The Compatibility Question on a Mixed-Material System

Why this matters

Compatibility gets checked one pair at a time: does this seal handle this fluid, does this metal touch that metal. That works on a single joint and it quietly fails on a system, because in a shared fluid loop the interactions do not require contact. A metal dissolving at one end of a building plates out on a component at the other end. A polymer that admits a trace of oxygen through its wall corrodes iron in a boiler it never touches. An additive protecting one metal gets consumed defending against something a third material introduced.

Every one of those failures passes a pairwise compatibility check, because every pair in the system is individually fine. The question a mixed-material system asks is different: what is in the loop, all of it, and what does the fluid carry between the parts.

The inventory, and what people leave off it

Before any compatibility judgment, list every material the shared fluid touches. Not the major components, everything, because the items that get left off are consistently the ones that cause trouble.

The list that gets written easily: pipe and tubing, the heat exchanger or coil, the pump body and impeller, valve bodies.

The list that gets left off, and should not be:

• Seals and elastomers. Pump seals, valve seats and stem packing, gaskets at every flange, the expansion tank bladder. These are the smallest items in the loop and the most chemically sensitive.
• Valve trim and internals, which are frequently a different alloy from the valve body they sit in.
• Sensor wells, probes and fittings, often stainless in a system that is otherwise not.
• The fluid's own additives. Glycol, inhibitor package, biocide, pH adjuster. These are materials in the loop with their own compatibility and their own lifespan.
• The fill water. Its hardness, chloride content and oxygen load are inputs, and a system topped up frequently is being dosed with fresh oxygen and fresh chlorides on every fill.
• Anything added since commissioning. A retrofit heat exchanger, a replacement pump, a new zone. The inventory that matters is today's, not the one on the original drawings.

Four routes between parts that never touch

Route one: ion transport and deposition. A metal that corrodes anywhere in the loop puts its ions into the fluid, and those ions can deposit on a more active metal downstream. Dissolved copper landing on aluminum is the textbook case, and the damage it does is not general thinning; it is localized pitting at each deposition site, because every deposited particle becomes a tiny cell with the base metal. A system that contains both copper and aluminum has this route open by definition, and the practical control is the fluid chemistry rather than the layout.

Route two: shared chemistry with a finite budget. An inhibitor package is consumed as it works. Anything in the loop that increases the demand, an oxygen leak, a large area of bare ferrous surface, a high-temperature zone, drains the same budget that is protecting everything else. The failure mode is remote and delayed: one zone's problem exhausts the protection on the whole system, and the first component to fail is whichever one was least tolerant, not the one causing the drain.

pH is the sharpest example, because in a mixed-metal loop it can be pulled in two directions at once. Ferrous protection generally favors holding the loop alkaline, while aluminum has a stated upper pH limit above which it is attacked, commonly in the neighborhood of 8.5 depending on the alloy and the manufacturer. When a system contains both, the aluminum manufacturer's stated limit governs, and that means the ferrous side must be protected by inhibitor chemistry rather than by pushing pH up. Get this backwards and you protect the steel by dissolving the heat exchanger.

Route three: permeation and ingress. Some polymers are permeable to oxygen. Tubing without an oxygen diffusion barrier admits oxygen continuously through the tube wall into a closed loop that is supposed to be oxygen-starved, and the oxygen then corrodes iron components elsewhere and consumes inhibitor everywhere. Nothing touches anything; the tubing is chemically fine; the boiler rusts. This runs the other way too, with a fluid or a vapor permeating out through an elastomer over time.

Route four: what one material gives up into the fluid. Plasticizers, processing aids and degradation products leach out of polymers, and glycol itself degrades over time into organic acids that drop pH and attack metals. Organic material in the loop can also feed microbial growth, and microbially influenced corrosion produces deep localized pitting under a deposit that looks like nothing much on the surface. A system with slow or dead-leg zones gives that growth somewhere to live.

Where the pairwise table is still the right tool

Do not throw the pairwise check away. It remains exactly right for the single-fluid, single-joint question: this seal against this chemical at this temperature, this fastener against this bracket. Use it whenever the two materials in question are the only two that matter, which covers most of the day.

The loop view earns its extra effort in four situations, and only these: the system is closed and shares a treated fluid, it contains more than two metal families, it contains any aluminum alongside copper, or it has been modified since commissioning. Outside those, running a full inventory is work without a finding.

Worked case: the loop that ate its own protection

A shop took over a closed hydronic system that had been retrofitted with an aluminum heat exchanger about a year and a half earlier. The complaint was discolored fluid and two failed pump seals in one season. The original piping was mostly copper with steel components, and one zone had been added years earlier in tubing without an oxygen diffusion barrier.

The fluid test told the story before any component was opened. The inhibitor charge, dosed to full concentration at the retrofit, tested at about 40 percent of its dose 8 months later. That is 60 percent consumed in 8 months, about 7.5 points per month, which puts full exhaustion around 13 months against a stated service interval of 24 months, so the package was being consumed at roughly 1.8 times the rate the interval assumes.

The shop's first instinct was to compare against a similar loop in a neighbouring building that held its charge across a full interval. They kept the comparison but stated the limitation in the same breath, because that loop has no aluminum and no non-barrier tubing, so it is not a like-for-like baseline; it establishes that a normal loop can hold its charge, not that this one should have.

Three routes were open at once, and the inventory made that visible. Route three was live through the non-barrier zone, admitting oxygen continuously and driving inhibitor consumption across the whole system, which explains the rate. Route one was live from the moment the aluminum exchanger went in, with copper already in the loop. Route two was live because the previous service had raised pH to protect the ferrous side, which is standard practice on a copper and steel loop and became the wrong move the day an aluminum component was added: the reading came back above the exchanger manufacturer's stated upper limit for that alloy.

The pitting found on inspection in the aluminum was consistent with deposition rather than general attack, which fits route one rather than a straightforward pH excursion, though with pH also out of range both were contributing and the shop did not claim to have separated them.

The corrective plan was ordered by which route kept the others alive. Closing the oxygen ingress came first, because leaving it open means any new inhibitor charge is consumed at the same 1.8 times rate and the fix has a shelf life. That zone was isolated behind a heat exchanger rather than repiped, which was the cheaper path in labor hours. Then the chemistry was reset to a program compatible with aluminum, held under the exchanger manufacturer's stated pH limit, with the ferrous protection carried by inhibitor rather than by alkalinity. Then the pump seals were replaced, last, because replacing them first would have put new seals into the fluid that killed the old ones.

The failure mode to notice here is not any one of the three routes. It is that the retrofit was a correct pairwise decision, an aluminum exchanger appropriate for the duty, installed into a loop whose existing chemistry and existing oxygen leak made it the wrong choice. Nobody checked the loop, because nobody was asked a loop question.

How to verify: the fluid is the instrument

You cannot see any of these four routes by inspecting components, and by the time you can, the damage is done. On a treated closed loop, the fluid test is the diagnostic, and it is worth running on a schedule rather than after a failure.

Sample from a live circulating point, not from a drain that has been sitting, and pull the sample at the same point every time so the readings are comparable. A reasonable starting cadence is at commissioning, again at about 3 months, and then every 6 months, tightened to quarterly on any system that has had a component change or a top-up history. Tune it to what your readings do; a loop that holds steady for two intervals has earned a longer one.

Read four things together, because each is ambiguous alone:

• Inhibitor concentration, compared against the dose and the elapsed time. A concentration falling faster than the stated interval implies is the single most useful early signal, and it points at demand somewhere in the loop rather than at a bad product.
• pH, against the limit of the least tolerant material in the inventory, not against a general target.
• Dissolved metals. Rising copper in a loop containing aluminum is route one running now. Rising iron is oxygen or acid attack.
• Appearance and particulates. Dark fluid, magnetic particulate, or a slime on a strainer each point somewhere specific, and a strainer is the cheapest sample point in the building.

One reading is a number. Three readings on the same loop are a trend, and the trend is what tells you whether the last fix closed the route or just topped up the fluid.

References

• Manufacturer documentation for heat exchangers, pumps and boilers stating permitted pH range, permitted inhibitor chemistries, and material limits
• Fluid treatment supplier documentation for inhibitor dose, service interval and test method
• See related: Chemical Compatibility and the Checks Worth Running; Galvanic Corrosion and the Metals That Fight; The Dielectric Union and What It Does Not Fix; Reading Rust and Corrosion Patterns