The disinfection dilemma: chlorine, ClO₂, ozone, or UV — the numbers that decide

Free chlorine (Cl₂ or NaOCl), chlorine dioxide (ClO₂), ozone (O₃), and ultraviolet light (UV, typically 254 nm) are the workhorses of primary disinfection. EPA guidance and WHO limits draw a stark contrast between what each does best, what each costs, and which by‑products (DBPs) each leaves behind (EPA CT tables; WHO bromate limit).

On microbial kill, ozone is a sledgehammer, UV is laser‑precise on protozoa but needs higher dose for viruses, chlorine dioxide sits in the middle, and free chlorine remains the cost‑effective workhorse for bacteria and many viruses. On DBPs, chlorine’s THMs and HAAs are the headline risks (EPA TTHM MCL = 80 μg/L), while ozone’s bromate risk climbs with bromide in source water, and ClO₂’s chlorite/chlorate demand tight control (EPA DBP limits).

Pathogen inactivation benchmarks (CT and dose)

CT (concentration×time, mg·min/L) and UV dose (mJ/cm²) are the hard metrics here. Free chlorine excels on bacteria and many viruses: EPA tables show a 4‑log (99.99%) virus inactivation at ~20 °C with chlorine CT on the order of 3–4 mg·min/L (EPA CT). But chlorine is much less effective on protozoa: at 10 °C and pH 7.0, CT₉₉.₉ (3‑log) for Giardia is ~134 mg·min/L, rising to ~205 mg·min/L near 0.5 °C (EPA CT).

Chlorine dioxide is roughly 4× more potent than free chlorine against Giardia at typical conditions: around ~15 mg·min/L for 3‑log at 20 °C (pH 6–9) versus ~73 mg·min/L for free chlorine (EPA comparison; ClO₂ CT; Cl₂ CT). For viruses, the advantage flips: 4‑log virus kill at 20 °C is ~12.5 mg·min/L for ClO₂ vs ~3 mg·min/L for chlorine (ClO₂ virus CT; Cl₂ virus CT).

Ozone delivers ultra‑low CTs: CT₉₉.₉ for Giardia is ~2.9 mg·min/L at 0.5 °C — orders of magnitude below chlorine — and ozone inactivates Cryptosporidium as well (O₃ Giardia CT; Cl₂ contrast; Crypto with ozone). EPA tables also put 4‑log virus kills with ozone at ~0.5 mg·min/L at 20 °C (EPA ozone virus CT).

UV is strong against protozoa and effective for bacteria/viruses when designed to dose: 3‑log Cryptosporidium ~12 mJ/cm² and 3‑log Giardia ~11 mJ/cm², while viruses need higher dose (~143 mJ/cm² for 3‑log) (EPA UV dose). In practice, many plants design around 30–40 mJ/cm², which typically achieves ≥4‑log for bacteria and enteric viruses (except adenovirus), and ≥2–3‑log for protozoa. USEPA tables also credit ~40 mJ/cm² for 3‑log Giardia and Cryptosporidium inactivation, even as raw studies report ~11–12 mJ/cm² yields 3‑log Giardia/Crypto (EPA UV dose).

Disinfection by‑products and regulatory limits

Free chlorine oxidizes natural organic matter to THMs and HAAs. EPA and WHO caps typically set total THMs around 80–300 μg/L; the US EPA TTHM MCL is 80 μg/L, and real‑world chlorinated water often ranges from tens to low‑hundreds μg/L depending on organic load (EPA DBP limits). Chlorine also forms haloacetonitriles and smaller amounts of chlorite from chlorite precursors.

Chlorine dioxide avoids THMs/HAAs but produces chlorite (ClO₂⁻) and chlorate (ClO₃⁻). The chlorite guideline is ~0.7 mg/L, and meeting it requires careful control; bromate from ClO₂ is minimal.

Ozone does not produce THMs/HAAs, but can oxidize bromide to bromate (BrO₃⁻), a regulated carcinogen with a WHO/EPA limit near 0.01 mg/L (10 μg/L). Ozone also yields aldehydes and other oxygenates, and it can create biodegradable DBPs that may fuel bacterial regrowth if not followed by a secondary disinfectant (WHO bromate limit).

UV adds no chemical disinfectant and produces essentially no new DBPs. A limited concern arises only when UV is combined with ozone (photolysis interactions), not with pure UV dosing.

Utilities with high organic carbon frequently balance disinfection vs DBPs; for instance, in water with 4–8 mg/L DOC, switching from Cl₂ to ClO₂ or O₃ can cut THMs by >50% (DBP overview). Indonesian surface waters, rich in humic matter, often trigger DBP concerns (context).

Capital and O&M cost ranges

Capital cost figures cited here are approximate per unit treatment capacity, expressed per MGD (million gallons per day). Operation costs noted for chemicals and electricity are those used for dosing or pumps; for chlorine, these costs are small relative to capital.

Free chlorine systems are low‑cost (feed pump and tank): roughly $10⁴–10⁵ per MGD, with very low O&M (mainly chemical and minimal power). A sodium hypochlorite or chlorine gas program that maintains ~0.5 mg/L residual costs only cents per 1,000 m³, and energy use is minimal.

Chlorine dioxide requires on‑site generation, placing it in the moderate–high capital bracket (~$10⁵–10⁶ per MGD), with higher O&M due to chlorite precursor chemicals and power (~$O(10^4)/yr per MGD). Monitoring of ClO₂ residual and chlorite adds to O&M.

Ozone sits at the high end: generator, contactor, off‑gas destruction, and support systems typically push systems to $10⁶+ at MGD scale. Electricity demand is high: published ranges include ~0.5–1.0 kWh/m³ for ozone processes, and producing 1 kg O₃ often takes 10–15 kWh. Achieving 1 mg/L O₃ dose at 10 MGD (~38,000 m³/d) needs ~30 kg O₃, or ~300–450 kWh/day, leading to power bills in the tens of thousands of USD per year for multi‑MGD plants (ozone energy LCA).

UV reactors (low‑pressure mercury lamp systems) are moderate–high in capital (~$10⁵–10⁶ per MGD) and moderate in O&M. A published study (Arora et al. 2001) estimated annual O&M at $22,600 for 1 MGD and $326,500 for 100 MGD, roughly $0.005–0.01 per m³ treated at medium dose. Newer medium‑pressure or LED units trend higher (UV cost study).

A life‑cycle assessment found UV‑LED and ozone setups had the largest environmental (and thus energy) footprints, whereas simple chlorination was least (LCA result).

Free chlorine: residual and DBPs

Free chlorine provides a lasting residual in distribution (typically ~0.2–0.5 mg/L), is simple to control, and is highly effective against bacteria and most viruses. Chlorine gas and hypochlorite do pose safety considerations and taste/odor if overdosed. For metering, utilities typically rely on equipment such as dosing pumps to deliver precise feed.

EPA data put 4‑log virus inactivation at ~3–4 mg·min/L at ~20 °C. For Giardia, the CT₉₉.₉ (3‑log) at 10 °C and pH 7.0 is ~134 mg·min/L, and near 0.5 °C it is ~205 mg·min/L — compared with ozone’s ~2.9 mg·min/L at 0.5 °C for similar kill (virus CT; Giardia CT; ozone contrast). DBPs are the downside: THMs, HAAs, and haloacetonitriles form in proportion to organic precursors, with EPA’s TTHM MCL at 80 μg/L (EPA DBP limits).

Costs are favorable: simple feed systems cost on the order of $10⁴–10⁵ per MGD; O&M is very low, with chemical purchases dominating and energy use negligible.

Chlorine dioxide: protozoa performance, chlorite control

ClO₂ is effective over a broad pH range and is markedly more potent on Giardia than chlorine (about 4× faster), but its virus performance is comparable to chlorine and sometimes behind it. EPA tables put 3‑log Giardia CT at ~15 mg·min/L at 20 °C (pH 6–9), and 4‑log virus kill at ~12.5 mg·min/L at 20 °C — compared to ~3 mg·min/L for chlorine under similar virus conditions. ClO₂ can inactivate Cryptosporidium at high doses, though it requires more than ozone (EPA efficacy; Giardia CT; virus CT; chlorine virus CT).

DBPs shift from halogenated organics to inorganic oxychlorine: no THMs/HAAs, but chlorite and chlorate arise, with a chlorite guideline near 0.7 mg/L. Bromate is minimal. Because ClO₂ leaves no long‑lived residual, utilities often maintain a secondary disinfectant.

Capital and O&M reflect on‑site generation: ~$10⁵–10⁶ per MGD for generators and feed equipment, with higher ongoing costs for chlorite precursors and power (~$O(10^4)/yr per MGD). Handling hazards exist (less than gas Cl₂, but not trivial), and monitoring of chlorite is mandatory.

Ozone: full‑spectrum kill, energy premium

Ozone is among the most powerful oxidants in water treatment. EPA data show CT₉₉.₉ for Giardia of ~2.9 mg·min/L at 0.5 °C, versus ~205 mg·min/L for chlorine, with ozone also delivering 4‑log virus kills at ~0.5 mg·min/L at 20 °C and inactivating Cryptosporidium effectively (Giardia CT; chlorine contrast; Crypto efficacy; virus CT).

On DBPs, ozone avoids THMs/HAAs but can form bromate (BrO₃⁻) when bromide is present, with a drinking‑water limit of ~0.01 mg/L (10 μg/L). It also yields aldehydes and other oxygenates. Ozone decays rapidly, leaving no residual, so plants commonly follow with chloramine or chlorine downstream (WHO bromate limit).

Expect high capital and energy costs: a generator, contactor, off‑gas destructor, and support systems (dry air or oxygen supply, cooling) often exceed $10⁶ at MGD scale, with electricity in the ~0.5–1.0 kWh/m³ range and frequent maintenance needs (energy footprint). Producing 1 kg of O₃ often requires 10–15 kWh; providing ~1 mg/L at 10 MGD (~38,000 m³/d) needs ~30 kg O₃ and ~300–450 kWh/day. Supporting equipment for ozone trains and off‑gas management typically sits alongside specialized ancillary systems.

UV light (254 nm): dose‑driven and DBP‑free

UV inactivates DNA/RNA with no chemicals added. It is very effective against protozoa (3‑log Crypto ~12 mJ/cm²; 3‑log Giardia ~11 mJ/cm²) and against bacteria/viruses when designed to dose, though viruses generally need higher dose (~143 mJ/cm² for 3‑log). Many plants design for 30–40 mJ/cm², typically achieving ≥4‑log for bacteria and enteric viruses (except adenovirus) and ≥2–3‑log for protozoa; USEPA credit tables also list ~40 mJ/cm² for 3‑log Giardia and Cryptosporidium (raw studies report ~11–12 mJ/cm²) (EPA UV dose).

DBP formation is essentially nil for pure UV. UV efficacy is immediate (no contact time), but water must be very clear (turbidity <1 NTU) and free of UV‑absorbing substances, and UV trains need periodic sleeve cleaning and lamp replacements as UV output declines. UV reactors used in municipal service align with systems such as ultraviolet units designed for drinking water duty.

Capital is typically ~$10⁵–10⁶ per MGD for low‑pressure lamp reactors, with O&M around $0.03–0.10/kL for lamp/electronics replacements (Arora et al. 2001). The same study estimated annual O&M at $22,600 for 1 MGD and $326,500 for 100 MGD, and newer medium‑pressure or LED units trend higher (UV costs).

Comparative profile and multi‑barrier practice

In aggregate: ozone provides the strongest kill (lowest CT) across pathogens, followed by chlorine dioxide, then free chlorine, and finally UV for viruses specifically (UV is stronger than low‑dose chlorine for protozoa). Only ozone, ClO₂, and UV can reliably inactivate Cryptosporidium; chlorine alone cannot at practicable doses. Many large plants run multi‑barrier lines — for example, pre‑ozonation or UV for cysts/precursors followed by chlorine to carry residual.

Costwise, chlorine is the cheapest on both capital and O&M: 1 mg/L as Cl₂ can cost <$0.02 per 1,000 gallons, and chlorination gear is often <$100k/MGD. UV capital and O&M sit in the middle, and ozone is energy‑intensive. A life‑cycle assessment reports UV‑LED and ozone systems with the largest energy footprints, with simple chlorination least (LCA).

Practical operating considerations matter. Safety: chlorine gas requires strict protocols; many plants prefer sodium hypochlorite despite higher chemical cost. Ozone gas handling requires containment and off‑gas destruction. Maintenance: UV lamps and sensors need cleaning/replacement; ozone systems need air prep, destructors, and cooling; chlorinators require calibration. Regulatory: coagulation, filtration, and a disinfectant residual (often ≥0.2 mg/L as chlorine at distribution entry) are typical, and Indonesia’s drinking water standards (Permenkes 492/2010, etc.) focus on microbial compliance and allow either residual chlorine or safe distance to tap; any chosen disinfection must meet coliform/virus criteria without violating chemical limits (e.g., THMs, microcystin if pretreated water).

Design takeaways and numeric guardrails

Chlorine remains attractive for low cost and residual protection, but high organic precursors or protozoa risk push many toward hybrid lines. ClO₂ fits when THM limits are tight and protozoa control is needed, at the price of chlorite management and on‑site generation. Ozone is favored for advanced goals (taste/odor removal, micropollutants, full‑spectrum disinfection) if budgets support capital and energy. UV is increasingly deployed — sometimes with low chlorine after — for its DBP‑free disinfection and robust protozoa inactivation.

Designers rely on jar tests, pilots, and modeling to size units and predict DBPs. As a numeric anchor, achieving 99.9% Giardia reduction with ozone might require only ~0.5 mg/L O₃ at warm temperatures, whereas chlorine would need ~5–10 mg/L (Giardia CT contrast: ozone; chlorine). Decision‑makers should weigh those targets against capital budgets and limits such as the WHO bromate limit of 0.01 mg/L (10 μg/L) for ozone and a chlorite limit near ~0.7 mg/L for ClO₂ (WHO bromate).

Sources: EPA disinfection guidance (CT tables for pathogens) (EPA CT table 1; EPA CT table 2; EPA ozone CT; Giardia CT chlorine), an LCA on disinfection energy (ozone/UV energy; LCA result), EPA Stage‑2 T&D cost guidance (Giardia/ClO₂; virus CT ClO₂), UV cost estimates (UNH UV costs), WHO bromate limits (WHO), and a Crypto/ozone CT correlation (ResearchGate).