Ozone is increasingly specified for effluent polishing — micropollutants, colour, pathogens. It is also a powerful enough oxidant to manufacture genotoxic by-products from constituents already in the water. In bromide-bearing UK effluent, bromate, not ozone demand, is what caps the dose.
Bromate: the dose-limiting by-product
Bromate is a genotoxic carcinogen regulated at 10 µg/L. It forms by two competing routes, and which one dominates decides how you control it.
The direct ozone pathway
Br− + O3 → OBr− → BrO2− → BrO3−Sequential oxidation by molecular ozone; the first step is slow and rate-limiting
The hydroxyl radical pathway
Br− + •OH → Br• → BrO• → BrO2• → BrO3−Radical-mediated; faster, and it opens whenever ozone decomposes to •OH
Formation is therefore the sum of both:
d[BrO3−]/dt = k1[HOBr/OBr−][O3] + k2[Br−][•OH]
The direct route dominates below about pH 8 with low radical exposure. The radical route takes over when hydrogen peroxide is present or at high pH, where ozone decomposition to •OH accelerates. That is the awkward part of designing an ozone stage for micropollutant removal: the radical exposure you want for oxidising pharmaceuticals is the same exposure that makes bromate.
| Control | Mechanism | Practical note |
|---|---|---|
| Depress pH to 6.5–7.0 | Suppresses O3 decomposition to •OH and shifts the hypobromite equilibrium towards the less reactive HOBr | Usually the cheapest lever; acid dosing and downstream pH correction |
| Ammonia addition | NH3 scavenges HOBr as bromamine, which oxidises to bromate only slowly | Adds nitrogen to the effluent — check the consent first |
| Peroxide quenching | H2O2 reduces HOBr back to bromide | Counter-intuitive: peroxide also generates •OH, so it must be dosed after the contactor, not into it |
N-nitrosamines
NDMA is a potent carcinogen with a guideline value of 0.01 µg/L — two orders of magnitude below the bromate limit, and correspondingly harder to hit. Unlike bromate it comes from organic precursors rather than an inorganic ion, which changes the control strategy entirely.
| Precursor | Reaction | NDMA yield |
|---|---|---|
| Dimethylamine (DMA) | Oxidation via the N-oxide | Low, below about 1% |
| Ranitidine-type pharmaceuticals | Direct ozonation of the tertiary amine | High, 50–90% |
Three controls, in order of preference:
- Remove the precursor. Granular carbon ahead of the contactor takes out ranitidine-type compounds and amines before they can react. See GAC breakthrough prediction.
- Photolyse afterwards. NDMA absorbs strongly near 228 nm; a UV fluence of the order of 1,000 mJ/cm² achieves better than 90 per cent destruction. That is roughly 25 times a normal disinfection dose, so it is a dedicated stage, not a bolt-on to UV disinfection.
- Limit radical exposure by holding the ozone-to-carbon ratio down — which is the same lever that controls bromate.
Aldehydes
Ozonating natural organic matter cleaves it into low-molecular-weight carbonyls — formaldehyde, acetaldehyde, glyoxal — at yields of roughly 5–20 µg per mg of ozone consumed. These are the most predictable by-products of the three and the least troublesome, because they are readily biodegradable.
That biodegradability is also the risk: assimilable organic carbon released into a distribution system feeds bacterial regrowth. The answer is to consume it deliberately on site rather than let the network do it — biological activated carbon downstream of the contactor removes aldehydes by assimilation in the biofilm while also destroying ozone residual. On a wastewater works discharging to a watercourse the regrowth concern largely falls away, but the BAC stage is usually justified anyway for residual ozone.
Design protocol
For UK effluent with bromide at 0.5–2.0 mg/L and DOC at 8–15 mg/L:
- Measure bromide and NDMA formation potential before designing anything. Both vary by catchment in ways that bulk parameters do not predict, and both can make an otherwise sound ozone scheme unviable.
- Above 1 mg/L bromide with pH over 7.5, specify pH depression or downstream peroxide quenching from the outset — retrofitting either is expensive.
- Above about 50 ng/L NDMA formation potential, specify carbon pre-treatment, or select a different advanced oxidation route.
- Design to an ozone-to-carbon ratio of 0.3–0.5 mg O3 per mg C. This is the master control variable: enough oxidant exposure for micropollutant removal, not enough radical exposure to run away on by-products.
- Provide carbon or biological carbon polishing as standard, for aldehydes and residual ozone.
- Monitor bromate online by ion chromatography where bromide is variable. A works that only samples weekly cannot demonstrate compliance against a 10 µg/L limit on a parameter that moves with the tide.
Related: ozonation in water treatment, disinfection by-products, advanced oxidation processes.
Frequently asked questions
Why does bromate limit the ozone dose rather than ozone demand?
Because bromide is not removed upstream and converts to bromate in proportion to oxidant and radical exposure. On a bromide-bearing effluent you reach the 10 µg/L bromate limit before you reach the dose that would satisfy ozone demand, so the by-product, not the target contaminant, sets the operating point.
Can hydrogen peroxide be used to control bromate?
Only downstream. Peroxide reduces hypobromite back to bromide, which suppresses bromate — but it also generates hydroxyl radicals, which is the other route to bromate. Dose it after the contactor as a quench, never into it.
Why do two works with the same DOC produce very different NDMA?
Because NDMA yield depends on which precursors are present, not how much organic carbon there is. Ranitidine-type pharmaceuticals convert at 50–90 per cent while bulk amines convert at under one per cent, so a catchment with pharmaceutical or personal-care discharge can produce an order of magnitude more NDMA at identical DOC.
Is granular carbon needed after ozonation?
Effectively yes. It removes the aldehydes ozonation creates from natural organic matter, and destroys residual ozone. On a works discharging to a watercourse the regrowth argument is weaker, but the residual-ozone duty alone usually justifies it.