Advanced oxidation and nanobubbles for effluent that will not settle

September 5, 2026

By Vishalkumar Suthar, Director

Row of four cylindrical filter vessels connected by blue pipework with a pressure gauge and valves at an outdoor water treatment installation — Photo: Alexey Demidov via Pexels

When a stream has been through settling and biology and still carries colour and refractory COD, what is left is chemistry. Advanced oxidation attacks it with hydroxyl radicals; nanobubbles on their own are not shown to. What they do is get gas into the water and keep it there. The two are often sold together, and confusing them is the usual reason a plant disappoints.

What an advanced oxidation process is

The United States Environmental Protection Agency’s 1998 handbook on advanced photochemical oxidation processes gives the working definition. These processes generate and use powerful but relatively non-selective transient oxidising species, primarily the hydroxyl radical. The handbook’s Table 1-1 of oxidation potentials in water puts the hydroxyl radical at the top of its list, 2.80 against 2.07 for ozone, on a scale whose column is labelled in electronvolts where the quantity is volts against the standard hydrogen electrode. Its Table 1-2 adds that most contaminants react one million to one billion times faster with the radical than with ozone.

The routes are familiar: ultraviolet light with ozone or hydrogen peroxide, ozone with hydrogen peroxide, Fenton’s reagent, electron beam. The handbook covers both photochemical and non-photochemical families and cites the 1987 paper by Glaze, Kang and Chapin on the chemistry of water treatment involving ozone, hydrogen peroxide and ultraviolet radiation.

Non-selectivity has a consequence. The radicals oxidise whatever is in front of them, including the biodegradable load you could have removed far more cheaply, and the carbonate that scavenges them. Advanced oxidation is a finishing step, not a substitute for biology.

What a nanobubble is, and what it does

Chen and colleagues, reviewing micro and nanobubbles for environmental remediation in Frontiers in Environmental Science in 2025, describe them as bubbles with diameters in the micrometre and nanometre ranges, and list what follows from that size: high internal pressure, large specific surface area, longer residence time in the liquid and a high surface zeta potential. They put a one micrometre bubble at about 390 kPa, roughly four times atmospheric, and note it is specific surface area that raises gas transfer.

Those are the defensible claims, and worth having: more interfacial area per unit of gas, slower loss to the surface and longer contact are what a gas-limited process lacks.

The micro and nanobubble generator that came out of our founder’s work is sold as NANOBBLE by Energenius Technologies, and is described there the same way: bubbles small enough to stay in suspension rather than rise, changing how gas transfers into a liquid.

What nanobubbles are not

They are not an oxidation process on their own, and the literature is careful about this even where enthusiastic.

Chen and colleagues set out the competing explanations for reactive oxygen species in these systems. One is interface ion accumulation as a bubble contracts, which they trace to Takahashi and colleagues in 2007 on collapsing microbubbles: the surface zeta potential rises as contraction accelerates, and hydroxide ions concentrate at the gas-liquid interface. The same review then states that there are theoretical disputes about the generation pathways, and that the link between a microbubble’s dynamics and its oxidation effect has not been systematically analysed.

Read that as an engineer. Radicals have been reported under particular conditions, the mechanism is contested, and nobody has shown a bubble generator alone doing an oxidant’s duty. If a supplier offers COD removal from bubbles with no oxidant, ask which oxidant is consumed and how it was measured.

Where the two belong together

The honest case for combining them is mass transfer, and ozone is where it matters most. Chen and colleagues describe conventional ozone as limited by low solubility, short half-life and a low mass transfer coefficient, with micro and nanobubbles improving its dissolution and extending its residence time.

The Central Pollution Control Board’s charter for water recycling in textile industries in the Ganga river basin, whose action plan runs in two phases to 31 December 2021 and 31 December 2022, describes the same problem from the plant side. The contactor exists to transfer ozone from the gas bubble into the bulk liquid with enough contact time, and because ozone is consumed quickly it must be contacted uniformly in a near plug flow vessel. The charter places ozonation, with activated carbon adsorption, as a way of making effluent fit for membrane filtration, not a bulk COD process.

Hutagalung and colleagues put numbers to it in Frontiers in Environmental Science in 2023, on textile wastewater at pH 7.93, COD 114.30 mg/L and TSS 33.73 mg/L. Over 300 minutes, ozone-based advanced oxidation alone removed about 10 per cent of COD; with nanobubble generation it removed about 80 per cent, the bubble population peaking at 216.9 nm. Treat it as one bench comparison at a low starting COD, not a design figure. The mechanism is worth testing: ozone was the oxidant in both runs, and the bubbles changed how much reached the pollutant instead of the vent.

What decides the cost

The handbook’s costing method is the useful part for a budget. It works from EE/O, the electrical energy for one order-of-magnitude removal per cubic metre treated, from treatability tests. The dose is EE/O times the log reductions needed, the electrical cost is that dose times the power price, and lamp replacement typically adds 30 to 50 per cent.

The spread across the handbook’s own case studies follows from that. Groundwater carrying RDX took 0.53 kWh per cubic metre with 10 mg/L of hydrogen peroxide for better than 82 per cent removal, at an estimated USD 0.02 per cubic metre at high flow. Wastewater from a display manufacturing plant at 3,000 mg/L COD took 160 kWh per cubic metre of photo-Fenton to reach about 50 mg/L, at USD 44. Same family of technology, three orders of magnitude apart, because the load and the decades to cross changed. The handbook also lists intermediates left behind in one case, including nitrate, nitrite and ammonia: oxidation ends somewhere, and the consent applies where it ends.

What this means for your plant

Ask what the radical source is. An advanced oxidation process is defined by hydroxyl radical generation; if nothing in the flowsheet makes them, it is not one.
Buy nanobubbles for mass transfer, flotation and residence time. Those claims are supported; radical generation from bubbles alone is reported under particular conditions and disputed.
Put oxidation last. Non-selective radicals spend themselves on biodegradable load and carbonate alkalinity before reaching the refractory fraction you are paying to remove.
Size from a treatability test, in log reductions. The dose, and so the cost, scales with the decades you have to cross.
Measure the scavengers and the products, not only the target. Carbonate and bicarbonate scavenge the radicals, which is why the handbook has some systems dosing acid ahead of the reactors; intermediates decide whether you meet consent.

Working out which applies is bench work before equipment, and where the answer is a machine that does not exist off the shelf, design, machining and fabrication sit together.

If a stream has stopped responding to your biology, tell us what it is doing and we will say what we would test first.