Marcus Theory

Two parabolas and a tunneling electron

Electron transfer is the simplest reaction in chemistry: a single electron hops from a donor to an acceptor, no bonds broken, no atoms moved. Yet for thirty years no one could explain why some of these hops are blisteringly fast and others, with even more thermodynamic push behind them, are sluggish. Rudolph Marcus solved it by realizing that the electron is the easy part — the hard part is rearranging everything around the electron.

An electron moves in about 10⁻¹⁵ s. The bonds and solvent molecules around the donor and acceptor move a thousand times slower, on the order of 10⁻¹³–10⁻¹² s. So the electron can only jump at an instant when the nuclei happen to be arranged so that the donor's "before" state and the acceptor's "after" state have exactly the same energy. That is the Franck–Condon principle applied to a reaction: the electron tunnels vertically on a frozen nuclear frame.

Marcus drew the system as two parabolas — one for the reactant (electron on the donor), one for the product (electron on the acceptor) — plotted against a single "reaction coordinate" that bundles together all the nuclear and solvent motion. Electron transfer occurs only where the two parabolas cross. The height of that crossing point above the reactant minimum is the activation free energy, ΔG‡.

Energy
  │   Reactant         Product
  │   parabola         parabola
  │      ╲               ╱
  │       ╲     ✕ ← crossing point = transition state
  │        ╲   ╱  ╲     (ΔG‡ measured up to here)
  │         ╲ ╱    ╲   ╱
  │      ────●──────●────
  │       reactant  product
  │        min       min   (vertical gap = ΔG°)
  └──────────────────────────────→
              nuclear / solvent coordinate

Reorganization energy λ: the cost of getting ready

The width and position of the parabolas are set by a single quantity, the reorganization energy λ. It is defined as the energy it would take to distort the reactant nuclei into the product's equilibrium geometry without letting the electron move. Geometrically, λ is the vertical height of the product parabola measured at the reactant's minimum — how far up the product curve you sit if the nuclei are still frozen in the reactant arrangement.

λ has two physically distinct contributions:

• Inner-sphere, λ_in. The cost of changing bond lengths and angles inside the redox partners. When [Fe(H₂O)₆]³⁺ gains an electron to become [Fe(H₂O)₆]²⁺, the Fe–O bonds lengthen by about 0.13 Å; squeezing or stretching those six bonds to the wrong length costs energy. λ_in is large for metal complexes with big geometry changes and small for rigid aromatic π-systems where the electron lands in a delocalized orbital.
• Outer-sphere, λ_out. The cost of reorienting the solvent dipoles around the new charge distribution. Marcus gave a closed-form dielectric-continuum expression:

λ_out = (Δe)² · [ 1/(2a₁) + 1/(2a₂) − 1/d ] · [ 1/ε_op − 1/ε_s ]

where a₁ and a₂ are the radii of donor and acceptor, d their center-to-center distance, ε_op the optical (high-frequency) dielectric constant and ε_s the static one. The bracket (1/ε_op − 1/ε_s) is the Pekar factor; it is large in polar solvents like water (ε_s ≈ 78) and small in nonpolar ones, which is why outer-sphere λ is much larger in water than in hexane. Total λ values typically run 0.5–2 eV (roughly 50–200 kJ/mol).

The Marcus equation

Because both curves are parabolas of the same curvature, finding where they cross is just algebra. The activation free energy comes out as a clean quadratic in the driving force:

ΔG‡ = (λ + ΔG°)² / (4λ)

and the rate constant is the usual barrier-crossing form, with an electronic-coupling prefactor that accounts for how strongly the donor and acceptor orbitals overlap:

k_ET = (2π/ℏ) · |H_AB|² · (1 / √(4π·λ·k_B·T)) · exp[ −(λ + ΔG°)² / (4λ·k_B·T) ]

Read the barrier formula as a function of the driving force −ΔG°, holding λ fixed, and three regimes fall out automatically:

• Normal region (−ΔG° < λ). Increasing the driving force lowers the barrier and speeds the reaction — the intuitive behavior everyone expects.
• Activationless / optimal (−ΔG° = λ). The numerator (λ + ΔG°)² is zero, so ΔG‡ = 0. The rate is at its maximum; every productive collision succeeds. The parabolas cross right at the bottom of the reactant well.
• Inverted region (−ΔG° > λ). Push the driving force past λ and the barrier reappears. The reaction is now so downhill that the crossing point climbs up the far wall of the reactant parabola. More thermodynamic push, slower kinetics.

The formula is symmetric: a reaction with −ΔG° = λ + x has exactly the same barrier as one with −ΔG° = λ − x. The whole story is a parabola of rate (or its log) plotted against driving force, peaking at −ΔG° = λ.

Worked example: tracking the barrier across the peak

Take a system with reorganization energy λ = 1.0 eV and watch ΔG‡ as the driving force grows. (We'll quote energies in eV since electron-transfer chemists do; 1 eV = 96.5 kJ/mol, and k_BT ≈ 0.0257 eV at 298 K.)

−ΔG° (eV)	Region	ΔG‡ = (λ+ΔG°)²/(4λ)	Relative rate ∝ exp(−ΔG‡/k_BT)
0.0	Self-exchange	0.250 eV	6.4 × 10⁻⁵
0.5	Normal	0.0625 eV	0.086
1.0	Optimal	0.000 eV	1.00 (maximum)
1.5	Inverted	0.0625 eV	0.086
2.0	Inverted	0.250 eV	6.4 × 10⁻⁵

Notice the perfect mirror image. The reaction with −ΔG° = 2.0 eV is enormously more favorable than the one with −ΔG° = 0.5 eV — yet it is over a thousand times slower, because its barrier (0.25 eV) is four times larger. Going from −ΔG° = 1.0 to −ΔG° = 2.0 eV doubles the driving force but drops the rate by a factor of ~16,000. This is the single most surprising prediction in all of chemical kinetics, and it is exactly what the experiments show.

Marcus theory vs Arrhenius and transition-state theory

	Marcus theory	Arrhenius / TST
What it predicts	How the barrier changes with driving force across a reaction family	The rate of one reaction from a measured/computed barrier
Barrier formula	ΔG‡ = (λ + ΔG°)²/(4λ)	Ea (empirical) or ΔG‡ (structural, not predicted from ΔG°)
Inputs needed	λ and ΔG° only	A and Ea, or partition functions for the TS
Can more driving force slow a reaction?	Yes — the inverted region	No — barrier monotonically depends on structure, not ΔG°
Best for	Outer-sphere electron transfer	Bond-making/breaking reactions generally
Quantum corrections	Nuclear tunneling softens the inverted-region falloff	Tunneling treated as an add-on (e.g. H/D KIE)
Distance dependence	|H_AB|² decays exp(−βr), β ≈ 1.0–1.4 Å⁻¹	Not a standard parameter

Marcus theory is not a rival to transition-state theory — it is a special case of it for electron transfer, in which the otherwise-mysterious ΔG‡ is given an explicit, predictive formula. Where TST asks you to find the barrier, Marcus tells you what the barrier has to be once you know λ and ΔG°.

Where Marcus theory shows up

• Photosynthesis. In the bacterial reaction center, an excited chlorophyll special pair hands an electron down a chain of cofactors in <3 ps with near-unit quantum yield. The forward steps are tuned close to the optimal −ΔG° ≈ λ, while the wasteful charge-recombination back-reactions are deliberately pushed deep into the inverted region — so they are slow, and the charge separation survives long enough to be useful. Nature uses the inverted region as a one-way valve.
• The Closs–Miller experiment (1984). Gerhard Closs and John Miller at Argonne attached a biphenyl donor and a series of acceptors (naphthalene, anthracene, benzoquinone…) to a rigid steroid spacer that fixed the distance. By swapping the acceptor they tuned −ΔG° from 0 to ~2.4 eV at constant λ and distance, and watched the rate rise, peak near −ΔG° ≈ 1.2 eV, then fall — the first clean, distance-controlled demonstration of the inverted region.
• Self-exchange reactions. [Fe(H₂O)₆]²⁺/³⁺ swaps an electron with itself at only ~4 M⁻¹s⁻¹ because its λ_in is huge (big Fe–O bond-length change), whereas [Ru(bpy)₃]²⁺/³⁺ self-exchanges near 10⁹ M⁻¹s⁻¹ thanks to a tiny λ_in — the electron lands in a delocalized π* orbital that barely perturbs the geometry. Same ΔG° = 0; rates differ by 10⁸ purely from λ.
• OLEDs and organic solar cells. Charge recombination at donor–acceptor interfaces is engineered to fall in the inverted region so that recombination is slow and current is collected before charges are lost.

Common misconceptions and pitfalls

• Thinking more exergonic always means faster. True for ordinary bond chemistry, false for electron transfer once −ΔG° > λ. The inverted region is the whole point of the theory.
• Confusing λ with ΔG° or with the barrier. λ is a reorganization cost (always positive); ΔG° is the thermodynamic driving force (negative when downhill); ΔG‡ is the kinetic barrier built from both. They are three different numbers on the same diagram.
• Expecting the rate to crash to zero in the deep inverted region. Classical Marcus theory predicts a sharp Gaussian falloff, but real systems fall off much more gently because high-frequency vibrations let the system tunnel into excited vibrational levels of the product (the semiclassical Marcus–Jortner treatment). The experimental inverted region is real but softer than the bare parabola.
• Forgetting electronic coupling. Marcus's barrier sets the nuclear part of the rate; the |H_AB|² prefactor sets the electronic part. For long-range transfer through protein or DNA, |H_AB|² decays roughly as exp(−βr) with β ≈ 1.0–1.4 Å⁻¹, so doubling the donor–acceptor distance can slow the rate by many orders of magnitude even at the optimal driving force.
• Treating λ as a property of one molecule. λ depends on both partners and the solvent. The same donor has a different λ in water versus acetonitrile, because λ_out tracks the solvent's Pekar factor.

Variants and refinements

• Marcus–Jortner (semiclassical) equation. Adds a sum over product vibrational quantum numbers, replacing the single classical parabola with a Franck–Condon-weighted set. This is what correctly reproduces the gentle, asymmetric inverted-region falloff seen in real data.
• Marcus–Hush theory. Extends the picture to mixed-valence compounds and intervalence charge-transfer bands, connecting λ to the energy of an optical absorption: the band maximum of a symmetric mixed-valence dimer sits at exactly λ.
• Proton-coupled electron transfer (PCET). When an electron and a proton move together, the Marcus framework is generalized with a separate reorganization term and proton tunneling — central to water oxidation in Photosystem II and to many redox enzymes.
• Marcus–Gerischer model. The electrode analogue, used in electrochemistry to predict how electron-transfer current depends on overpotential, integrating over the continuum of electronic states in the metal.