The Franck-Condon Principle: Why Electronic Transitions Are Vertical

The core idea: freezing the nuclei during an electronic leap

The Franck-Condon principle states that an electronic transition happens so much faster than nuclear motion that the nuclei — their positions and their momenta — are effectively frozen for the duration of the jump. The absorption of a visible or UV photon promotes an electron in ~10⁻¹⁵ s; a molecular vibration, even a fast X–H stretch, takes ~10⁻¹⁴ s, and a heavy-atom stretch like I–I takes ~10⁻¹³ s. On the potential-energy-versus-internuclear-distance diagram, the transition therefore proceeds vertically: the system moves straight up (or down) at fixed nuclear geometry from one Born-Oppenheimer surface to another.

The physical consequence is that a molecule is almost never 'born' into the vibrationless minimum of its excited state. If the excited electronic state has a longer equilibrium bond length than the ground state — which is typical, because promoting an electron into an antibonding-character orbital weakens bonding — then the vertical line launched from the ground-state minimum arrives on the upper surface at a point displaced from that surface's minimum, i.e. into a vibrationally excited level. The molecule finds itself compressed relative to its new preferred geometry and starts to vibrate.

James Franck articulated this in 1925 in semiclassical terms, reasoning about momentum conservation and the fact that nuclei dwell longest near the classical turning points of a vibration. Edward Condon supplied the quantum-mechanical machinery in a pair of 1926 and 1928 papers, recasting the qualitative rule as an overlap integral between vibrational wavefunctions. The two descriptions agree in the classical limit but the quantum version is essential to get the intensity of the crucial v″ = 0 band right.

The quantum derivation: factoring the transition dipole

Start from Fermi's golden rule: the intensity of a transition between an initial vibronic state |Ψ_i⟩ and final |Ψ_f⟩ is proportional to |⟨Ψ_f|μ̂|Ψ_i⟩|², where μ̂ is the electric dipole operator. Invoking the Born-Oppenheimer approximation, each vibronic wavefunction factors into an electronic part depending parametrically on nuclear coordinates and a nuclear (vibrational) part: Ψ = ψ_el(r; R)·χ_vib(R). The dipole operator splits into electronic and nuclear pieces, μ̂ = μ̂_el + μ̂_nuc.

Because the electronic states are orthogonal, the nuclear-dipole term vanishes between different electronic states, and the surviving matrix element is ⟨χ_v′|⟨ψ_el′|μ̂_el|ψ_el″⟩|χ_v″⟩. The Condon approximation now assumes the electronic transition dipole μ_el(R) = ⟨ψ_el′|μ̂_el|ψ_el″⟩ varies slowly with R and can be pulled out at the equilibrium geometry R₀. The matrix element then factorizes cleanly:

• μ_fi ≈ μ_el(R₀) · ⟨χ_v′ | χ_v″⟩ — an electronic dipole times a pure vibrational overlap integral.
• The line intensity is I ∝ |μ_el(R₀)|² · |⟨χ_v′|χ_v″⟩|²; the second factor is the dimensionless Franck-Condon factor, FCF = |⟨χ_v′|χ_v″⟩|².

Crucially, there is no Δv selection rule. Any v″ → v′ is allowed; its brightness is simply set by how much the two vibrational wavefunctions overlap. Summed over all final vibrational levels, the overlaps obey a completeness (sum) rule, Σ_v′ |⟨χ_v′|χ_v″⟩|² = 1, so the vibronic structure redistributes a fixed total electronic oscillator strength among the vibronic lines — it never creates or destroys it.

Displaced harmonic oscillators and the Huang-Rhys factor

To turn the overlap integral into numbers, model both electronic states as harmonic oscillators of the same frequency ω but with minima displaced along a normal coordinate by ΔQ. Starting from the vibrationless ground state (v″ = 0), the Franck-Condon factors to the upper levels v′ follow a Poisson distribution:

• |⟨v′|0⟩|² = e^(−S) · Sⱽ′ / v′! , where S is the dimensionless Huang-Rhys factor, S = ½ μω(ΔQ)²/ℏ (in ordinary normal coordinates); in mass-weighted coordinates the reduced mass is absorbed into Q and drops out, giving S = ½(ω/ℏ)(ΔQ)².

S is the average number of vibrational quanta deposited in the excited state by a vertical transition — physically, the reorganization energy λ = Sℏω expressed in quanta. The distribution peaks near v′ ≈ S, so the brightest vibronic band is generally not the 0–0 line whenever the excited-state geometry is displaced (S > 0). When S ≈ 0 (no geometry change, ΔQ ≈ 0), overlap is perfect only for v′ = v″, ⟨v′|0⟩² = δ_{v′,0}, and the 0–0 band carries essentially all the intensity — a sharp, structureless origin.

The shape of the progression thus reads out the excited-state geometry change directly. A small S (say 0.3) gives a spectrum dominated by the 0–0 band with a weak first member; a large S (say 5) gives a broad, bell-shaped envelope of many lines whose maximum sits five quanta up. This is exactly why the strongly displaced B ³Πₒᵤ⁺ ← X ¹Σᵍ⁺ absorption of I₂ shows dozens of resolved members, whereas a rigid aromatic like benzene shows a comparatively compact vibronic pattern.

Worked example: I₂, the textbook long progression

Molecular iodine is the canonical demonstration. Its ground state X ¹Σᵍ⁺ has an equilibrium bond length R_e ≈ 2.666 Å and a vibrational spacing ω_e ≈ 214 cm⁻¹. The bound excited state B ³Πₒᵤ⁺, reached in the visible near 500–650 nm, has a substantially longer bond, R_e′ ≈ 3.02 Å, and a softer spring, ω_e′ ≈ 125 cm⁻¹, because the promoted electron occupies an orbital with more antibonding character. The bond lengthens by ΔR ≈ 0.35 Å on excitation.

That displacement makes the vertical transition from the v″ = 0 level land high on the B-state ladder — around v′ ≈ 25–27 near the absorption maximum — producing the famous violet-to-red banded absorption you can resolve with a benchtop grating spectrometer. Because ΔR is large, the Huang-Rhys factor is large, the Poisson envelope is broad, and the 0–0 band is vanishingly weak while the intensity maximum sits far up the progression. As one climbs toward the dissociation limit, the vibrational spacings converge (anharmonicity), and a Birge-Sponer extrapolation of those shrinking intervals yields the B-state dissociation energy, D₀′ ≈ 4381 cm⁻¹ above the potential minimum — a classic undergraduate lab result.

Contrast this with a nearly rigid chromophore. In formaldehyde, the S₁ ← S₀ (n→π*) transition puts an electron into the antibonding π*, pyramidalizing the carbon and lengthening C=O only modestly; the dominant activity is in the out-of-plane bending mode and a short C=O stretch progression. The same principle — vertical transition, overlap-weighted intensity — explains both the sprawling I₂ envelope and the compact H₂CO pattern; only the magnitude of ΔQ (hence S) differs.

Stokes shift, mirror symmetry, and the origin of the 0–0 offset

The Franck-Condon principle governs emission just as it governs absorption, and the two together explain the Stokes shift — the fact that fluorescence appears at lower energy than absorption. On absorption, the molecule rises vertically into a vibrationally hot upper state, then relaxes via internal conversion and vibrational cooling (~10⁻¹²–10⁻¹¹ s) to the relaxed v′ = 0 minimum of the lowest excited singlet, consistent with Kasha's rule. From there it emits, again vertically, down onto a vibrationally excited level of the ground state. Both vertical steps overshoot the target minimum, so the emitted photon is redshifted from the absorbed one by roughly twice the reorganization energy, Δ(Stokes) ≈ 2λ = 2Sℏω, for displaced harmonic oscillators.

A second, beautiful consequence is mirror-image symmetry. When the excited- and ground-state vibrational frequencies are similar and only the origin is displaced, the absorption vibronic progression and the fluorescence progression are approximate mirror images reflected about the shared 0–0 line. Deviations from mirror symmetry are diagnostic: they flag frequency changes, large geometry distortions, or excited-state chemistry (proton transfer, twisting) between absorption and emission.

These ideas connect directly to Marcus theory of electron transfer, where the same displaced-oscillator, overlap-weighted formalism reappears: the classical Marcus reorganization energy λ is precisely the Sℏω of the accepting vibrational modes, and the quantum (Marcus-Levich-Jortner) rate expression contains an explicit sum of Franck-Condon factors, e^(−S)Sⁿ/n!, coupling reactant and product nuclear wavefunctions. Nonradiative decay rates in the energy-gap law likewise hinge on Franck-Condon overlaps between initial and final surfaces separated by a large energy gap.

Limits, breakdown, and where the principle earns its keep

The Franck-Condon factorization is not exact — it rests on two approximations that can fail. The first is the Condon approximation itself: pulling μ_el(R) out of the integral assumes the electronic transition dipole is R-independent. When the pure electronic transition is symmetry-forbidden, μ_el(R₀) = 0 and the naive Franck-Condon intensity vanishes. Intensity is then borrowed through the linear term in μ_el(R), i.e. through vibronic (Herzberg-Teller) coupling: a non-totally-symmetric vibration lowers the instantaneous symmetry and 'switches on' the transition. The forbidden ¹B₂ᵤ ← ¹A₁ᵍ transition of benzene near 260 nm is the archetype — it is nominally symmetry-forbidden by g/u and orbital-symmetry rules, and appears only because the e₂ᵍ ring-distortion mode (ν₆) supplies a false origin. Its progression is built on that vibronic origin, not the true 0–0.

The second limit is the Born-Oppenheimer approximation breaking down near conical intersections, where two electronic surfaces touch and nuclear and electronic motions become strongly coupled. There the clean separation into electronic dipole × nuclear overlap loses meaning, and ultrafast nonadiabatic dynamics (the workhorse of photochemistry — vision, DNA photostability, photoswitches) takes over. Franck-Condon thinking still sets the initial condition: a pulse prepares a wavepacket on the upper surface at the vertical (Franck-Condon) geometry, which then evolves.

Where the principle earns its keep is enormous. It quantifies vibronic band intensities in UV-vis absorption and fluorescence; it explains the vibrational structure of photoelectron spectra, where the geometry difference between neutral and cation prints a Franck-Condon envelope (a single sharp line means the ionized orbital was nonbonding); it underpins resonance Raman intensities, single-molecule spectral line shapes, OLED emission profiles, and the calculation of nonradiative rates. Modern quantum-chemistry packages compute multidimensional Franck-Condon factors from ab initio ground- and excited-state geometries and Hessians to simulate whole spectra — a direct, still-thriving descendant of Condon's 1928 overlap integral.

Franck-Condon principle: classical (Franck) versus quantum-mechanical (Condon) formulations of the same physics

Aspect	Classical picture (Franck)	Quantum picture (Condon)
Central idea	Nuclei don't move during the electronic jump; position and momentum are conserved	Transition dipole factors into electronic × nuclear overlap: μ = μ_el·⟨χ_v′|χ_v″⟩
Geometric rule	Transition is a vertical line on the E-vs-R diagram	Intensity ∝ |⟨χ_v′|χ_v″⟩|², the Franck-Condon factor
Most probable region	Turning points of the vibration (nuclei linger there longest)	Regions of maximal wavefunction amplitude overlap
Best-intensity band	Vertical line from the lower-state minimum (ground v″=0 max density is at center)	The v′ whose χ_v′ has a lobe over the v″=0 Gaussian, i.e. v′ ≈ S
Handles zero-point motion	No — misplaces intensity for v″=0 (density peaks at center, not turning points)	Yes — correctly gives 0–0 dominance when ΔR ≈ 0
Isotope / mode structure	Qualitative only	Quantitative progressions, hot bands, mode-specific factors