A fossil, combined-cycle, or nuclear plant’s steam-water cycle typically runs at purity levels most industries never touch — total dissolved solids in the low parts-per-billion range, iron held below a couple of micrograms per kilogram, cation conductivity fractions of a microsiemens per centimeter. It’s tempting to read that as overkill. It isn’t: the cycle spends its whole life at high pressure, high temperature, and high heat flux, and every one of those conditions is a mechanism for turning a trace impurity into a concentrated one. A contaminant that’s genuinely negligible in the bulk feedwater can end up thousands to a million times more concentrated at a tube wall, under a deposit, or in the first droplets condensing inside a turbine stage. Flow-accelerated corrosion, under-deposit corrosion, stress corrosion cracking, and turbine blade fouling all trace back to that same amplification step. Cycle chemistry control is the engineering discipline built specifically to stay ahead of it.
How a trace impurity becomes a metal-loss mechanism
Carbon steel, low-alloy steel, and the nickel-based alloys used in nuclear steam generators all depend on a passive oxide film for their corrosion resistance — magnetite (Fe₃O₄) under reducing conditions, hematite (Fe₂O₃) under mildly oxidizing ones. Cycle chemistry control is, at its core, the business of keeping that film intact.
Flow-accelerated corrosion (FAC) is the most consequential failure mode in an all-ferrous system, and it isn’t erosion in the mechanical sense — it’s continuous chemical dissolution of the protective magnetite layer at the flow interface, exposing fresh metal to further electrochemical attack. Magnetite’s solubility in water is governed by the reductive dissolution reaction
and that solubility depends strongly on local pH at operating temperature (pH), oxidation-reduction potential, dissolved oxygen, and flow shear. In a reducing, low-ORP environment with a low pH — single-phase water or two-phase wet steam alike — magnetite solubility climbs sharply, and wall-thinning rates as high as 3 mm/year have been documented. That’s fast enough to perforate a deaerator outlet line, a feedwater heater, an HRSG economizer, or a moisture separator reheater in a single operating cycle if the geometry and chemistry line up wrong — FAC failures are the leading cause of sudden, unplanned tube ruptures in all-ferrous cycles.
Caustic gouging and acid chloride attack both start the same way: local boiling under a porous deposit — typically iron corrosion products that have themselves built up into a scale layer — concentrates whatever’s dissolved in the bulk water by a factor of 10³ to 10⁶ right at the tube wall, because water evaporates out of the deposit faster than dissolved solids can diffuse back into the bulk flow. If the impurity that concentrates is free NaOH, it dissolves the protective oxide outright, forming soluble sodium ferrite and gouging the tube wall into a distinctive channeled groove. If instead the impurity is a hydrolyzable chloride salt — magnesium chloride or calcium chloride, commonly from condenser tube leaks on brackish or seawater cooling — high-temperature hydrolysis generates a locally strongly acidic environment, driving pitting that develops into acid chloride corrosion.
Stress corrosion cracking (SCC) shows up wherever this same concentration mechanism meets a crevice geometry and residual tensile stress — most consequentially in PWR steam generators, where the narrow gap between heat-transfer tubing (Alloy 600 MA or Alloy 690) and tube support plates concentrates chloride, sulfate, and caustic species exactly the way a porous deposit does elsewhere. That combination is what actually threatens the radiological barrier between primary and secondary loops in a pressurized water reactor, and it’s a large part of why PWR secondary-side chemistry gets treated as a safety-class concern, not just an efficiency one.
Turbine blade deposition works through the same underlying physics from the opposite direction: as steam expands stage by stage through a turbine, both its temperature and pressure drop sharply, and the solubility of dissolved salts and silica in that steam falls exponentially along with them. Once expansion carries the steam into the phase transition zone — the point where the first condensate droplets begin to form — supersaturated salts and silica deposit directly onto stationary and rotating blade surfaces. That changes blade aerodynamic profile, increases flow resistance, and triggers boundary-layer separation, cutting turbine thermal efficiency by 1-2%. Soluble salts that concentrate at blade roots and in fir-tree slots go further, driving pitting and corrosion fatigue that can end in blade fracture — a failure mode with obvious safety consequences well beyond the efficiency loss.
Why some impurities travel in steam and others don’t
Predicting and controlling all of this starts with understanding how an impurity gets from boiler water into steam in the first place. There are two distinct mechanisms, and they behave completely differently.
Mechanical carryover is simple entrainment: tiny water droplets, produced when steam bubbles burst at the boiling surface, get physically swept along with the steam flow. It’s governed almost entirely by the mechanical separation efficiency of the drum’s steam-water separators, and well-designed separators hold mechanical carryover below roughly 0.05% of the boiler water flow.
Vaporous carryover is a genuinely different mechanism — a thermodynamic equilibrium in which a compound dissolves directly into the vapor phase itself, at the molecular level, independent of any droplet entrainment. The partitioning coefficient that governs it is defined simply as the ratio of a compound’s concentration in steam to its concentration in boiler water:
and that ratio is driven largely by the density difference between the two phases, plus a compound-specific constant tied to its molecular structure and dissociation behavior. As operating pressure climbs toward the critical point (22.1 MPa) — commonly by 16.5 MPa (2400 psi) in subcritical or near-critical units — the density gap between saturated steam and saturated water shrinks dramatically, and for many compounds climbs by orders of magnitude as a direct consequence.
That distinction has real practical teeth. Weak electrolytes like silica and small-molecule organic acids (acetic, formic) have inherently high and travel into steam readily through vaporous carryover alone — silica scaling on turbine blades is a vaporous-carryover problem first and foremost. Strong electrolytes — sodium chloride, sodium sulfate, sodium hydroxide — have working against them (their partitioning strongly favors the liquid phase), so even a mechanical carryover fraction as small as 0.05% still delivers more total sodium and chloride to the steam than vaporous carryover does. Which mechanism dominates for a given contaminant is exactly what determines which lever — separator performance versus boiler water chemistry — actually fixes the problem.
Early condensate acidity is where these two transport mechanisms collide in a way that specific conductivity monitoring completely misses. In the turbine’s phase transition zone, the very first microscopic condensate droplets form well before bulk moisture appears. Volatile acidic species — chloride, sulfate, and low-molecular-weight organic acids — have partitioning behavior sharply different from the neutralizing amine used to alkalize the cycle, and they preferentially concentrate into those first droplets. The result is a local pH crash: main steam running comfortably alkaline (pH 9.0-9.5) can produce early condensate at strongly acidic pH, while the bulk steam’s conductivity monitoring shows nothing unusual at all, because the acidic species are present in such small absolute quantity that they don’t move the whole-stream average. That early condensate is disproportionately responsible for pitting and corrosion fatigue in last-stage turbine blades and localized attack in superheater/reheater sections — exactly the failure modes a plant’s routine conductivity trending won’t see coming.
Matching the chemistry regime to the metallurgy
IAPWS and EPRI have codified several distinct cycle chemistry treatment regimes, and picking the wrong one for a given system’s metallurgy and configuration doesn’t just underperform — it can actively accelerate the corrosion it’s meant to prevent.
Reducing all-volatile treatment, AVT(R), dosed with a volatile alkalizer (ammonia or an organic amine) plus a strong reducing agent/oxygen scavenger such as hydrazine, was designed for older units with copper-alloy heater tubing, where a low oxidation-reduction potential protects the copper. Applied to an all-ferrous system, that same low ORP does the opposite of what’s needed: it accelerates magnetite dissolution on carbon steel and drives exactly the single- and two-phase FAC the regime was never meant to cause there.
Oxidizing all-volatile treatment, AVT(O), is the standard answer for all-ferrous systems: dose only the volatile alkalizer to hold pH in the 9.2-9.8 range, drop the reducing agent and oxygen scavenger entirely, and let a small, controlled dissolved-oxygen level — 5-30 µg/kg, whether from controlled condenser in-leakage or a deliberate trace addition — do the protective work. That trace oxygen oxidizes the porous outer magnetite layer into a dense, low-solubility hematite passivation layer, cutting single-phase FAC by more than an order of magnitude. The tradeoff is tighter feedwater purity discipline: cation conductivity after hydrogen-form ion exchange (CACE) has to stay at or below 0.2 µS/cm for the regime to hold.
Oxygenated treatment (OT) goes a step further and is reserved for supercritical and ultra-supercritical once-through units running exceptionally pure feedwater. High-purity oxygen is injected directly into the feedwater — 30-150 µg/kg — while amine dosing is trimmed to the minimum needed to hold pH around 8.0-8.5. Done correctly, OT builds a dual-layer oxide (a magnetite inner layer under a hematite outer layer) that drives dissolved iron in the water phase down below roughly 2 µg/kg, meaningfully extending boiler chemical-cleaning intervals and effectively eliminating economizer FAC — but it demands even tighter feedwater purity than AVT(O), with CACE held below 0.15 µS/cm.
Phosphate treatment (PT) and caustic treatment (CT) are drum-boiler-specific regimes. Phosphate treatment — commonly equilibrium phosphate treatment (EPT) — doses small amounts of trisodium and disodium phosphate to hold boiler water pH in the 9.0-9.8 range, buffering against hardness contamination and acid or caustic ingress. Caustic treatment doses a small amount of free NaOH directly. Both suppress vaporous chloride carryover effectively, but if underdosed or overdosed, excess free alkali raises sodium carryover to steam and can itself trigger caustic embrittlement or gouging — the same failure mode the treatment exists to prevent, just from the opposite direction.
Advanced neutralizing amines solve a specific problem in PWR secondary loops that ordinary ammonia can’t: ammonia’s partitioning coefficient is so high that most of it volatilizes straight into the vapor phase, leaving the liquid phase in two-phase regions — wet steam piping, moisture separator reheaters, low-pressure heaters — with a depressed pH and correspondingly severe FAC. Lower-volatility, high-alkalizing-power amines such as ethanolamine (ETA) or morpholine stay preferentially in the liquid phase, holding two-phase-region pH above roughly 9.2 and substantially suppressing the transport of iron corrosion products into the steam generator.
Film-forming substances (FFS/FFA) — long-chain aliphatic amines or film-forming polymers — serve a different purpose entirely: for cycling and peaking units that start up and shut down frequently, or during extended layup, they adsorb onto metal surfaces as a self-assembled hydrophobic monolayer, physically isolating the metal from the corrosive medium when normal operating chemistry isn’t sustained.
| Regime | Chemistry environment | Feedwater CACE limit | Feedwater Fe target | FAC control | Typical application |
|---|---|---|---|---|---|
| AVT(R) | Reducing, mildly alkaline, low ORP | ≤ 0.2 µS/cm | < 2 µg/kg | Poor — promotes carbon steel FAC | Older units with copper-alloy components |
| AVT(O) | Weakly oxidizing, mildly alkaline | ≤ 0.2 µS/cm | < 2 µg/kg | Excellent — eliminates single-phase FAC | All-ferrous drum boilers and HRSGs |
| OT | Neutral/weakly alkaline, strongly oxidizing | < 0.15 µS/cm | < 2 µg/kg | Outstanding — builds hematite passivation | Supercritical/ultra-supercritical once-through units |
| PT / CT | Strongly buffered alkaline (boiler water) | System-dependent | Feedwater-regime-dependent | Depends on feedwater regime | Subcritical/high-pressure drum boilers |
| Advanced amines | High liquid-phase pH, low vapor loss | ≤ 0.2 µS/cm | < 2 µg/kg | Excellent for wet-steam-region FAC | PWR secondary loops |
| FFS / FFA | Hydrophobic self-assembled monolayer | Formulation-dependent | Significantly reduced | Good | Cycling/peaking units, layup protection |
Why acid conductivity, not raw conductivity, is the diagnostic that matters
Direct specific-conductivity measurement can’t reliably catch trace strong-electrolyte contamination, because the alkalizing amine deliberately dosed into the cycle (ammonia, morpholine, ETA) produces its own conductivity background that swamps the signal from whatever trace salt actually needs catching. Cation conductivity after hydrogen-form exchange — CACE, sometimes called acid conductivity — is the diagnostic built specifically to strip that background out.
The sample stream passes through a column of strong-acid cation exchange resin in the hydrogen form, which exchanges every cation present — sodium, ammonium, amine cations alike — for the far more mobile H⁺ ion:
The originally neutral trace salt comes out the other side as a strong acid, and because H⁺‘s limiting molar conductivity (349.8 S·cm²/mol) runs roughly seven times higher than the alkalizing cation it displaced, the conductivity contribution of trace corrosive anions — chloride, sulfate — gets amplified several-fold at the same time the alkalizer’s own conductivity signature is completely neutralized out. What’s left is a clean, sensitive read on exactly the contaminants that matter for corrosion risk. When dissolved CO₂ (forming carbonic acid) is also present and needs to be excluded from the reading, degassed CACE (DCACE) — heating or inert-gas-stripping the sample to remove CO₂ before the ion-exchange step — isolates the strong-acid-anion signal from the weak-acid carbonic contribution, giving a cleaner read on genuine corrosion risk.
Closing the loop: makeup water purity and zero-liquid-discharge blowdown recovery
None of the cycle chemistry regimes above can hold their targets against a feedwater source that isn’t already extremely pure. Makeup water treatment, condensate polishing, and continuous online monitoring form the three-part front line that keeps contamination out of the cycle in the first place, and the purity targets involved are strict at every sampling point: makeup water outlet limits typically require CACE at or below 0.2 µS/cm alongside tight sodium, chloride, and silica ceilings; condensate pump discharge adds dissolved oxygen and total iron to that list; main and high-pressure steam carry the same CACE, sodium, and chloride limits through to the turbine inlet. Mixed-bed ion exchange (MBIX) and electrodeionization (EDI) are the standard condensate polishing technologies built to hold those numbers. Readers sizing mixed-bed or EDI capacity for makeup or condensate polishing duty can work the ion loading and resin/regeneration demand through this site’s Ion Exchange Resin & CEDI Designer, and dosing rates for the amines, oxygen scavengers, and phosphate compounds discussed above can be worked through the Chemical Dosing Calculator.
On the discharge side, tightening wastewater regulations have pushed plants toward recovering rather than simply blowing down cycle water, using largely the same high-recovery membrane and electrodialysis toolkit covered in depth in this site’s zero liquid discharge piece: high-pressure RO, osmotically assisted RO, electrodialysis reversal, and membrane distillation, sequenced to push blowdown recovery toward zero liquid discharge. Electrodialysis reversal is a particularly good structural fit for this specific water — its periodic polarity reversal gives it strong resistance to the high-silica, high-calcium-sulfate scaling that a blowdown stream concentrated by a boiler’s own cycle chemistry tends to carry, at a specific energy consumption around 3-7 kWh/m³.
What actually prevents the failure
The unifying idea across every mechanism here is the same: bulk-water concentration numbers are almost never what causes the damage. A boiler tube doesn’t fail because feedwater iron averaged 3 µg/kg over the quarter — it fails because a deposit concentrated whatever was dissolved by five or six orders of magnitude at one specific spot, or because early condensate in a turbine stage ran acidic for a few seconds every startup while bulk steam conductivity looked perfectly normal. Cycle chemistry control that actually works starts from that premise: match the treatment regime to the system’s actual metallurgy (AVT(O) or OT for all-ferrous systems, advanced amines for PWR secondary loops), hold feedwater purity — measured by CACE, not raw conductivity — inside the narrow band each regime requires, and build the makeup-water and condensate-polishing capacity to keep it there continuously rather than catching excursions after the fact.
Further reading
- Electric Power Research Institute, Cycle Chemistry Guidelines for Fossil Plants, EPRI cycle chemistry guideline series — the primary industry reference for AVT/OT/PT regime selection and control limits.
- International Association for the Properties of Water and Steam (IAPWS), Technical Guidance Document on Steam Purity and TGD8-16 (film-forming substances) — iapws.org/documents/techguide.
- U.S. NRC, Pressurized Water Reactor Secondary Water Chemistry Guidelines, Revision 6 (ML0508/ML050840532) — regulatory basis for PWR secondary-side amine treatment and SCC control.
- EPRI, The Volatility of Impurities in Water/Steam Cycles — the foundational reference on vaporous versus mechanical carryover and partitioning coefficient behavior near the critical point.