What a Corrosion Inhibitor Actually Does

Why this matters

The single most expensive belief in water treatment is that an inhibitor is a protective additive, so more of it is better and some of it is better than none. For one whole class of inhibitor the second half of that sentence is false in a way that puts holes in pipe: below its own threshold concentration an anodic passivating inhibitor concentrates attack into fewer, deeper sites than an untreated surface would have suffered. A loop that reads a little low is not a little protected. It may be in the worst condition it has ever been in.

Getting this right is the difference between a reserve reading you act on immediately and one you write down and revisit next quarter.

Before you handle any of these products

Read Section 8 of the product's safety data sheet and use the glove class and eye protection it names, per product. These packages span corrosive, oxidising and skin sensitising classes and the glove that resists one may be degraded by another; guessing a glove from the last drum you opened is how a chemical burn happens through intact PPE. The employer's obligation to make that data sheet available to the tech using the product sits in 29 CFR 1910.1200.

Never bring an acid product and a hypochlorite product together, in a system, in a pot feeder, or in a bucket, in either order. The reaction releases chlorine gas within seconds. That is an inhalation route, so the control is not gloves: it is keeping the two apart, flushing thoroughly between chemical steps, and if a respirator is ever part of your answer, it is only under a written program with fit testing per 29 CFR 1910.134.

The jobs an inhibitor cannot do

Most disappointment with treatment comes from assigning it work that belongs elsewhere. An inhibitor does not:

• Remove scale that already formed. A film forming chemical needs to reach bare metal. Under an existing deposit it reaches nothing, which is why a scaled system must be cleaned before it is inhibited, not instead of.
• Stop erosion corrosion from excessive velocity. Where flow is scouring the film off faster than the chemistry rebuilds it, more chemical does not win the race. That is a hydraulic problem with a hydraulic fix.
• Undo a galvanic couple. Two dissimilar metals in metallic contact in conductive water still drive current. Inhibitor slows the reaction at both electrodes; it does not remove the driving voltage. Isolation does.
• Kill anything. Several common inhibitors are nutrients. Nitrite is a nitrogen source that nitrifying organisms will oxidise to nitrate, and phosphate feeds a wide range of them. Biological control is a separate chemistry with separate testing.
• Keep oxygen out. An inhibitor manages what oxygen does when it arrives. An oxygen scavenger consumes it. A tight system stops it entering. Those are three different jobs and only the third is permanent.

That last one is worth sitting with. If a loop is consuming inhibitor faster than the metal surface can account for, the answer is usually upstream of the chemistry entirely.

What it actually does: three mechanisms, three failure behaviours

Anodic passivating inhibitors (nitrite and molybdate are the ones you will meet; chromate is the historical one and is not used now for good reason) work by repairing and maintaining the oxide film on the anodic sites, the places where metal wants to dissolve. They are extremely effective at or above their threshold concentration and they are the class with the dangerous underdose behaviour, because a partially covered anode leaves a small bare area facing a large cathodic area. The same total corrosion current now has far less metal to come out of, so it comes out fast and locally. Suppliers publish a minimum for this reason, and that minimum is typically well above the level at which you can still get a positive result on a test strip. Detecting nitrite is not the same as having enough of it.

Cathodic inhibitors (zinc salts, polyphosphate, and molybdate in part of its behaviour) suppress the reduction reaction instead, usually by depositing a barrier where the cathodic reaction happens. Their failure behaviour is far gentler: protection fades roughly in proportion as concentration falls, without the pitting mechanism. A loop drifting low on a cathodic package is genuinely a little protected.

Adsorption and film forming inhibitors are metal specific by design. The azoles (benzotriazole and tolyltriazole are the common names) form a molecular layer on copper alloys and do very little for steel. They matter far more than their small dose suggests, because copper that dissolves does not stay dissolved: it plates out on steel and aluminum downstream and creates a fresh galvanic couple at every landing site. A steel corrosion problem in a mixed metal loop is sometimes a copper protection failure that moved.

Metal specificity is the part that gets skipped

A package protects the metals it was formulated for. The mismatch that costs the most is aluminum, because aluminum's stable band is narrow, roughly pH 7 to 8.5, and it corrodes at both ends of it rather than only in acid. Steel loops are conventionally run alkaline, often pH 9 to 10, precisely because steel is happier there. Put a cast aluminum heat exchanger into a system running the steel target and the chemistry that is protecting the pipe is dissolving the appliance. The pH target for that loop comes from the heat exchanger manufacturer, and it constrains everything else you can do.

The reverse error also exists. A package chosen for an aluminum appliance and held at pH 8 leaves steel in a much less favourable band than it would be at 9.5, so a mixed loop is always a compromise struck deliberately, not a target inherited from habit.

Worked example: a reserve reading that is not a little low

A closed heating loop measured at 400 gallons. The fluid supplier's data sheet, and use your own product's figures rather than these, calls for a target of 1,200 ppm and a minimum of 700 ppm expressed as nitrite ion. The quarterly sample reads 300 ppm.

The first instinct, that this is roughly a quarter of target so add roughly a quarter of a charge, is wrong twice. It is wrong on the arithmetic, because 300 against 1,200 is a deficit of 900 ppm, three quarters of the charge and not a quarter. And it is wrong on the class, because at 300 ppm the loop is at 43% of the published minimum and has been sitting inside the band where an anodic package does harm.

Compute the dose from volume and deficit rather than from habit. Parts per million on a water solution is milligrams per litre, and 400 gallons is about 1,514 litres. A 900 mg/L deficit across 1,514 litres is roughly 1.36 kg of nitrite ion. Convert that to product using the active fraction on the data sheet: a package that is 10% active nitrite by weight needs about 13.6 kg of product. Cross check that against the supplier's own dosing table before you add anything, because active fraction varies widely between products and a table that disagrees with your arithmetic means one of you has the wrong number.

Add the product through a pot feeder or a designed injection point with the eye and hand protection from Section 8 of that product's data sheet, never by opening a hot loop. Adding chemical to a system in service is also a hazard to people who are not standing where you are: the loop feeds a building, and if any part of that path can reach a potable fixture, the connection protecting it is a backflow assembly required by the local plumbing code, not an optional fitting.

Then watch the right number. Nitrite is now the number you restored, not the number that tells you whether it worked. Dissolved iron and turbidity in the following samples are what tell you the film re-formed: iron falling across the next two samples means the surface has re-passivated. Iron holding or rising with nitrite now at target means the inhibitor is not reaching the metal, which points at a deposit shielding the surface or at organisms consuming the nitrite. Those two forks separate on the nitrate result: nitrite falling with nitrate appearing and no makeup water to explain dilution is a biological finding, and no amount of additional nitrite fixes it.

What would change the recommendation

• A cathodic or all organic package instead of a nitrite one. The urgency argument above is specific to the anodic passivating class. On a package whose protection fades in proportion, a reading at 43% of minimum is a schedule problem, not a same day problem, and you top up at the next visit.
• A loop that has just been opened for repair. Fresh metal surface consumes inhibitor rapidly for the first weeks after any work that exposed bare steel. A low reading three weeks after a pipe repair is expected and is not evidence of a leak.
• Aluminum anywhere in the wetted path. The dose calculation above does not change, but the pH you are allowed to sit at does, and some high alkalinity nitrite borate packages are not permitted at all. That is a data sheet question before it is a dosing question.

How to verify you got this right

Take the confirming sample after the loop has circulated long enough to be homogeneous, not immediately after dosing at the feeder. A sample drawn from the feeder's own leg minutes after a charge reads the charge, not the loop.

Then check that three things agree with each other on the following visit: the reserve is at or above target, the makeup meter has not moved much, and dissolved iron has fallen. Any one of those alone can mislead. Reserve at target with iron still climbing means the chemical is in the water but not on the metal. Reserve at target with the makeup meter running means you are topping up a loop that is diluting itself, and you will be back. The set of three is the reading; a single reserve number is not.

References

• 29 CFR 1910.1200, Hazard Communication, for the employer's duty to provide the safety data sheet whose Section 8 sets the glove class and eye protection for each product
• 29 CFR 1910.134, Respiratory Protection, for the written program and fit testing required before a respirator is any part of a chemical control
• Fluid and inhibitor supplier documentation for target concentration, published minimum, active fraction and test method for the specific package
• Boiler or heat exchanger manufacturer documentation for the permitted pH range where aluminum is in the wetted path
• See related: Galvanic Corrosion and the Metals That Fight; The Water Chemistry That Attacks a System