The Water Chemistry That Attacks a System

Why this matters

A water report is a materials specification that nobody reads as one. Six numbers on it decide whether the copper in a building lasts forty years or starts weeping pinholes in eight, whether stainless is the safe choice or the worst one, and whether a softener is going to help or quietly make things worse. Shops that cannot read the report end up replacing the same run of pipe twice and telling the customer they have "bad water," which is not a diagnosis and does not point at a fix.

The important part is that one water gives different verdicts for different metals in the same building. There is no such thing as water that is simply corrosive or simply fine.

Before you handle treatment chemicals

Never mix a hypochlorite (bleach-type) product with an acid. The reaction releases chlorine gas, and it happens in seconds in a small mechanical room, a pump pit, or a drain where someone poured one after the other. If both are in use on a site, store them apart, label them, and flush thoroughly between any two chemical steps.

Handle acid and caustic descaling and neutralizing chemicals with the eye and skin protection named in Section 8 of the product's safety data sheet, and add acid to water rather than water to acid. The reverse boils and spatters. The employer's obligation to make that safety data sheet available to the tech using the product sits in 29 CFR 1910.1200.

Any temporary connection made to a potable system for treatment, flushing, or testing needs backflow protection appropriate to the chemical. A hose from a chemical drum to a building's water is a cross-connection, and the local plumbing code, not preference, sets what is required.

The six numbers, and what each one attacks

Two more are worth requesting when the report does not include them. Dissolved oxygen drives general corrosion of steel, and in a closed loop it should be near zero, so a closed loop that keeps testing oxygenated has a leak with makeup water refilling it, which is a mechanical finding hiding in a chemistry report. Ammonia causes stress corrosion cracking of brass and other copper alloys, which is a fast, brittle failure with almost no corrosion product.

For hardness, the commonly used classification (as milligrams per liter of calcium carbonate) runs: soft below 60, moderately hard 61 to 120, hard 121 to 180, and very hard above 180. That scale is about scaling tendency, not about corrosivity, and reading it as a corrosion scale is the mistake this article exists to prevent.

What the balance index does and does not tell you

The Langelier Saturation Index combines pH, temperature, calcium hardness, alkalinity and total dissolved solids into a single number describing whether the water is saturated with calcium carbonate. Positive means a scaling tendency. Negative means the water is undersaturated and will tend to dissolve any protective carbonate film, which is what "aggressive" means. Near zero is balanced.

Two limits on it, both important. It describes calcium carbonate behavior only, so it says nothing about chloride pitting of stainless or ammonia cracking of brass, and a perfectly balanced water can still be the wrong water for a specific alloy. And it is a tendency, not a rate: a mildly negative index in a system with a mature protective film behaves very differently from the same water in new pipe.

One report, four verdicts

A small commercial building on a municipal supply. Illustrative figures from the report, all in milligrams per liter except pH and temperature:

• pH 6.6
• Alkalinity 25 as CaCO3
• Calcium hardness 40 as CaCO3
• Total dissolved solids 90
• Chloride 35
• Water heater setpoint 140 F

Read the water first. Hardness of 40 puts it in the soft band, below 60. Alkalinity of 25 is low, so the water has very little buffering and its pH moves easily. pH 6.6 is on the acid side of neutral. Soft, poorly buffered, slightly acidic water is undersaturated with calcium carbonate, so the balance index runs negative and the water will tend to strip rather than lay down a protective film. There is nothing wrong with this water to drink and it will not scale anything.

Verdict for copper, the domestic distribution. At risk. Aggressive, poorly buffered water is the classic setting for copper pinholes and for elevated copper at the tap. The Lead and Copper Rule sets the action level for copper at 1.3 mg/L and for lead at 0.015 mg/L in tap samples (40 CFR Part 141, Subpart I), and where a building has any lead-bearing solder or brass, this water is exactly the kind that mobilizes it. Sampling is the next step, not speculation.

Verdict for the stainless in the commercial equipment. Fine. Chloride at 35 mg/L is low, far below the 250 mg/L secondary drinking water standard for chloride (40 CFR Part 143) and below the ambient-temperature guidance figures commonly quoted for the standard austenitic grades. Nothing in this report threatens stainless. If the same building had a chloride-bearing sanitizer used on stainless surfaces and left to dry, that would be an entirely separate exposure and the water report would not show it.

Verdict for the ordinary brass fittings and valves. Elevated risk. Low alkalinity and soft water are the conditions associated with dezincification of high-zinc brass, which turns yellow brass pink and porous while leaving its shape intact. Where the building has ordinary brass in wetted service, replacements should be a dezincification-resistant brass or a bronze rather than like for like.

Verdict for the legacy galvanized steel. Poor. Zinc dissolves readily in soft acidic water, so the sacrificial coating is being consumed faster than it would be in balanced water, and once it is locally gone the steel underneath rusts and the failures start. This section is on a shorter clock than anything else in the building.

The correction, and the one that would backfire. Raising pH and alkalinity toward a balanced index, through a neutralizing filter or chemical feed, addresses the copper, the brass, and the galvanized at once, because all three of those verdicts trace to the same aggressive character. Sizing and chemical choice belong to the treatment supplier and to the local code.

The thing not to do, and the reason this water gets mishandled so often: installing a softener does not fix aggressive water and generally makes it more aggressive. A softener exchanges calcium for sodium, so hardness goes down while alkalinity and pH do not improve, and the balance index moves further negative. A customer whose water is already soft and acidic and who has been sold a softener for "bad water" now has a machine treating a problem they did not have, aggravating the one they did.

The reverse case, in one paragraph

Change five of those six numbers and every verdict flips. Take pH 7.8, alkalinity 240, hardness 320, chloride 480, TDS 1,100, same 140 F. Hardness of 320 is very hard, well above the 180 threshold, so this water scales hot surfaces and the balance index runs positive. Copper is comparatively protected here, because the film it needs is being laid down rather than stripped. Stainless is now the exposed material, because 480 mg/L of chloride is nearly double the 250 mg/L secondary standard and is into the range where pitting and, at 140 F with stress present, chloride stress corrosion cracking become live concerns. Brass is at risk again but by a different route. And the high TDS makes every dissimilar-metal joint in the building conduct better, so galvanic couples that would be slow in the first building's water are fast in this one.

Same two words, "bad water," two opposite chemistries, and the materials list you should specify is different in almost every line.

Verifying a water finding

• Sample where the problem is, not at the meter. Water changes as it moves through a building: it warms, it sits, it picks up whatever the pipe gives it. A report from the supply tells you what arrived, not what is attacking the third-floor return.
• Match the sampling protocol to the question. For lead and copper at the tap, the regulated protocol is a first-draw sample after a minimum stagnation period of six hours, because a flushed sample measures the main rather than the plumbing. Flush the line first and you will get a clean result on a building with a real problem.
• Test the hot side separately when the failures are on the hot side. Temperature changes the chemistry, and a cold-water report cannot settle a question about a heater or a recirculation loop.
• Get chloride specifically, not just TDS. Conductivity and TDS lump everything together and cannot distinguish a harmless mineral load from the one ion that pits stainless.
• Re-test after any treatment change, at the same points. Treatment shifts the balance in both directions, and a system corrected too far positive starts scaling. One before-and-after pair at the same sampling location is worth more than any number of one-off reports.

References

• 40 CFR Part 141, Subpart I (Lead and Copper Rule), for the tap action levels of 0.015 mg/L lead and 1.3 mg/L copper and the first-draw sampling protocol
• 40 CFR Part 143, National Secondary Drinking Water Regulations, for the secondary standards including chloride at 250 mg/L, sulfate at 250 mg/L and total dissolved solids at 500 mg/L, which are aesthetic rather than corrosion limits
• 29 CFR 1910.1200, Hazard Communication, for the safety data sheet governing handling of treatment chemicals
• USGS hardness classification (soft, moderately hard, hard, very hard) for the hardness bands used above, and alloy supplier documentation for the chloride limit applicable to a specific stainless grade at a specific temperature
• See related: Pitting, Crevice and Uniform Corrosion; How Scale Forms and What It Actually Costs