Advanced oxidation processes (AOPs) generate the hydroxyl radical (•OH), a near-universal oxidant (E0 ≈ 2.8 V), to destroy refractory COD and trace micropollutants that biology cannot touch. Because •OH reacts at near diffusion-limited rates (k ≈ 108–1010 M-1 s-1), the design problem is really one of radical yield versus scavenging by the water matrix.
What are advanced oxidation processes?
Advanced oxidation processes are a family of treatments that produce hydroxyl radicals (•OH) in situ to oxidise organic pollutants that are non-biodegradable, toxic or present at trace concentrations. Unlike selective oxidants such as chlorine, the hydroxyl radical is powerful and essentially non-selective, attacking almost any C–H or C=C bond by hydrogen abstraction or electrophilic addition.
They are used where conventional biology fails: refractory COD (dye house, landfill leachate, pharmaceutical and pesticide effluents), micropollutants (endocrine disruptors, PFAS precursors, antibiotics) and as a pretreatment to convert non-biodegradable organics into biodegradable fragments ahead of a biological stage. Quantifying that refractory fraction first — see our note on COD and BOD wastewater characterisation — is the essential precursor to specifying an AOP.
Why is the hydroxyl radical so effective?
The hydroxyl radical sits second only to fluorine among practical oxidants. Its standard reduction potential is E0 ≈ 2.8 V (versus 1.36 V for chlorine and 2.07 V for ozone), and it reacts by three fast routes: hydrogen abstraction, electrophilic addition to unsaturated bonds, and electron transfer.
where [P] = pollutant concentration (M), [•OH] = instantaneous radical concentration (M) and k•OH,P = second-order rate constant. For most organics k•OH,P = 108–1010 M-1 s-1, i.e. near the diffusion limit.
Because the radical is consumed in microseconds, its steady-state concentration is minuscule (typically 10-12–10-14 M). What actually governs removal is the time-integral of that concentration, the •OH exposure ∫[•OH] dt. Integrating the rate law gives a clean pseudo-first-order form:
The exposure ∫[•OH] dt links to ozone exposure through the Rct = ∫[•OH] dt / ∫[O3] dt concept — a matrix-specific constant (often 10-8–10-9) that lets a measurable ozone exposure predict radical exposure.
What are the main AOP routes?
All AOPs share the •OH endpoint but differ in how they make it, their energy source and their optimum chemistry:
- Fenton and photo-Fenton — ferrous iron catalyses hydrogen peroxide: Fe2+ + H2O2 → Fe3+ + •OH + OH-. Cheap and dark-driven, but demands an acidic optimum near pH 3 (above pH ~4 iron precipitates as hydroxide) and generates iron sludge. Adding UV/visible light regenerates Fe2+ via the photoactive Fe(OH)2+ complex, cutting iron dose.
- Ozone and peroxone (O3/H2O2) — molecular ozone is selective, but adding peroxide accelerates its decomposition to •OH. The optimum is around H2O2:O3 ≈ 0.5 mol/mol; excess peroxide scavenges radicals.
- UV/H2O2 — 254 nm photons photolyse peroxide (H2O2 + hν → 2 •OH) with a primary quantum yield Φ ≈ 0.5 mol/einstein (two radicals per photon absorbed). Reliable and residual-free, but H2O2 absorbs weakly (ε254 ≈ 18–19 M-1 cm-1), so it is UV-energy intensive.
- Heterogeneous photocatalysis (TiO2/UV) — band-gap illumination (<387 nm for anatase, Eg ≈ 3.2 eV) creates electron–hole pairs that generate •OH at the surface. Works near neutral pH with no chemical dosing, but photon efficiency and catalyst recovery limit full-scale use.
The choice between them is rarely made on radical chemistry alone. Fenton wins on capital cost and raw oxidation capacity for coloured, high-COD industrial effluents where an acid/neutralise cycle is tolerable. Ozone-based routes dominate potable and reuse polishing because they leave no persistent residual and simultaneously oxidise taste, odour and colour. UV/H2O2 is favoured where a validated, residual-free barrier that also disinfects is required — groundwater remediation and water reuse in particular. Photocatalysis remains largely a research and niche technology, held back by the difficulty of recovering fine catalyst and the low fraction of incident photons usefully absorbed.
How does the water matrix limit performance?
The single biggest design error is ignoring radical scavenging. Because •OH is non-selective, it is consumed not only by the target pollutant but by every reactive constituent of the matrix — and the background usually dwarfs the target.
Carbonate and bicarbonate are the classic scavengers: HCO3- (k ≈ 8.5 × 106 M-1 s-1) and CO32- (k ≈ 3.9 × 108 M-1 s-1) convert •OH into the far less reactive carbonate radical. Dissolved organic matter (background COD/DOC) competes similarly. The fraction of radicals reaching your target is:
where the denominator sums the scavenging rate of every matrix species Si. High alkalinity, high DOC and nitrite all push fP down, wasting oxidant. This is why AOPs suit low-matrix polishing duties, not raw high-COD streams.
The practical consequence: soften or pre-treat to strip alkalinity where possible, and target AOPs at the refractory residual after biology rather than the full organic load.
Two further matrix effects deserve a bench check before design. First, turbidity and colour attenuate photons in every UV-driven route — a coloured landfill leachate can absorb most of the incident light within millimetres, so the usable optical path and hence reactor geometry are dictated by the water's UV transmittance, not just its chemistry. Second, bromide in the raw water is oxidised by ozone to bromate, a regulated (10 µg/L) carcinogen; peroxone and O3 routes must be screened for bromide and, if present, may be ruled out in favour of a UV route. These interactions are why a validated pilot on the actual effluent, rather than literature rate constants alone, underpins any defensible AOP specification.
Worked example: UV/H2O2 dose and electrical energy per order
Consider polishing a groundwater carrying 5 µg/L of a pesticide (k•OH,P = 5 × 109 M-1 s-1) down to 0.5 µg/L — one order of magnitude removal — by UV/H2O2. The standard scale-up metric is the electrical energy per order, EE/O: the kWh needed to reduce a pollutant by 90% in 1 m³ of water.
- Definition: EE/O = P / (Q · log10([P]0/[P])), with lamp power P in kW and flow Q in m³/h.
- Suppose a validated pilot on this water removes one log with a 12 kW lamp bank treating 15 m³/h. Then EE/O = 12 / (15 × 1) = 0.80 kWh/m³ per order.
- Two logs (5 → 0.05 µg/L) needs 2 × EE/O × Q = 2 × 0.80 × 15 = 24 kW of delivered UV power.
- Peroxide demand: at a typical 5 mg/L H2O2 residual target and 15 m³/h, dose = 5 g/m³ × 15 = 75 g/h of H2O2, plus any quenched downstream to protect a following stage.
At an energy price of £0.20/kWh, the two-log duty costs ≈ 24 kW × £0.20 = £4.8/h in UV energy alone, or about £0.32/m³ — a figure that makes clear why AOPs are reserved for low-flow, high-value polishing. EE/O values above ~1–2 kWh/m³/order usually signal a scavenging-dominated matrix that should be pre-treated first. For the downstream UV hardware itself, our guide to UV reactor and dose validation covers lamp selection and validated dose delivery.
How do AOP routes compare?
No single AOP wins on all counts; selection turns on matrix, flow, target and whether a chemical residual is acceptable.
| Route | Oxidant / driver | Optimum pH | Strengths | Limitations |
|---|---|---|---|---|
| Fenton | Fe2+ + H2O2 | ~3 | Low capital, no energy input, high COD capacity | Acidification + neutralisation, iron sludge |
| Photo-Fenton | Fe / H2O2 + UV/solar | ~2.8–3 | Lower iron dose, solar-drivable, fast | Still acidic, needs photons + sludge handling |
| O3/H2O2 (peroxone) | Ozone + peroxide | 7–8 | No pH shift, good for taste/odour and micropollutants | Ozone generation energy, bromate risk |
| UV/H2O2 | 254 nm + peroxide | Broad (5–9) | Residual-free, reliable, dual disinfection | UV-energy intensive, weak ε of H2O2 |
| TiO2/UV | Photocatalyst + UV-A | ~5–7 | No chemical dosing, ambient conditions | Low photon efficiency, catalyst recovery |
Where an AOP doubles as a disinfection barrier, the UV routes integrate naturally with UV disinfection design, sharing the same reactor and dose-validation logic.
When should you use partial oxidation instead of full mineralisation?
Driving an AOP to complete mineralisation — every organic carbon converted to CO2 and water — is thermodynamically possible but rarely economic, because the last increments of COD are the most oxidant-hungry and the intermediates (short-chain carboxylic acids: oxalic, acetic, formic) react slowly with •OH.
The smarter strategy is usually partial oxidation as pretreatment: apply just enough AOP to break refractory rings and cleave toxic structures, raising the BOD5/COD ratio from a non-biodegradable <0.2 up to >0.4, then hand the water to a cheap biological stage. This biodegradability uplift is monitored via BOD5/COD (or a respirometric/Zahn–Wellens test) and typically minimises total oxidant cost per unit COD removed. Integrating an AOP as a targeted pre- or post-treatment within the wider industrial wastewater treatment train — rather than as a standalone polisher — is what makes the economics work.
Frequently asked questions
What is an advanced oxidation process?
An advanced oxidation process (AOP) is any treatment that generates hydroxyl radicals (•OH) in situ to oxidise pollutants. The radical is a powerful, non-selective oxidant (E0 ≈ 2.8 V) that attacks organics conventional biology cannot degrade, including refractory COD and trace micropollutants such as pesticides and pharmaceuticals.
When should you choose an AOP over biological treatment?
Use an AOP when the organics are non-biodegradable, toxic to biomass, or present at trace levels that biology cannot reach. AOPs are energy- and chemical-intensive, so they suit low-flow polishing or targeted pretreatment of a refractory residual — not raw, high-COD streams that a cheaper biological stage can handle.
Why does the water matrix matter so much in AOP design?
Because the hydroxyl radical is non-selective, it is scavenged by matrix constituents — especially carbonate, bicarbonate and background dissolved organics — before it reaches the target pollutant. High alkalinity and DOC waste oxidant and raise energy cost sharply, so matrix characterisation and, where possible, alkalinity reduction are essential first steps.
What is EE/O and why is it used?
EE/O, the electrical energy per order, is the kWh needed to reduce a pollutant by 90% in one cubic metre of water. It normalises AOP energy demand across flows and log-removals, making it the standard metric for comparing UV-based processes and for scale-up from validated pilot data.
Why is Fenton oxidation run at acidic pH?
The Fenton reaction needs dissolved ferrous/ferric iron to catalyse peroxide decomposition to •OH. Above roughly pH 4 the iron precipitates as ferric hydroxide, losing catalytic activity and forming sludge. The optimum is near pH 3, which means acidification before and neutralisation after treatment — a key operating cost of the process.
Can AOPs fully mineralise organic pollutants?
In principle yes, to CO2 and water, but full mineralisation is rarely economic because the final COD increments and slow-reacting intermediates such as oxalic and acetic acid consume disproportionate oxidant. Partial oxidation to raise the BOD/COD ratio above 0.4, followed by biological treatment, is usually the lower-cost route.