Franck-Condon Factors and Vibronic Intensities

Franck-Condon analysis remains a working tool wherever a vibronic spectrum needs to be turned into structural or dynamical information about an electronically excited state. In organic light-emitting diode (OLED) and organic photovoltaic design, chemists tune the Huang-Rhys factor of emissive chromophores deliberately: a small $S$ concentrates emission into a narrow 0-0 line, sharpening color purity for display applications, while materials with large $S$ spread emission broadly, useful for white-light or broadband absorber design. In photosynthesis research, Franck-Condon-weighted density of states calculations, built directly from the overlap integrals derived here, feed into Marcus theory rate expressions for electron transfer between chlorophylls and quinones, connecting vibronic spectroscopy directly to charge-separation kinetics in the reaction center. Resonance Raman spectroscopy exploits the same overlap integrals from the opposite direction: the relative intensities of overtone progressions in resonance Raman spectra encode the same excited-state displacement parameters that govern absorption band shapes, and the two techniques are routinely combined to extract multidimensional excited-state geometry changes for photoactive molecules such as retinal in rhodopsin or the chromophores of fluorescent proteins. In astrochemistry, Franck-Condon factors computed for diatomic radicals and small polyatomics (CN, C₂, CH) determine which vibronic bands dominate cometary and interstellar emission spectra, allowing abundance and rotational-temperature determinations from telescope data. Modern computational chemistry packages compute multidimensional Franck-Condon factors, including Duschinsky rotation, directly from excited-state potential energy surfaces obtained via time-dependent density functional theory, making the vibronic envelope a routine predictive output rather than only a fitting tool — a natural bridge toward the excited-state decay processes treated under Fluorescence and Phosphorescence: The Quantum Picture.

Ask the tutor: How do Franck-Condon factors change when more than one normal mode is vibronically active and Duschinsky rotation mixes the modes between electronic states?

Consider a diatomic molecule undergoing an electronic transition from a ground state $S_0$ to an excited state $S_1$. Within the Born-Oppenheimer picture, each electronic state has its own potential energy curve $V_0(R)$ and $V_1(R)$, and superimposed on each curve is its own ladder of vibrational levels with wavefunctions $\chi''_{v''}(R)$ (ground electronic state) and $\chi'_{v'}(R)$ (excited electronic state). The electronic transition is fast; the nuclei do not move during the instant of absorption. What changes is which potential curve governs the nuclear dynamics after the jump. If the equilibrium bond length does not change between the two states ($R_e'=R_e''$), then the $v''=0$ wavefunction — a Gaussian centered at $R_e''$ — sits directly on top of the excited-state potential minimum. The vertical transition lands almost entirely on $v'=0$, and the spectrum is dominated by one sharp 0-0 line with a rapidly decaying progression above it. If, however, the excited state has a substantially different equilibrium bond length (common when the promoted electron occupies an antibonding orbital), the $v''=0$ wavefunction is no longer centered on the new potential minimum. Classically, the vertical line from $R_e''$ up to the excited curve lands on the repulsive wall of $V_1(R)$ at some higher vibrational level; quantum mechanically, the overlap integral $\langle \chi'_{v'}|\chi''_{0}\rangle$ is largest for some $v' > 0$, and the whole progression shifts its intensity maximum away from the origin. The Franck-Condon factor formalizes exactly this geometric argument as a number between 0 and 1: the squared overlap of two harmonic-oscillator wavefunctions living on displaced parabolas. The key parameter is the dimensionless displacement, usually packaged as the Huang-Rhys factor $S$, which is proportional to the square of how far the two minima are shifted, measured in units of the vibrational zero-point spread. Large $S$ means a large equilibrium shift (or, entering through the same combination, a low vibrational frequency with a small absolute shift) and produces long, Poisson-shaped vibronic progressions; small $S$ collapses the progression onto the origin band.

The lever that connects the picture to numbers is the vibrational overlap integral itself, and to compute it precisely we must write down the full molecular transition dipole moment and factor it under the Born-Oppenheimer approximation — which is exactly what the governing equations do next.

The full transition dipole moment between vibronic states $|\Psi_g\rangle = |\psi_g\rangle|\chi''_{v''}\rangle$ and $|\Psi_e\rangle = |\psi_e\rangle|\chi'_{v'}\rangle$, factorized under the Born-Oppenheimer approximation, is:

where:

• $\hat{\vec\mu}_{el}$, $\hat{\vec\mu}_{nuc}$ — electronic and nuclear parts of the dipole operator,
• $\langle \psi_e|\hat{\vec\mu}_{el}|\psi_g\rangle_{R_0}$ — the electronic transition dipole moment, evaluated at a fixed reference geometry $R_0$ (the Condon approximation),
• $\langle \chi'_{v'}|\chi''_{v''}\rangle$ — the vibrational overlap integral between initial and final nuclear wavefunctions,
• the second term vanishes exactly because $\langle\psi_e|\psi_g\rangle=0$ for orthogonal electronic states.

Reading: the entire vibronic transition dipole is the electronic transition dipole (a single number set by orbital symmetry and electron redistribution) multiplied by a purely nuclear overlap integral that depends only on which vibrational levels are involved. The electronic part sets the overall strength of the transition; the vibrational overlap distributes that strength across the vibronic progression. Squaring the transition moment and integrating over the emission or absorption lineshape gives the intensity of an individual vibronic line, and this is where the Franck-Condon factor enters explicitly:

where:

• $\nu^3$ applies for spontaneous emission rates (Einstein $A$ coefficient), $\nu^1$ for absorption cross sections in the usual oscillator-strength convention,
• $\big|\langle \chi'_{v'}|\chi''_{v''}\rangle\big|^2 \equiv q_{v'v''}$ is the Franck-Condon factor, dimensionless and bounded between 0 and 1,
• the electronic factor is treated as constant across the vibronic progression (Condon approximation): it does not vary line to line.

Reading: because the electronic dipole factor is common to every line in a progression, relative vibronic intensities within one electronic band are governed entirely by Franck-Condon factors. This is the single most important operational fact in vibronic spectroscopy — measure relative peak heights, and you are measuring $q_{v'v''}$ directly. The Franck-Condon factors obey a completeness (sum rule) relation that follows from the fact that the excited-state vibrational wavefunctions form a complete orthonormal set:

where:

• the sum runs over the complete set of excited-state (or ground-state, depending on direction) vibrational levels,
• normalization of $\chi''_{v''}$ makes the right-hand side exactly 1.

Reading: the Franck-Condon factors emanating from a single initial vibrational level must add up to unity — intensity is conserved and merely redistributed among the vibronic lines, never created or destroyed by the nuclear factor. Finally, for the workhorse model — two harmonic oscillators of the same force constant (same vibrational frequency $\omega$) but displaced equilibrium positions — the Franck-Condon factor for the 0→v′ progression collapses to a Poisson distribution in a single dimensionless parameter:

where:

• $S$ — the Huang-Rhys factor, $S = \dfrac{1}{2}\left(\dfrac{\Delta R}{\ell}\right)^2$ with $\ell=\sqrt{\hbar/(\mu\omega)}$ the harmonic oscillator length and $\Delta R$ the displacement between the two equilibrium geometries,
• $v'$ — vibrational quantum number of the final (excited) state,
• the distribution is properly normalized: $\sum_{v'=0}^{\infty} S^{v'}e^{-S}/v'! = 1$.

Reading: $S$ is literally the mean number of vibrational quanta excited by the transition (the Poisson mean), and it packages the geometry change, the reduced mass, and the vibrational frequency into one number that a single absorption spectrum can measure.

Example 1 — Foundational: Franck-Condon factors for a modestly displaced diatomic-like mode. Suppose a chromophore's active mode has vibrational frequency $\omega$ corresponding to $\tilde\nu = 1200\ \text{cm}^{-1}$, reduced mass $\mu = 6.855$ u (appropriate to a C=O stretch-like coordinate), and the excited-state equilibrium geometry is displaced by $\Delta R = 0.028$ Å relative to the ground state (same frequency assumed in both states, for this illustrative case). Step 1 — oscillator length. Convert to SI: $\omega = 2\pi c\tilde\nu = 2\pi (3.00\times10^{10}\,\text{cm s}^{-1})(1200\,\text{cm}^{-1}) = 2.262\times10^{14}\,\text{s}^{-1}$. $\mu = 6.855\,\text{u} \times 1.6605\times10^{-27}\,\text{kg/u} = 1.1383\times10^{-26}\,\text{kg}$.

Step 2 — dimensionless displacement and Huang-Rhys factor.

Step 3 — Franck-Condon factors for the first four bands.

Check the sum rule to three terms plus remainder: $0.9087+0.0870+0.00416+0.000133 = 0.99999$, consistent with unity once the negligible $v'\geq4$ tail is included — the sum-rule constraint from the core equations is satisfied to within rounding. Interpretation. With $S\approx 0.096 \ll 1$, this is the small-displacement regime: the 0-0 band carries roughly 91% of the total intensity, the 0-1 band about 9%, and higher bands are essentially invisible. This is the pattern expected for a transition that barely perturbs the bond length — e.g., an $n\to\pi^*$ transition localized on a lone pair not strongly involved in that particular stretching coordinate.

Example 2 — Applied: extracting the excited-state bond length change from a real absorption progression. A gas-phase electronic absorption spectrum shows a vibronic progression in $\tilde\nu = 1580\ \text{cm}^{-1}$ (the totally symmetric mode of the excited state) with relative intensities (v′ = 0,1,2,3,4) of 1.00 : 2.35 : 2.76 : 2.16 : 1.27 (normalized so the 0-0 line is 1.00). The reduced mass along this mode is $\mu = 6.860$ u. Find the Huang-Rhys factor and the equilibrium displacement $\Delta R$. Step 1 — fit $S$ from the intensity ratio. For a Poisson distribution, consecutive ratios obey $q_{v+1,0}/q_{v,0} = S/(v+1)$. Using the 0→1 ratio:

Cross-check with the 1→2 ratio, which should give $S/2$:

And the 2→3 ratio, expected to be $S/3$:

All three independent ratios converge on $S = 2.35$, confirming the data really do follow a single-mode Poisson (harmonic, equal-frequency) Franck-Condon envelope — an internal consistency check that a real spectroscopist performs before trusting the fit. Step 2 — locate the envelope maximum as a sanity check. For $S=2.35$ the Poisson maximum sits near $v'\approx S = 2.35$, i.e., between $v'=2$ and $v'=3$; the tabulated intensities indeed peak at $v'=2$ (2.76) with $v'=3$ close behind (2.16), matching the qualitative shape expected. Step 3 — convert $S$ to a dimensionless displacement, then to $\Delta R$.

Oscillator length, following the same route as Example 1 but with $\tilde\nu=1580\ \text{cm}^{-1}$:

Interpretation. A bond-length change of roughly 0.12 Å between ground and excited state — comparable to what is observed for π→π* transitions in aromatic carbonyls where the C=O bond weakens substantially upon excitation — is extracted purely from relative peak heights in an absorption spectrum, with no need for any excited-state structural probe.

Example 3 — Exam/competition level: two-mode Franck-Condon problem with a frequency change (beyond the simple Poisson model). A polyatomic chromophore has one totally symmetric mode active in its $S_0\to S_1$ transition. The ground-state frequency is $\tilde\nu'' = 1400\ \text{cm}^{-1}$ and the excited-state frequency is $\tilde\nu' = 1000\ \text{cm}^{-1}$ (a substantial frequency change, reflecting bond weakening upon excitation), with a dimensionless displacement (referenced to the ground-state oscillator length) of $d=1.60$. Because the frequencies differ, the simple equal-frequency Poisson formula does not apply exactly; estimate the 0-0 and 0-1 Franck-Condon factors using the generating-function method for displaced oscillators of unequal frequency, and compare with the naive (frequency-unchanged) Poisson estimate to quantify the error introduced by ignoring the frequency change. Step 1 — set up the frequency ratio. Define $\rho = \omega'/\omega'' = \tilde\nu'/\tilde\nu'' = 1000/1400 = 0.7143$. The unequal-frequency overlap integrals require the full generating-function (Manneback-type) treatment; here we use the leading-order results for the 0-0 and 0-1 factors, which are exact closed forms even for $\rho\neq1$:

This reduces correctly to $q_{00}=e^{-S}=e^{-d^2/2}$ when $\rho\to1$ (limiting-case check performed explicitly below). Step 2 — evaluate the prefactor.

Step 3 — evaluate the exponent.

Step 4 — compare with the naive equal-frequency Poisson estimate. If one incorrectly ignored the frequency change and used the ground-state $S$ computed from $d$ alone, $S = d^2/2 = 1.28$, the naive estimate would be

The unequal-frequency result (0.340) is about 22% larger than the naive equal-frequency Poisson estimate (0.278) — the frequency contraction upon excitation ($\omega'<\omega''$) increases the 0-0 overlap relative to what a pure-displacement model would predict, because the excited-state wavefunction spreads over a larger oscillator length and covers more of the ground-state wavefunction's spatial extent. This is the qualitative reason spectroscopists must fit both displacement and frequency change simultaneously (a two-parameter, sometimes Duschinsky-mixed, fit) rather than assuming a pure Poisson progression whenever the vibrational frequency itself is observed to change between absorption and emission spectra (a mirror-symmetry breakdown that is itself diagnostic). Step 5 — limiting case check. Setting $\rho=1$ exactly recovers $2\sqrt{1}/(1+1)=1$ and exponent $d^2(1)/(1+1) = d^2/2 = S$, giving back $q_{00}=e^{-S}$ precisely — confirming the unequal-frequency formula correctly contains the equal-frequency Poisson result as the $\rho\to1$ limit, exactly as required for internal consistency.

The conceptual seed was planted by James Franck in 1925, arguing qualitatively from the relative timescales of electronic and nuclear motion that an electronic transition should be "vertical" on a potential energy diagram. Edward Condon supplied the rigorous quantum-mechanical treatment in 1926-1928, showing that the intensity distribution among vibronic bands is governed by the overlap integral of vibrational wavefunctions — precisely the object this article has derived and computed. Condon's 1926 paper, "A Theory of Intensity Distribution in Band Systems," is the direct ancestor of the Franck-Condon factor as used today; his 1928 paper made the connection to the classical "vertical transition" picture fully explicit, showing how the classical limit emerges from the quantum overlap integral for high vibrational quantum numbers.