Introduction to Molecular Fluorescence

Fluorescence is the working currency of modern molecular biology and clinical diagnostics: fluorescently labeled antibodies and nucleic acid probes underlie ELISA immunoassays, fluorescence in situ hybridization (FISH), and quantitative PCR, where a reporter dye's fluorescence increases in real time as amplicon accumulates. DNA sequencing by Sanger and next-generation methods reads bases by their distinct fluorescence emission colors. Flow cytometry sorts and counts individual cells at thousands per second by measuring the fluorescence of surface-bound fluorophore-tagged antibodies as each cell streams past a laser. Green fluorescent protein (GFP) and its engineered relatives, discovered in jellyfish and refined through decades of protein engineering, let biologists watch gene expression and protein trafficking live inside cells — work recognized by the 2008 Nobel Prize in Chemistry. Environmental and industrial laboratories use fluorescence to detect trace polycyclic aromatic hydrocarbons in water at parts-per-trillion levels, well below what direct absorption spectrophotometry could ever resolve, precisely because of the dark-background sensitivity advantage described here. Pharmaceutical quality control exploits fluorescence quenching and lifetime measurements to probe drug-protein binding, and clinical pulse oximeters and blood-gas analyzers use fluorescence quenching by oxygen (the same Stern–Volmer chemistry worked through above) as an optical sensing mechanism. The instrumentation that makes all of this practical — light sources, monochromators arranged at right angles, photomultiplier detectors, and the calibration strategies that handle the inner-filter nonlinearity seen above — is the natural next topic, developed in Fluorescence Instrumentation and Quantitation, while the slower, spin-forbidden sibling process sketched at the bottom of the Jablonski diagram is the subject of Phosphorescence and Delayed Emission.

:::quiz[In a Stern–Volmer plot of $F^0/F$ versus quencher concentration $[Q]$, the slope is equal to]

• $k_f$
• $\Phi_f$
• $K_{SV} = k_q \tau_f^0$ ✓
• $1/\varepsilon b$

Ask the tutor: How does the choice of excitation and emission monochromator bandwidths affect the trade-off between fluorescence sensitivity and spectral resolution in a real instrument?

Absorption promotes an electron from the ground electronic state to an excited electronic state, and each electronic state carries a manifold of vibrational sublevels riding on top of it, exactly as introduced in Absorption and Emission Processes. At room temperature, thermal population overwhelmingly favors the lowest vibrational level of the ground state ($S_0$, $v=0$), so essentially every photon absorbed starts from there. The Franck–Condon principle says that because nuclei are so much heavier than electrons, an electronic transition occurs faster ($\sim 10^{-15}$ s) than the nuclei can move, so absorption is a vertical line on an energy diagram: the molecule lands in whatever vibrational level of the excited state has the best overlap with the ground-state nuclear geometry, not necessarily $v=0$ of $S_1$. Once in an excited vibrational level of $S_1$ (or a higher singlet $S_2$, $S_3,\dots$), the molecule does not sit still. Collisions with solvent molecules drain away the excess vibrational energy as heat in roughly $10^{-12}$ s, a process called vibrational relaxation. If the molecule was excited into $S_2$ or higher, an additional radiationless process called internal conversion (IC) carries it down to $S_1$ — this happens because high vibrational levels of $S_1$ often overlap in energy with low vibrational levels of $S_2$, allowing an isoenergetic radiationless crossover, followed by more vibrational relaxation. Both processes are so much faster than fluorescence ($10^{-12}$–$10^{-11}$ s versus $10^{-9}$–$10^{-7}$ s) that by the time any photon is emitted, the molecule has already relaxed to the zero-vibrational level of $S_1$. This is Kasha's rule: fluorescence emission spectra are essentially independent of the exact excitation wavelength, because emission always starts from $S_1(v=0)$. From $S_1(v=0)$ several things can happen. The molecule can emit a photon and drop to some vibrational level of $S_0$ — this radiative, spin-allowed ($S_1 \to S_0$, singlet-to-singlet) transition is fluorescence. Alternatively it can undergo intersystem crossing (ISC), a spin-forbidden but not-impossible radiationless hop to the excited triplet state $T_1$, made possible by spin–orbit coupling (strongly enhanced by heavy atoms like Br, I, or transition metals). From $T_1$, emission back to $S_0$ is phosphorescence — slow ($10^{-6}$–$10^1$ s or longer) precisely because it is spin-forbidden. Finally, the excited molecule can lose energy entirely non-radiatively via collisional deactivation or quenching by another species (dissolved $\mathrm{O_2}$ is a notoriously efficient quencher because it is a ground-state triplet itself).

Two spectral consequences follow directly from this picture. First, the Stokes shift: because vibrational relaxation happens in both the excited state (before emission) and the ground state (after emission, as the molecule lands in a vibrationally excited level of $S_0$ and then relaxes to $v=0$), the emitted photon always carries less energy than the absorbed one. Fluorescence spectra are therefore red-shifted relative to absorption spectra, and the shift is a signature of how much geometric reorganization the excited state undergoes. Second, mirror-image symmetry: if the vibrational spacing of $S_0$ resembles that of $S_1$ (common for rigid, planar aromatics), the fluorescence emission spectrum is approximately a mirror image of the lowest-energy absorption band, reflected about the wavelength where they overlap (the $0$–$0$ transition).

The absorption event is unimolecular and essentially instantaneous, so the physical chemistry of fluorescence lives entirely in the rate competition among the pathways out of $S_1$. Every excited-state process is treated as a first-order (or pseudo-first-order) rate:

where:

• $[S_1]$ — concentration (or population) of molecules in the excited singlet state at time $t$
• $k_f$ — first-order rate constant for radiative decay (fluorescence), units s⁻¹
• $k_{ic}$ — rate constant for internal conversion, $S_1 \to S_0$ (radiationless), s⁻¹
• $k_{isc}$ — rate constant for intersystem crossing, $S_1 \to T_1$ (radiationless), s⁻¹
• $k_q$ — bimolecular quenching rate constant, M⁻¹ s⁻¹; $[Q]$ is quencher concentration, M

Reading: all four channels drain the same reservoir simultaneously and independently; the sum of their rates sets how fast $S_1$ empties, while the fraction going through $k_f$ sets how much light comes out. The two headline quantities of fluorescence spectroscopy fall directly out of this rate law: the fluorescence quantum yield and the fluorescence lifetime.

where:

• $\Phi_f$ — fluorescence quantum yield, the fraction of absorbed photons re-emitted as fluorescence (dimensionless, $0 \le \Phi_f \le 1$)

Reading: $\Phi_f$ is a branching ratio — the radiative "escape rate" divided by the total rate at which the excited population is destroyed by every channel combined.

where:

• $\tau_f$ — observed (or "measured") fluorescence lifetime, s; the time for $[S_1]$ to fall to $1/e$ of its initial value

Reading: $\tau_f$ is set by the same total decay rate that determines $\Phi_f$; the two are linked through $k_f$ alone, $\Phi_f = k_f \tau_f$.

Example 1 (foundational): computing $\Phi_f$ and $\tau_f$ from rate constants. Anthracene in cyclohexane at room temperature has approximate rate constants $k_f = 4.3\times10^7\ \text{s}^{-1}$, $k_{ic} = 1.6\times10^6\ \text{s}^{-1}$, and $k_{isc} = 8.0\times10^6\ \text{s}^{-1}$ (no added quencher, $[Q]=0$). Find $\Phi_f$ and $\tau_f$. Total decay rate:

Quantum yield:

Lifetime:

Both values agree with anthracene's well-known photophysics: high quantum yield (~0.8, it's a classic fluorescence standard) and a lifetime in the tens-of-nanoseconds range typical of allowed $\pi\to\pi^*$ singlet fluorescence. Checking self-consistency: $\Phi_f = k_f \tau_f = (4.3\times10^7\,\text{s}^{-1})(1.90\times10^{-8}\,\text{s}) = 0.817$ — matches exactly, as the algebra in Step 2 of the first derivation guarantees.

Example 2 (applied): quantifying riboflavin by fluorescence, and checking linearity. Riboflavin (vitamin B₂) fluoresces near 525 nm after excitation at 445 nm, with $\varepsilon_{445} = 1.06\times10^4\ \text{M}^{-1}\text{cm}^{-1}$ and $\Phi_f = 0.26$ in aqueous buffer. A calibration standard at $c = 4.00\times10^{-7}$ M is run in a $b = 1.00$ cm cuvette with incident power $P_0$ corresponding to an instrument reading of 500.0 arbitrary units when $K$ and $P_0$ are folded together (i.e., using $F = 2.303\,K\Phi_f P_0\,\varepsilon b c$, with $K P_0 = 500.0/(\,2.303\times0.26\times1.06\times10^4\times1.00\times4.00\times10^{-7}\,)$ back-solved from a known point). We want (a) the absorbance of the standard, confirming the linear regime applies, and (b) the predicted signal for an unknown at $c_x = 1.50\times10^{-6}$ M. First, absorbance of the standard:

This is far below the $A \approx 0.05$ threshold where the linear approximation of Step 5 breaks down, so treating $F \propto c$ is fully justified for this standard. From $F_{std} = 500.0$ at $c_{std} = 4.00\times10^{-7}$ M, the proportionality constant is

Now check the unknown's absorbance before trusting linearity:

still comfortably linear ($A_x < 0.05$). Predicted signal:

If a laboratory measurement instead gave, say, $1.70\times10^3$ units, that ~9% low bias relative to the linear prediction would be a warning sign of inner-filter absorption or a slight excitation/emission overlap self-absorption effect — worth investigating with a dilution check (halving $c$ should double $F$ exactly in the linear regime).

Example 3 (exam/competition level): Stern–Volmer quenching, extracting $k_q$, and a mixed static/dynamic sanity check. A fluorophore has an unquenched lifetime $\tau_f^0 = 5.80$ ns and unquenched intensity $F^0 = 862$ (arbitrary units). Dissolved oxygen is added as a quencher, and the following intensities are measured at 25 °C:

(a) Determine $K_{SV}$, (b) determine the bimolecular quenching rate constant $k_q$, and (c) compare $k_q$ to the diffusion-controlled limit in water ($k_{diff} \approx 7.0\times10^9\ \text{M}^{-1}\text{s}^{-1}$) to judge whether quenching here is dynamic (collisional), static, or reaction-limited.

Step 1 — Build the Stern–Volmer data table. Compute $F^0/F$ at each point:

Step 2 — Linear regression of $F^0/F$ vs. $[Q]$. Using the three nonzero points plus the trivial point $(0, 1)$, a least-squares fit of $y = 1 + K_{SV}x$ (forcing the physically required intercept of 1.00) gives slope

Step 3 — Extract $k_q$ from $K_{SV} = k_q \tau_f^0$.

Step 4 — Interpret against the diffusion limit. The extracted $k_q = 6.6\times10^{10}\ \text{M}^{-1}\text{s}^{-1}$ exceeds the quoted diffusion-controlled limit of $7.0\times10^9\ \text{M}^{-1}\text{s}^{-1}$ by nearly an order of magnitude. Since a purely collisional (dynamic) mechanism cannot have $k_q$ larger than the diffusion limit, this signals that the quenching is not purely dynamic — a static component (ground-state complex formation between fluorophore and $\mathrm{O_2}$ that is fluorescence-dark from the start) is contributing, inflating the apparent $K_{SV}$ beyond what collision alone would produce. A full treatment would need time-resolved lifetime measurements at each $[Q]$ (checking whether $\tau_f^0/\tau_f$ tracks $F^0/F$, the hallmark test distinguishing dynamic from static quenching) rather than steady-state intensities alone.

Limiting-case check: setting $[Q]\to 0$ in the fitted line recovers $F^0/F \to 1.00$ exactly, matching the forced intercept and confirming the regression is self-consistent with the unquenched reference point.

The Jablonski diagram is named for Aleksander Jabłoński, the Polish physicist who systematized the kinetic scheme of excited-state relaxation pathways in the 1930s while studying the polarization and quenching of fluorescence in solutions and vapors. Working largely in Poland and later at the Nicolaus Copernicus University in Toruń, Jabłoński was concerned with a very concrete puzzle: why does the degree of polarization of fluorescent light depend on solvent viscosity, concentration, and temperature in ways that simple absorption theory could not explain? His answer required carefully separating the fast photophysical processes (excitation, vibrational relaxation) from the comparatively slow processes competing for the excited population (fluorescence itself, intersystem crossing to a metastable state, and quenching by molecular collisions) — precisely the rate-competition framework this article derives. The diagrammatic scheme that now bears his name became the standard visual language of photophysics because it cleanly separates what energy level a molecule occupies from what is competing to empty that level, which is exactly the distinction needed to reason quantitatively about quantum yields and lifetimes.