The Disinfection Trade‑Off: Chlorine, ClO₂, Ozone, or UV?

Chlorination has kept city water safe for over a century, but it comes with hard trade‑offs: regulated disinfection by‑products (DBPs) and weak performance against protozoa. Alternatives—chlorine dioxide (ClO₂), ozone, and ultraviolet (UV) light—promise faster inactivation and fewer halogenated DBPs but add complexity, energy, and the need for a secondary residual.

Numbers tell the story. At 10°C, achieving 1‑log (90%) Giardia inactivation takes roughly ~42 mg·min/L with free chlorine, versus ~0.48 mg·min/L with ozone (EPA guidance). Chlorine is cheap and reliable but forms trihalomethanes (THMs) and haloacetic acids (HAAs) when it meets natural organic matter—compounds WHO and EPA regulate tightly (WHO | NRC).

The upside for operators: each tool has a clear profile on pathogens, DBPs, and cost. The downside: there is no single winner.

Pathogen inactivation benchmarks

Free chlorine readily inactivates bacteria and many viruses at modest doses—on the order of 1–4 mg·L⁻¹ for minutes (NRC). It is essentially ineffective against chlorine‑resistant protozoan cysts (e.g., Cryptosporidium) at practical doses (WHO), with EPA CT values at 10°C of ~42 mg·min/L for 1‑log Giardia (e.g., 1 mg·L⁻¹ for 42 minutes) (EPA). In contrast, ozone achieves the same 1‑log Giardia at ~0.48 mg·min/L under the same conditions (EPA).

ClO₂ sits between chlorine and ozone. EPA data indicate it is about 4× more effective than free chlorine for Giardia at 10°C (1‑log CT ~21 mg·min/L for Cl₂ vs ~5.7 mg·min/L for ClO₂) and remains favorable at 20°C (chlorine ~21 mg·min/L vs ClO₂ ~7.7 mg·min/L for 1‑log) (EPA). It is generally superior to chlorine for protozoa, while chlorine can be somewhat better for certain viruses (EPA).

Ozone is the most potent broad‑spectrum disinfectant of the four. Small concentrations can reduce E. coli by 10⁴‑fold within minutes; 0.2–1.5 mg/L ozone over ~40 seconds delivered 2‑log poliovirus kill, and 1.45 mg/L ozone achieved >5‑log coxsackie inactivation (EPA). Protozoa remain toughest: roughly, CT for 2‑log Cryptosporidium is ~1.2 mg·min/L (e.g., ~5 min at 1.2 mg/L ozone), ~25× the CT for Giardia (EPA | EPA), consistent with the general resistance order: bacteria < viruses < protozoa (EPA).

UV inactivates by damaging microbial DNA/RNA. It is particularly effective against protozoa: a UV dose of only 1.6–2.5 mJ/cm² yields ~1‑log kill of Cryptosporidium or Giardia (EPA; WHO notes the same here). Many viruses are UV‑resistant: EPA tables cite ~58 mJ/cm² for 1‑log generic viruses, while bench studies show coliphage MS‑2 needed ~25 mJ/cm² for ~0.3‑log and adenoviruses ~24–30 mJ/cm² for 1‑log (EPA). In practice, UV reactors are sized for 2–4 log viruses (≈40–100 mJ/cm²)—and designs often cite 40–60+ mJ/cm²—ensuring >4 log for Giardia/Crypto (EPA).

DBP formation and limits

Free chlorine forms halogenated DBPs when it reacts with natural organic matter, most notably THMs (e.g., chloroform) and HAAs. WHO cites an upper safety range of a few hundred μg/L for THMs—e.g., 200 μg/L for chloroform—while the US EPA’s 80 μg/L total THM limit illustrates the concern (WHO). Hundreds of other halogenated DBPs (haloacetaldehydes, halonitriles, chloral hydrate, etc.) are also known (WHO). In practice, controlling total organic carbon or using precursor removal is required to meet THM/HAA standards (chloramines are generally a secondary disinfectant and can minimize THMs but form different DBPs like nitrosamines) (WHO). Many plants pair that strategy with targeted adsorption; “precursor removal” often includes options like activated carbon.

ClO₂ does not chlorinate organics, so it does not form THMs or HAAs. Its principal by‑products are chlorite (ClO₂⁻) and chlorate (ClO₃⁻), with WHO provisional guideline values of 0.7 mg/L each for drinking water (WHO; see also EPA). Typical applications (a few mg/L dose) produce chlorite on the order of a few tenths of mg/L or less, but overdosing or impurities from generation (e.g., unreacted Cl₂) can elevate chlorite/chlorate, so dose and purity must be controlled (WHO | EPA).

Ozone avoids THMs/HAAs but can form bromate (BrO₃⁻) when bromide is present; WHO sets a 0.01 mg/L (10 μg/L) guideline (WHO). It may also create aldehydes, ketones, and organic acids—largely unregulated relative to bromate. Running at lower ozone dose, shorter contact time, or lower pH helps keep bromate below 0.01 mg/L (WHO). Ozone can make some organics more biodegradable; post‑ozonation biological filtration is often used.

UV produces essentially no regulated chemical DBPs. It does not generate THMs or HAAs (EPA). At very high energies it may oxidize nitrite/nitrate, and UV‑chlorine combinations can form nitrosamines, but with plain UV there are no regulated DBPs.

Residual management in distribution

Only chlorine provides a lasting disinfectant residual in the network. Residual chloride ~0.2–0.5 mg/L is typical to protect against recontamination in pipes. ClO₂ has no long‑lived residual (it quickly degrades to chlorite), ozone decays within minutes, and UV leaves no residual—so ClO₂, ozone, and UV installations generally add a small chlorine or monochloramine dose after primary disinfection (WHO).

Indonesian regulations broadly follow WHO/ISO guidance (e.g., requiring ≥0.5 mg/L residual), so most PDAM utilities use chlorination or chloramination for residual protection (WHO).

Capital and O&M cost signals

Chlorine systems are the lowest‑cost to build and run. As the NRC puts it, “chlorine…is inexpensive and relatively convenient to produce, store, transport, and use” (NRC). Basic chlorinator installations—gas feed or bleach dosing and a contact tank—can cost only a few thousand dollars for small plants; for a 1 ML/d capacity, a simple feed system can be under $10,000. Annual chemical costs are modest, on the order of a few hundred dollars per million gallons treated; operational energy is minimal. Typical chemical cost lands around $0.01–0.05 per m³ treated. Chlorine feed is commonly automated using a dosing pump. Historically, ~95% of US systems used chlorine (NRC), and most Indonesian PDAM utilities likewise rely on it.

ClO₂ must be generated on‑site. Systems include a generator (e.g., using chlorite plus acid) and dosing equipment; capital is moderate—e.g., a small 780 gpm lab system was ~US$9,670 in 1979 dollars—scaling to tens–hundreds of thousands for a municipal plant (EPA). O&M includes sodium chlorite, acids or chlorine feed stocks, electricity, and routine labor/maintenance (pumps, analyzers); consumption is on the order of a few kg per 10⁶ m³ treated. Relative to ozone or UV, ClO₂ labor and maintenance are modest.

Ozone is capital‑ and energy‑intensive. A complete installation includes an ozone generator (oxygen or dried air fed), injection/contact chambers, and an off‑gas destructor. Capital often falls on the order of USD 2–5 per cubic meter of treated water capacity (modern terms), so a 100 MLD plant can run several million USD. Typical generators require ~0.4–0.6 kWh/kg O₃. Achieving ~2 mg/L residual ozone may need ~5–10 Wh per liter of water, which drives electrical costs into the tens to hundreds of $/ML depending on power price (EPA). Maintenance (oxygen supply, vacuum pumps, corrosion‑resistant lines, corona chamber replacement) adds to O&M, though there is no ongoing chemical feed cost. Very large plants (≥100,000 m³/h) are in operation (EPA).

UV capital and O&M are moderate. A UV reactor uses lamp banks in a flow‑through vessel with sensors and wipers. Costs scale sub‑linearly with size: one 11 MGD (≈42 ML/d) installation was estimated at ~US$1.2 million (≈$11,000 per MGD), while a 210 MGD plant cost ~$17 million (UNH | UNH). Annual O&M for the 11 MGD case was ~$23,000/yr, or about $0.0037/m³; in big systems, lamp power can be ~50–60% of O&M (the 210 MGD plant had ~61% of O&M for power) (UNH). Lamp life is ~9–12 months, and energy use is typically 0.02–0.05 kWh/m³—cheaper than ozone, costlier than chlorine. Compact, modular reactors ease retrofits; see ultraviolet systems.

Design choices and combinations

Because ClO₂ has no lasting residual, it is typically applied as a primary disinfectant after filtration—then followed by a small chlorine dose for residual (WHO). Plants commonly position this step after filtration to stabilize upstream water quality. Ozone and UV likewise require a post‑disinfection residual (usually chlorine or monochloramine) because neither provides lasting protection in distribution (WHO).

Controlling DBPs drives upstream choices. For chlorine, meeting THM/HAA limits typically means cutting organics; “controlling total organic carbon or using precursors removal is required” (WHO)—often via dedicated steps such as precursor removal. For ozone, operating strategies (lower dose, shorter contact, lower pH) minimize bromate to ≤0.01 mg/L (WHO).

A compact snapshot of the trade‑space

Chlorine: bacteria very susceptible; viruses readily inactivated; Cryptosporidium essentially not (WHO). Typical 1‑log Giardia CT is ~21–42 mg·min/L at 10–20°C (EPA | EPA). Pros: easiest and cheapest to implement; provides corrosion‑suppressing residual via chloramine formation. Cons: forms THMs/HAAs (WHO), poor against Cryptosporidium, gas‑handling hazards.

ClO₂: broad‑spectrum; stronger on protozoa (≈4× more effective vs Giardia than Cl₂ at 10°C; CT ~5.7 vs ~21 mg·min/L) and effective vs viruses (with some needing higher dose) (EPA). Minimal THMs, but generates regulated chlorite/chlorate (0.7 mg/L each, WHO) (WHO). No residual; on‑site generation carries maintenance and on‑site explosion risk.

Ozone: most powerful overall; 1‑log Giardia CT ~0.2–0.5 mg·min/L (EPA). Inactivation resistance: Cryptosporidium > viruses > bacteria (EPA). No residual; watch bromate (0.01 mg/L, WHO) (WHO). Corrosive gas demands special materials.

UV: effective on all classes; exceptionally strong on protozoa (≈2–5 mJ/cm² for 1‑log Crypto/Giardia) and decent on bacteria; viruses require higher UV dose (design to 40–60+ mJ/cm²; many plants design to ~40–100 mJ/cm² for 2–4 log viruses) (EPA | EPA). No chemical DBPs.

Capex at 1 ML/d: chlorine very low (simple feed/contact can be <$10,000); ClO₂ moderate (on‑site generator: tens of thousands for a few ML/d); ozone high (>$1–2 per L/day of capacity; often $Ms); UV moderate (~$0.5–1 per L/day; e.g., 11 MGD ≈$1.2M; 210 MGD ≈$17M) (UNH | UNH). Opex per ML: chlorine low (~$0.01–0.05/m³; energy negligible); ClO₂ medium (chemicals plus power); ozone high (energy often ~$5–15/ML); UV medium (power and lamps; e.g., 11 MGD ≈$23,000/yr; 50–60% of O&M is power) (UNH).

What utility managers are choosing

Each utility must weigh these trade‑offs. A small system with little organic load may stay with chlorine for simplicity and low cost (NRC). A large city targeting maximum pathogen removal and DBP control may opt for UV or ozone despite higher capex and O&M (EPA). For ClO₂ implementations, primary dosing typically follows filtration and is backed by a chlorine residual; in chlorine systems, maintaining a residual of ~0.2–0.5 mg/L is common practice in distribution (WHO).

Across all options, designers should base choices on source water quality (organics, bromide, turbidity), target inactivation (e.g., Giardia removal), DBP regulations, and life‑cycle cost. Where emerging contaminants or strict DBP limits drive design, utilities are increasingly considering ozone or UV as the primary barrier. Safety and operability considerations also matter: chlorine gas is hazardous; ClO₂ generation entails on‑site explosion risk; ozone is a corrosive, toxic off‑gas requiring destruction; UV performance is sensitive to UV transmittance.

Sources: WHO, EPA, NRC, and engineering cost studies underpin the figures above, including EPA Disinfection/DBP Rule technical tables for CT comparisons and virus/UV dose data (EPA | EPA | EPA), WHO guideline values for bromate (10 μg/L) and chlorite/chlorate (0.7 mg/L each) (WHO | WHO), the NRC on chlorine’s practicality (NRC), and UV capital/O&M benchmarks (UNH | UNH).