Born–Huang Expansion in Molecular Quantum Mechanics

• Born–Huang expansion is a multi-channel adiabatic method that represents the full molecular wavefunction using a complete set of electronic eigenfunctions and corresponding nuclear coefficients.
• It generalizes the Born–Oppenheimer approximation by retaining non-adiabatic derivative couplings, which facilitate transitions between potential energy surfaces and model phenomena like avoided crossings.
• The framework is crucial in quantum chemistry and has been extended to areas such as cavity QED, relativistic formulations, and adiabatic spin dynamics for describing electron–nuclear entanglement.

The Born–Huang expansion is the general adiabatic expansion of a molecular wavefunction in a basis of electronic states that depend parametrically on nuclear coordinates. In its canonical form, the full electron–nuclear wavefunction is written as a sum over adiabatic electronic eigenfunctions, with nuclear coefficient functions determined by coupled equations that retain non-adiabatic derivative couplings. It therefore extends the single-product Born–Oppenheimer approximation from nuclear motion on one potential energy surface to coupled nuclear dynamics on multiple surfaces, and it provides the standard coupled-channel framework for avoided crossings, conical intersections, vibronic mixing, and several modern generalizations in cavity QED and related settings (Kerley, 2013).

1. Formal definition and basis structure

In the molecular setting, the starting point is the nonrelativistic Hamiltonian

$$\hat{H}=\hat{T}_e+\hat{T}_N+\hat{U}(\mathbf{x},\mathbf{R}),$$

with electronic coordinates $\mathbf{x}$, nuclear coordinates $\mathbf{R}$, and electronic and nuclear kinetic-energy operators $\hat T_e$ and $\hat T_N$. The adiabatic electronic basis is obtained from the clamped-nuclei eigenproblem

$$\left[\hat{T}_e+\hat{U}(\mathbf{x},\mathbf{R})\right]\phi_n(\mathbf{x};\mathbf{R}) = E_n(\mathbf{R})\,\phi_n(\mathbf{x};\mathbf{R}),$$

where $E_n(\mathbf{R})$ defines an adiabatic potential energy surface. In the Born–Oppenheimer approximation one keeps a single channel and writes

$$\Psi^{\mathrm{BO}}_{n,m}(\mathbf{x},\mathbf{R}) = \phi_n(\mathbf{x};\mathbf{R})\,\chi_{n,m}(\mathbf{R}),$$

with $\chi_{n,m}$ solving a nuclear Schrödinger equation on the single surface $E_n(\mathbf R)$ (Mosquera et al., 29 Aug 2025).

The Born–Huang expansion removes that single-channel truncation and writes the total state as

$\mathbf{x}$0

where the $\mathbf{x}$1 form a complete set of adiabatic Born–Oppenheimer electronic eigenfunctions and the $\mathbf{x}$2 are nuclear coefficient functions. In equivalent notation, exact molecular eigenfunctions may be written as

$\mathbf{x}$3

with electronic basis functions $\mathbf{x}$4 orthonormal in the electronic coordinates and nuclear coefficients $\mathbf{x}$5 that encode both motion on and transitions between surfaces (Kerley, 2013).

The formal distinction is therefore precise. The Born–Oppenheimer approximation is a single-channel truncation of the adiabatic expansion together with neglect of interchannel couplings, whereas the Born–Huang expansion is the multi-channel adiabatic expansion itself. In that sense, Born–Huang is not merely a correction term but the general coupled-state ansatz from which the Born–Oppenheimer approximation is recovered only after further approximations.

2. Coupled nuclear equations and non-adiabatic coupling operators

Substituting the Born–Huang ansatz into the full Schrödinger equation and projecting onto a chosen electronic state yields coupled nuclear equations. In one standard form,

$\mathbf{x}$6

where $\mathbf{x}$7 is the total vibronic energy and $\mathbf{x}$8 is the non-adiabatic coupling operator (Mosquera et al., 29 Aug 2025).

For the adiabatic electronic basis, the Hermitian coupling operator is

$\mathbf{x}$9

with first- and second-derivative couplings

$\mathbf{R}$0

$\mathbf{R}$1

The first-order derivative coupling $\mathbf{R}$2 is the usual non-adiabatic derivative coupling. The second-order term $\mathbf{R}$3 is kept in the formal derivation, although it is described as usually ignored in practical computations (Mosquera et al., 29 Aug 2025).

A convenient computational representation expands each nuclear coefficient in vibrational eigenfunctions on its own adiabatic surface,

$\mathbf{R}$4

which converts the coupled differential system into a matrix eigenvalue problem,

$\mathbf{R}$5

The full Born–Huang wavefunction then becomes

$\mathbf{R}$6

so the formalism is explicitly a superposition over both electronic states $\mathbf{R}$7 and vibrational states $\mathbf{R}$8, with coefficients determined by vibronic non-adiabatic couplings (Mosquera et al., 29 Aug 2025).

Kerley’s formulation uses equivalent structures. After projecting onto the electronic basis, the derivative terms appear as a first-derivative operator $\mathbf{R}$9 and a second-derivative term $\hat T_e$0, with diagonal matrix elements yielding diagonal Born–Huang corrections and off-diagonal matrix elements generating transitions between surfaces. In this language, the coupled equations are the precise realization of adiabatic and non-adiabatic corrections beyond the single-surface picture (Kerley, 2013).

If all non-adiabatic couplings are neglected, the channels decouple and the Born–Huang expansion reduces to an uncoupled sum of Born–Oppenheimer terms. Physically, this is the limit of independent vibronic manifolds on separate potential energy surfaces.

3. Relation to Born–Oppenheimer, diabatic representations, and entanglement

The Born–Huang expansion extends the Born–Oppenheimer approximation in two distinct senses. First, it retains the complete adiabatic electronic basis instead of a single product state. Second, it retains the derivative-coupling terms that couple nuclear motion across electronic channels. The resulting dynamics is therefore motion on coupled surfaces rather than motion on a single potential energy surface (Mosquera et al., 29 Aug 2025).

The distinction between adiabatic and diabatic pictures is central. In the adiabatic Born–Oppenheimer picture, the electronic states are eigenstates of the clamped-nuclei Hamiltonian and non-adiabatic couplings become large near avoided crossings. In a diabatic picture, the electronic states are obtained by unitary rotations of adiabatic states near avoided crossings; derivative couplings are transformed away and replaced by electrostatic couplings $\hat T_e$1, while the diabatic potentials $\hat T_e$2 cross. The Born–Huang framework itself is built in the adiabatic basis, so it makes the non-adiabatic couplings explicit rather than transforming them away (Mosquera et al., 29 Aug 2025).

A notable modern development is the analysis of electron–nuclear entanglement within Born–Oppenheimer and Born–Huang states. The molecular Hilbert space is treated as a bipartite space,

$\hat T_e$3

and the Schmidt decomposition

$\hat T_e$4

yields the entanglement entropy

$\hat T_e$5

Within this analysis, even a Born–Oppenheimer product $\hat T_e$6 is a conditional product rather than a separable $\hat T_e$7, and is therefore generically entangled in the electron–nuclear bipartition. The ground Born–Oppenheimer vibronic state of a molecule may be regarded as almost separable and non-entangled, but this property worsens with vibrational excitation, which increases the entanglement monotonically. Avoided crossings strongly enhance entanglement in the adiabatic picture, whereas a diabatic representation substantially reduces it for many states (Mosquera et al., 29 Aug 2025).

In this interpretation, the nuclear wavefunction acts as a tester of the variation of the electronic wavefunction with geometry. If the electronic state remains nearly unchanged over the nuclear support, entanglement is small. If the electronic character changes strongly with geometry, reduced density matrices develop several significant eigenvalues and the entanglement grows. The Born–Huang expansion then reveals “synergistic vibronic states” whose entanglement is larger than that of any Born–Oppenheimer component in the superposition. In the Shin–Metiu model, Born–Huang states 95 and 96 at $\hat T_e$8 a.u. and $\hat T_e$9 a.u. are approximately equal mixtures of $\hat T_N$0 and $\hat T_N$1, and their entanglement exceeds that of either dominant Born–Oppenheimer component (Mosquera et al., 29 Aug 2025).

This also corrects a common oversimplification: the Born–Huang expansion is not only a formal device for computing non-adiabatic spectra, but also a natural framework for quantifying how coupled-surface structure reorganizes electron–nuclear correlations.

4. Extensions to cavity QED, geometric gauge structures, and relativistic formulations

The Born–Huang construction has been generalized beyond the standard electron–nuclear problem. In vibro-polaritonic chemistry, the slow variables are taken to be nuclei together with cavity coordinates, and the adiabatic basis is formed by electron–photon states. The generalized expansion reads

$\hat T_N$2

where $\hat T_N$3 are adiabatic electron–photon states and the coefficient functions $\hat T_N$4 describe nuclear–photon motion. Projection yields coupled nuclear–photon equations with non-adiabatic operators from both nuclear and cavity kinetic-energy terms. In this framework, generalized Hellmann–Feynman relations express cavity derivative couplings in terms of transition dipoles between adiabatic electron–photon states and adiabatic energy gaps. A crude VSC Born–Huang expansion built on adiabatic electronic states leads to a crude cavity Born–Oppenheimer approximation, and the crude CBO ground state is identified as a first-order approximation to the full CBO ground state (Fischer et al., 2023).

That extension sharpens the distinction between correlated and projected descriptions. The fully correlated CBO picture uses adiabatic electron–photon states and retains electron–photon entanglement in the ground state, whereas the crude CBO picture neglects cavity-induced non-adiabatic transition-dipole couplings to excited states. The neglected terms are precisely the channels through which electron–photon entanglement enters beyond the projected ground-state Hamiltonian (Fischer et al., 2023).

A different extension appears in magnetic chip traps for atoms with internal spin. There the Born–Huang expansion is performed in a local adiabatic spin basis,

$\hat T_N$5

and projection of the kinetic energy produces diagonal terms of the form

$\hat T_N$6

where $\hat T_N$7 is the Berry–Mead vector potential and $\hat T_N$8 is the Born–Huang scalar potential. Off-diagonal terms encode non-adiabatic couplings between trapping, flat, and anti-trapping channels. In this setting, the Born–Huang potential modifies the trap and counteracts the adiabatic potential, while the off-diagonal terms are responsible for Majorana losses when the control parameter $\hat T_N$9 is not small (Gürkan et al., 2016).

A further generalization is the relativistic Born–Oppenheimer–Huang approach based on Dirac electronic states. The total wavefunction keeps the same expansion structure,

$$\left[\hat{T}_e+\hat{U}(\mathbf{x},\mathbf{R})\right]\phi_n(\mathbf{x};\mathbf{R}) = E_n(\mathbf{R})\,\phi_n(\mathbf{x};\mathbf{R}),$$0

but the adiabatic electronic states are replaced by Dirac spinors, so the potential energy surfaces and non-adiabatic coupling terms inherit relativistic corrections. In Baer’s field-dressed formulation, time-dependent dressed non-adiabatic coupling terms satisfy wave equations reminiscent of Maxwell–Lorentz equations, motivating the terminology “molecular fields.” The formal structure of the Born–Huang nuclear equations is unchanged; relativity enters through the electronic ingredients used to construct the adiabatic basis and the coupling matrices (Baer, 2017).

5. Representative models and computational realizations

Reduced-dimensional models provide explicit realizations of the Born–Huang formalism. In a one-dimensional $$\left[\hat{T}_e+\hat{U}(\mathbf{x},\mathbf{R})\right]\phi_n(\mathbf{x};\mathbf{R}) = E_n(\mathbf{R})\,\phi_n(\mathbf{x};\mathbf{R}),$$1 model, Born–Oppenheimer electronic and nuclear problems were solved with a Fourier Grid Hamiltonian, and the entanglement of Born–Oppenheimer vibronic states was evaluated from reduced density matrices. In that study the Born–Huang expansion was not explicitly constructed for $$\left[\hat{T}_e+\hat{U}(\mathbf{x},\mathbf{R})\right]\phi_n(\mathbf{x};\mathbf{R}) = E_n(\mathbf{R})\,\phi_n(\mathbf{x};\mathbf{R}),$$2, but its formal structure remained applicable in principle (Mosquera et al., 29 Aug 2025).

The Shin–Metiu model offers a complete Born–Huang implementation. For a parameter choice yielding three adiabatic potential energy surfaces $$\left[\hat{T}_e+\hat{U}(\mathbf{x},\mathbf{R})\right]\phi_n(\mathbf{x};\mathbf{R}) = E_n(\mathbf{R})\,\phi_n(\mathbf{x};\mathbf{R}),$$3, $$\left[\hat{T}_e+\hat{U}(\mathbf{x},\mathbf{R})\right]\phi_n(\mathbf{x};\mathbf{R}) = E_n(\mathbf{R})\,\phi_n(\mathbf{x};\mathbf{R}),$$4, and $$\left[\hat{T}_e+\hat{U}(\mathbf{x},\mathbf{R})\right]\phi_n(\mathbf{x};\mathbf{R}) = E_n(\mathbf{R})\,\phi_n(\mathbf{x};\mathbf{R}),$$5, the model exhibits a sharp avoided crossing between $$\left[\hat{T}_e+\hat{U}(\mathbf{x},\mathbf{R})\right]\phi_n(\mathbf{x};\mathbf{R}) = E_n(\mathbf{R})\,\phi_n(\mathbf{x};\mathbf{R}),$$6 and $$\left[\hat{T}_e+\hat{U}(\mathbf{x},\mathbf{R})\right]\phi_n(\mathbf{x};\mathbf{R}) = E_n(\mathbf{R})\,\phi_n(\mathbf{x};\mathbf{R}),$$7 at $$\left[\hat{T}_e+\hat{U}(\mathbf{x},\mathbf{R})\right]\phi_n(\mathbf{x};\mathbf{R}) = E_n(\mathbf{R})\,\phi_n(\mathbf{x};\mathbf{R}),$$8 a.u. and a broader avoided crossing between $$\left[\hat{T}_e+\hat{U}(\mathbf{x},\mathbf{R})\right]\phi_n(\mathbf{x};\mathbf{R}) = E_n(\mathbf{R})\,\phi_n(\mathbf{x};\mathbf{R}),$$9 and $E_n(\mathbf{R})$0 at $E_n(\mathbf{R})$1 a.u. The first-order non-adiabatic couplings are obtained via Hellmann–Feynman or from analytic derivatives of diabatization angles, the second-order couplings are reconstructed through completeness, and the resulting vibronic coupling matrix is diagonalized to obtain Born–Huang energies and coefficients $E_n(\mathbf{R})$2. Numerically, most Born–Huang states lie close to one of the Born–Oppenheimer entropy curves, reflecting dominance by a single Born–Oppenheimer component, while a subset of strongly mixed states lies significantly above those curves and realizes the synergistic regime (Mosquera et al., 29 Aug 2025).

Kerley’s exactly solvable coupled-oscillator model serves a different purpose: it illustrates the perturbative mechanics of the Born–Huang framework. The model mimics a light electronic degree of freedom coupled to a heavier nuclear one, and the analysis shows that the adiabatic result is good when the mass-ratio parameter is small, that second-order non-adiabatic corrections greatly improve the energies, and that a fuller perturbative treatment reproduces the exact spectrum nearly perfectly. The example is used to separate adiabatic corrections that modify a surface from non-adiabatic corrections that mix motion across surfaces (Kerley, 2013).

These examples highlight a practical pattern. In weakly coupled regimes, a single Born–Oppenheimer component may dominate and the Born–Huang state mainly refines energies. Near avoided crossings, near-degeneracies, or strong transition-dipole couplings, the same formalism becomes essential for the structure of the eigenstates themselves.

6. Interpretive issues, variants, and terminological ambiguity

One methodological variant is Kerley’s “xiabatic” reformulation. Instead of taking clamped-nuclei electronic eigenfunctions as the adiabatic basis without modification, he defines electronic states by requiring

$E_n(\mathbf{R})$3

so that all adiabatic corrections that do not explicitly mix nuclear wavefunctions are absorbed into redefined electronic states and “xiabatic surfaces.” The remaining off-diagonal terms are then treated as non-adiabatic corrections by perturbation theory. In this construction, the Born–Oppenheimer approximation is recovered by using clamped-nuclei eigenfunctions, ignoring $E_n(\mathbf{R})$4, $E_n(\mathbf{R})$5, and $E_n(\mathbf{R})$6, and truncating to a single electronic state (Kerley, 2013).

The formal perturbative status of the adiabatic Born–Oppenheimer/Born–Huang expansion has also been challenged. Zakharov argues that standard perturbation theory requires identical operator domains, $E_n(\mathbf{R})$7, whereas the usual Born–Oppenheimer partitioning takes the unperturbed operator to be an electronic Hamiltonian acting on $E_n(\mathbf{R})$8 and the full operator to act on $E_n(\mathbf{R})$9. On this basis, he concludes that the adiabatic Born–Oppenheimer expansion does not satisfy the necessary condition for applicability of perturbation theory and proposes an alternative scheme in which both the unperturbed and full Hamiltonians act on the full molecular Hilbert space, with auxiliary binding potentials used to define a bound unperturbed problem (Zakharov, 2017).

This critique does not eliminate the operational role of the Born–Huang expansion in molecular quantum mechanics, but it does distinguish two questions that are often conflated: whether the coupled-channel adiabatic expansion is the correct formal representation of the molecular problem, and whether its usual mass-ratio interpretation qualifies as regular perturbation theory in the operator-theoretic sense. A plausible implication is that the Born–Huang expansion should be viewed primarily as the exact adiabatic coupled-basis framework, with perturbative interpretations depending on the precise partition used.

There is also a terminological ambiguity across subfields. In the dilute Fermi-gas literature, “Born–Huang” or “Huang–Yang” can refer to a low-density asymptotic expansion of the ground-state energy in powers of density or $$\Psi^{\mathrm{BO}}_{n,m}(\mathbf{x},\mathbf{R}) = \phi_n(\mathbf{x};\mathbf{R})\,\chi_{n,m}(\mathbf{R}),$$0, rather than to the molecular coupled-surface adiabatic expansion. In that usage, the expansion concerns the universal three-term asymptotics of a dilute spin-$$\Psi^{\mathrm{BO}}_{n,m}(\mathbf{x},\mathbf{R}) = \phi_n(\mathbf{x};\mathbf{R})\,\chi_{n,m}(\mathbf{R}),$$1 Fermi gas and depends on the interaction only through the scattering length (Giacomelli et al., 28 May 2025). This suggests that, outside molecular quantum mechanics and related adiabatic theories, the label “Born–Huang expansion” may designate a mathematically distinct asymptotic series.

Taken in its standard molecular sense, however, the Born–Huang expansion remains the canonical multi-channel adiabatic decomposition of the electron–nuclear wavefunction: exact as a basis expansion, nontrivial because of derivative couplings, and indispensable whenever the single-surface picture fails.