In the early 1980s, researchers had concluded that dissolved oxygen (DO) ingress into condensate/feedwater was a prime factor for carbon steel corrosion (and unarguably corrosion of copper alloys) during normal operation of utility steam generators.

So, both mechanical deaeration and chemical methods (i.e., oxygen scavenger feed) were typically utilized to reduce DO concentrations to near zero as measured at the economizer inlet of conventional fossil-fired utility boilers. Oxygen scavenger (the more correct term is reducing agent) treatment combined with ammonia or a neutralizing amine for pH control came to be known as all-volatile treatment reducing [AVT(R)]. The chemistry induces formation of the familiar gray-black magnetite (Fe3O4) layer on carbon steel surfaces.

In 1986, the foundations of this chemistry received a severe jolt.

 “On December 9 of that year, an elbow in the condensate system ruptured at the Surry Nuclear Power Station (Virginia),” according to a 2005 report from the Electric Power Research Institute (EPRI). “The failure caused four fatalities and tens of millions of dollars in repair costs and lost revenues.”

Researchers learned from that accident and others since then that the reducing environment produced by oxygen scavengers is the prime ingredient for single-phase flow-accelerated corrosion (FAC) of carbon steel. The attack occurs at flow disturbances such as elbows in feedwater piping and economizers, feedwater heater drains, locations downstream of valves and reducing fittings, attemperator piping; and, most notably for combined-cycle heat recovery steam generators (HRSGs), in low-pressure evaporators, where the waterwall tubes, aka harps, have many short-radius elbows. In fact, FAC is typically the leading on-line corrosion mechanism in HRSGs.

Gradual metal loss occurs at these locations, as is illustrated in the figure below:

Figure 1: Photo of tube-wall thinning caused by single-phase FAC.

Wall thinning progresses until the remaining material at the affected location can no longer withstand the process pressure, with sudden failure as the end result.

Figure 2: Catastrophic failures induced by FAC. Photos courtesy of Steve Shulder, Electric Power Research Institute (EPRI), per the 35^th Annual Electric Utility Chemistry Workshop

Two other factors, temperature and pH, also strongly influence FAC.

Figure 3: Influence of temperature and pH on iron dissolution from carbon steel. (Photo courtesy of Shulder, per the 35^th Annual Electric Utility Chemistry Workshop)

The temperature aspect is why FAC is typically most pronounced in the feedwater system and LP evaporator of HRSGs. Based on the method of oxygenated treatment (OT) that arose in Europe in the early 1970s, EPRI developed a program to replace AVT(R) for drum units, known as AVT(O), which stands for all-volatile treatment oxidizing. If the condensate/feedwater system contains no copper alloys, which is true for virtually all HRSGs, then AVT(R) is not recommended, rather AVT(O).

In brief, with AVT(O) chemistry the oxygen scavenger feed is eliminated, and a small residual concentration [5 to 10 parts-per-billion (ppb)] of dissolved oxygen is maintained at the economizer inlet. Ammonia or an ammonia/neutralizing amine blend is still utilized for pH control. High-purity condensate (cation conductivity ≤0.2 μS/cm) is a requirement for AVT(O), but when proper conditions are established, magnetite becomes overlaid and interspersed with a tighter-bonding oxide, known variously as hematite, or ferric oxide hydrate. It is noticeable for its distinct red color.

Figure 4: Properly passivated surfaces in a unit with AVT(O). (Photo courtesy of Dan Dixon, Lincoln Electric System)

Also gaining momentum is use of film-forming products (FFP) to protect all metal surfaces from both on-line and off-line corrosion. A discussion of FFP chemistry is beyond the scope of this article, but the subject will be reviewed in more detail at the 39th Annual Electric Utility Chemistry Workshop, June 4-6 in Champaign, Illinois. For more information, go to www.conferences.illinois.edu/eucw.

The key point of this article, is that whether a unit is on AVT(O), or AVT(R) for those older, conventional units with copper-alloy feedwater heater tubes, or film-forming chemistry, monitoring of condensate/feedwater iron concentrations is key towards proper chemistry control and minimization of FAC, including another mechanism known as two-phase FAC that occurs in certain locations of both HRSGs and conventional units where a water/steam environment co-exists.

A Key to Success is Iron Monitoring

For years, EPRI-established guidelines suggest that with proper chemistry, the total iron concentration in boiler feedwater can be maintained below 2 parts per billion, even for those units on AVT(R). But if FAC is underway, the iron concentration is typically well above that value. Thus, regular iron analyses are critical for establishing and maintaining the correct chemistry to control FAC. A complicating factor, however, is that typically 90 percent or greater of the steel corrosion products exist as particulate iron, with only a small fraction of dissolved iron. So, any analyses obviously need to account for the particulate iron. Two common techniques are on-line particle counting, and corrosion product sampling. The latter incorporates both filtration and ion exchange to capture particulate and dissolved metals, which are then analyzed to determine the metal concentrations. Common is to collect a sample for a week or so, and then have the filters and ion exchange resin analyzed for results. While these techniques have been successful, other methods have emerged that offer good results at reasonable cost.

Simple colorimetric lab methods have traditionally been used to monitor dissolved iron contamination. The common colorimetric method for dissolved iron is based on the extremely sensitive ferrozine-ferrous iron complex described by Stookey in his 1970 piece for Analytical Chemistry. Ferrozine complexes with dissolved ferrous iron to form an intensely colored purple complex. The dissolved ferrous iron concentration may be determined by measuring the absorbance of this complex. Modifications of this traditional method now allow for the determination of both dissolved iron and particulate iron oxides at very low concentrations. (Kuruc, K., Johnson, L. “Further Advances in Monitoring Low-Level Iron in the Steam Cycle.” PowerPlant Chemistry, 2015)

The reductive dissolution of iron oxides via thiol-containing compounds has been thoroughly investigated by Waite et al. Thioglycolic acid (TGA) has been used to successfully dissolve and reduce various iron oxides. While magnetite is dissolved relatively easily with TGA, hematite has been shown to be much more resistant to this method, according to work by Waite, Torikov and Smith and by Baumgartner, Maroto and Blesa.

However, TGA is compatible with the sensitive ferrozine reagent and is commercially available as a combined reagent. This combination digestion-reduction-detection reagent is particularly useful for simplifying analysis and minimizing contamination.

Figure 5: Combination reagent, digestion vials and heater block (left); 1” sample cell (center) and spectrophotometer (right)

Complete dissolution of particulate magnetite and hematite can be achieved with a 135°C, 30 min closed vessel digestion using 240 μL of combination reagent and 12 mL of sample. The digestion is carried out in a 20 mL glass vial heated in an aluminum block. After the sample has cooled the absorbance is read with a spectrophotometer and a 1” cell. The calibrated range using this procedure is 1-100 μg/L with a method detection limit (MDL) of 0.3 μg/L.

A combination of a simple colorimetric total iron laboratory analysis with a sensitive laser nephelometric analyzer can also provide a method for cost effective, quantitative, real-time corrosion monitoring. When properly calibrated, the nephelometric units provided by a nephelometer can be correlated to total iron concentration values. The iron concentration of the process water is a direct indicator of steel corrosion.

As the process waters used in power generation are extremely pure, it can be assumed that almost all insoluble matter present in a ferrous metallurgy process stream is due to steel corrosion in the form of particulate or colloidal iron oxides. Corrosion of steel components in power generation is generally found as iron oxides and hydroxides, primarily, iron (II, III) oxide (magnetite), α-iron (III) oxide (hematite), or dissolved iron. Each of these species produces a different nephelometric response to visible light. Black magnetite absorbs more and reflects less light than red hematite. Dissolved iron does not produce any nephelometric response.

Figure 6: Suspended particles of magnetite (black) and hematite (red) in water

Corrosion products range in size from sub-micron to 10 μm in diameter, with an average diameter of 1 μm (Kuruc and Johnson, 2015). This range of diameters poses another challenge for particle monitoring because nephelometers respond differently to different particle sizes.

The variables associated with iron corrosion products (species, color, particle size) make it impossible to create a universal nephelometric calibration for quantification of corrosion products. A nephelometric calibration which is appropriate for a particular sample location with particular corrosion characteristics will not be accurate for a different location with different corrosion characteristics.

Therefore, quantification of total iron via nephelometry must be accomplished through site-specific calibration. Variability in water chemistry, phase, and local piping configurations contribute to variable corrosion characteristics. Site-specific calibration ensures that nephelometric response is correlated to the specific corrosion characteristics present at each installation.

Figure 7. Laser nephelometer mounted on a water panel/view of the sample cell