The Sabatier Principle and the Catalyst Volcano Plot

What the Sabatier Principle Is and Where It Applies

The Sabatier principle states that a catalyst works best when its interaction with reaction intermediates is of intermediate strength. If binding is too weak, the reactants never adsorb or activate — the surface stays bare and the rate is throttled by adsorption. If binding is too strong, intermediates or products cling to the surface, poisoning active sites and throttling the rate by desorption. The optimum sits between these failures.

• Heterogeneous catalysis: ammonia synthesis (Fe, Ru), CO oxidation, hydrogenation on Pt/Pd/Ni.
• Electrocatalysis: the hydrogen evolution reaction (HER), oxygen reduction (ORR), CO₂ reduction — the modern home of volcano analysis.
• Even biology: enzymes bind substrate tightly but transition states more tightly still, a Sabatier-like compromise.

The principle is powerful because it reduces a many-step surface reaction to a single descriptor — typically the adsorption energy of one rate-controlling intermediate — turning catalyst screening into the search for a metal whose descriptor sits at the peak.

Deriving the Volcano: Two Competing Rate Legs

Consider a two-step surface reaction: (1) A adsorbs, A + * → A*, and (2) A* reacts and the product desorbs, A* → products + *. The overall rate depends on surface coverage θ, which is set by how strongly A binds.

Weak-binding leg: when adsorption is weak, coverage θ is small and the rate rises as binding gets stronger (more sites occupied). Rate increases with |ΔG_ads|. Strong-binding leg: once θ saturates, the bottleneck becomes removing A* — desorption/product release slows exponentially as binding tightens, so the rate falls. The two opposing trends cross at an optimum.

The descending legs follow the Brønsted–Evans–Polanyi (BEP) relation, which linearly links an elementary step's activation energy to its reaction energy: Ea = α·ΔE + β, with a transfer coefficient α typically 0.3–1. Because activation barriers scale linearly with ΔE, and rate depends exponentially on Ea (Arrhenius, k = A·exp(−Ea/RT)), plotting log(rate) versus the binding descriptor yields two straight lines of opposite slope — the two sides of the volcano.

The Governing Equations and a Worked HER Example

For the hydrogen evolution reaction (2H⁺ + 2e⁻ → H₂), the descriptor is the hydrogen adsorption free energy ΔG_H*, and the Sabatier optimum is simply ΔG_H* = 0 eV — binding neither uphill nor downhill. A common activity model is:

i₀ = −e·k₀ / (1 + exp(|ΔG_H*| / kT))

where i₀ is the exchange current density, e the elementary charge, k₀ a rate prefactor, k Boltzmann's constant (8.617×10⁻⁵ eV/K), and T ≈ 298 K so kT ≈ 0.0257 eV. The activity peaks when ΔG_H* = 0 and falls symmetrically as |ΔG_H*| grows.

• Platinum: ΔG_H* ≈ −0.09 eV → essentially at the peak, exchange current density ~10⁻³ A/cm², the benchmark HER catalyst.
• Gold: ΔG_H* ≈ +0.30 eV → weak leg, orders of magnitude slower.
• Tungsten: ΔG_H* ≈ −0.55 eV → strong leg, poisoned by adsorbed H.

Plugging ΔG_H* = ±0.30 eV into the model gives exp(0.30/0.0257) ≈ e^11.7 ≈ 1.2×10⁵ in the denominator — roughly a 10⁵ activity penalty versus the optimum, quantitatively reproducing the volcano's steep flanks.

How It's Measured and Used in Catalyst Discovery

Building a volcano requires two axes: an activity metric and a binding descriptor.

• Activity (y-axis): turnover frequency (TOF, s⁻¹), exchange current density i₀, or overpotential at a fixed current (e.g. η at 10 mA/cm² for electrocatalysts).
• Descriptor (x-axis): adsorption energy of the key intermediate, obtained from DFT calculations (typically ~0.1–0.2 eV accuracy) or estimated from temperature-programmed desorption (TPD) and microcalorimetry.

The workhorse of the modern approach is Nørskov's computational descriptor screening (2000s): compute one adsorption energy per candidate, place it on the known volcano, and predict activity without synthesizing anything. This drove discovery of MoS₂ edges (ΔG_H* ≈ +0.08 eV) as a cheap Pt substitute for HER, and Pt-skin alloys with tuned d-band centers for ORR. Scaling relations — near-linear correlations between adsorption energies of related intermediates (e.g. E_OH vs E_OOH differ by a nearly constant ~3.2 eV) — let a single descriptor stand in for several, which is what makes one-dimensional volcanoes valid in the first place.

Sabatier vs. Its Close Cousins

The Sabatier principle is often confused with related ideas that describe different things:

• Sabatier principle vs. the Sabatier reaction: the reaction (CO₂ + 4H₂ → CH₄ + 2H₂O over Ni, also from Sabatier) is a specific methanation process — not the general binding rule. Same chemist, different concept.
• Sabatier vs. Brønsted–Evans–Polanyi: BEP is the linear Ea–ΔE relation that gives the volcano its slopes; Sabatier is the higher-level statement that an optimum exists. BEP is the mechanism behind the metaphor.
• Sabatier vs. the d-band model: Hammer–Nørskov's d-band center theory explains why a metal binds as it does (a higher d-band center → stronger adsorption); Sabatier tells you what binding strength you should aim for.
• Sabatier vs. Curtin–Hammett / Bell–Evans–Polanyi in homogeneous catalysis: volcano relationships also appear in molecular catalysis and enzymology, but there the descriptor is a binding constant or ΔG of a resting state, not a surface adsorption energy.

Exceptions, Limits, and Famous Cases

The volcano is an idealization, and its cracks are scientifically important:

• Scaling-relation limits: because intermediates' binding energies are correlated, you often cannot independently optimize them. For ORR/OER, the fixed ~3.2 eV gap between E_OOH and E_OH imposes a minimum theoretical overpotential of ~0.3–0.4 V — a fundamental ceiling no single-site catalyst on the volcano can beat. Breaking scaling relations (bifunctional sites, confinement) is a major research frontier.
• Multiple descriptors: when two intermediates matter independently, the 1-D volcano becomes a 2-D activity map with a peak plateau.
• Famous case — ammonia synthesis: the N₂ dissociation volcano peaks near Fe and Ru; industrially, Haber–Bosch uses promoted Fe precisely because it sits near the summit at reasonable cost, while Ru is more active but pricier.
• Famous case — HER: Pt's near-zero ΔG_H* is why it defines the benchmark, and why the entire non-precious-metal catalyst field is framed as "climbing to Pt's spot" on the volcano.

The principle earned Sabatier the 1912 Nobel Prize in Chemistry and remains, a century on, the compass of rational catalyst design.

Hydrogen adsorption free energy (ΔG_H*) and relative HER activity across metals — a classic volcano series. Values are approximate DFT/experiment figures near 0 V vs RHE.

Metal	ΔG_H* (eV)	Binding regime	HER activity
Pt	≈ −0.09	Near-optimal	Highest (volcano peak)
Ni	≈ −0.28	Slightly too strong	Good
Au	≈ +0.30	Too weak	Low (weak-binding leg)
W	≈ −0.55	Too strong	Low (strong-binding leg)
Mo (as MoS₂ edge)	≈ +0.08	Near-optimal	High (non-precious)
Hg	≈ +1.0	Far too weak	Negligible