Carbonic Acid Corrosion

A common issue in condensate systems is carbonic acid corrosion. Carbon dioxide that enters systems through air in-leakage or comes from carryover of decomposed bicarbonate alkalinity with steam will dissolve to form carbonic acid. 

CO₂ + H₂O ⇌ H₂CO₃                                                   (1)   

While H₂CO₃ is not a strong acid, it can cause considerable damage, as evidenced in Figure 1.

Not only is direct corrosion a result, but iron oxide particles generated by the attack will travel to steam generators and precipitate on boiler tubes. The precipitates increase the potential for under-deposit corrosion. 

Alkalizing (formerly known as neutralizing) amine injection is a straightforward method for pH adjustment in industrial boiler feedwater and condensate systems. The amines are small-chain organic molecules with an ammonia group attached to or embedded within the compound. The most well-known are shown in Figure 2.

Factors that influence amine properties and efficacy include volatility, basicity and, in high-pressure boilers with superheaters and reheaters, the tendency to decompose at high temperatures.³ Every application must be evaluated individually to select the correct program for protecting both the water and steam sides of the steam-generating network. In some cases, the amine concentration in steam may be limited to a low parts-per-million (ppm) level per concerns about the potential for contamination of food or consumer products. Consultation with a reputable water-treatment vendor is critical for selecting the correct treatment program.

Reference 4 provides a real-world example of some of these improvements, made at a major college campus (with two hospitals), which included installation of alkalizing amine feed systems for condensate return pH conditioning and replacement of makeup treatment sodium softeners with reverse osmosis (RO) units.

In some cases, ion-exchange polishing can provide the required protection for condensate return streams.

IX condensate polishing is an absolute requirement for some facilities such as supercritical power plants, while other applications may require specialty resins to remove unusual ions.

Dissolved Oxygen on the Attack

Oxygen attack in industrial water systems induces localized corrosion that can generate through-wall penetration of tubes, piping, and vessel walls. 

In water networks, the rust may deposit at corrosion sites to form tubercles that restrict flow and increase corrosion rates underneath the deposits.

Obviously, corrosion control is immensely important to maintain equipment and performance reliability. The root-cause method to prevent attack is to eliminate oxygen ingress to steam/condensate systems. But with the extensive piping and myriad equipment connections typical at large plants, complete elimination of air ingress is often a practical impossibility. Two primary options to protect carbon steel and copper alloys are:

• Mechanical and chemical methods to remove dissolved oxygen
• Chemical programs to treat metal surfaces and reduce the corrosion potential 

For boiler feedwater networks, the mechanical control device is the deaerator.

Makeup water and return condensate enter near the top of the deaerating compartment, and, via a series of headers/nozzles, the condensate is sprayed counter-currently onto steam injected from below. Frequently, the compartment will include trays to enhance steam-water contact. The deaerator and deaerator storage tank are typically placed far above the boiler feed pump to provide adequate suction head for the pump and minimize cavitation. A properly operating deaerator will bring the condensate temperature to within a few degrees of saturation, with a standard goal of 7 parts-per-billion (ppb) D.O. in the deaerator storage tank.  

Supplemental oxygen scavenger/reducing agent feed is typical to reduce D.O. concentrations to near zero and to passivate metal surfaces. Some of the most common compounds include catalyzed sodium sulfite (Na₂SO₃), carbohydrazide [(NH₂NH)₂CO], diethylhydroxylamine [DEHA, (C₂H₅)2NOH] and erythorbate. Factors that influence the choice of reducing agent include boiler pressure, ease of application, and safety. 

For example, sodium sulfite is an inexpensive, non-toxic compound, but it adds dissolved solids to boiler water, and, at pressures above 600 psi, begins to decompose into corrosive products that include sulfur dioxide (SO₂) and hydrogen sulfide (H₂S). It is also not a good passivating agent. Carbohydrazide is a volatile oxygen scavenger/reducing agent; developed as a replacement for hydrazine (N₂H₄), which was straightforward and effective to use but became a suspected carcinogen. Carbohydrazide is a safe-to-handle chemical that breaks down to hydrazine with rising temperature in the feedwater system. 

At plants where steam could potentially contact food or consumer products, many of the common reducing agents are prohibited. A potential alternative is erythorbate, where erythorbic acid has an almost identical structure to ascorbic acid, Vitamin C.⁶ 

Another approach, which is enjoying a renaissance due to improved chemistry, is the use of film-forming products that directly protect metal surfaces.

Influence of Process Impurities

A number of years ago, I visited an organic chemicals plant with a colleague. The plant had four 550 psi package boilers with superheaters. The steam provided energy to multiple plant heat exchangers, with recovery of most of the condensate. Each of the boiler superheaters failed, on average, every 1.5–2 years from internal deposition and subsequent overheating of the tubes. Our inspection of an extracted superheater tube bundle revealed deposits of approximately ⅛–¼ inches in depth. 

Additional inspection revealed foam issuing from the saturated steam sample line of each boiler.  The cause of the foam formation became quickly apparent, as water/steam chemistry analyses performed by the facility’s water treatment vendor included data showing total organic carbon (TOC) concentrations of up to 200 mg/L in the condensate return. Contrast that with the <0.5 mg/L feedwater TOC recommendation from the American Society of Mechanical Engineers.⁸

No treatment processes or condensate polishing systems were in place to remove these organics (five phenol derivatives) upstream of the boilers.

Based on the TOC data alone, it became quite understandable why foam was issuing from the saturated steam sample lines and why the superheaters rapidly accumulated deposits and then failed from overheating. Furthermore, the facility’s rudimentary water laboratory did not have the instrumentation for differentiating between the five compounds, making it more difficult to find the source of leaking equipment.

The frequent superheater replacements represented a reactive not preventive approach. Admittedly, proactive solutions would have been expensive, starting with a root-cause effort to find leaks and make repairs. Condensate dumping would have required retrofit of a much larger makeup water production system and an upgrade to the plant wastewater-treatment system. Activated carbon filtration, which is quite common for makeup water treatment may have been viable, per issues related to molecular characteristics of the impurities and reaction kinetics, but laboratory and pilot testing were warranted, in our opinion. Several raw materials serve as the starting source for activated carbon, including coal, coconut shells and others. One or more of these materials may have been effective as determined by test results. We were not involved in any test development because it became clear that management was looking for a “magic pixie dust” solution and did not want to spend the money, at least at that time, for thorough investigation. 

Other Return-Condensate Impurities

While the condensate impurities in the above case history were all within the same chemical family, a much larger variety may exist at other facilities such as refineries and petrochemical plants. The figure below outlines, in schematic format, typical refinery unit operations.