Carbon steel corrosion in steam condensate return systems is controlled by maintaining a thin, adherent iron-silicate film at the metal surface. The film is generated when reactive silica, carried into the steam by mechanical carryover or intentional feed of sodium silicate, adsorbs onto the hydrated iron oxide surface at pH 8.0 and polymerizes through siloxane condensation. The protection is not comparable to magnetite passivation in high-purity boiler water; it depends on continuous repair of silica-rich deposits from the condensate. At pH 8.0, monosilicic acid is the dominant dissolved silica species, with only a small fraction present as the silicate anion because the first dissociation constant of silicic acid is near pKa 9.8 at 25 °C. Film formation therefore proceeds through adsorption and polymerization of neutral silicic acid rather than through electrostatic deposition of charged silicate oligomers. Any local acidification from dissolved carbon dioxide, any oxygen ingress, or any excessive silicate solubility shift disrupts the film and exposes carbon steel to the underlying corrosion circuit. The corrosion rate is governed by the film’s ability to suppress both anodic iron dissolution and cathodic oxygen or hydrogen reduction, a behavior that is measured in plant systems by coupon mass loss in accordance with ASTM G1-03(2017)e1 and ASTM G31-72(2004). Bulk pH 8.0 is not a direct indicator of film integrity because the interfacial pH beneath a porous silica layer can be depressed by hydrolysis of ferrous ions and carbonic acid, while the bulk water remains within specification.

Sodium silicate used for this application is usually an N-grade liquid with a SiO₂/Na₂O weight ratio of 3.22 and a total solids content near 37–40 wt%. The product is diluted to a 1–3 wt% solution before injection. The silicate residual required to maintain film repair in a clean low-pressure steam condensate system is commonly held between 1.0 mg/L and 3.0 mg/L as SiO₂ by ASTM D859-16; above this range, deposition on heat-transfer surfaces and downstream strainers increases. Below 0.5 mg/L as SiO₂, the repair rate may be lower than the film dissolution rate. However, the exact residual boundary is site-specific because it depends on condensate temperature, CO₂ partial pressure, iron loading, and flow velocity. Published data for a universal minimum residual applicable to all steam condensate configurations is limited. Measurement of reactive silica alone can overstate the film-forming capacity because colloidal or polymeric silica is not molybdate-reactive and does not repair the film with the same efficiency. Total silica by digestion followed by ASTM D859-16 can be used to estimate the unreactive fraction. A high total-to-reactive silica ratio indicates formation of colloidal silica in the condensate and usually precedes loss of film coherence and deposition in low-velocity sections.

The hydrolysis of carbon dioxide is the principal acid load in steam condensate. At 25 °C the equilibrium ratio of bicarbonate to dissolved CO₂ is approximately 45:1 at pH 8.0, leaving a small but pH-active CO₂ reservoir. In a long carbon steel return line, this reservoir can feed acidification at the metal surface even when bulk pH is 8.0, because the local pH under a porous silicate film follows the anodic and cathodic reaction stoichiometry rather than the mixed bulk condition. The film is therefore most vulnerable at locations with high CO₂ partial pressure, low temperature, high condensation rate, and low alkalinity. In subcooled condensate pockets, CO₂ absorption increases and the pH can drift below 7.5 while the return-header sample reads 8.0. Carbon steel corrosion in carbon dioxide-containing condensate without an intact filming inhibitor typically ranges from 0.25 mm/y to 1.0 mm/y, while systems with an intact silicate film and residual alkalinity generally exhibit time-averaged corrosion rates below 0.025 mm/y when evaluated by ASTM G31-72(2004) coupons. This range is not a film guarantee; it is an operational envelope that assumes oxygen is below 0.02 mg/L and the condensate is not contaminated by chloride or hardness.

Does the 8.0 pH Setpoint Remain Valid for Saturated, Subcooled, and Flashing Condensate Returns?

The validity of pH 8.0 as a control target is influenced by temperature and phase behavior. Silica solubility increases with temperature, and the polymerization rate of silicic acid is temperature-sensitive. In a saturated condensate line at 95–100 °C, film deposition can be rapid, but the film may be gelatinous rather than dense if the reactive silica residual is too high. In subcooled returns at 40–60 °C, the film forms more slowly and may be thinner. Flashing condensate creates two-phase flow with cavitation and mechanical shear that can erode the film. The wall shear stress in a two-phase line is not described by single-phase velocity alone; slug flow and bubble collapse at fittings create local stresses that exceed the cohesive strength of the silica-rich layer. Pipe geometry, therefore, determines whether pH 8.0 is a sufficient setpoint. A long horizontal return with laminar or low-velocity stratified flow will accumulate a thicker but potentially porous silicate scale, while a short vertical riser with slug flow will show film thinning at the top and at elbows. Corrosion coupon data from such systems must be interpreted in conjunction with pipe wall inspection by ultrasonic thickness measurement, because a coupon rack placed in a low-shear bypass may not reproduce the erosion-corrosion environment of the main line. The chemical setpoint alone cannot compensate for mechanical film removal. If the return system includes an atmospheric flash vessel, the sudden pressure drop releases CO₂ and raises the pH above 8.5, where silicate dissolution accelerates and the film may partially strip. Under those conditions, the control strategy should be shifted toward total alkalinity and condensate receiver pH rather than a single upstream pH target.

The pH 8.0 setpoint also assumes that the alkalinity present in the condensate is predominantly bicarbonate/carbonate and not generated by strong alkali carryover. Boiler water carryover containing sodium hydroxide or phosphate can raise the local pH above the film stability window, producing soluble silicate and leaving the metal surface unprotected. Conversely, the absence of alkalinity causes pH to be governed entirely by CO₂ and the silicate feed, which has a relatively weak buffering effect. A practical control envelope is pH 8.0 ± 0.2 with total alkalinity of 10–30 mg/L as CaCO₃ in low-pressure condensate systems, though the upper alkalinity limit depends on boiler water limits and steam purity requirements. The pH measurement should be performed on a cooled flowing sample conforming to ASTM D1293-18, with temperature compensation reported. Conductivity measured by ASTM D1125-14 should remain below 25 µS/cm at 25 °C in high-purity condensate; increasing conductivity in a silicate-treated system can indicate accumulation of dissolved acid gases, ingress of cooling water, or breakdown of organic contaminants.

Dissolved oxygen ingress converts a silicate film from a protective layer into a cathodic area adjacent to exposed steel. In the presence of oxygen, any small breach in the film becomes an anode, while the surrounding silica-coated surface supports cathodic oxygen reduction. This produces intense localized attack, often in the form of pits at flow disturbances, gasket crevices, or under deposited iron oxide. Oxygen concentrations above 0.1 mg/L measured by ASTM D888-18 are sufficient to accelerate pitting in low-flow condensate, but the exact threshold depends on chloride, pH, and temperature. Dissolved oxygen also oxidizes ferrous iron released from the corrosion process to ferric iron, which precipitates as hydrated iron oxide and becomes embedded in the silicate film. The resulting iron-silicate scale is less protective because it is more brittle and more porous than a pure silica film. A condensate system operating at pH 8.0 with a silicate residual of 1.0–3.0 mg/L as SiO₂ may still show a high pitting rate if the deaerator or condensate receiver is not sealed. Vented receivers, pump packing leaks, and vacuum portions of the return system are common oxygen ingress points. Mechanical deaeration combined with an oxygen scavenger is therefore required before the silicate film can perform as the primary corrosion control barrier. Published data for the maximum dissolved oxygen that can be tolerated without pit initiation in silicate-treated carbon steel condensate is limited because pit initiation is a probabilistic function of surface condition and flow.

When Ammonia and Neutralizing Amines Raise Condensate pH Without Supplying Film-Forming Silica

Neutralizing amines such as cyclohexylamine, morpholine, diethylaminoethanol, or ammonia are often used to elevate condensate pH and counteract CO₂. These treatments do not create a silicate film, but they can destabilize an existing one if the pH excursions exceed the film stability window. At pH 8.5 and above, the solubility of amorphous silica increases as the silicate anion fraction increases, and the film can begin to dissolve. In systems where an amine is applied in parallel with sodium silicate, the control logic must separate the pH contribution of the amine from the film repair contribution of the silicate. A film that is thinned by amine-induced dissolution may still produce an acceptable bulk pH reading, especially if the amine is volatile and accumulates in the condensate. The proper diagnostic is not pH alone but the combination of reactive silica residual, total iron, dissolved oxygen, and corrosion coupon mass loss. In a system with ammonia at pH 8.0, the ratio of free ammonia to ammonium ion is low, but at higher temperatures the volatility of ammonia changes, producing uneven pH control across long headers. Amine treatment can also increase the electrical conductivity of condensate and interfere with online sodium and silica analyzers if the sample conditioning system is not designed for amine-laden water.

The use of film-forming amines with silicate is an operational incompatibility in some systems. Film-forming amines create hydrophobic organic films on metal surfaces that can prevent direct contact between the silica species and the metal oxide surface, resulting in a patchy, non-repairable silicate film. The organic film may also trap water beneath it and create under-deposit corrosion at the interface. If a plant switches from a film-forming amine program to a silicate program, the line must be cleaned to remove the organic film; otherwise, the corrosion rate measured by coupons may be misleadingly low while insulated piping undergoes localized attack. The cleaning procedure should be validated by petrographic or FTIR analysis of pipe surface deposits, not by visual inspection alone. Because field data on mixed amine-silicate interactions show strong system dependence, published data for the exact compatibility threshold of each amine is limited; a site-specific compatibility test in a bypass coupon rack is the only defensible method before full-scale conversion.

The presence of hardness in condensate is usually caused by condenser leakage or make-up water contamination. Magnesium and calcium react with silicate and hydroxide to form metal-silicate precipitates that are less protective than the iron-silicate film. Magnesium silicate in particular produces a porous scale that can promote under-deposit corrosion and interfere with heat transfer. Hardness ingress above 0.5 mg/L as CaCO₃ into a silicate-treated condensate return is a known incompatibility at pH 8.0 because the low solubility of magnesium silicate shifts the deposit from protective to obstructive. Condensate polishers or improved condenser integrity are required before the silicate program can be considered stable. Similarly, chloride ingress from cooling water cross-contamination can initiate pitting at film defects even when the bulk pH is 8.0. A chloride limit of 5 mg/L in condensate is often used as an action level for industrial systems, but the exact pitting threshold depends on dissolved oxygen, temperature, and surface condition. Sulfate and nitrate ingress can also increase conductivity and interfere with silica polymerization. The chemical control program must therefore include cation conductivity, chloride, sulfate, and hardness monitoring in addition to pH and silica.

Surface analysis of carbon steel coupons from silicate-treated systems is performed when the corrosion rate exceeds 0.025 mm/y or pitting is observed. A failed coupon should be cross-sectioned and examined by scanning electron microscopy with energy-dispersive X-ray spectroscopy to determine whether silicon is present at the metal surface or only in an outer deposit. X-ray photoelectron spectroscopy can identify bridging and non-bridging oxygen in the silicate network, but this method is not routine. The presence of a silicon-rich layer is not sufficient to claim passivation; the film must be continuous and free of cracks at the magnification used. The surface should also be checked for chloride and sulfur, which indicate ingress or oxygen-scavenger breakdown products. In a plant failure investigation, the most common root causes are oxygen ingress at pump seals, low-flow dead legs with under-deposit corrosion, and pH excursions caused by loss of alkalinity. The corrective action is not automatically to increase silicate feed; if the film is porous and iron-rich, increasing silica can accelerate deposition without restoring protection.

Online Analyzer Cross-Validation, Coupon Rack Placement, and Condensate Return Sampling Frequency

The sampling point location determines whether the pH and silica data represent the bulk return or a stagnant by-pass dead leg. The sample line should be a continuous flowing line from the return header to an isokinetic sample cooler, with a sample velocity not less than 1.5 m/s in the sample tube and a delay time documented. Sample coolers should reduce the temperature to 25 ± 5 °C before pH and conductivity measurement because high-temperature pH electrodes drift and silica polymerization continues in a hot sample. The sample cooler outlet should feed a side-stream filter and then split to the silica analyzer, dissolved oxygen analyzer, and grab-sample point. A grab sample for total iron should be collected after a 15-minute flush at the sample point, acidified to pH <2 with nitric acid, and analyzed by ASTM D1068-15. The grab-sample container must be polyethylene or PTFE, not glass, if low-level silica analysis is to be performed. Sampling frequency should follow the plant’s chemical control procedure, but during startup and after any contamination event, the interval should be shortened to hourly for pH and reactive silica until the system returns to the control range.

Thermal Stability Limits of Sodium Silicate Feed and the Consequences of Unintentional Polymerization

Sodium silicate feed solution is thermally sensitive in stagnant injection lines and pump heads. Concentrated silicate exposed to temperatures above 60 °C can undergo accelerated polymerization, forming a viscous gel that plugs metering pumps, check valves, and quill tips. Dilution with demineralized water reduces the gelation rate but does not eliminate the risk if the line is stagnant. The injection quill should be installed so that the tip is cooled by the flowing steam or water and does not sit in a dead cavity. A 316L stainless steel quill is commonly used, although crevice corrosion at sealing faces has been observed if chloride contamination is present. Batch-to-batch variation in commercial sodium silicate can change the SiO₂/Na₂O ratio and the polymerization state; each tote should be analyzed for density and alkalinity before use. The feed tank and lines must be flushed with demineralized water after the silicate feed is interrupted, because concentrated sodium silicate forms a cementitious scale in stagnant pumps and check valves. The injection line should be fitted with a rotameter calibrated for the diluted silicate viscosity, and the pump should be a low-pulsation metering pump with stroke-length adjustment. If the silicate feed is interrupted for more than 24 hours, the film in the condensate system may slowly dissolve, especially if the pH remains above 8.5 due to amine carryover or if the condensate temperature exceeds 80 °C. Return to service should include a gradual increase in silicate residual from 0.5 mg/L to the target range over 48 hours, followed by coupon inspection at the next scheduled interval.

The same silica residual that protects the return piping can foul heat exchangers, steam traps, and condensate coolers. Silicate scale on heat-transfer surfaces reduces the overall heat-transfer coefficient and creates a thermal barrier that raises the metal temperature at the tube surface. In a steam-to-water condensate cooler with a heat flux above 30 kW/m², the surface temperature can exceed the bulk condensate temperature by 5–10 °C, changing the local silica solubility and deposition rate. The risk is highest in low-velocity calm zones where colloidal silica can settle and dewater. Plant experience shows that narrow plate-and-frame heat exchangers are more sensitive to silicate fouling than shell-and-tube exchangers with wide channels, but published data for site-specific fouling factors is limited. The remedy is not necessarily a lower silicate residual; it is better pH control, removal of oxygen and iron, and maintenance of sufficient velocity to keep colloidal silica in suspension. The minimum flow velocity to avoid deposition is typically above 0.6 m/s in horizontal lines, but the value depends on pipe diameter, temperature, and the coagulating effects of iron and magnesium. A silicate treatment program in a system with heat-transfer surfaces therefore requires a deposition risk assessment separate from the corrosion rate assessment.

The selection of an oxygen scavenger in a silicate-treated system must account for the interaction with the film. Catalyzed sulfite is common in low-pressure steam plants because it reacts quickly with dissolved oxygen at ambient condensate temperature. Excess sulfite contributes to conductivity as sulfate after oxidation, and the sulfate can compete with silicate at the iron oxide surface. Carbohydrazide or erythorbate-based organic oxygen scavengers do not add sulfate but may alter the redox potential of the condensate. If the condensate is used for food contact or humidification, the oxygen scavenger and silicate feed must comply with applicable regulations, such as FDA 21 CFR 173.310 for secondary direct food additives or local boiler water standards. In any case, the oxygen scavenger should not be used as a substitute for mechanical deaeration; a residual oxygen scavenger concentration alone is not sufficient to protect a breached silicate film from pit initiation.