Redox Stability in Water

Corrosion engineering leans on the water stability window every day: a metal whose native oxide-free state sits below the H⁺/H₂ line is a "corrosion-prone in acid" metal (like iron or zinc), while one that sits comfortably inside or above the window without being water-reactive (like gold or platinum) is termed "noble." Passivation — the formation of a protective oxide layer, as in aluminum, titanium, or stainless steel — is what allows metals that are thermodynamically well outside the window to survive in practice, because the oxide film imposes an enormous kinetic overpotential barrier on further reaction. Water electrolyzers for green hydrogen production are engineered explicitly around minimizing the overpotential gap above 1.23 V, since every extra volt applied is wasted energy that shows up as heat rather than stored chemical energy in H₂; state-of-the-art proton-exchange-membrane (PEM) electrolyzers using iridium oxide anodes and platinum cathodes now operate around 1.7–1.8 V at industrially relevant current densities, a direct descendant of the Example 3 analysis above. Battery chemistry is dominated by this window from the opposite direction: lithium-ion cells operate at cell voltages of 3.0–4.2 V, far outside water's 1.23 V window, which is exactly why lithium-ion cells cannot use aqueous electrolytes at all — water would be electrolyzed continuously at such potentials. This has driven the entire field toward non-aqueous, aprotic solvents (carbonates, ethers) whose own decomposition windows are much wider, and more recently toward engineered "water-in-salt" superconcentrated aqueous electrolytes that kinetically suppress water electrolysis even at cell voltages nominally outside the thermodynamic window, achieving safer aqueous batteries with respectable voltage. Biological electron transport chains operate almost entirely within or just at the edges of the water window: the terminal step of cellular respiration, O₂ + 4H⁺ + 4e⁻ → 2H₂O at cytochrome c oxidase, is literally the upper boundary reaction of the window itself, run in reverse-favorable direction by biology to harvest the 1.23 V worth of driving force (partitioned across the electron transport chain into a proton gradient) that would otherwise be released as bulk chemical energy. Photosystem II, meanwhile, does the reverse — it is one of the very few biological catalysts on Earth capable of oxidizing water to O₂, requiring the ~1.23 V-plus-overpotential energy of absorbed photons to do a job that is chemically identical to the O₂-evolving half of a water electrolyzer. Instrumentally, the stability window sets the practical limits of aqueous voltammetry: cyclic voltammograms run in water always show a "solvent window" beyond which the current rises sharply due to H₂ or O₂ evolution swamping the signal of interest, which is why electrochemists routinely switch to non-aqueous solvents (acetonitrile, DMF) or use mercury electrodes (which suppress HER kinetically) to study redox couples that fall outside water's accessible range.

Ask the tutor: How do I add a third redox couple (say, Cl⁻/Cl₂) onto this same potential–pH picture to see whether chlorine evolution competes with oxygen evolution during brine electrolysis?

Water is thermodynamically metastable with respect to its own elements combined with oxygen — pure H₂O sitting in a beaker is not sitting at the bottom of every possible energy well available to the atoms H and O. There are two ways water can decompose electrochemically, and both involve electron transfer:

Reduction of water (or H⁺) to hydrogen gas:

Oxidation of water to oxygen gas:

Both are written, by convention, as reductions. The first says: if some species is a strong enough reducing agent (i.e., if the couple you compare it to has a low enough reduction potential), it will happily reduce H⁺ to H₂ rather than do anything useful, releasing hydrogen gas and getting oxidized itself. The second says: if some species is a strong enough oxidizing agent (high reduction potential), it will rip electrons out of water itself, releasing O₂ and getting reduced. Between these two potentials — 0.00 V at the bottom and 1.23 V at the top, both referenced to standard hydrogen conditions at pH 0 — lies the stability window of water: the range of electrode potentials in which water is inert to any redox couple with a potential inside that range, thermodynamically speaking. A couple whose standard potential lies inside the window can, in principle, exist in water without spontaneously oxidizing or reducing the solvent. A couple whose potential lies above the O₂/H₂O line will spontaneously oxidize water (evolve O₂) as it is itself reduced. A couple whose potential lies below the H⁺/H₂ line will spontaneously reduce water (evolve H₂) as it is itself oxidized. Crucially, both boundary lines are pH-dependent, and they shift together, so the vertical extent of the window (1.23 V) never changes — only its absolute position on the potential axis does, sliding down as pH rises. This gives the window the shape of a diagonal band on a potential-vs-pH plot, the immediate precursor to a full Pourbaix Diagram where actual species stability regions are drawn inside (or straddling) this band.

Notice that both lines have identical slope, −0.0592 V per pH unit, because each involves the transfer of one H⁺ per electron (4 H⁺/4 e⁻ for oxygen, 2 H⁺/2 e⁻ for hydrogen — both ratios reduce to 1:1). Any redox couple whose own H⁺-dependence differs from 1:1 per electron will cut across the window at an angle rather than staying parallel to it — this is exactly the geometrical richness that a full Pourbaix diagram exploits.

The window's boundaries come directly from applying the Nernst Equation to the two water half-reactions, exactly as done generally when discussing the pH Dependence of Redox Potentials. For the hydrogen couple,

where: $E^\circ_{\mathrm{H^+/H_2}}=0$ V by definition (SHE reference); the factor 0.0592 V is $\dfrac{RT\ln 10}{F}$ at 298 K; the $[\mathrm{H^+}]^2$ in the log accounts for two protons per two-electron step, giving a clean 1:1 electron:proton ratio and hence a 59.2 mV/pH slope. For the oxygen couple,

where: $E^\circ_{\mathrm{O_2/H_2O}}=1.23$ V is the standard reduction potential at pH 0; again the four-electron, four-proton stoichiometry collapses to a 1:1 ratio, giving the same 59.2 mV/pH slope as the hydrogen line. The width of the window at any given pH is simply the difference:

where: the pH terms cancel exactly because both slopes are identical, leaving a pH-independent window width of 1.23 V — this is nothing but $\Delta G^\circ = -nFE^\circ$ for the overall decomposition reaction $2\mathrm{H_2O} \rightarrow 2\mathrm{H_2} + \mathrm{O_2}$, restated as a potential difference. Reading it: no matter what pH you run your chemistry at, water always has exactly a 1.23 V "safe zone" for redox couples — the zone just slides up or down together as pH changes, like a ladder sliding down a wall at constant length. A convenient rule-of-thumb version, useful for quick screening of a species against water, is to compare $E^\circ$ of the couple in question (at the relevant pH, computed via its own Nernst equation) against these two moving lines:

where: if the couple's own potential falls outside either bound, spontaneous water decomposition (coupled to reduction or oxidation of the species) is thermodynamically favorable; whether it is kinetically fast is a separate question addressed via overpotential.

Example 1 (foundational): Confirm the window width is unchanged at pH = 9.30, an odd non-round value. Using the two boxed Nernst expressions:

The width reproduces 1.23 V exactly (to rounding), confirming the cancellation derived symbolically above — even at this deliberately non-round, non-"nice" pH value the arithmetic still gives exactly 1.23 V, because the cancellation is structural, not numerical coincidence. Limiting case check: as pH → 14 (strongly basic), both lines continue sliding down together — $E_{\mathrm{O_2/H_2O}} \to 1.23 - 0.829 = 0.401$ V and $E_{\mathrm{H^+/H_2}} \to 0.00 - 0.829 = -0.829$ V — and the difference remains $0.401-(-0.829)=1.230$ V, confirming the width is truly pH-independent even at the extreme edge of the aqueous pH scale.

Example 2 (applied): Is aqueous Fe²⁺ stable in air-saturated water at pH 5.0, or does it oxidize? The relevant couple is

This half-reaction involves no H⁺, so its potential is pH-independent at the level of the simple Fe³⁺/Fe²⁺ couple (ignoring hydrolysis effects, which would matter at higher pH but are neglected here for clarity): $E = 0.771$ V regardless of pH. We are not asking whether Fe²⁺ reduces water (it doesn't — 0.771 V is far above the H⁺/H₂ line) but whether dissolved O₂ can oxidize Fe²⁺ to Fe³⁺ — i.e., whether Fe²⁺ is stable with respect to air, using the O₂/H₂O line as the relevant oxidant boundary at pH 5:

Since $E_{\mathrm{O_2/H_2O}}(5.0) = 0.934\ \mathrm{V} > E_{\mathrm{Fe^{3+}/Fe^{2+}}} = 0.771\ \mathrm{V}$, the oxygen couple's potential is higher: dissolved O₂ is thermodynamically capable of oxidizing Fe²⁺ to Fe³⁺, releasing water as O₂ is reduced. This matches everyday observation — ferrous solutions left open to air slowly turn rust-colored as Fe²⁺ is oxidized to Fe³⁺, which then hydrolyzes and precipitates as Fe(OH)₃. The window/Nernst framework doesn't merely say "iron rusts" qualitatively; it locates exactly how much thermodynamic driving force exists (0.934 − 0.771 = 0.163 V, i.e., about 15.7 kJ/mol per electron) at this specific pH.

Limiting case check: at pH 0 (strongly acidic), $E_{\mathrm{O_2/H_2O}}(0)=1.23$ V, an even larger gap above 0.771 V — Fe²⁺ is even more thermodynamically prone to air-oxidation in strong acid than at pH 5 (though in practice strongly acidic ferrous solutions are often kinetically more air-stable, illustrating again the split between thermodynamic prediction and kinetic reality).

Example 3 (exam/competition level): Why does practical water electrolysis require ≈1.8–2.0 V rather than the thermodynamic 1.23 V, and how much of that extra voltage is assignable to which electrode? The thermodynamic minimum voltage to split water, derived above, is 1.23 V at pH 0 (and the same 1.23 V at any pH, since both electrode reactions shift identically). In a real electrolyzer, however, the applied cell voltage $V_{\text{applied}}$ must overcome three things:

where: $E^\circ_{\text{cell}} = 1.23$ V is the thermodynamic requirement; $\eta_{\text{anode}}$ and $\eta_{\text{cathode}}$ are kinetic overpotentials — the extra potential needed to drive the O₂ and H₂ evolution reactions at a measurable rate — and $iR$ is the ohmic drop through electrolyte and contacts. For O₂ evolution (oxygen evolution reaction, OER) on typical electrodes (e.g., Pt, or even better catalysts), the overpotential is large because the mechanism requires four sequential proton-coupled electron transfers and O–O bond formation, a kinetically demanding multi-step surface process. Typical OER overpotentials at practical current densities are 0.4–0.6 V even on good catalysts, and can be 0.6–0.9 V on poor ones. For H₂ evolution (hydrogen evolution reaction, HER), the overpotential is much smaller on Pt (as low as 0.01–0.05 V) because the mechanism (adsorption of H, then either electrochemical or chemical desorption as H₂) is fast on Pt's surface, but can be substantial (0.3–0.5 V) on poor HER catalysts like mercury or lead — this is precisely why mercury cathodes historically allowed other reductions (like Na⁺ → Na amalgam) to proceed preferentially over H₂ evolution, despite Na⁺/Na sitting far below the H⁺/H₂ line thermodynamically. Taking representative practical values — $\eta_{\text{anode}} \approx 0.45$ V (OER on a modest catalyst), $\eta_{\text{cathode}} \approx 0.10$ V (HER on a reasonable catalyst), and a modest $iR \approx 0.10$ V for a well-conducting cell:

This lands squarely in the commonly quoted practical range of 1.8–2.0 V for water electrolysis, and it demonstrates numerically that most of the "extra" voltage beyond 1.23 V is attributable to the oxygen side of the reaction, not the hydrogen side — a direct consequence of the greater kinetic complexity of forming an O=O bond compared to an H–H bond. This is exactly why virtually all electrocatalysis research aimed at "cheaper hydrogen" focuses disproportionately on OER catalyst design (e.g., IrO₂, RuO₂, Ni-Fe layered double hydroxides) rather than HER catalyst design.

Limiting case check: if a hypothetical perfect catalyst existed with $\eta_{\text{anode}}=\eta_{\text{cathode}}=0$ and superconducting leads ($iR=0$), the equation correctly collapses to $V_{\text{applied}} = 1.23$ V — the pure thermodynamic limit — confirming the overpotential terms are additive corrections layered on, not replacements for, the underlying Nernstian window.

The modern language of standard reduction potentials, the SHE reference, and tabulated E° values traces most directly to the electrochemical convention-building of Wilhelm Ostwald and, later, the systematic tabulations organized under the IUPAC-endorsed "reduction potential" sign convention formalized in the mid-20th century. But the person whose name is most tightly bound to turning these numbers into a diagram — the geometric object that visualizes exactly the window discussed here, extended into full species-stability maps — is the Belgian electrochemist Marcel Pourbaix. Pourbaix, working at the Free University of Brussels in the 1930s–1960s, systematized the practice of plotting electrode potential against pH for all the relevant redox and acid–base equilibria of an element simultaneously, producing what are now universally called Pourbaix diagrams. His 1963 Atlas of Electrochemical Equilibria in Aqueous Solutions remains the standard reference compilation. The water stability window described in this article is, in Pourbaix's own diagrams, always drawn as two dashed lines (traditionally labeled "a" for the H⁺/H₂ line and "b" for the O₂/H₂O line) superimposed on every single-element diagram, precisely so a reader can see at a glance which of that element's stable species, if any, survive within the aqueous window and which are only stable outside it (and hence only observable under kinetic protection or in non-aqueous media).