Modern concepts in Makeup Water Treatment

Water and steam are vital components of virtually all plants in the chemical process industries (CPI). And, a very important operation is the treatment of makeup water to produce the purity required for certain applications, most notably for use in heat exchangers and steam turbines. An important consideration in water treatment is that many industrial facilities are, or will soon be, dealing with tightening restrictions on wastewater discharges. These may include cooling tower blowdown, rainwater runoff and other aqueous discharges. As with other technologies, the processes utilized to produce high-purity water have been greatly improved over the last decades. This article examines several of the most modern techniques for makeup water production. These technologies are also being employed to reduce waste water volume and to recover pure water for reuse in the plant. (For more on reuse strategies, see Strategies for Water Reuse, Chem. Eng. September 2009, pp. 34–39.)

Think twice about a clarifier Typically, the general makeup-treatment process involves:
1) large solids removal by screens or settling basins;
2) suspended solids removal; and 3) dissolved solids removal to produce the water needed for the process.


We will focus on the last two process steps. Clarification and media filtration were commonly used in the past for suspended solids removal. Where lime softening is needed to reduce raw water hardness, clarification is still a viable process. However, when the primary focus is removal of suspended solids only, micro- or ultra-filtration offers a very reliable alternative. My own personal experience is with microfiltration as a clarifier/filter replacement for makeup water pretreatment to an 800- MW supercritical utility boiler [1]. All major microfilter designs utilize hollow-fiber membranes. A very common design uses pressurized systems, where thousands of spaghetti-like membranes are packed into pressure vessels. The number of pressure vessels then determines system production capabilities. This microfiltration process, like others of its type, operates in what can be thought of as a combination of crossflow and dead-end modes. Raw water flows parallel to the membrane surface in a crossflow pattern (Figure 1), but unlike reverse osmosis (RO) where most impurities are carried away with the concentrated stream (called the reject or concentrate), suspended solids collect on the membrane. Water that passes through the membranes and is purified is known as permeate and, of course, is sent onward, while in many designs the reject returns to the unit inlet. So, this begs the question of how suspended solids are purged from the system. Typically, a regular backwash with an accompanying air scrub dislodges particulates and discharges them to a waste stream. One microfiltration system has, in five years of operation, consistently produced water with turbidity levels below 0.04 nephelometric turbidity units (NTU). The results on downstream RO operation have been predictable in the best sense of the word, where the frequency of cartridge filter replacements went from weeks to months; and where microfiltration combined with an improved chemical-treatment program corrected a microbiological fouling problem that once plagued the RO membranes. An aspect to note is that this microfilter treats lake water in which turbidity rarely, if ever, exceeds 15 or 20 NTU. This is certainly not a highly demanding application, but microfiltration technology has advanced to the point that systems are now in place to treat river water where the turbidity sometimes exceeds 2,000 NTU. For applications like this, close monitoring is imperative. For applications where high suspended-solids loading is frequent, it may be necessary to employ clarification ahead of the microfilter to prevent excessive membrane fouling. For even finer particulate-matter control, ultrafiltration (UF) is an option. The minimum particle size screened by microfiltration (MF) is in the range of 0.05–5 microns, whereas for UF the range is 0.005–0.1 microns. UF will even remove viruses from water, and thus can serve in potable water applications. The common material of construction for MF and UF membranes is polyvinylidene fluoride (PVDF), which is quite chemically resistant. Accordingly, the membranes are very tolerant of oxidizing biocides, and can be aggressively cleaned for removal of fouling and scaling impurities.


Reverse osmosis: The first step in demineralization Before the development of reliable RO systems, the common makeup-water treatment process utilized ion exchange immediately after the clarification/ filtration step. Thus, the lead ion-exchange resins were subjected to water that contained roughly the original concentration of dissolved solids. The end result was restricted ion-exchanger run times and frequent regenerations with sulfuric acid and sodium hydroxide. Then RO came into the picture. Reverse osmosis is a filtration technique, but unlike conventional filtration, RO has the ability to remove dissolved solids down to the smallest ions. As the name RO indicates, the process utilizes pressure to force water through membrane materials, producing high purity permeate on one side of the membrane and a concentrated reject on the other. Although the passages within RO membranes are often referred to as pores, they more resemble very tiny maze-like tunnels. The passages range from one Angstrom (10–8 cm) to ten Angstroms in diameter. Because the pathways are so narrow, the molecular layer of water that attaches to the membrane surface inhibits ions from passing through the pores. Even so, some ions and molecules may still find their way through the RO membrane. These include monovalent ions, such as sodium and chloride, and small organic molecules. By far the most-common RO membrane design is the spiral-wound configuration (Figure 2). These membranes are manufactured in flat sheets, which are wound around a central core to produce a membrane element. Several elements are placed in series and are sealed in a pressure vessel (Figure 3). Within the vessel, the feed enters the forward end of each element and flows to the opposite end. Permeate passes to the central core of the element, while the concentrate is collected and discharged at the element end cap.


The feature of spiral-wound membranes that makes them most practical is the multiple wrap within an element. This allows an element to process much more water than would be possible through a flat sheet. The once-common standard element surface area was 365 ft², but advances in membrane technology now allow a larger surface area of perhaps 400 ft². However, larger surface area requires a smaller spacer size, which makes the element more susceptible to suspended solids fouling. Proper pretreatment becomes even more important in these applications. Regarding element loading in pressure vessels, a very common configuration for the large units at power plants and other industrial complexes is five or six elements per pressure vessel. An RO unit has been called nothing more than a high-pressure pump, some pressure vessels and pipe. In truth, the operation is a bit more complex than this description. Spiralwound- membrane elements can come in several different sizes, but the most popular size by far is 8-in. dia. by 40-in. length. The rate at which water passes through the membrane is known as the flux, and is measured in gallons per square foot per day (GFD). The general purity of the water in large part dictates the flux rate. A general guideline suggests the following flux rates:

• Surface water: 8 to 14 GFD
• Well water: 14 to 18 GFD

For normal surface and ground waters, each pressure vessel will produce about 50% purified water and 50% concentrated reject. This may not seem very efficient, but the concentrate is normally still pure enough to be treated again at another 50/50 split to produce 75% permeate. This design is the very common single-pass, twostage configuration (Figure 4).


Membrane technology has developed such that 99% or greater dissolved-solids removal is practical. Even higher purity can be obtained if the firstpass permeate is treated in a second pass, and, as we shall examine shortly, two-pass RO followed by a single ionexchange polishing stage is becoming increasingly popular. By far the most common material utilized for RO membrane construction is polyamide, which is always layered with other materials for structural support. These membranes are most often known by the name of thin film composite (TFC).

Pretreatment Proper pretreatment ahead of an RO membrane is vital to ensure reliable operation. Excessive concentrations of suspended solids will quickly foul RO membranes, and for this reason RO units are almost always equipped with inlet prefilters. The installation of a micro- or ultra-filter ahead of an RO element can greatly improve RO run times between membrane cleanings. Polyamide is quite intolerant toward oxidizing biocides, including chlorine and bromine, so these compounds must be removed with a reducing agent prior to membrane contact. The most common reducing agent is liquid sodium bisulfite (NaHSO₃), injected via a basic chemical-feed system. Pretreatment chemicals, such as those used in clarifiers, can negatively impact RO membranes. Coagulating agents of the cationic variety are particularly troublesome to RO membranes, especially to polyamide membranes whose surface is negatively charged. Aluminum gels that carry over from alum treatment in clarifiers can also cause problems. Such fouling is often overlooked until it occurs. For those who may eventually be tasked with specifying, purchasing, or installing an RO system, the importance of obtaining accurate influent water-quality data cannot be overemphasized. Ideally, historical data would be available. If not, then analyses should be collected as far in advance as possible of the decision to purchase an RO unit.

Scale formation Prevention of scale deposits is very important for RO operation. As permeate is produced by the successive membranes, the reject ion concentration continually increases as the water passes from element to element. This increases the scaling potential. Calcium carbonate, calcium sulfate and other compounds can build up to a point where precipitation begins to occur. Additional potential scales include silica and alkaline metal silicates, strontium sulfate, barium sulfate and calcium fluoride. While pretreatment will reduce the concentrations of many scale-forming compounds, the remainder may still cause problems. Barium- and strontium- sulfate scales are especially difficult to remove. Reputable membrane manufacturers have developed programs that will calculate the solubility limits for scale-forming compounds. The program warns the user if any solubility limit is exceeded. Because scale formation can quickly, and often irreversibly degrade membrane performance, anti-scalant feed is commonly used for industrial RO units. Common anti-scalant chemicals include polyacrylates and phosphonates. The correct anti-scalant or blend can control calcium sulfate at up to 230% above the saturation limit, strontium sulfate at 800% above the saturation limit, and barium sulfate at 6,000% above the saturation limit. The chemicals function by sequestering cations or modifying crystal growth, such that adherent scales do not form. Just a few parts per million of the treatment is usually sufficient to prevent abnormal scaling. (For more on scale inhibitors, see Biodegradation and Testing of Scale Inhibitors, Chem. Eng. April 2011, pp. 49–53.)

Polishing the water While single-pass, and especially twopass RO permeate will perform satisfactorily in low-pressure steam generators and many heat exchangers, an extra treatment step is necessary if the water serves as feed to high-pressure boilers or highly refined applications, such as those in the pharmaceutical or electronic industries. Single-pass RO plus a downstream cation-anion, mixed-bed ion-exchange unit is one such scenario. However, a technique whose popularity is increasing is two-pass RO followed by portable mixed-bed ion-exchange polishers. A primary motivating factor for this choice is that expensive and hazardous regeneration chemicals, such as sulfuric acid and caustic, may be eliminated from the plant site by having the mixed-bed “bottles” regenerated off-site by a contract supplier. Another polishing option that is gaining popularity due to technology improvements is electrodionization (EDI) for polishing. EDI is one of the newest developments in water purification. It is based on ion exchange principles (Figure 5), but there are two significant differences:

1. Membranes are used, but they are actually ion exchange materials in very thin, flat-sheet form
2. Regular mixed-bed resins are also utilized, but they are regenerated by electricity, not by acid and caustic


EDI is based on an older technology called electrodialysis (ED). With EDI, water is introduced into compartments that have a cation-exchange membrane on one side and an anion-exchange membrane on the other. Under the influence of an applied direct-current voltage, cations migrate through the cation-exchange membrane and anions migrate through the anion-exchange membrane. As the water flows down the chambers, the columns from which the ions migrate become progressively demineralized, while water in the remaining columns becomes increasingly concentrated. The product is delivered to the process, while the concentrate discharges to waste. One of the initial shortcomings of the old ED technology was its limitation in removing silica and CO₂, even though the anion membrane is a strong base exchanger. This deficiency is overcome in EDI by the mixed-bed resins within the compartments. Another ion-exchange option — especially useful when space is limited for equipment installation — is known generically as short-bed demineralization In this option, compact exchangers use very fine mixed-bed resin for high-purity water production. These units typically operate on short run times of perhaps 20 minutes or so, and then are automatically regenerated. Because the resin bed completely fills the compartment, thus prohibiting backwash for solids removal, the feed to these ion exchangers must be quite free of suspended solids. A number of my colleagues in the power industry have installed or operated these systems, some with limited success and some with excellent results. Good pretreatment is a key issue.

---

Reference: 1. Buecker, B., Microfiltration: An Up and Coming Approach to Pre-Treatment for the Power Industry, presented at the 26th Annual Electric Utility Chemistry Workshop, Champaign, Illinois, May 9–11, 2006.
Author: Brad Buecker is a process specialist with Kiewit Power Engineers (9401 Renner Boulevard, Lenexa, KS 66219; Phone: 913-928-7000; Fax: 913-689-4000; Email: brad. [You must be registered and logged in to see this link.]).