Boiler Feed Water—Reducing Scale and Corrision, Part 1 of 2
By C.F. “Chubb” Michaud, CWS-VI
Summary: A big part of ion exchange’s use in commercial/industrial applications involves boiler feed water. This series takes a look at treatment methods using salt regeneration in Part 1 and more advanced chemical regeneration and treatment techniques in Part 2.
While taste, color and odor are the primary customer concerns with regard to residential water treatment, water quality is more than a matter of aesthetics when it comes to commercial/industrial uses.
What is clean water?
The truth is, our municipal water is safe for the most part and enhancement is a matter of taste. However, the human body is far more tolerant of many “impurities” contained in our water than are many industrial applications that use water in their manufacturing processes. This series addresses some of the limitations of raw water for use as boiler feed and common methods of treatment for reduction of scale and corrosion. Discussions are limited to treatment by ion exchange.
Water treatment needs
Manufacturers who heat or cool water in their processes soon discover strange things can happen to water when solubility parameters of salts and gasses it contains are exceeded. Ice becomes cloudy, cooling loops fill with sludge, soap curdles, boilers develop an insulating layer of scale, condensers plug with corrosion and foods develop an off taste and color. Most, if not all, industrial processes require some form of water treatment for system efficiency.
Volumes have been written on the various methods of pre-treating industrial water and none are complete. This article addresses only a few treatment processes—namely softening, dealkalizing and desilicizing raw feed water. The techniques described in this article are also applicable to laundries, ice and beverage production, food processing and others. It basically addresses removal of calcium, magnesium, alkalinity and silica to prevent scale and corrosion.
Causing the problem
The two most important reasons for pre-treating boiler feed water are for reduction of scale and prevention of corrosion. Scale is caused by the precipitation of hardness and/or silica that’s present in feed water. Corrosion is generally caused by presence of alkalinity that converts to carbonic acid in the steam. While corrosion can also stem from dissolved oxygen and/or other acid formers, we won’t address those here.
Scale formation can stem from temporary hardness—that is 1) calcium, magnesium or ferrous bicarbonate; 2) permanent hardness, such as from calcium sulfate; or 3) hardness caused by precipitation of silica (see Reactions 1, 2&3).
These are explained as follows. Calcium and magnesium bicarbonate decompose to carbonates upon heating. Calcium sulfate di-hydrate (natural gypsum)—which has fair solubility in cold water (2,400 milligrams per liter or mg/L)—can precipitate out as gypsum scale in hot water above 100oF as it converts to the anhydrite form. Silica, which exists as a hydrated substance at room temperature, can form a glassy-like precipitate at elevated temperatures. Silica can also volatilize and carry over with steam.
Further, carbonic acid (from Reaction 1) volatilizes as CO2 gas and water, which can re-condense as carbonic acid in the steam (see Reaction 4). Even softened water containing alkalinity (as bicarbonate) will produce corrosive steam, although the resulting Na2CO3 won’t form scale (see Reaction 5). The carbonate salt can further react to form the hydroxide and CO2 (see Reaction 6).
If we consider the purpose of treating boiler feed is to produce high purity steam, we can more readily appreciate that the higher the pressure and temperature the more likely we are to carry over contaminants to the steam. In addition, we have more energy invested in producing higher temperature steam and would like to minimize the “blow down”—dumping part of the boiling water to reduce total dissolved solids (TDS)—to save money. It follows that the higher the boiler pressure the more we must limit the contaminants in the feed water. Keep in mind that boiling water can concentrate residual salts by a factor of 20:1 or higher. If we can only tolerate 1 part per million (ppm) of a particular contaminant in the boiler, we can only tolerate 0.05 ppm in the feed to the boiler.
Table 1 demonstrates the guideline requirements recommended by the American Boiler Manufacturers Association—see www.abma.com—for boiler feed water based on a blow down of 5 percent (20:1 concentration):
It can be noted that low and medium high-pressure boilers will not require complete demineralization. Scale control via softening or chemical injection with dealkalization will usually suffice. Silica can also be controlled by chemical addition or desilicization by ion exchange.
Hardness & alkalinity sources
Falling rain and surface waters absorb carbon dioxide and other gasses to form dilute acids (see Reactions 7 & 8). As this water percolates into the soil, it will dissolve minerals and build in TDS (see Reaction 9).
Alkalinity represents acid neutralizing or buffering capabilities in water. It can come from the presence of CO3-2 or OH– ions as well as HCO3– in water. With the exception of HCO3–, alkaline salts of divalent ions such as Ca+2 and Fe+2 are almost completely insoluble. Therefore, the alkalinity we experience in most feed water with a pH of 6.5 to 8.5 will be HCO3– if hardness is present. Natural water rarely contains hydroxyl alkalinity (OH–). Depending upon pH, the relationship between CO2, HCO3– and CO3-2 changes (see Figure 1).
The hardness associated with bicarbonate alkalinity is termed temporary hardness. While this may sound like a minor problem, it isn’t. When this water is heated, the HCO3– ion decomposes to CO2 and CO3-2 ion and the CO2 goes off with the steam and becomes carbonic acid (see Reaction 1). The carbonate that’s left behind precipitates as insoluble calcium carbonate scale directly onto the heating surfaces. You now have a scaled up boiler and a corroded heat exchanger (but very little soluble hardness in the blow down). Needless to say, this isn’t a good thing. Scale buildup in any heat transfer vessel can reduce the heat exchange efficiency by 50 percent or more. In hot water heaters, the build up of scale will reduce the heater volume over time.
Silica comes from the partial breakdown of natural alumino-silicates contained in the soil. Silica is more soluble in hot water than in cold and generally doesn’t present a problem in water heaters or low-pressure boilers. At elevated temperatures, however, silica actually volatilizes with the steam. It can then deposit in condensers or on turbine blades causing mechanical problems and costly shutdowns. The amount of silica contained in the steam becomes significant, above 600 pounds per square inch (psi) of steam pressure. Silica is less soluble in low TDS waters than in high TDS water, thus a boiler requiring deionized (DI) water—generally above 1,000 psi—may also have limits for silica lower than might otherwise be expected.
There are both economic as well as efficiency reasons to soften and dealkalize boiler feed water. Unless there’s a need for TDS or silica reduction along with hardness and alkalinity reductions, simple salt regenerated ion exchange provides a complete and economical choice. Silica, however, cannot be removed with a salt regenerated system.
Hardness removal uses a strong acid cation exchanger. Sizing of the softening unit will vary with the flow rate and desired run length (see Reaction 10). Calculate grain removal capacity as: gpm × gpg × 60 × hrs = grains removal capacity between cycles. Divide this number by the capacity of the softening resin to get the number of cubic feet of resin required (see Reaction 11).
The ability of a softener to produce very low hardness leakage will depend upon the TDS and regeneration salt level. Ten pounds of salt will produce 1 ppm of hardness leakage in water up to about 700 ppm in TDS, and 5 ppm leakage in water up to about 1,600 ppm in TDS. Fifteen pounds or more can be used for lower leakages or to treat water at higher TDS (see Figure 2).
Co-current strong acid cation (SAC) softeners are limited to TDS in the feed of about 3,000 ppm but can be run effectively at TDS in excess of 5,000 ppm with high brine doses—30 pounds per cubic foot (lbs/ft3) or more.
Counter-flow regenerated SAC allows for more effective usage of brine and produces lower leakage in higher TDS feed water. Because of the complexity and expense of a proper counter-current system, units are often designed with only the polisher of a two-in-tandem softener design. This is referred to as series softening, which utilizes a co-flow primary and a counter-flow polisher. This design has been effectively used for softening “produced waters” from oil field steam floods that are 5,000+ TDS and deliver leakages of less than 1 ppm hardness (see Figure 3).
Strong brine at 15-to-18 percent is pumped into the bottom of the polisher. A blocking flow of primary softened water is pumped to the top. The partially spent and diluted brine is taken off at the regenerant collector (now at about 10 percent strength) and pumped to the top of the primary. The total amount of brine is based on a stoichiometric quantity for the polisher plus the normal level for the primary. The polisher, however, sees all the brine, which is what produces very low leakages. Additional efficiency benefits can be gained by running both the primary and the polisher in a counter-flow mode. Most of the benefit of counter-flow regeneration is lost if the resin bed isn’t held in place during brining.
Hardness salts don’t form scale unless they have the appropriate counter-ions present (CO3-2, SO4-2). The process of using a salt-regenerated strong base anion exchanger to remove those ions has been termed “anion softening.” Here, strong base anion (SBA) resins—usually a Type II—in the chloride form will exchange bicarbonates and carbonates (alkalinity) along with sulfates for chlorides (see Reaction 12).
When an anion softener is used purely for the reduction of alkalinity, it’s referred to as a dealkalizer. Here, our service reaction involves only the removal of bicarbonates, although CO3-2 and SO4-2 will still be removed (with reactions similar to Reaction 12). Capacity can be enhanced by adding a small amount of caustic soda (NaOH) to the brine. A comparison of capacities for a typical Type II SBA is shown in Figure 4.
Use of a small amount of caustic with regenerant brine will improve performance of a dealkalizer by elevating pH slightly during service. Some HCO3– will convert to CO3-2, which is picked up better by the chloride form SBA resin. Feed water should be pre-softened if your design includes use of caustic with brine for regeneration.
Inclusion of SAC and SBA in the same vessel for simultaneous reduction of hardness and alkalinity in a single tank, although commonly done, is not recommended. The waste regenerant, which will be high in both hardness and alkalinity, will surely precipitate and cause fouling. The same holds true if the same regenerant is used for regenerating two separate vessels simultaneously. Acidified brine has been successfully employed. If the regenerant pH is held below 5.5 with addition of citric acid (1.0 lbs citric acid/100 lbs of salt), CO3-2 ions will be converted to HCO3– and carbonate precipitation can be avoided. HCO3– to SO4-2 ratios should be at least 10:1. Otherwise, CaSO4 precipitation may still occur (see Reaction 2).
This article dealt with boiler feed water treatment and scale prevention methods using ion exchange that involve salt regeneration only. Part 2 of this series next month will review use of weak acid resins, de-carbonators and silica removal.
About the author
C.F. “Chubb” Michaud, CWS-VI, holds bachelor’s and master’s degrees in chemical engineering from the University of Maine and has more than 30 years of professional experience in water and fluid treatment processes. Michaud is technical director for Systematix Inc. of Buena Park, Calif. He also is chairman of the Water Quality Association’s Ion Exchange Task Force, sits on the Science Advisory Committee and is a founding member of the WC&P Technical Review Committee. Michaud can be reached at (714) 522-5453, (714) 522-5443 (fax) or email: cmichaud@systematixUSA.com