BlogApplication notes
Why are my cooling water pipes corroding?
- Type
- Application notes
- Published
- 20 November 2026
- Prepared by
- Technical team
- Product group
- Heat Transfer & Cooling
Adding more corrosion inhibitor is the usual response to rusting pipe, but the right fix depends on whether the attack is even or confined to a few deep pits. What actually corrodes the pipe can be even oxygen attack, under-deposit pitting, microbial corrosion, aggressive water or dissimilar metals, and some of these, such as deposits and biofilm, feed one another.
On some sites water starts to bead round the flanges, sample bottles come back cloudy red, and each shutdown adds more plugs to the exchanger tubes. In the same system, one pipe thins evenly all over while another looks sound outside yet perforates at a single point. The rust may look alike, but what is happening underneath can differ from pipe to pipe.
As corrosion advances, pipes and exchanger tubes perforate, causing leaks and unplanned stops, and the dissolved iron settles out again as deposits that cut heat transfer. Those deposits then become the sites for the next round of localised corrosion (Buecker, 2025a). Hygiene is part of the same problem. The US Centers for Disease Control and Prevention names sediment and biofilm as key factors in Legionella growth in cooling towers, and treats scale, corrosion and sediment control plus system cleaning as critical both to tower operation and to preventing Legionnaires' disease (CDC, 2025). The US Occupational Safety and Health Administration likewise notes that cooling towers and evaporative condensers can create ideal conditions for microbial growth (OSHA, n.d.). Corrosion, deposits and microbes are not separate problems; each feeds the others.
The basic principle is that of a battery. At one part of the metal surface, the anode, iron dissolves as ions and gives up electrons; at another, the cathode, dissolved oxygen takes those electrons and becomes hydroxide. In neutral and mildly alkaline cooling water this oxygen reduction drives the corrosion reaction, and protecting carbon steel from oxygen attack in an open recirculating system is regarded as the most basic aim of a treatment programme (Buecker, 2025a). How much oxygen arrives, and where the anodes and cathodes form and how large they are, set both the rate and the shape of the attack, so whether a pipe thins evenly or pits deeply in a few places narrows the candidates a great deal.
In such cases the technical team sets up the five hypotheses below and rules them out one by one. The first check never changes: hang corrosion coupons, small pre-weighed metal strips, in a side-stream rack, a test loop carrying part of the main flow, to put numbers on the corrosion rate and pitting, and on the same day record pH, conductivity, oxidant residual, the chlorine-type biocide left in the water, and turbidity against the targets of the treatment programme.
Hypothesis 1: even oxygen attack
In a system showing red water and broad, even rust together, oxygen is the first candidate. It shows up where inhibitor is added by hand and its level swings, where some days pass with the chemical tank empty, in closed loops whose make-up has risen, and in hot zones such as exchanger outlets. Take a carbon steel coupon out of the side-stream rack after its set exposure and the whole face is evenly red, while the feed log for the same period includes days when the tank ran dry. Open the pipe and the rust lies in a broad layer with no mounds.
While a tower cools water by spraying it through air, the circulating water takes up oxygen close to saturation, and oxygen in the make-up water adds to it. As that oxygen takes electrons at the steel surface, iron dissolves as ions and turns through iron hydroxide into red rust. In neutral and alkaline cooling water this reaction is the main driving force of corrosion (Buecker, 2025a), and inhibitor slows it by forming a thin film on the surface. Whenever that film lapses, oxygen can attack a wide area again.
Temperature changes the rate too. In a system closed to the atmosphere the corrosion rate of iron rises almost linearly with temperature from about 40 to 160°C, while in an open system it rises up to about 80°C and then falls (DoD, 2003). Higher temperature speeds oxygen diffusion, but in an open system less oxygen stays dissolved, and that effect eventually wins. A closed loop should hold little oxygen, yet oxygen can keep entering through saturated make-up, the air-water surface in an expansion tank and leaking pump seals, so a closed loop that takes a lot of make-up can corrode much like an open one.
Start by weighing the coupons. ASTM D2688 hangs metal coupons in a side-stream rack, takes the average corrosion rate from weight loss and checks for pitting, and it covers open recirculating cooling water and closed chilled water systems (ASTM International, 2023). Repeat it at the same spot over the same period so results compare. A high average rate with no pits, over a period when inhibitor strayed out of range, makes this hypothesis the strong one. In a closed loop, also check whether the make-up meter reading has risen; more make-up is read as more oxygen coming in.
Inhibitor level is the first thing to put right. If the coupon's average rate is above the programme target, bring the inhibitor level back first, and automate the feed if readings swing. CDC recommends automating the addition and monitoring of corrosion and scale inhibitors, and checking their levels as often as the water data call for (CDC, 2025). If make-up has risen in a closed loop, find the leak or the pump seal before adding more chemical. Target corrosion rates and inhibitor levels follow the manufacturer's recommended range for the programme in use.
Hypothesis 2: under-deposit pitting
When the coupon's average rate is low but the pipe still pits deeply in places, it is time to look under the deposits. This starts where solids settle: dead legs, closed-off branches where water sits still, slow horizontal runs, the bottoms of exchangers, and systems that run with high turbidity. Cut a slow horizontal run during a shutdown and rust mounds sit in a row along the bottom of the bore; scrape one off and a deep pit lies beneath. The bare metal between the mounds is in fairly good shape.
A spot oxygen cannot reach becomes an anode. When scale, sludge, silt or rust mounds settle on the pipe wall, oxygen has trouble getting underneath, and the oxygen-poor area under the deposit becomes anodic to the oxygen-rich metal around it, so corrosion concentrates there. This is called a differential aeration cell, and even ordinary deposits cause trouble because they set one up (Buecker, 2025a). A large cathode surrounds a small anode, so one point can pit deeply even while the average rate stays low.
Iron released at an active site piles up over it into a blister-like rust mound called a tubercle, and chloride and sulfate gather underneath to push corrosion further (Buecker, 2025a). Chloride is known to promote pitting and crevice corrosion (NIH, 2025). As the tubercle grows, the space beneath becomes a small, more acidic, oxygen-free chamber where corrosion goes on even when the bulk water looks fine. The same structure keeps inhibitor from getting through the tubercle to the bottom of the pit.
The surest evidence is the tubercle and the pit beneath it. Rows of tubercles when the pipe is opened, with a deep pit under each one removed, mean under-deposit corrosion. When the pipe cannot be opened, narrow it down with coupons and turbidity. Coupons that pit despite a low average rate, or turbidity that stays high, point to deposits (ASTM International, 2023); mark the leak points on the system drawing and see whether they cluster in low-flow runs.
Cutting solids and ending stagnation come first. CDC recommends designing piping without stagnant runs or dead legs, flushing low-flow runs and dead legs at least weekly, and considering filtration to reduce suspended solids, with towers taken out of service for cleaning and disinfection at least once a year (CDC, 2025). Once tubercles have grown, inhibitor alone cannot reach beneath them, so remove them by mechanical or chemical cleaning, re-establish the inhibitor level and check with coupons that the pitting has stopped.
Hypothesis 3: microbial corrosion
A rotten-egg smell together with black deposits makes microbes the likely suspect. It tends to appear in condensers and pipes left full of water during an outage, in systems where the biocide residual swings, and in stagnant zones such as dead legs. Restart a condenser that has stood for a while, and several tubes begin to weep; pull one and the area under the tubercles is stained black, and a foul smell rises the moment the pipe is opened. A condenser left with cooling water in its tubes during an outage has indeed been reported to develop thousands of pinhole leaks (Buecker, 2025a).
Microbiologically influenced corrosion, or MIC, arises from specific interactions between metal, microbes and water, so any mechanism has to consider all three (Little & Lee, 2014). Bacteria attach to the pipe wall in a biofilm held together by their own slime, and its inner layer turns oxygen-free. Among the anaerobes in that slime, sulfate-reducing bacteria produce hydrogen sulfide, which is highly corrosive to many metals (Buecker, 2025a).
Patchy biofilm is itself a differential aeration cell: oxygen is depleted beneath it, making that area the anode, while the uncovered surroundings act as the cathode. Tiny manganese sulfide particles in carbon steel can become local anodes and pit when sulfate-reducing bacteria are present, and the slime bacteria release has been reported to start corrosion on its own (Knisz et al., 2023). Iron-related bacteria build tubercles that in turn shelter more microbes (NIH, 2025). The once-popular cathodic depolarisation theory is now rejected by many researchers, so no single mechanism should be assumed (Knisz et al., 2023).
Smell and colour are clues, not a verdict. A rotten-egg smell on opening and black material under tubercles suggest sulfide; mark separately the sections that held standing water and the pinholes that appeared after an outage. Diagnosing MIC after the fact needs microbiological, metallurgical and chemical analyses together (Little et al., 2006), and finding bacteria alone does not prove the cause. Take samples as a set: the deposit under a tubercle, the metal at that spot and the water from the same time.
Control centres on residual, stagnation and deposits. Keep the biocide residual steady, cut stagnant zones and remove deposits. Sediment and biofilm use up disinfectant and harbour growth (CDC, 2025), so when biofilm is thick, clean first rather than simply raising the biocide dose. OSHA notes data suggesting that quaternary ammonium compounds, widely used against cooling tower biofouling, may not fully control Legionella (OSHA, n.d.), so confirm that a biocide works against the target organisms. When a system stands for maintenance, either drain it or keep it circulating and treated.
Hypothesis 4: aggressive water
When corrosion speeds up from a certain point, and the water record or chemical feed changed around then, look at the water itself. Typical cases are systems fed acid to control scale, operation where the blowdown setting has slipped and concentration has climbed, and systems just switched to softened or demineralised make-up. In a system that has only recently started acid feed, a short run just downstream of the injection nozzle can thin markedly while the rest of the pipe stays sound. Over the same period the continuous pH trace shows a brief dip each time the acid pump runs.
The thin film that protects carbon steel depends on pH. In neutral and mildly alkaline water, carbon steel slows corrosion by forming thin oxide and inhibitor films; as pH falls those films dissolve, and hydrogen ions start taking electrons directly in addition to oxygen. Below pH 4 the corrosion rate of mild steel rises, and at pH 3 it climbs sharply (DoD, 2003). The bulk cooling water rarely gets that low, but acid fed to control scale can reach that range locally when it is overfed or poorly mixed near the injection point.
Evaporation concentrates salts as well. How many times stronger the circulating water is than the make-up is called the cycles of concentration, and it is controlled by blowdown, draining off part of the water. The higher the cycles, the more chloride and sulfate, ions that promote pitting and crevice corrosion and gather under tubercles (Buecker, 2025a; NIH, 2025). At the other end, softened or demineralised water with almost no hardness cannot form a protective calcium carbonate film and can act corrosively. Whether water tends to dissolve or deposit calcium carbonate is estimated with a saturation index such as the Langelier index.
Read pH as a trend rather than a single value. Check the continuous record for dips each time the acid pump runs, and if pH is read by hand, read it downstream of the injection point too. Confirm the cycles from the ratio of conductivity or chloride between circulating and make-up water, and check the blowdown setting is actually being held. Severe attack confined to a few metres below the acid nozzle points to local low pH; coupon rates that rose across the system from the day the make-up changed put the change in water first.
Re-establish the pH range first. Hold pH in a non-corrosive range set by the disinfectant in use and the manufacturer's recommendations (CDC, 2025). Tie acid feed to a pH signal, and move the injection point to where flow is fast enough to mix quickly. When chloride climbs high, raising blowdown to lower the cycles is a more basic fix than adding inhibitor. If corrosion rose right after the make-up was switched to softened water, redesign the treatment programme for that water.
Hypothesis 5: dissimilar metals
When corrosion gathers at steel ends joined to brass valves or copper-tube exchangers, the pairing of metals is the next suspect. Steel pipe with copper-alloy parts fitted straight onto it, systems with copper-tube exchangers, and plants where process leaks can bring in ammonia all fit this pattern. Lay out the steel pipe removed from either side of a copper-tube exchanger at replacement, and the end nearest the joint is far more pitted, while copper-coloured metal glints at the bottom of pits even in steel well away from the joint.
Two metals with different potentials, electrically joined in the same water, form a cell. The less noble one becomes the anode and corrodes, and the larger their potential difference, the faster the attack (Buecker, 2025b; DoD, 2005). In cooling systems the most common case is mild steel touching a copper alloy, such as a brass valve fitted straight onto steel pipe. A small anode joined to a large cathode concentrates the current on a small area and perforates the steel quickly; steel rivets holding copper-alloy plates are one example (Buecker, 2025b).
The same happens without direct contact. Copper dissolved from copper alloys plates back onto mild steel as metal, and each deposit sets up a new small cell (DoD, 2005). Copper alloys can be oxidised by dissolved oxygen in mildly alkaline cooling water, and where oxygen and ammonia occur together, ammonia dissolves the copper's oxide layer and makes the environment particularly aggressive (Buecker, 2025b). That is why copper release and steel pitting can rise together in systems where process leaks bring in ammonia.
Hanging a copper coupon beside the carbon steel one puts numbers on this hypothesis (ASTM International, 2023); measure copper and ammonia in the circulating water alongside. Heavy corrosion at steel ends joined to brass valves or copper-tube exchangers means contact galvanic corrosion; copper-coloured metal at the bottom of pits in steel away from any joint suggests copper has plated out. Set the time the copper coupon's rate rose beside the time steel pitting increased.
A dielectric union that breaks metal-to-metal contact is the common fix, and where the materials cannot change, arrange them so that the larger area is the anode (Buecker, 2025b). In systems with copper alloys, include a copper corrosion inhibitor in the programme so copper does not dissolve in the first place, and if ammonia appears, look for the process leak first. When the copper coupon's rate rises, act before steel pitting follows.
Testing the hypotheses
The biggest clue for separating the five hypotheses is the shape of the corrosion. Even thinning across the pipe points to oxygen with too little inhibitor, hypothesis 1, or changed water, hypothesis 4. Deep pits in a few places narrow it to deposits, hypothesis 2, microbes, hypothesis 3, or dissimilar metals, hypothesis 5. Coupons show the shape too, so hang carbon steel and copper coupons together in the side-stream rack and record the average rate and pitting side by side (ASTM International, 2023).
Next come time and place. Overlay the trends for pH, conductivity, chloride, oxidant residual, inhibitor level and turbidity on the time corrosion increased, and it shows whether that was a spell with an empty chemical tank, a change of make-up or the start of acid feed. Place is read the same way: dead legs and slow runs mean deposits and microbes, downstream of acid feed means pH, and joints between metals mean galvanic attack (Buecker, 2025a, 2025b; DoD, 2003).
When two remain, a control test separates them. Hang carbon steel and copper coupons side by side in the same rack, or hang coupons for the same period near and far from the acid feed point, or in a fast main and a slow branch. If rate or pitting differs sharply on one side only, that side's condition is the candidate. At shutdown, open pipes or exchangers to look for tubercles, black sulfide and plated copper, and photograph them as the baseline for the next comparison.
MIC is confirmed last, from combined evidence. If it is suspected, commission microbiological, metallurgical and chemical analyses together, and do not conclude on a bacteria count alone (Little et al., 2006). Whatever the hypothesis, judge the job done only when coupons at the same spot over the same period show rate and pitting back within target (ASTM International, 2023).
When hypotheses overlap, set the order for untangling them. On site, deposits, biofilm and tubercles can appear together, each feeding the others (Buecker, 2025a; NIH, 2025). Clean out deposits and biofilm first, then re-establish inhibitor and biocide levels, and finally confirm the effect with coupons. Changing only the chemicals while tubercles remain makes it hard to tell which step worked.
Limits and cautions
Target corrosion rates, pH ranges and inhibitor levels vary with the treatment programme and equipment design, so the numbers follow the manufacturer's recommended range for the programme in use. Narrowing the cause is not the same as judging remaining wall thickness or when to replace pipe. If leaks have already appeared, or thinning is suspected in pressure piping or exchanger tubes, hand it to specialists who measure thickness and run non-destructive tests.
Acids, biocides and corrosion inhibitors can be hazardous to handle. Check the safety data sheet, the SDS, for each chemical and wear protection such as goggles and gloves. Work near the tower and cleaning can expose people to Legionella through aerosols, so tower care belongs inside a Legionella management plan (ASHRAE, 2020; CDC, 2025). ASHRAE Guideline 12 gives detailed guidance for building such a water management plan (ASHRAE, 2020).
Sources
- ASHRAE. (2020). Guideline 12-2020, Managing the risk of legionellosis associated with building water systems. https://www.ashrae.org/about/news/2020/ashrae-publishes-updated-legionella-guideline
- ASTM International. (2023). ASTM D2688-23, Standard test method for determination of corrosion rate in a water system in the absence of heat transfer (weight loss method). https://store.astm.org/d2688-23.html
- Buecker, B. (2025a, June 25). Cooling towers: Corrosion mechanisms to critical failures, part 3. Chemical Processing. https://www.chemicalprocessing.com/asset-management/reliability-maintenance/article/55298865/cooling-towers-corrosion-mechanisms-to-critical-failures-part-3
- Buecker, B. (2025b, October 31). Cooling towers: Managing galvanic corrosion and white rust formation, part 4. Chemical Processing. https://www.chemicalprocessing.com/asset-management/reliability-maintenance/article/55325456/cooling-towers-managing-galvanic-corrosion-and-white-rust-formation
- Centers for Disease Control and Prevention [CDC]. (2025). Controlling Legionella in cooling towers. Legionella Control Toolkit. https://www.cdc.gov/control-legionella/php/toolkit/cooling-towers-module.html
- Knisz, J., Eckert, R., Gieg, L. M., Koerdt, A., Lee, J. S., Silva, E. R., Skovhus, T. L., An Stepec, B. A., & Wade, S. A. (2023). Microbiologically influenced corrosion: More than just microorganisms. FEMS Microbiology Reviews, 47(5), fuad041. https://doi.org/10.1093/femsre/fuad041
- Little, B. J., & Lee, J. S. (2014). Microbiologically influenced corrosion: An update. International Materials Reviews, 59(7), 384–393. https://doi.org/10.1179/1743280414Y.0000000035
- Little, B. J., Lee, J. S., & Ray, R. I. (2006). Diagnosing microbiologically influenced corrosion: A state-of-the-art review. Corrosion. NACE International. https://apps.dtic.mil/sti/pdfs/ADA470200.pdf
- National Institutes of Health [NIH], Office of Research Facilities. (2025, May). Corrosion in closed-loop water systems, part 1: Understanding corrosion. Technical Bulletin. https://orf.od.nih.gov/TechnicalResources/Documents/Technical%20Bulletins/25TB/Corrosion%20in%20Closed-Loop%20Water%20Systems%20Part%201%20-%20Understanding%20Corrosion%20-%20May%202025%20TB%20_508.pdf
- Occupational Safety and Health Administration [OSHA]. (n.d.). Legionellosis, Legionnaires' disease and Pontiac fever: Control and prevention. U.S. Department of Labor. https://www.osha.gov/legionnaires-disease/control-prevention
- U.S. Department of Defense [DoD]. (2003). Unified Facilities Criteria UFC 3-570-06. https://ufgscriteria.tpub.com/UFC3-570-06/UFC3-570-060035.htm
- U.S. Department of Defense [DoD]. (2005). Unified Facilities Criteria UFC 3-240-13FN, 4-5.1 Galvanic corrosion. https://ufgscriteria.tpub.com/3_240_13/3_240_130143.htm