately describes any photochemical process; many photochemical reactions specifically involve the nuclei crossing between two different, closely spaced electronic surfaces, precisely where the approximation is least reliable.
Forgetting that the electronic wavefunction depends parametrically, not dynamically, on \(\mathbf{R}\) — it is recalculated fresh at each geometry, not evolved forward in time alongside the nuclei.
Confusing the potential energy surface \(E_{\text{elec}}(\mathbf{R})\) (Step 3) with the full molecular energy \(E\) of Step 4, which also includes the nuclear kinetic and vibrational/rotational energy.

Discussion

Max Born and J. Robert Oppenheimer introduced this separation in 1927, shortly after the Schrödinger equation itself was formulated, recognising that the large electron-nucleus mass disparity permitted exactly this kind of decoupling. The approximation is arguably the single most consequential simplification in the entire history of quantum chemistry, since without it essentially no practical molecular calculation, from the simplest diatomic to the largest protein, would be computationally feasible at all.

Corrections beyond the Born-Oppenheimer approximation (non-adiabatic coupling terms, describing genuine mixing between different electronic states as nuclei move) become essential precisely in photochemistry, where a molecule absorbing light is promoted to an excited electronic surface and must eventually find its way back to the ground surface, very often through exactly the kind of near-degeneracy the approximation assumes does not occur.

Common misconception: that the Born-Oppenheimer approximation means nuclei are simply "frozen" or ignored throughout a calculation. In fact nuclei are central to the second half of the approximation (Step 4): they move fully quantum mechanically, just on an effective potential surface supplied by a separate, prior electronic calculation, rather than being solved simultaneously and self-consistently with the electrons in one single step.

Worked examples

\text{H}_2^+ \text{: electronic energy computed at several fixed internuclear separations } R

For the simplest possible molecule, a single electron bound to two fixed protons, the electronic Schrödinger equation (Step 2) can be solved exactly at each chosen \(R\); repeating this at many values of \(R\) traces out the potential energy curve directly. A

E_{\text{elec}}(R) \text{ has a minimum at } R\approx1.06\,\text{Å, the equilibrium bond length}

The resulting curve falls from a high energy at very short \(R\) (nuclear-nuclear repulsion dominating), passes through a minimum, and rises again at large \(R\) (approaching the separated-atom limit); the minimum is, by Step 5, the predicted equilibrium bond length, closely matching the experimentally measured value. A

\text{Potential energy curve, } E_{\text{elec}}(R) \text{: minimum defines equilibrium bond length and curvature sets vibrational frequency}

Reading. Even the simplest possible molecular potential energy curve already illustrates the two central practical outputs of the Born-Oppenheimer approach: an equilibrium geometry, and (from the curve's curvature at that minimum) a vibrational force constant.

Scope. The same procedure, computationally far more demanding but conceptually identical, underlies every modern computed potential energy surface for larger polyatomic molecules (potential-energy-surfaces).

Problems

Explain briefly why the Born-Oppenheimer approximation would be expected to work far less well for a hypothetical system in which the "nuclei" and "electrons" had comparable mass.
Solution
The approximation rests entirely on electrons responding essentially instantaneously to nuclear motion because they are so much lighter (Hypotheses); if the two masses were comparable, neither particle type would move on a much faster timescale than the other, so there would be no justification for solving the electronic problem at a frozen nuclear geometry and treating the nuclear motion as a separate, subsequent step. Both particle types would need to be treated fully dynamically and simultaneously.
A molecule absorbs a photon and is promoted to an excited electronic state whose potential energy surface crosses closely with the ground-state surface at a particular geometry. Explain why the Born-Oppenheimer approximation is especially likely to break down near that crossing.
Solution
The approximation assumes the nuclei remain on a single, well-separated electronic surface throughout their motion (Hypotheses' third point). Near a close approach or crossing of two electronic surfaces, the energy gap between the two electronic states becomes small, and nuclear motion can genuinely couple the two electronic states together (non-adiabatic coupling), allowing population to transfer between surfaces — precisely the situation the single-surface, factorised wavefunction of the Result cannot describe.
Explain why the equilibrium bond length of a molecule is defined as a property of the electronic potential energy surface \(E_{\text{elec}}(R)\) alone, not of the full nuclear wavefunction \(\chi_{\text{nuc}}(R)\).
Solution
By Step 5, the equilibrium geometry is the geometry at which the net force on the nuclei is zero, i.e. the minimum of \(E_{\text{elec}}(R)\); this is a property of the surface itself, computed from the electronic problem alone (Step 2-3), before the nuclear motion on that surface is even considered. The nuclear wavefunction \(\chi_{\text{nuc}}(R)\), by contrast, describes how the nuclei actually move and vibrate around that minimum, including quantum zero-point motion that spreads the nuclear probability density around, but not exactly at, the strict potential-energy minimum.