Ozone vs UV for Drinking Water Disinfection

UV for drinking water is typically specified at 30–40 mJ/cm² at 254 nm, at end of lamp life and with a fouling allowance. The dose is delivered in seconds; contact time is not a design variable in the same way as it is for ozone. The key variables are flow, UV transmittance, turbidity and lamp output.

Ozone for drinking water is typically 1–3 mg/L with 5–15 minutes of contact. Bottle rinsing uses 2–4 mg/L. The dose must cover the ozone demand of the water plus the residual needed for disinfection. CT, the product of dissolved residual and contact time in mg·min/L, is the design metric. A tower holding 5 m³ at 10 m³/h gives a nominal 30 minutes; a residual of 0.4 mg/L after demand gives a CT of 12 mg·min/L.

Ozone. It is a strong oxidant that breaks down colour-causing organics, destroys many taste and odour compounds, and converts soluble iron and manganese into particles that a multimedia filter can capture. The oxidation step must be followed by filtration; ozone alone converts the metals but does not remove the resulting solids. Typical doses for iron and manganese removal start at 1–3 mg/L, set by the stoichiometric demand plus the organic demand.

UV has no effect on colour, taste, iron or manganese at the doses used for disinfection. It cannot replace an oxidation step, and it cannot compensate for a filter that is not removing the oxidised particles. If the problem is aesthetic or metallic rather than microbiological, UV is the wrong tool no matter how large the reactor.

UV leaves no residual. Protection ends at the reactor outlet, and any microorganism introduced in the storage tank, the recirculation loop or a dead leg can grow without further inhibition. That is why a UV system depends on hygienic storage: a sealed tank, a 0.2 µm hydrophobic vent, a conical bottom short enough to prevent stratification, and no dead legs longer than three pipe diameters.

Ozone leaves a short-lived residual, typically 0.2–0.5 mg/L in the tank and a half-life of 20–30 minutes in clean water at 20 °C. That is enough to protect a recirculated tank and loop, but not a long distribution network. Monitor dissolved ozone or ORP, and do not assume the residual reaches a branch that is not circulating. For the last metre, UV at the point of use is often added.

UV at 254 nm produces no known disinfection by-products at drinking-water doses, because it does not add a chemical. The practical risks are electrical, optical and maintenance-related: never look at an energised lamp, handle quartz sleeves carefully, and replace lamps and sleeves on schedule. UV does not remove nitrate, arsenic or other dissolved contaminants, so it cannot be used to fix a chemical water-quality problem.

Ozone can produce bromate when bromide is present, and bromate is regulated in many drinking-water standards. Check the bromide concentration and the ozone dose before design, and confirm the bromate result in the finished water. Ozone is also a workplace hazard: the generator room needs ventilation, ozone detection and an off-gas destructor on the contact tower vent. Never discharge ozone-containing gas to an occupied space.

Yes, and the combination is common in bottling and building drinking-water plants. Ozone is dosed into the storage tank recirculation loop to oxidise organics and maintain a residual; UV is installed at the point of use to protect the last metre of pipe without adding a residual to the drinking water. The order matters: ozone first, then storage, then UV, so that the UV sees water with a low oxidant demand.

A common mistake is to place UV before the ozone contact tank. Ozone in the water absorbs 254 nm light, so the UV dose falls even though the lamp output is normal. Another is to use ozone for disinfection and then install a carbon filter downstream without controlling the carbon bed; the carbon removes the residual and can support bacterial growth if it is not maintained.

Start from the water problem, not from the equipment. If the risk is microbiological and the water is clear, UV is simpler, leaves no residual and needs no contact tower. If the water has colour, taste, iron or manganese, or if a storage tank and loop need protection, ozone does more work. If both are true, use both in the right order.

UV is electrically simple but has consumables: lamps at 8,000–9,000 hours, quartz sleeves, sensors and cartridges. Ozone has a generator, an air or oxygen feed, a contact tower, an off-gas destructor and more instrumentation. Ozone generally needs more space, more maintenance skill and more safety management; UV generally needs cleaner water and a disciplined lamp and sleeve replacement routine.

Neither choice is permanent. A plant may start with UV for a clear municipal supply and add ozone later if the source changes or a storage tank is added. Design the pipework with future sample points and a bypass so that the second barrier can be added without cutting the main line. The water analysis should be repeated after any source change, because the disinfection choice follows the water, not the other way round.