Formicary corrosion: the coil leak that starts indoors

Formicary corrosion attacks copper from the outer face of the tube and works inward. Under a microscope the damage shows as a network of fine tunnels running through the wall of the metal. That pattern named the fault. The word comes from the Latin for ant, and the failure literature also calls it ant-nest corrosion.

The attack needs three things at once. Oxygen, moisture and an organic acid all have to reach the metal together. Published failure analysis of copper tubing in aircon units names formic and acetic acid as the usual corrodents. Propionic and butyric acid, along with several other volatile compounds, are implicated too. Take away any one of the three conditions and the process stops.

The share of failures involved is small and steady. One published review of copper tube failures in this industry attributes roughly one in ten early failures to this mechanism. The copper industry's own trade body puts it lower still, describing it as rarer than ordinary pitting and confined to heat exchangers that meet particular chemicals. Both readings agree on the same two points. The fault is genuine, and it is uncommon.

Tube walls are thin because thin metal moves heat well. That leaves almost nothing between a tunnel and the gas inside. These pits do not spread sideways across the surface. They dig, they branch, and eventually one of them breaks through. The hole it makes is a pinhole.

Copper is also not uniformly exposed. The failure literature associates the pattern with thin-walled tube, and with certain copper chemistries used in heat exchanger work. That is a record of where the damage has been observed. It is not a ranking of suppliers, and nothing in it points at one maker over another. Every brand in this market draws on the same short list of tube specifications.

The damage sits inside the wall, so the outside can look untouched. Industry research on indoor coil failures describes these leaks as tiny pinholes at the tube surface that the human eye cannot pick out. A coil in that state still photographs as new.

Other corrosion patterns do announce themselves. Chloride attack leaves larger pits, shaped something like bite marks, and those can often be seen without magnification. Salt damage on an outdoor unit is louder again, with white bloom on the fins and metal that crumbles at the edges. That is the subject of coastal aircon servicing, and it reads nothing like this one.

This is also a separate problem from moisture in the refrigerant circuit, which is contamination sealed inside the pipework at installation. Formicary corrosion works on the outside face of the tube, in the fin pack, where room air passes over it. Copper and aluminium in contact corrode each other as well where water bridges the two metals, and that is a third mechanism with its own signature.

What all of that costs a homeowner is proof. Looking at a coil cannot clear it of this fault, and a coil pronounced clean by eye has not been tested for anything. Locating a leak of this kind is detection work, and the leak detection methods that do it are a separate subject from how a coil looks.

The second cost lands on the conversation instead of the coil. Confirming this mechanism means sectioning the metal and studying the wall under magnification, which is laboratory work on a part that has already come out. Almost no domestic job travels that far. The name therefore stays a description of what fits the evidence, and a technician who states it as a settled finding is claiming more than a site visit can carry.

Ordinary indoor air carries them, and building materials are the largest identified source. Formic and acetic acid are the two named most often. Formaldehyde in the air converts to formic acid, and then to formate once moisture is present. Acetic acid becomes acetate in water. Every one of those forms is aggressive toward copper.

The measurement work behind that claim is public. A Lawrence Berkeley National Laboratory study of new houses measured airborne compounds indoors and found acetic acid among the most abundant, alongside formaldehyde. Plywood flooring, latex paint and sheet vinyl flooring were the major sources it identified. Indoor levels ran well above the outdoor air at the same houses.

A tighter building holds more of it. Homes built to lose less cooled air also exchange less air with outside, so what is emitted indoors sits at higher concentration. The industry research on coil failures treats that construction trend as part of the cause, not a footnote to it.

There is direct evidence that the acid comes from the home and not from the coil. Condensate was sampled at homes where coils had failed, and the water coming off the coil carried significant formate and acetate. Right after the coil was replaced, levels in the fresh condensate dropped sharply. Once the new coil had been running, they climbed back to where they had been. The room refilled them.

What none of that does is name the item in any one home. The published source lists are catalogues of what can emit these acids, assembled from the literature. They run to adhesives, cabinetry, laminates, particle board, plywood, paints, wallpaper, silicone sealant, cleaning solvents and vinegar. Which of them matters at a given address is unsettled, and no one has offered a measurement inside a single flat that settles it.

None of the measurement work above was carried out in Singapore, and that limit is worth stating plainly. What travels is the finding that freshly finished interiors give off the most, because the named sources are the materials that go into a new fit-out. Singapore housing turns its interiors over often, and a flat can change hands with new boards, new adhesive and new paint throughout. That makes the general finding relevant here. It does not make it measured here.

None of this is a reason to strip a room of its paint or its furniture. The exposure described here is ordinary, the failures are uncommon, and no published list ranks these items by how much they contribute. Reading a leak backwards to one product in the room goes well past what anybody has shown.

The indoor coil is the one place in the system where all three conditions hold at once. It runs colder than the room, so water condenses on it whenever the unit cools. That water then sits on copper in open air. Oxygen, moisture and metal are together there continuously, by design.

It is also where room air meets water. Every hour the unit runs, air passes over the fins in bulk, and anything in it that dissolves collects in the film on the metal. Weak acid concentrates on the coil rather than passing through with the air. That concentration is what the condensate sampling was picking up.

The leaks turn up in the fin pack. Industry research places these failures in the length of tube buried inside the aluminium fins. That is also the least reachable part of the coil, so a leak there is costly in labour before it is costly in parts.

Water lingers longest exactly where the tube sits. The coil fins pack tightly around it to move heat, and the narrow gaps between them hold moisture after the unit stops. Anything dissolved in that water keeps contact with the copper through every off cycle. The face of the coil dries first. The metal buried behind it dries last.

The pipe run between the two units escapes for a plain reason. It is wrapped, so room air never touches the copper, and no condensate forms along it while the lagging stays intact. Damage to that pipework starts somewhere else, usually with water getting in under insulation that has split. A coil carries no such wrapping, because its entire job is to hold bare metal open to the air.

The outdoor coil lives in a different chemistry. Its load is salt, exhaust and grit, its damage lands first on the aluminium fins, and it is visible from the front. The indoor coil looks like the sheltered one, and on this particular measure it is the more chemically active of the two.

The presentation is gas loss on a system too young to expect it. Cooling fades, the gas is topped up, and it goes again. Nothing outside the flat explains it, because nothing outside the flat is involved. The signs of refrigerant leak that reach a homeowner look the same here as anywhere else.

The first explanation offered is usually the installer. Flare joints are the standard suspect for early gas loss, and they earn that place, because they are the most common source and the quickest to test. Location is what separates the two cases. A joint leaks at a joint. This leaks in the middle of a tube run, deep in the fins, where nobody ever made a connection.

The second explanation is that the machine was bad when it arrived. That one is harder to sustain. The industry research report examined failed indoor coils from a dozen different suppliers and found the same corrosion signature in every one of them. A pattern that crosses an entire market is describing the environment those coils sit in.

The position that survives both readings is a narrower one. The tube met an environment it had no protection against, and the outcome looks identical whoever built the coil and whoever fitted it. That is unsatisfying as a story and useful as a diagnosis, because it changes what gets checked next instead of who gets blamed.

A second opinion earns its name by settling three things before it names any cause. Where along the circuit the loss is happening. Whether the joints hold when they are properly pressurised and watched. Whether the same part has now lost gas twice after a repair. Those three answers narrow the field without a laboratory, and they are worth asking for by name instead of waiting to be offered.