Bleach to chlorine dioxide in cooling towers: switching risks
Summary
Bleach suits cooling towers run near or below pH 7.5 with little ammonia. Chlorine dioxide suits alkaline or ammonia-bearing water, provided on-site generation runs reliably. A switch goes wrong in three ways: the generator adds the chloride you wanted to remove, the plan leaves out acid supply and residual monitoring, or copper behavior is assumed instead of measured.
Hypochlorous acid, the stronger biocidal form of bleach, has a pKa of about 7.5, and above pH 8.5 it is fully dissociated to the weaker hypochlorite ion (US EPA, 1999).
Chlorine dioxide is not pH dependent between pH 5 and 10 in how it inactivates microorganisms, but above pH 9 it disproportionates to chlorite and chlorate (US EPA, 1999).
After replacing bleach with chlorine dioxide, the High Desert plant cut chloride loading on its zero liquid discharge system by at least 15% (Combined Cycle Journal, 2017 pilot).
In one maker's 90-day trial, effluent chloroform fell from 16.3 ppb under bleach to below 0.5 ppb under chlorine dioxide at 0.2-0.4 ppm residual (Nalco Water, 2016).
Chlorine dioxide gas is explosive above 10% by volume in air, so it must be generated on site rather than shipped (US EPA, 1999).
This comparison covers open recirculating cooling towers that use bleach (sodium hypochlorite) as the main oxidizing biocide. It leaves out drinking water disinfection limits, once-through seawater cooling, and generator design, which is the generator maker's job.
How do bleach and chlorine dioxide differ in a cooling tower?
The two biocides differ mainly in three things: how pH and ammonia affect them, what they leave behind, and what equipment they need.
Bleach works through hypochlorous acid, its stronger biocidal form, which dissociates into the weaker hypochlorite ion as pH rises. With a pKa of about 7.5 (US EPA, 1999), bleach loses strength in alkaline water. Chlorine dioxide stays in molecular form across the pH range of natural waters, and its inactivation of microorganisms is not pH dependent between pH 5 and 10 (US EPA, 1999). Above pH 9, though, chlorine dioxide disproportionates, splitting into chlorite and chlorate (US EPA, 1999), so a tower run above pH 9 loses part of each dose to byproducts.
That stability comes at a price. The gas is explosive above 10% by volume in air (US EPA, 1999), so chlorine dioxide needs a generator on site.
Most of the risk sits with ammonia, chloride and copper, and each gets its own section below. Work through the table one check at a time to decide whether a switch is worth a trial at your tower.
Check | Bleach (sodium hypochlorite) | Chlorine dioxide |
|---|---|---|
Effect of pH | Weaker as pH rises; pKa about 7.5 (US EPA, 1999) | Not pH dependent between pH 5 and 10 (US EPA, 1999); disproportionates to chlorite and chlorate above pH 9 (US EPA, 1999) |
Reaction with ammonia | Forms chloramines at pH 7.0 to 8.5 (US EPA, 1999) | Not consumed by ammonia-nitrogen (World Fertilizer, 2019) |
Chloride added | Adds chloride to the tower (Combined Cycle Journal) | Depends on the generator route |
Byproducts | Chlorinated organics; chloroform 16.3 ppb in one trial (Nalco Water, 2016) | Chlorite and chlorate (US EPA, 1999) |
Feed equipment and acid supply | Tank and metering pumps | Generator plus acid and precursor storage |
Exposure limit | Follow the bleach SDS | OSHA PEL 0.1 ppm as an 8-hour TWA (29 CFR 1910.1000, Table Z-1) |
Reverting | Restart existing feed | Requires the bleach system to be kept |
The row that matters most is ammonia. In a tower contaminated with ammonia, bleach is used up forming chloramines while chlorine dioxide is not, and that is the strongest reason to switch.
Does chlorine dioxide react with ammonia in cooling water?
Bleach reacts rapidly with ammonia, while chlorine dioxide is reported not to be consumed by it in cooling water. Between pH 7.0 and 8.5, hypochlorous acid "reacts rapidly with ammonia to form inorganic chloramines" (US EPA, 1999). So ammonia from process leaks, makeup water or grey water uses up bleach before it reaches the microbes.
A Nalco Water article on ammonia plant cooling systems states that chlorine dioxide "is not consumed by organic contaminants such as ammonia-nitrogen" and that its required dose stays fairly constant over a wide range of cooling water conditions (World Fertilizer, 2019). The claim comes from a chlorine dioxide supplier, so treat it as a reason to trial, not proof for your tower.
Chlorine dioxide still has a demand of its own: other substances in the water consume part of each dose. The EPA manual notes that it reacts with many organic and inorganic compounds when added to water (US EPA, 1999). A tower with heavy organic loading, oil leaks or sulfide will still consume chlorine dioxide, so measure the dose in an oxidant demand test instead of assuming it.
Chlorine dioxide also stays in water as a dissolved gas that is extremely volatile and easily removed by aeration (US EPA, 1999), so measure the residual at the tower return as well as at the feed point.
Before a switch, log ammonia, organic load and bleach demand over several weeks. If bleach demand climbs with the ammonia readings, chlorine dioxide is likely to help. If demand follows organics instead, the gain is less certain.
Which chlorine dioxide generator type adds chloride to a cooling tower?
Generators that react sodium chlorite with chlorine or hydrochloric acid add chloride to the tower. Systems based on sodium chlorate, hydrogen peroxide and sulfuric acid add sulfate instead. The EPA manual gives the reactions behind this. Chlorine gas with sodium chlorite yields 2 moles of sodium chloride per 2 moles of chlorine dioxide, and direct acidification with hydrochloric acid yields 5 moles of sodium chloride per 4 moles of chlorine dioxide (US EPA, 1999). The chlorate-based route uses concentrated hydrogen peroxide and sulfuric acid (US EPA, 1999).
The route matters most at plants that switch to cut chloride. The High Desert plant replaced bleach partly because excessive chloride from bleach affected its zero liquid discharge cycles, and it chose a system that makes chlorine dioxide from sulfuric acid and a proprietary precursor (Combined Cycle Journal). A generator running on chlorite and hydrochloric acid would have kept adding chloride. At a Texas plant, the Nalco trial reported lower chloride and sulfate in the tower through lower bleach and sulfuric acid use (Nalco Water, 2016). Before you sign, ask the generator supplier for the byproduct mass per kilogram of chlorine dioxide and add it to your tower's chloride and sulfate balance.
Is chlorine dioxide harder on copper than bleach?
In the one published side-by-side trial cited here, chlorine dioxide was gentler on copper than bleach. Nalco Water reported a "dramatic reduction in copper corrosion rates" during its chlorine dioxide trial compared with bleach fed before and after, with mild steel corrosion rates unaffected (Nalco Water, 2016). The plant itself had no copper metallurgy in its system, so the copper result came from corrosion rate measurements, not from equipment life. The trial is also a maker's case study, so it does not settle the question for every tower.
The High Desert account also reports reduced corrosion potential in the condenser after the switch (Combined Cycle Journal). Neither source offers an independent, controlled comparison on admiralty brass or other copper alloys under your water chemistry, inhibitor program and residual.
If your tower has copper alloy condensers or heat exchangers, the safe course is to measure rather than assume. Install copper and mild steel coupons (small test pieces of each metal) on a side-stream rack and determine corrosion rate by weight loss, as ASTM D2688-23 describes for water systems without heat transfer (ASTM D2688-23, scope). Run coupons through the bleach baseline and again through the chlorine dioxide trial, at the residual you plan to hold. Compare pitting as well as average rate.
When should you keep bleach instead of switching to chlorine dioxide?
Keep bleach when the tower runs near or below pH 7.5, ammonia is low, the bleach feed is reliable and chloride is not limiting cycles of concentration. Under those conditions bleach keeps most of its strength (US EPA, 1999), and a switch would add a generator, a second chemical supply and new monitoring without fixing a real problem.
Both case studies in this comparison started from bleach systems that were already failing: High Desert had pumps losing prime, cracked tubing and probe failures (Combined Cycle Journal), and the Texas plant struggled to keep bleach feed working at three locations (Nalco Water, 2016).
Bleach is also the right answer when the site cannot supply what a generator needs. The High Desert turnkey package did not include the sulfuric acid tank the system required (Combined Cycle Journal), and the Nalco trial needed 78% sulfuric acid for full generation efficiency (Nalco Water, 2016). Every acid and precursor has to be stored and handled as its SDS directs, and workplace exposure to chlorine dioxide must stay under the OSHA limit of 0.1 ppm as an 8-hour TWA (29 CFR 1910.1000, Table Z-1).
Lubinpla can lay out keeping bleach and each chlorine dioxide route as separate scenarios, each with its early warning signals, and asks you for any premise your records do not hold, such as acid storage space.
How long should a chlorine dioxide field trial run before switching?
A chlorine dioxide field trial should run long enough to cover the tower's worst season. In the published cases that meant 90 days to six months. High Desert ran a six-month pilot in 2017 before committing (Combined Cycle Journal). The Nalco trial at the Texas plant ran 90 days starting on 7 August 2014, beginning in the summer months when manganese in its lake water spiked (Nalco Water, 2016). Byproducts in that trial were measured after two weeks of treatment at 0.2-0.4 ppm chlorine dioxide residual (Nalco Water, 2016).
Write the success criteria down before the trial starts: microbial counts and biofilm, residual stability, chloride and sulfate in blowdown, copper and steel coupon rates, byproducts at the outfall, and operator hours on the feed system. Criteria picked after the results arrive tend to confirm whatever the plant hoped for. The same discipline applies to any product change, as covered in the risks of formulation changes and how to decide.
Keep the bleach system intact and ready for the whole trial. If chlorine dioxide fails a criterion, reverting then means restarting a system that still works, not rebuilding one.
What drives bleach and chlorine dioxide cost over the service life?
Over the service life, what either biocide costs comes from chemical deliveries, equipment ownership, monitoring, downstream effects and labor, not from the chemical price alone. Bleach degrades in storage: a 12.5% solution can fall to 10% in 30 days under best-case conditions (US EPA, 1999), so dose drift and storage time cost money. The Texas plant projected bleach at about 20 trucks a year against about 3 trucks of chlorine dioxide precursor (Nalco Water, 2016).
Chlorine dioxide has costs of its own: generator ownership or rental, maintenance contracts and testing. The EPA manual lists high costs for training, sampling and laboratory testing for chlorite and chlorate, with equipment typically rented, in a drinking water context (US EPA, 1999). High Desert contracted maintenance to the supplier (Combined Cycle Journal).
Downstream effects can outweigh both. High Desert reported RO and microfilter membrane cleanings halved and a lower total operating cost than bleach (Combined Cycle Journal). Count these effects for your own plant: a tower without membranes or a zero liquid discharge system will not see that saving.
Frequently asked questions
Can bleach and chlorine dioxide be fed to the same cooling tower?
Yes, with planned feed points and timing: in drinking water systems, chlorate forms when residual chlorite meets free chlorine (US EPA, 1999), and the same reaction can occur in a tower. Agree the feed points and timing with the generator supplier, and monitor chlorate in blowdown if your discharge permit limits it.
Does chlorine dioxide leave a residual you can measure in the tower?
Chlorine dioxide provides a residual (US EPA, 1999), and the Nalco trial held 0.2-0.4 ppm (Nalco Water, 2016). Agree the residual test method with the generator supplier before the trial so that baseline and trial readings are comparable.
Is chlorine dioxide a disinfection byproduct risk in the blowdown?
Chlorine dioxide forms chlorite and chlorate as endproducts (US EPA, 1999), although it does not form trihalomethanes (NHMRC, Australian Drinking Water Guidelines). Before the trial, check whether your discharge permit or receiving water authority sets limits for chlorite or chlorate, and include them in the trial sampling plan.