Coal, Gas

The Evolution of Power Plant Water/Steam Chemistry

Issue 10 and Volume 116.

By Brad Buecker, Contributing Editor

Water and steam are the lifeblood of any steam-driven power facility, and proper chemistry control is critically important to plant operation, reliability, the bottom line, and especially personnel safety.

In the last three decades, our understanding of high-pressure water/ steam chemistry has greatly advanced through direct experience and from excellent research and funding by organizations such as the Electric Power Research Institute (EPRI).

In spite of this research and numerous publications regarding steam generation chemistry, many in the power industry have not yet been properly informed of new developments. This problem is exacerbated by the retirement of many power plant Baby Boomers, which is creating a void in operational knowledge.

Also, it is becoming quite evident that natural gas combined-cycle power production will be replacing coal-fired generation at a rapidly increasing pace. Many of the lessons learned from the coal-fired power industry are directly applicable to high-pressure heat recovery steam generators (HRSGs), but sometimes the message is not transferred adequately. So, let’s dig into the most important issues.

Condensate/Feedwater Treatment

When I began my utility career in 1981, widely-accepted condensate/feedwater chemical treatment consisted of ammonia or amine feed to establish an alkaline environment. This treatment was combined with feed of an oxygen scavenger/metal passivator to eliminate any oxygen not removed by a deaerator and/or a deaerating condenser. The combination of ammonia and oxygen scavenger feed came to be called all-volatile treatment reducing, [AVT(R)]. The thinking was that all oxygen must be eliminated from the feedwater system.

“This changed in 1986. On December 9 of that year, and elbow in the condensate system ruptured at the Surry Nuclear Power Station [near Rushmere, Virginia, USA]. The failure caused four fatalities and tens of millions of dollars in repair costs and lost revenues.” [1]. Researchers learned from this accident and others that the reducing environment produced by oxygen scavenger feed results in single-phase flow-accelerated corrosion (FAC).

The corrosion is most prevalent at directional changes and flow disturbances, e.g., feedwater and economizer elbows, valve perturbations, feedwater heater drains, and other locations. The pipe wall gradually thins, and catastrophic failures occur when the corroded area can no longer withstand the internal pressure. The chart in Figure 1 shows that the phenomenon is primarily influenced by pH and temperature.

The corrosion mechanism typically reaches a maximum at a temperature of 300o F, which in the case of HRSGs can initiate severe FAC in the LP evaporator circuit with its many short-radius elbows. IP and HP evaporator elbows can also suffer from FAC at times even though the temperature is greater than in the LP. In addition to the accident noted above at the Surry Nuclear Plant, during the last decade a FAC-induced attemperator line failure occurred at a sister plant to one at which I worked, in which the failure killed two plant personnel. I cannot overemphasize the importance of this issue.

The feedwater treatment program now employed for most drum units is all-volatile treatment oxidizing [AVT(O)]. In this program, the (typically) small amount of air that leaks in through the condenser is allowed to remain, with ammonia injection to raise the pH to a range of 9.2 to 9.6. If the cation conductivity of the feedwater is maintained below 0.2 µS/cm, the magnetite (Fe3O4) layer that naturally develops on carbon steel becomes interspersed and covered with a coating of ferric oxide hydrate (FeOOH), which is not susceptible to FAC. AVT(O) is an offshoot of oxygenated treatment (OT), which was developed in Europe almost 40 years ago as a condensate/treatment method for once-through steam generators. With OT, pure oxygen is deliberately injected into the condensate and feedwater to induce formation of the FeOOH protective coating. The condensate must be very pure (cation conductivity ≤0.15 µS/cm) for this treatment to work properly.

An important point to note is that if the feedwater system contains any pre-heaters with copper-alloy tubes, AVT(O) or OT cannot be employed, as the oxygen would quickly corrode the copper. AVT(R) is the only choice when copper alloys are present, but chemistry must be carefully monitored and controlled to achieve a balance between minimized copper corrosion and FAC. [2]

With regard to HRSGs, single-phase FAC can be addressed in the design phase by fabricating evaporator elbows from 2.5 percent chromium steel as opposed to straight carbon steel. Even a small amount of chromium imparts great resistance to FAC. Admittedly, however, this method will increase material costs for the LP steam generator.

Another phenomenon that may affect high-pressure steam generators is two-phase FAC. This corrosion occurs in locations where water and steam exist simultaneously, such that any protection offered by AVT(O) or OT is lost when the oxygen flashes away with the steam. A primary location is the deaerator in both conventional units and HRSGs. Two-phase FAC is not easy to control, so regular inspections of locations susceptible to the corrosion are important. I have personally inspected or observed situations of two-phase FAC corrosion of deaerators and feedwater heater drains in coal-fired units. The latter actually caused a failure, fortunately when no personnel were in the area at the time of the occurrence.

Other location(s) for two-phase FAC are the distribution headers and cooling circuits of air-cooled condensers. Two-phase condensate from the LP turbine exhaust can be quite damaging in these locations.

A different issue regarding oxygen corrosion must be noted. For low-pressure boilers such as industrial units or power plant auxiliary boilers, an oxygen scavenger/metal passivator is often quite beneficial, as the makeup may contain high levels of dissolved oxygen. For units at pressures below 600 psig, sodium sulfite (Na2SO3) is a practical choice. The compound removes oxygen thusly,

2Na2SO3 + O2 → 2Na2SO4

For higher pressures, one of the substitutes for hydrazine (N2H4) is a better choice, as sodium sulfite will break down to produce sulfur dioxide (SO2) and perhaps some hydrogen sulfide (H2S). Both of these decomposition products can cause considerable damage in boilers.

Boiler Water Treatment

Boiler water chemistry for drum units has evolved for two primary reasons. First, much more is known about waterwall corrosion mechanisms, particularly with regard to traditional treatment programs. Secondly, issues related to carryover of products from boiler drums to steam turbines are also much better understood.

Phosphate treatment of drum boilers is still a common method to protect against corrosion, but gone are the days of coordinated and congruent phosphate treatment, where the chemical feed consisted primarily of tri-sodium phosphate (Na3PO4) but with some di-sodium phosphate (Na2HPO4) and even sometimes mono-sodium phosphate (NaH2PO4) blended in. It is now well known that phosphate will precipitate on boiler internals at high temperatures (in a phenomenon known as hideout), and where the old programs can produce low sodium-to-phosphate ratio deposits that directly corrode steel. At present, the most common phosphate program, in the U.S. at least, is the Electric Power Research Institute’s (EPRI) phosphate continuum (PC). Only tri-sodium phosphate (Na3PO4) is utilized, with perhaps a bit of caustic (less than 1 part-per-million [ppm]). While the recommended phosphate range is 0.2 ppm (absolute minimum) to 10 ppm, most chemists tend to control chemistry at the lower end of the range. This relates back to the phosphate hideout issue mentioned above and to the potential for solids carryover into steam, which we will soon examine. The typical pH range for this program is 9.0 to 10.0, with no more than 1 ppm of free caustic. A very critical aspect of these programs is that even minor in-leakage of contaminants from a condenser tube leak may introduce chlorides and sulfates to the boiler water. These contaminants, and chlorides in particular, can concentrate underneath deposits and initiate hydrogen damage of boiler tubes.

An alternative to phosphate is straight caustic treatment of evaporator circuits. Two features of caustic treatment stand out. One, the elimination of phosphate also eliminates phosphate carryover to the turbine, which has been a problem at numerous plants. Secondly, caustic is best at neutralizing chloride or sulfate salts that may enter the system. However, NaOH chemistry must be carefully monitored as caustic can directly attack boiler tubes, particularly under deposits.

Once-through steam generators must operate on some form of AVT, with OT being the preferred choice. OT is also possible for drum units, in which case the only other chemical feed besides oxygen is ammonia. A point to be noted about AVT, whether it be AVT(R), AVT(O), or OT, is that the chemistry offers no protection towards an influx of impurities, say from a condenser tube leak, makeup treatment system upset, or other source. Continuous, on-line chemistry monitoring is vital to detect problems and make quick corrections. I hope to also address this issue in more detail in a future Power Engineering article.

Steam Purity

The steam turbine is, of course, a highly-tuned piece of equipment that has to operate under high stress conditions. Even in the absence of ingress of catastrophic impurities, or concentrations thereof, the recirculating nature of drum boilers will cause an increase in dissolved solids within the boiler water. If these are not controlled properly, the solids will carry over with steam to foul downstream components. Carryover occurs by two mechanisms, either mechanical, where impurities enter the steam path via water droplets entrained in steam, or vaporous, where contaminants volatilize. Mechanical carryover typically results from poor boiler operation, improper chemistry control, or poor design/maintenance of the drum and its water/steam separating devices. Even in the best systems, a minor fraction of water will carry over with steam. Boiler pressure plays an important part in mechanical carryover, as the density of water and steam approach unity as the pressure approaches the critical point of 3208 psia. With vaporous carryover, the impurities volatilize, and are not entrained in moisture droplets as they enter the steam. Boiler pressure also is very important with regard to vaporous carryover issues, as higher pressure typically equates to greater carryover of these products.

magnetit dissolution

The list below outlines many of the major carryover products, and the difficulties they cause in the steam system. The list also includes common limits for reheat steam purity. These limits have been lowered over the decades, and may be lowered even more, as knowledge improves regarding the influence of impurities on turbine components.

  • Sodium phosphates. Phosphates carry over mechanically and settle out in the superheater and reheater, particularly the latter. Many plants have had to deal with overheating in reheater U-bends due to sodium phosphate deposition. Sodium limit, ≤ 2 parts-per-billion (ppb).
  • Chlorides. The majority of chlorides will carry over mechanically although some vaporous carryover may occur as ammonium chloride (NH4Cl) or even some hydrogen chloride (HCl). Chloride’s most notorious reputation in steam circuits is as a pitting and stress corrosion cracking agent in low-pressure turbines. Limit, ≤ 2 ppb.
  • Sulfates. Sulfate is another anion that will carry over as sodium sulfate or in AVT units as ammonium sulfate [(NH4)2SO4]. Volatile carryover is lower than from chlorides. Sulfate causes similar problems as chloride, but is not quite as nasty. Limit, ≤ 2 ppb.
  • Silica. Vaporous carryover is silica’s primary transport mechanism. Silica will precipitate in the high-pressure end of the turbine as steam pressure decreases during its path through the turbine blades. Silica forms a tenacious and hard deposit that is difficult to remove. Limit, ≤ 10 ppb.
  • Copper. Copper carries over vaporously, particularly in steam generators operating at 2400 psi and above. Like silica, copper precipitates on high-pressure turbine blades, where only a few pounds of deposition will alter the aerodynamics to the point that megawatts will be lost. Steam purity limits for copper are not standard, but a common limit for feedwater is 2 ppb.

The allowable concentration of impurities varies from unit to unit and is dependent upon operating pressure, system configuration, and other factors. While general charts and graphs are available to calculate maximum impurity control limits, proper monitoring of each unit is necessary to determine the limits for that particular steam generator.

This article only briefly touched upon many important issues related to high-pressure water steam chemistry. The reader should refer to the literature and consider attending events such as the Electric Utility Chemistry Workshop for additional information.


1. Guidelines for Controlling Flow-Accelerated Corrosion in Fossil and Combined-Cycle Plants, EPRI Report 1008082, Final Report, March 2005, the Electric Power Research Institute, Palo Alto, California.

2. Buecker, B., and S. Shulder, “The Basics of Power Plant Cycle Chemistry”; pre-conference seminar for the 27th Annual Electric Utility Chemistry Workshop, May 15-17, 2007, Champaign, Illinois, USA.

Many of the lessons learned from the coal-fired power industry are directly applicable to high-pressure heat recovery steam generators.


Brad Buecker is a contributing editor for Power Engineering and also serves as a process specialist with Kiewit Power Engineers, in Lenexa, Kan.