Metal Adsorption on Iron Oxide: Surface Complexation and Surface Precipitation

Why is hydroxide precipitation not the whole story?

The classical account of metal removal is bulk hydroxide precipitation: raise pH until the metal hydroxide solubility product is exceeded, M²⁺ + 2 OH⁻ → M(OH)₂(s), and the solid settles or is floated out. That model predicts a residual dissolved concentration governed only by K_sp and pH. In practice, plants dosing ferric coagulant routinely reach residual metal concentrations one to three orders of magnitude below the hydroxide solubility limit, and they do so at pH values where the free metal is thermodynamically undersaturated with respect to M(OH)₂(s).

The missing mechanism is adsorption onto the hydrous ferric oxide (HFO) generated when ferric salts hydrolyse. Freshly precipitated HFO — two-line ferrihydrite — has an enormous specific surface area (typically 600–800 m²/g) and a high density of reactive surface hydroxyl groups. Dissolved metals coordinate directly to these groups, a process described quantitatively by surface complexation theory rather than by a solubility product. This is why coagulant dosing and clarification equipment removes trace metals that simple pH adjustment leaves behind, and it underpins the practice covered in our heavy metals removal and compliance guide.

What is a surface complex on iron oxide?

The HFO surface is populated by amphoteric hydroxyl groups written generically as ≡FeOH, where ≡ denotes the three lattice bonds fixing the iron to the solid. Each site can protonate or deprotonate depending on solution pH, so the surface is amphoteric:

A dissolved metal cation forms an inner-sphere complex by displacing the proton(s) and coordinating directly to the oxygen — there is no water molecule interposed between metal and surface oxygen, distinguishing it from weak electrostatic (outer-sphere) attraction. Spectroscopic and modelling work by Benjamin and others confirms this covalent, specific binding. Because a proton is released, adsorption of a cation lowers the pH; conversely raising the pH drives more adsorption — the origin of the adsorption edge described below.

How does the mass-law describe adsorption?

Surface complexation is treated as a chemical equilibrium with a mass-action expression, exactly like an aqueous reaction, but with an electrostatic correction because charged ions must cross the charged surface–solution interface. For a divalent cation binding a single site:

The generalised two-layer (diffuse-layer) model of Dzombak & Morel is the reference parameterisation for HFO: it uses two site types — a small set of high-affinity ("strong") sites and a large set of low-affinity ("weak") sites — with a single self-consistent database of intrinsic constants fitted to a vast body of experimental data. The constant-capacitance model is the high-ionic-strength simplification in which ψ₀ is linear in surface charge (σ = C·ψ₀). Both are solved simultaneously with the aqueous speciation and the mass and charge balances.

What sets the total number of binding sites?

Adsorption capacity is finite: it is capped by the total number of surface sites, which scales with the mass of HFO precipitated and therefore with the ferric dose. The site balance is the conservation equation that closes the model:

The practical consequence is direct: the coagulant dose is not just a means of forming a settleable floc, it is what provisions the adsorption sites. Under-dose ferric and you run out of high-affinity strong sites, the isotherm saturates, and residual metal rises sharply. This is the quantitative link between coagulant chemistry and metal compliance — see our coagulation and flocculation guide for the dosing and rapid-mix context.

What is the pH adsorption edge?

The single most important behaviour to internalise is the adsorption edge: the S-shaped rise of fractional adsorption over a narrow (1–2 unit) pH window. It follows directly from the mass law, because both surface deprotonation and cation binding release protons.

Cations (Cu, Zn, Pb, Cd, Ni, Co) adsorb increasingly with rising pH: as pH climbs, ≡FeOH sites deprotonate to ≡FeO⁻, the surface becomes more negative, and cation uptake sweeps from near-zero to near-complete. The stronger the metal–oxygen bond, the lower the pH₅₀ (the pH of half-adsorption): Pb and Cu adsorb at lower pH than Cd or Ni.

Oxyanions (arsenate AsO₄³⁻, chromate CrO₄²⁻, selenite, phosphate, molybdate) show the opposite edge: they adsorb best under acidic conditions where the surface is positively charged (≡FeOH₂⁺) and desorb as pH rises. This is why arsenic removal on iron media is favoured at pH 6–7 and falls off above pH 8. The contrast is summarised below.

The engineering implication is that there is no single "correct" pH for a mixed metal stream: a pH optimal for cationic Zn and Cu may be poor for anionic As(V). Multi-metal effluents often need staged pH or a compromise setpoint identified from the composite edge.

When does adsorption become surface precipitation?

Adsorption isotherms on HFO do not saturate cleanly at a fixed monolayer. At low sorbate-to-sorbent ratios the data fit a Langmuir-type surface complexation isotherm; at high metal loading the uptake continues to climb far beyond monolayer coverage and the apparent partitioning approaches that of a pure metal hydroxide solid. Benjamin (1983) documented exactly this continuum for metals on amorphous iron oxyhydroxide.

The Farley–Dzombak–Morel surface precipitation model (1985) unifies the two regimes as a thermodynamic continuum. As surface sites fill, the sorbing metal begins to form its own hydroxide, which itself presents fresh sites; the sorbent surface effectively becomes a solid solution between the host oxide, Fe(OH)₃(s), and the sorbate hydroxide, M(OH)₂(s). The governing condition couples adsorption equilibrium to solid-solution equilibrium:

This is more than a curiosity. It explains why removal efficiency is a smooth function of loading rather than a hard capacity cliff, why aged flocs behave differently from fresh ones, and why co-precipitation — adding ferric while the metal is still in solution — outperforms adsorption onto pre-formed oxide: the metal is incorporated into the growing solid as it forms, occupying interior as well as surface positions.

Worked example: ferric dose to provision sorption sites

A metal-finishing effluent contains 2.0 mg/L dissolved Zn (as Zn²⁺) at a flow of 20 m³/h, at pH 7.0 where free Zn is undersaturated with respect to Zn(OH)₂. The target is 0.5 mg/L. We estimate the ferric dose that provisions enough HFO sites to adsorb the 1.5 mg/L that must be removed, then check against the strong-site budget.

• Zn to be adsorbed: 1.5 mg/L ÷ 65.4 g/mol = 2.29 × 10⁻⁵ mol/L = 22.9 µmol/L.
• Site requirement: assume, from the edge, that at pH 7 the available high-affinity fraction lets each mole of Zn occupy roughly one strong site; strong sites are only ~0.005 mol per mol Fe. Required strong sites ≈ 22.9 µmol/L, so required Fe ≈ 22.9 / 0.005 = 4,580 µmol Fe/L if relying on strong sites alone.
• Include weak sites: total sites are ~0.2 mol/mol Fe. At pH 7 Zn also uses weak sites; taking an effective usable capacity of ~0.05 mol Zn per mol Fe at this pH and loading gives Fe ≈ 22.9 / 0.05 = 458 µmol Fe/L.
• Convert to dose: 458 µmol/L × 55.85 g/mol = 25.6 mg Fe/L. As ferric chloride (FeCl₃, 162.2 g/mol), dose ≈ 25.6 × (162.2 / 55.85) = 74 mg/L FeCl₃ (about 1.5 kg/h at 20 m³/h).

So a ferric dose of order 25 mg Fe/L supplies the surface sites to pull Zn from 2.0 to 0.5 mg/L by adsorption/co-precipitation at a pH where hydroxide precipitation alone would not. The spread between the strong-site-only (4,580 µmol) and mixed-site (458 µmol) estimates shows why bench jar testing is essential: the true operating point on the isotherm, and hence the real dose, must be fixed experimentally. Raising pH toward 8–8.5 moves up the Zn adsorption edge and reduces the dose needed — the classic dose–pH trade-off.

What does this mean for real coagulation plants?

Reading real ferric-dosed clarification through the surface-complexation lens changes several design decisions:

• Dose sizes the sorbent, not just the floc. Metal removal below the hydroxide floor is capacity-limited by HFO surface sites, so the ferric dose is a sorption parameter. Marginal dose increases can yield large residual-metal reductions until sites saturate.
• pH is a lever on the edge. Nudging pH to sit higher on a cation edge (or lower for arsenate) is often cheaper than more coagulant. Mixed cation/anion streams need a deliberate compromise or staging.
• Co-precipitate, do not post-adsorb. Dosing ferric upstream of, or into, the metal-bearing stream (rapid mix) lets metals incorporate into the forming solid — more effective than contacting pre-formed oxide, per the surface-precipitation continuum.
• Floc age and competition matter. Competing sorbates (phosphate, silicate, natural organics, Ca) occupy sites; recrystallisation of ferrihydrite to goethite over time lowers area and can re-release weakly held metals.

These mechanisms are the scientific backbone of the operational guidance in the heavy metals removal and compliance article; here we have set out why the chemistry behaves as it does so the dosing and pH strategy can be reasoned about rather than merely tabulated.