Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
4 State of the Technology Desalination technologies and their application have evolved sub- stantially over the past 50 years. The five key elements of a desalination system (see Figure 4-1), for either brackish water or seawater desalina- tion, are as follows: 1. Intakesâthe structures used to extract source water and convey it to the process system; 2. Pretreatmentâremoval of suspended solids and control of bio- logical growth, to prepare the source water for further processing; 3. Desalinationâthe process that removes dissolved solids, primar- ily salts and other inorganic constituents, from a water source; 4. Post-treatmentâthe addition of chemicals to the product water to prevent corrosion of downstream infrastructure piping; and 5. Concentrate managementâthe handling and disposal or reuse of waste residuals from the desalination system.1 Depending on the source water and the desalination technology used, specific elements may vary in their importance in the overall system. For example, inland brackish groundwater desalination facilities will use wells and pumps to bring the source water to the facility, and these sys- tems may need little or no pretreatment. In contrast, seawater reverse osmosis (RO) desalination may use more elaborate intake structures, de- pending on the specific site conditions, and may require extensive pre- treatment. The state of the technology for each of these elements and 1 The broader term concentrate management (as opposed to concentrate disposal) is used throughout the report in recognition that waste management options are not limited only to disposal or discharge. However, it is worth noting that few economically viable con- centrate reuse applications currently exist and that nearly all desalination concentrate in the United States is disposed rather than beneficially reused. 59
60 Desalination: A National Perspective Seawater or brackish water intakes Post-treatment Pre-treatment Desalination step: Membrane modules Concentrate discharge FIGURE 4-1. Key elements of a desalination system. Although shown here for a membrane-based system, these steps also describe the major components of non-membrane systems. SOURCE: Modified from Buros et al. (1980). their current technical barriers are discussed in this chapter. The focus of the chapter is on technologies that are commercially available or in the late stages of development, although some emerging technologies that are still in the early phases of research and development are discussed in boxes. FEEDWATER INTAKE OPTIONS Desalination facilities require a reliable supply of feedwater. Feed- water quantity and quality vary based on the specifics of the site and of- ten determine the feasibility of siting a plant at a given location. Intake designs can affect feedwater quality and the environmental impacts of a desalination facility at a given site. Current technologies and issues with desalination intakes are discussed in this section. Brackish water desalination facilities can utilize feedwater from sur- face water sources or wells. Inland desalination plants use intake tech- nology that is no different from traditional water-treatment plants de- pendent on surface water or groundwater, and this technology is well developed. Therefore, these technologies will not be described in detail here. There are important environmental issues, however, associated with
State of the Technology 61 sustainable brackish groundwater withdrawals for inland systems that are discussed further in Chapter 5. Seawater desalination intakes generally fall into one of two major categoriesâsurface intakes located above the seafloor and subsurface intakes located beneath the seafloor or sandy beach. Surface intakes, of- ten called open intakes, are in direct communication with the ocean or sea, where feedwater can be taken directly from the surface or below the surface through submerged intakes. Subsurface intakes take their feed- water from below the floor of the ocean using naturally occurring sand and geologic formations to provide filtration. Subsurface intakes can be horizontally drilled from central wells, slant drilled from onshore beaches, or excavated to create infiltration beds. Subsurface intake op- tions can produce higher-quality feedwater and thereby reduce the pre- treatment necessary for membrane desalination systems. In contrast, thermal seawater desalination systems require less pretreatment than RO and require only coarse screens to protect the process equipment. Ther- mal desalination plants commonly use surface (open) intakes. Design engineering, equipment procurement, and construction spending on intakes and outfalls are estimated to total 5 to 7 percent of capital costs for RO and thermal desalination plants (GWI, 2006a). Other costs, such as monitoring and permitting, may add to the overall costs of the intakes. There are several factors affecting the final cost for con- structing and operating an intake system; among these are the type of intake being used, the type of coastal conditions, and the distance from the intake to the plant itself. The following section focuses on the latest design and engineering options for coastal intakes. Surface Intakes Thermal seawater desalination and large seawater RO facilities (>38,000 m3/day; >10 million gallons per day [MGD]) predominantly use open-water intakes. Screens are added to the intake structures to re- duce the number of marine organisms taken in with the source water (re- ferred to as entrainment; see also Chapter 5). Application of screen tech- nology to the power industry has existed since the early twentieth cen- tury. Early screens included a front-end âtrash rackâ consisting of fixed bars to prevent large debris from entering the water intake system. Trav- eling screens are rotated and washed intermittently with a high-pressure wash. Alternative screen technology includes modified traveling screens with fish handling systems, fine-mesh screens, cylindrical wedge wire screens, fish net barriers, louvers, angled traveling screens, and velocity caps (California Water Desalination Task Force, 2003).
62 Desalination: A National Perspective Impingement, defined as the pinning and trapping of fish or other larger organisms against the screens of the intake structures, can cause severe injury and death to organisms. State-of-the-art intake systems have been developed to greatly reduce impingement. For example, some intake screens can be backflushed with compressed air. These screens have no moving parts, operate with a very low velocity (to mitigate im- pingement), and are generally referred to as âpassive screens.â A Ris- troph screen is a modified traveling screen with water-filled lifting buck- ets that collect impinged organisms and transport them to a bypass, trough, or other protected area. Similarly, fish baskets consisting of framed screen panels can be attached to the vertical traveling screens. Fish that are removed are typically returned to the water via sluiceway or pipeline. Fine-mesh screens have a mesh size of 5 mm or less and are designed to exclude larger eggs, larvae, and juvenile fish from the in- takes. Cylindrical wedge wire screens will exclude organisms larger than the nominal screen opening of 0.5 to 10 mm. Their cylindrical shape dis- sipates the velocity, allowing organisms to escape the flow field, al- though adequate countercurrent flow is needed to transport organisms away from the screen (California Water Desalination Task Force, 2003). An emerging screen technology that is currently in development to ad- dress impingement and entrainment is described in Box 4-1. Louvers are a series of vertical panels placed perpendicular to the in- take approach flow. They create a new velocity field that carries fish away from the intake and toward a fish bypass system. Louvers rely pri- marily on a fishâs ability to recognize the new flow field and swim away. They have been successful in reducing impingement but are not effective against the entrainment of eggs and larvae (California Water Desalina- tion Task Force, 2003). Shipboard seawater desalination approaches that situate the water treatment facility in the deeper ocean far from environmentally sensitive coastal areas could also reduce impingement and entrainment. One recent approach uses telescoping source water intakes to bypass the photic zone, where most marine organisms reside. This deeper water also con- tains fewer suspended solids, thus reducing the pretreatment required. Subsurface Intakes Coastal subsurface intakes include beach wells, radial wells, horizon- tal directionally drilled (also called slant-drilled) wells, and infiltration galleries. By taking advantage of the natural filtration provided by sedi- ments, subsurface seawater intakes can reduce the amount of total or- ganic carbon and total suspended solids, thereby reducing the pretreat-
State of the Technology 63 BOX 4-1 Emerging Technologies to Reduce Impingement and/or Entrainment A Marine Life Exclusion System has been designed to reduce impingement as well as entrainment. This water-permeable barrier (see Figure 4-2) is spread around the intake structure, preventing aquatic organisms from approaching the water intake point. The curtain is either suspended by flotation billets and an- chored in place, or integrated into existing shoreline intake structures. Sealed against the seafloor and shoreline structures, it completely surrounds the intake structure, preventing targeted planktonic and neustonic organisms from entering the system. Because the surface area of the curtain is large compared to an in- take screen, the water velocity through the curtain is up to 98 percent less than the velocity near the intake structure. Low water velocity enables even small fish larvae to drift away from the boom. This technology has primarily been used in riverine environments, although it is currently being tested in marine settings to examine its durability, susceptibility to fouling, and cleaning requirements (McCusker et al., 2007; Mirant Lovett, 2006; San Francisco Bay Conservation and Development Commission, 2005). FIGURE 4-2. Curtain designed to reduce intake velocities and minimize im- pingement and entrainment. SOURCE: McCusker et al. (2007) ment required for membrane-based desalination systems and lowering the associated operations and maintenance costs. Pumping from subsur- face intakes may also under some conditions dilute the seawater with less saline groundwater, thereby reducing the total dissolved solids (TDS) in the intake water. Vertically drilled beach wells are typically used for small (<19,000 3 m /day; <5 MGD) systems where the local hydrogeology (e.g., aquifer transmissivity) will permit it. Beach wells have been used effectively in
64 Desalination: A National Perspective the Caribbean and Mediterranean and are the intakes of choice for pro- posed plants in Hawaii. They pose minimal environmental concerns be- cause benthic communities remain undisturbed and entrainment and im- pingement of marine organisms are eliminated. In most cases, no further pretreatment is needed. One of the potential disadvantages of beach wells is that deep wells may result in lower water temperature and thus higher viscosity; hence, higher pressure (and increased energy) will be required to pump the water through the RO membranes. Care should also be taken to ensure that the source water withdrawals do not cause deleterious ef- fects on local aquifers. Horizontal directionally drilled (or slant-drilled) wells, shown in Figure 4-3, are increasingly being considered for use in large seawater desalination facilities. Although more expensive to construct than beach wells, they can minimize shoreline structures. Slant-drilled wells are un- der study in Dana Point, California, for example. They are also currently in use at several seawater RO plants in Spain, including the facility at San Pedro del Pinatar, which has a capacity of over 170,000 m3/day (Pe- ters et al., 2006). An alternative approach recently used at the Fukuoka, Japan plant is a seabed infiltration gallery (Figure 4-4). This intake system requires iso- lating a section of beach so that sand can be removed to the desired depth. Varying grades of small rock and gravel are placed into the exca- vation and perforated pipes are installed to convey source water to the plant. The rock and gravel are covered with the same sand material that was excavated from the beach, before the ocean is allowed to resume its FIGURE 4-3. Slant-drilled well concept. SOURCE: Adapted from Richard Bell (2006).
State of the Technology 65 FIGURE 4-4. Seabed filtration system. SOURCE: Missimer (1994). normal shoreline. The benefits of this approach are that the source water is nearly all seawater and not diluted by freshwater aquifer contributions and that it greatly mitigates entrainment and impingement (Wright and Missimer, 1997). Although proven technically successful in Japan, theenvironmental impacts will need to be better understood and, if needed, mitigated. PRETREATMENT Pretreatment is generally required for all desalination processes. Pre- treatment ensures that constituents in the source water do not reduce the
66 Desalination: A National Perspective performance of the desalination facility. Thermal processes require pre- treatment to avoid scaling and to control corrosive constituents of the source water. Some removal of sand or gritlike suspended solids may also be necessary to avoid pipe erosion. In membrane desalination, pre- treatment involves these considerations as well as further pretreatment to remove suspended solids of both biological and mineral origin to avoid membrane fouling. Biological growth may need to be inhibited by a dis- infectant or biocide. Pretreatment is a critical step in seawater and brackish water mem- brane desalination systems that utilize feedwater from surface water sources, because the suspended and colloidal particles, organisms, and natural organic matter need to be removed before the feedwater reaches the membranes. Indeed, proper pretreatment of feedwater is the most important factor in the successful operation of an RO plant, and pilot testing of the pretreatment process is a critical part of plant design. Brackish water desalination systems that treat groundwater require very minimal, if any, pretreatment to remove particulates because the water typically contains very low concentrations of suspended solids and organic matter. Nevertheless, brackish groundwater may require pre- treatment to remove selected constituents such as dissolved iron, manga- nese, and sulfides, which, if oxidized, create particulates that can foul RO membranes (USBR, 2003). The quality of source water available at a particular site will also af- fect the extent of pretreatment needed for membrane desalination. Source water quality will depend on local site factors such as source water depth, turbidity, boat traffic, oil contamination, nearby outfalls, wind conditions, tides, and the influence of runoff. As discussed previously, subsurface seawater intakes, aquatic filter barriers, and deep ocean water intakes can greatly reduce the need for pretreatment. Due to permitting regulations and available land, however, desalination plants cannot al- ways be sited where they will have the lowest pretreatment costs. Fur- thermore, because the United States employs more rigorous accounting of environmental costs, siting options in the United States may lead to greater pretreatment and greater pretreatment residuals handling needs compared to global standards of practice. The most common pretreat- ment processes are discussed below. Scaling and Corrosion Control Scaling is caused by the precipitation of minerals, such as calcium car- bonate, from solution. Calcium sulfate scaling can be controlled via tem- perature control or through pretreatment by nanofiltration to remove the calcium ions. Acidification of the feedwater can prevent calcium carbon-
State of the Technology 67 ate or magnesium hydroxide formation and scaling. Finally, the use of chemical antiscalants such as sodium hexametaphosphate or polymeric acids can sequester the cations that can lead to scaling problems (Table 4-1). Corrosion can be reduced by removing corrosive gases in pretreat- ment. Carbon dioxide can be controlled through acidification, and oxy- gen can be controlled with an oxygen scavenger such as sodium bisulfate or ferrous sulfate. Alternatively, corrosion can be controlled in some sys- tems through the formation of a protective film within the system by add- ing zinc orthophosphate (Table 4-1). Conventional Solids Removal Methods Conventional solids removal methods such as coagulation and sedi- mentation followed by media filtration are still the predominant pre- treatment processes for seawater RO. Chemicals such as ferric chloride or polyelectrolytes are added to enhance the coagulation of suspended solids prior to settling and filtration (Table 4-1). Traditional gravity flow filtration has been successfully used at many seawater RO plants around the world. At Point Lisas, Trinidad, gravity filters with greater-than- normal depth proved to be successful in pretreating seawater that en- counters severe spikes in turbidity due to the intake location in a ship turning basin (Jacangelo and Grounds, 2004). These are mature tech- nologies, although novel approaches to conventional filtration continue to be examined. For example, at Tampa Bay, an upflow dual sand proc- ess was installed that had previously only been used for industrial and wastewater applications (see Box 4-2). Nevertheless, there is a need to improve the quality and stability of influent to RO membranes; thus, other pretreatment options continue to emerge. Microfiltration and Ultrafiltration Microfiltration (MF) and ultrafiltration (UF) membranes are increas- ingly being used in the pretreatment processes for membrane desalina- tion. Water molecules and salts are free to pass through, and water is pushed (or pulled) through the membrane at very low pressures. Particles larger than the membrane pore size (0.03-10 Âµm for MF and 0.002-0.1 Âµm for UF) are removed. Membranes are commercially available in flat- sheet, tubular, hollow-fiber, and spirally wound configurations. Among the benefits of MF/UF pretreatment compared to conventional pretreat-
68 Desalination: A National Perspective TABLE 4-1 Reported Dosing Concentrations of Pretreatment Chemical Additives in Reverse Osmosis and Multistage Flash Desalination REVERSE OSMOSIS DESALINATION Reported Dosing Chemical Additive (mg/L) References Biocide Chlorine 0.5-6 Abart, 1993; Redondo and Lomax, 1997; Morton et al., 1996; Woodward Clyde Con- sultants, 1991 Chlorine removal Sodium bisulfite 3-19 Morton et al., 1996; Redondo and Lomax, 1997; Woodward Clyde Consultants, 1991 Coagulants Ferric chloride 0.8-25 Baig and Kutbi, 1998; Woodward Clyde Consultants, 1991 Polyelectrolyte 0.2-4 Ebrahim et al., 1995; DuPont, 1994; Hus- sain and Ahmed, 1998 Antiscalants Sulfuric acid 6.6-100 Al-Shammiri et al., 2000; Morton et al., 1996; Al-Ahmad and Aleem, 1993 Sodium hexametaphos- 2-10 Al-Ahmad and Aleem, 1993; Al-Shammiri et phate (SHMP) al,. 2000 FilmTec, 2000 Polyacrylic acid 2.9 Woodward Cycle Consultants, 1991 Phosphonate 1.4 Al-Shammiri et al., 2000 MULTISTAGE FLASH DISTILLATION Biocide Chlorine 0.25-4 Iman et al., 2000; Shams El Din and Mak- kawi, 1998; Khordagui, 1992; Abdel-Jawad and Al-Tabtabaei, 1999 Hypochlorite 2 Burashid, 1992 Antiscalants Polyphosphate 2.2-2.5 Hamed et al., 2000; Abdel-Jawad and Al- Tabtabaei, 1999 Polycarboxylic acid 1.5-2 Hamed et al., 1999 Polyphosphonate 1-3 Hamed et al., 1999, 2000 Antifoaming agents Polypropylene glycol 0.035-0.15 Imam et al., 2000 Corrosion control Sodium bisulfite Not given Imam et al., 2000 Ferrous sulfate 1-3 Shams El Din and Makkawi, 1998 Ð-ethyl phenyl ketocyclo- 25 Andijani et al., 2000 hexylamino hydrochloride NOTE: The types and concentrations of pretreatment chemicals vary with plant design and source water conditions. Thus, some or all of these chemicals may not be used at all, or they may only be used intermittently. Some added chemicals can be recovered or re- moved (e.g., chlorine). SOURCE: Adapted from Lattemann and HÃ¶pner (2003).
State of the Technology 69 BOX 4-2 Pretreatment Changes at the Tampa Bay Seawater Desalination Plant The largest seawater desalination plant in the United States (95,000 m3/day or 25 MGD) is located in the Tampa Bay region of Florida. The Tampa Bay Sea- water Desalination project obtains its source water from a âonce-throughâ cooling system at the Tampa Electric Company (TECO) Big Bend Power Station, which withdraws its cooling water from Tampa Bay. After coming online in March 2003, the plant experienced performance problems that significantly stemmed from an inadequate pretreatment system. The desalination plant did not include any addi- tional screens beyond those that existed within the power plant, and the upflow sand filter system was inadequate to produce pretreated seawater adequate to sustain reverse osmosis process operation. Although the warmer water from the power plant requires less energy for the desalination process, higher water tem- perature introduces greater potential for biological growth and pretreatment chal- lenges. The original pretreatment process (Figure 4-5) resulted in severe prema- ture fouling of the cartridge filters after 3-4 days of operation and premature membrane cleaning immediately following a 2-week acceptance test. FIGURE 4-5. Original pretreatment design for the Tampa Bay Water Desalination Plant. SOURCE: Adapted from figure courtesy of Tampa Bay Water. The new design (Figure 4-6) is a much more robust treatment process that includes significant changes to the head works of the plant. Changes include the addition of 1/16-inch traveling screens for debris removal, ferric chloride as a coagulant, the use of chlorine dioxide for biological growth control, mechanical chemical mixing and coagulation basins to achieve better floc formation, and extended flocculation zones to aggregate and settle suspended particles up continued
70 Desalination: A National Perspective stream of the sand filters. Water distribution to the individual sand filters is now individually piped, valved, and metered to allow remote control of specified flow to each filter. The upflow sand filters were converted from a dual-stage, roughing and polishing configuration to a single-stage system, which reduced the hydraulic loading rate to each filter. Diatomaceous earth filters were added as the polishing step after the upflow sand filters to complete the particulate removal process immediately prior to the 5-Âµm cartridge filters. Development of a robust pretreat- ment process will reduce the fouling rate of the cartridge filters and cleaning re- quirements of the reverse osmosis membranes. Reducing the membrane clean- ing frequency will extend the useful life of the roughly 10,000 reverse osmosis membrane modules at the plant, protecting the membrane installation cost of approximately $6 million. The Tampa Bay pretreatment system depicted in Figure 4-6, the first of its kind for seawater desalination, is not a proven pretreatment technology. Because pretreatment is a key step to successful operation of seawater desalination plants, it is critical that new pretreatment approaches be tested systematically before implementation. The capital cost of the Tampa Bay pretreatment system is estimated to be approximately 30 percent ($32 million) of the total engineering and construction cost ($108 million). The pretreatment operating costs, including power, chemicals, and diatomaceous earth, is $0.12 per cubic meter ($0.44 per thousand gallons) (Joe Dysard, R.W. Beck, personal communication, 2007). FIGURE 4-6. Modified pretreatment process at Tampa Bay Desalination Plant. SOURCE: Adapted from figure courtesy of Tampa Bay Water. ment technologies are (1) production of feedwater to the RO system of constant and high quality regardless of source water fluctuations; (2) re- duced RO fouling, which results in less cleaning and longer membrane life; (3) smaller footprint; and (4) lower consumption of chemicals. Po- tential disadvantages include higher costs and negative environmental impacts of concentrate from these membranes. MF and UF membrane processes have been developed and piloted for application in seawater pretreatment. Several studies highlight the benefits of using MF or UF as pretreatment for RO desalination (e.g.,
State of the Technology 71 Allam et al., 2003; Galloway and Mahoney, 2004; Latorre, 2001; Pearce, 2007; Taniguchi et al., 1995; Xu et al., 2007, Zhang et al., 2006). Biofouling Control Chlorine or hypochlorite have been the standard oxidants used for biofouling control, but thin-film composite polyamide membranes com- monly used in RO desalination cannot tolerate oxidants like chlorine, and the chlorine needs to be removed in pretreatment by addition of a reduc- ing agent such as sodium bisulfite. Sodium bisulfite and copper sulfate can also be used as biocides in membrane systems. Ultraviolet (UV) and ozone treatment are being considered potential replacements for chlorine-based biological growth control of RO feedwa- ter. Both UV and ozone have merit and have been successfully used in small to midsized drinking water and water reuse applications globally. UV will not cause problems with oxidant-sensitive membranes. Ozone is a much more effective disinfectant, but it poses a problem to the oxidant- sensitive RO membranes (Cotruvo, 2005; Glater et al., 1983). Assessment of Pretreatment Technology Conventional pretreatment for RO (i.e., coagulation and sand filtra- tion) is a mature technology and has been proven effective for seawater desalination plants with surface water intakes. MF and UF pretreatment systems for RO are likely to gain more popularity due to the superior quality of water that such systems can produce, although their robustness under such conditions remains unproven. High-quality pretreatment will be of greater importance with the advent of âhigh-fluxâ (high- permeability) RO membranes, as the propensity for fouling is signifi- cantly increased at higher water fluxes. To realize the benefits of high- flux RO membranes, more effective pretreatment systems will be needed. DESALINATION PROCESSES The desalination process represents the step in which dissolved sol- utes are substantially removed from the feedwater to create the desired product water. A number of technologies exist to accomplish this objec- tive, including the more commonly used membrane, thermal, and ion- exchange processes.
72 Desalination: A National Perspective Membranes can be designed to selectively permit or prohibit the passage of certain ions, including salts. Membranes play an important role in the separation of salts in the natural processes of dialysis and osmosis, and this natural principle has been adapted in two commercially important desalting processes: electrodialysis (ED) and RO. In recent years, significant advances in RO technology have been achieved and, globally, more new membrane desalination capacity is now added annually than distillation capacity (Figure 4-7). Currently, membrane processes, including RO, nanofiltration (NF), ED, and electrodialysis reversal (EDR), account for 56 percent of the online capacity for desalination worldwide. The basic concept of thermal distillation is to heat a saline solution to generate water vapor. If this vapor is directed toward a cool surface, it can be condensed to liquid water containing very little of the original salt. Water will boil under atmospheric pressure at 100Â°C, but thermal processes can also be designed to boil water in a series of vessels operating at successively lower temperatures and pressures. At one- quarter of normal atmospheric pressure, water will boil at 65Â°C, and it will boil at only 45Â°C if the pressure is decreased to one-tenth normal 30 Thermal 25 Membrane Cumulative capacity (million m /d) Current online thermal 3 capacity 20 Current online membrane capacity 15 10 5 0 1950 1960 1970 1980 1990 2000 2010 Year FIGURE 4-7. Cumulative global capacity of installed desalination plants for ther- mal and membrane technology. Thermal technology includes MED, MSF, and MVC. Membrane technology includes RO, NF, ED, and EDR. Points reflect cur- rent online (or presumed online) capacity of both technologies. SOURCE: GWI (2006b).
State of the Technology 73 atmospheric pressure. The concept of distilling water with a vessel operating at a reduced pressure has been used for well over a century.Thermal processes, such multistage flash (MSF), multiple effect distillation (MED), and mechanical vapor compression (MVC), account for 43 percent of the online capacity for desalination worldwide. Water can be desalted through many other processes, including small-scale ion-exchange resins and hybrid processes. None, however, has achieved the commercial success of membranes or thermal distillation. Together these other processes account for less than 1 percent of total desalination capacity worldwide (GWI, 2006b). There is no single "best" method of desalination. Globally, thermal and membrane technologies are both used widely for seawater desalina- tion. Both processes require energy to effect separation of salts, and vari- ous energy sources can be used. Brackish water is typically desalinated using RO, NF, or ED. Significant numbers of smaller plants have also been built that use vapor compression and other methods (GWI, 2006b; USAID, 1980). The dramatic growth in membrane and thermal desalina- tion processes over the past four decades is shown in Figure 4-7. The objective of this section is to present the state of the technology of currently applied desalination processes that might have potential for addressing water needs in the United States. Energy use, performance, theoretical limitations, and current technical issues that need to be im- proved will be addressed in this section. The specific energy for desalina- tion (i.e., the energy needed to produce a unit volume of fresh water from a saline feedwater) is an important consideration when comparing the various desalination technologies in this section. Potential improvements of these processes are discussed, considering that several of the tech- nologies are mature with respect to approaching the value of the theoreti- cal minimum energy for desalination (Box 4-3). Membrane Desalination Processes Commercially available membrane technologies for desalination in- clude RO, NF, and ED or EDR. RO and NF are classified as pressure- driven membrane processes, whereas ED and EDR are electrically driven (see Box 4-4). Membrane technologies can be used not only for desalting brackish water and seawater sources but also for treating wastewater in reuse and recycling applications, because of their ability to provide re- moval of nonsalinity contaminants (e.g., organic contaminants, bacteria, and viruses; see Figure 4-9). Typically, 35 to 60 percent of the seawater fed into a membrane process is recovered as product water. For brackish water desalination, water recovery can range from 50 to 90 percent, de-
74 Desalination: A National Perspective BOX 4-3 Theoretical Minimum Work (Energy) for Desalination Desalination requires an intrinsic minimum available energy. Available en- ergy, often called exergy, is mechanical, electrical, or any other energy, which in practice can be nearly completely converted into mechanical work (Spiegler and El-Sayed, 1994). Any desalination process will be most efficient if it involves a reversible thermodynamic process. In any reversible desalination process, the same energy is needed to desalt water, and this energy is independent of the technology or device employed and the exact mechanism of desalination. Thus, all desalination systems share a theoretical minimum work (available energy) that is independent of the system used. There are several approaches for calculating the theoretical minimum en- ergy for desalination. Stoughton and Lietzke (1965) based their analysis on the fact that the free energy change involved in removing a small amount of pure water from a mixture of water and salt is equal in magnitude but opposite in sign to that for adding the same amount of pure water to the mixture. They have fur- ther shown that the free energy change is equal to the minimum energy needed for desalination. The minimum energy for desalination for any percent of water recovery (up to saturation of salt) and a wide range of temperatures (from 25 to 200Â°C) was calculated. For zero percent recovery, that is, the removal of a rela- tively small amount of water from a very large amount of seawater, the calculated theoretical minimum energy for desalination was 0.70 kWh/m3 of freshwater pro- duced. Another approach which resulted in similar results for the theoretical mini- mum work for desalination was presented by Spiegler and El-Sayed (1994). In this approach, compartments containing saltwater and freshwater are consid- ered. The compartments are separated by a semipermeable membrane that permits water but not salt to pass through it. At equilibrium, hydraulic pressure must be exerted on the salt solution to prevent spontaneous transport of water from the freshwater to the salt solution. This hydraulic pressure is equal to the osmotic pressure, âos. To separate water from the salt solution, the pressure on the salt solution is increased under reversible conditions, so that the applied pressure exceeds the osmotic pressure by only an infinitesimal amount. A small volume of pure water, dV, is passed through the membrane, but because of the increase in the salt solution concentration and osmotic pressure, there is a need to increase the applied pressure again to ensure transport of water from the salt solution to the freshwater. This process continues until the desired amount of water has been removed from the salt solution (i.e., when a desired water recov- ery is attained). The differential work, dW, needed for obtaining a differential amount of fresh water, dV, is given by (Spiegler and El-Sayed, 1994) dW = Î os dV (1) Denoting the initially volume of the salt solution as V and the final volume as V, the total work per volume of freshwater produced, W, is continued
State of the Technology 75 V2 1 W= V1 â V2 â«Î V1 os dV (2) The osmotic pressure is a function of the activity of water, and it decreases with increasing salinity of the salt solution. These activity coefficients can be deter- mined from the ratio of the vapor pressure above the salt solution to the vapor pressure of pure water. Following this analysis, Spiegler and El-Sayed (1994) showed that the theoretical minimum work for zero recovery at 25Â°C is 0.70 kWh/m3, similar to the value reported by Stoughton and Lietzke (1965) as dis- cussed earlier. Furthermore, the theoretical minimum energy increases for salt solutions of higher temperatures. For instance, the theoretical minimum energy for zero recovery is 0.72, 0.82, and 0.87 kWh/m3 for seawater temperatures of 50, 75, and 100Â°C, respectively (Stoughton and Lietzke, 1965). It was further shown that, for any water recovery, the minimum work is re- lated to the theoretical minimum work for zero recovery, W0, via (Spiegler and El- Sayed, 1994): V1 V W = W0 ln 1 (3) V1 â V2 V2 It is useful to express the minimum work for desalination in terms of the water recovery, r, noting that r = (V1 â V2 )/V1, as follows: W0 1 W= ln (4) r 1â r The minimum work for desalination as a function of water recovery, calculated by the preceding analysis, is presented in Figure 4-8. As seen, the theoretical mini- mum energy for desalination increases dramatically at high water recoveries. 3.5 Minimum Energy (kW-h/m ) 3 3.0 O 100 C O 25 C 2.5 2.0 1.5 1.0 0.5 0 20 40 60 80 100 Percent Recovery FIGURE 4-8. Minimum work (energy) for desalination of seawater per cubic me- ter of produced water. Calculations are based on a seawater solution (3.45 wt% of salts) and for temperatures of 25 and 100Â°C.
76 Desalination: A National Perspective pending on initial salinity and the presence of sparingly soluble salts and silica, although recovery is typically between 60 and 85 percent (Sethi et al., 2006a). The remaining reject salt solution becomes more concen- trated and must be disposed. For both brackish water and seawater, membrane processes can reduce salinity in the product water to levels less than 500 ppm TDS. Reverse Osmosis The RO process uses semipermeable membranes and a driving force of hydraulic pressure, in the range of about 1,000 to 8,300 kilopascals (kPa) (150 to 1200 pounds per square inch [psi]; 10 to 83 bar), to remove BOX 4-4 Membrane Systems The major membrane types that can be used for desalination and/or pretreat- ment are the folllowing: Electrodialysis (ED) is an electrochemical separation process in which ions are transferred through ion-exchange membranes by a direct current voltage, leaving desalinated water as the product. Reverse osmosis (RO) membranes desalinate both brackish water and sea- water by applied pressure using a solution/diffusion mechanism whereby the water dissolves into and diffuses through the nonporous membrane, leaving the majority of the salts behind in the concentrate. RO membranes are also capable of removing some larger organic contaminants. Small uncharged species can pass through the membrane. Nanofiltration (NF) membranes are used for water softening, organics and sul- fate removal, and some removal of viruses. Pressure-driven removal is by com- bined particle size-based sieving and solution/diffusion. Pores in NF membranes are usually smaller than 0.001 Î¼m and a molecular weight cutoff (MWCO) of 1,000 to 10,000 daltons. Ultrafiltration (UF) membranes are used for removal of contaminants that affect color, high-weight dissolved organic compounds, bacteria, and some viruses. UF membranes operate via a pressure-driven size-based sieving mechanism through a membrane with pores in the range of 0.002 to 0.1 Î¼m with an MWCO of 10,000 to 100,000 daltons. Microfiltration (MF) membranes are used to reduce turbidity and remove sus- pended particles, algae, and bacteria. MF membranes operate via a sieving mechanism under a lower pressure than either UF or NF membranes, through membrane pores of 0.03 to 10 Î¼m and an MWCO of greater than 100,000 daltons. continued
State of the Technology 77 Suspended particles Algae Protozoa Microfiltration (MF) Bacteria Large macromolecules Small colloids Ultrafiltration (UF) Viruses Natural organic matter Nanofiltration (NF) Hardness (Ca , Mg2+ ) 2+ Dissolved contaminants Reverse osmosis (RO) Salt (Na+ , Cl-) Water FIGURE 4-9. Substances and contaminants nominally removed by pressure- driven membrane processes. Modified from Trussell and Trussell (2005). SOURCES: NRC (2004b), Cooley et al. (2006), Sedlak and Pinkston (2001), Heberer et al. (2001), AWWA (1999), AWWARF et al. (1996), and NRC (1997). dissolved solids from brackish water or seawater. The process can be described as solution/diffusion controlled, because the ions move through RO membranes via the process of diffusion (Lonsdale et al., 1965). Salts do permeate the membrane but at permeabilities that are or- ders of magnitude lower than that of water; thus, the majority of dis- solved salts are removed by the process. RO can also remove synthetic organic chemicals and disinfection by-product precursors. However, dis- solved gases such as hydrogen sulfide and carbon dioxide and some pes- ticides or low-molecular-weight organics pass through RO membranes. The specific energy for RO desalination varies with the system used, the operational conditions (e.g., flux, recovery), and the quality of feed- water to the RO system. For seawater RO, the specific energy usage is typically about 3-7 kWh/m3 with energy recovery devices (Alonitis et al., 2003; Miller, 2003; see Table 4-2). For brackish water RO, energy usage is comparatively lower, about 0.5-3 kWh/m3, because the energy re- quired for desalination is proportional to the feedwater salinity (Sethi et al., 2006b; see Table 4-3 and Figure 4-10). Energy usage values should be taken cautiously because the "system" for which desalination energy use is calculated and reported (i.e., basic RO process only, or including other ancillary equipment or processes) varies in the literature. RO membrane formulations include cellulose acetates, polyamides,
78 Desalination: A National Perspective TABLE 4-2 Comparison of Predominant Seawater Desalination Processes MED Seawater (with RO MSF TVC) MVC Operating <45 <120 <70 <70 temperature (Â°C) Pretreatment High Low Low Very low requirement Main energy form Mechanical Steam Steam Mechanical (electrical) (heat) (heat and (electrical) energy pressure) energy Heat consumption NA 250-330 145-390 NA (kJ/kg) Electrical energy 2.5-7 3-5 1.5-2.5 8-15 use (kWh/m3) Current, typical <20,000 <76,000 <36,000 <3,000 single train 3 a capacity (m /d) Product water 200-500b < 10 < 10 < 10 quality (TDS mg/L) c Typical water 35-50% 35-45% 35-45%c 23-41%c recovery Reliablility Moderate Very high Very high High a For the purpose of this table, a train is considered a process subsystem which includes the high-pressure pump, the membrane array(s), energy recovery de- vices, and associated instrumentation/control. However, larger facilities may group pumps, membranes, and energy recovery into process or pressure centers to lower capital costs and improve operating costs. b Product water quality for RO is a design variable. Each pass through an RO plant typically removes 99 to 99.5 percent of dissolves salts in the feedwater. Successive passes using additional membranes can be added along with other design optimizations to achieve permeate with the TDS required for a target wa- ter use. Potable water requirements can readily be met with 200-500 mg/L TDS water, which can be achieved from seawater with a single RO pass. c Cooling water is not factored into these recovery calculations. The MVC proc- ess does not require cooling water and typically operates at 23 to 41 percent recovery with seawater desalination, but recoveries can reach as high as 95 per- cent for industrial concentration/ZLD applications. âApparent recoveryâ of thermal desalination that uses intake water used as cooling water can be 10 to 20 per- cent. However, cooling water volumes can be substantially reduced by employing other cooling mechanisms, such as cooling towers. SOURCES: Wangnick (2002), German Aerospace Center (2007), GWI (2006a), USBR (2003); Spiegler and El-Sayed (1994); Thomas Pankratz, GWI, personal communication, 2008.
State of the Technology 79 TABLE 4-3 Comparison of Predominant Brackish Water Desalination Processes Brackish water RO ED/EDR NF Operating <45 <43 <45 temperature (Â°C) Pretreatment High Medium High requirement Electrical energy 0.5-3 ~0.5 kWh/m3 <1 3 use (kWh/m ) per 1,000 mg/L of ionic species removed Current, typical <20,000 <12,000 <20,000 single train 3 capacity (m /d) Percent ion 99-99.5% 50-95% 50-98% removal removal of divalent ions; 20-75% removal of monovalent ions Water recovery 50-90% 50-90% 50-90% SOURCES: Anne et al. (2001), Wangnick (2002), Kiernan and von Guttberg (2005), Reahl (2006), Sethi et al. (2006b), USBR (2003). FIGURE 4-10. Comparison of energy consumption by process for the desalina- tion of brackish feedwater across a range of TDS concentrations. SOURCE: USBR (2003).
80 Desalination: A National Perspective polyetheramides, and polyethersulfones. The present most widely used membrane material is a thin-film composite polymer combining a micro- porous polysulfone support layer with a thin polyamide layer. The mem- brane, commercialized in 1980, has been hugely successful and shaped the course of membrane technology. RO membranes have matured sig- nificantly over the past few decades, with exceptional improvement in RO membrane costs, water flux/permeability, membrane life, and salt rejection capability. Inflation-corrected membrane costs, for example, dropped by a factor of about 4 between 1975 and 1990 and by roughly another 75 percent between 1990 and 2002 (Birkett and Truby, 2007). These improvements in performance, in combination with advances in technologies to recover the unused energy from the concentrate stream, have led to dramatic reductions in energy costs and capital expenses re- quired to desalinate seawater and brackish ground waters. Although RO technology appears to be maturing, several major chal- lenges remain, including membrane fouling, which leads to increases in energy use and poor resistance to chlorine and other oxidants. Mem- branes have shifted from the original cellulose acetate membranes to thin-film composite (TFC) membranes. In the past few years, several variations of TFC membranes have been commercialized in an attempt to reduce fouling. Many of these developments have resulted from the addi- tion of polymer to smooth the surface or surface modifications such as addition of different functional groups to change the surface charge. While these improvements reduce fouling, truly fouling-resistant mem- branes are yet to be realized. Thus, opportunities exist to modify existing or create new membrane formulations or alter surface characteristics to reduce fouling. Currently, the most direct and effective way to protect against foul- ing is with effective pretreatment to remove suspended/colloidal matter and dissolved organic matter. As an alternative to fouling-resistant mem- branes, fouled membranes that could be cleaned easily with low-cost oxidants (e.g., chlorine) would be desirable. However, the state-of-the-art RO membranes for seawater and brackish water desalination cannot tol- erate oxidants such as free chlorine, and they require chlorine removal from the feedwater before being processed by the RO modules. Conse- quently, biofouling can be another challenge that limits the performance of RO membranes. Another limitation in RO desalination is the relatively low recovery rate in seawater and brackish water desalination (up to about 60 percent and 50-90 percent, respectively), which results in large volumes of con- centrate. The maximum recovery is limited by the mechanical pressure limitations of the materials in the membrane element whereas practical recoveries (typically 45 percent for seawater) consider optimization of other parameters such as solubility product limits and energy consump-
State of the Technology 81 tion. Improved membrane materials or configurations may increase the pressure tolerance of the RO modules and/or reduce osmotic pressure of the concentrate due to reduced concentration polarization (i.e., the buildup of dissolved salt near the membrane surface). Several approaches for improvements to overall RO recovery are currently being investi- gated (Mickley, 2007; Sethi et al., 2008). For example, demonstration testing is under way on a dual RO system with intermediate chemical precipitation to further enhance recovery in brackish water desalination by addressing scaling concerns (Williams et al., 2002). A patented high- efficiency RO technology has also recently been developed that com- bines a two-phase RO process with chemical pretreatment of primary RO, intermediate ion-exchange treatment of the primary RO concentrate, and high-pH operation of secondary RO to allow operation of the secon- dary RO at high recoveries (Jun et al., 2004; Mukhopadhyay, 1999). Other examples of high-recovery technologies being developed or tested include dual RO with intermediate biological treatment (Williams and Pirbazzi, 2003), treatment of RO concentrate via EDR with intermediate chemical precipitation (Sethi et al., 2008), and alternative technologies such as such as forward osmosis, membrane distillation, and dewvapora- tion (see Box 4-5). Improved module and membrane configuration and system design are also imperative to avoid operation under the phenomenon of âther- modynamic restrictionâ (Song et al., 2003a). This phenomenon, which is most likely to occur with the use of high-permeability RO membranes, results from significant buildup of salt concentration down the membrane channel (i.e., along the modules in the pressure vessel), such that midway through the channel the osmotic pressure of the concentrate increases to a level equal to the applied hydraulic pressure (Song et al., 2003a, 2003b). Under these conditions, all product water produced by the RO process would permeate out of the membrane before the flow reaches the end of the channel; the rest of the membrane channel would not produce any more water. Song et al. have demonstrated that many RO systems with high-permeability membranes are operated at or near the regime of thermodynamic restriction. Under thermodynamic restriction, increasing the applied hydraulic pressure has little effect on the overall product wa- ter flux or water recovery of the RO system. Thermodynamic restriction can be avoided or minimized by proper module configuration and system design, such as membrane modules with higher channel height or the reduction of the number of modules in a pressure vessel. RO membranes have been dominated by 2.5-, 4-, and 8-inch- diameter spiral-wound thin film composite configurations with a stan- dard length of 40 inches for many years. RO plants that produce between 250 and 330,000 m3/day currently utilize the 8-inch-diameter membrane,
82 Desalination: A National Perspective BOX 4-5 Research on Alternative Desalination Technologies to Improve Energy Efficiency Several alternative approaches have been proposed to reduce the energy requirements of desalination (see Miller and Mayer, 2005). Some approaches, such as forward osmosis, membrane distillation, and dewvaporation, present opportunities to improve the use of low-grade or waste heat, whereas other ap- proaches, such as freeze desalination and capacitative deionization, offer the potential to reduce overall energy use. Forward Osmosis Forward osmosis is a membrane-based separation process that uses os- motic pressure difference between a concentrated âdrawâ solution and a feed stream to drive water flux across a semipermeable membrane. Given sufficient difference in osmotic pressure, the magnitude of water flux and degree of salt rejection can be competitive with RO (McCutcheon et al., 2005). The primary challenge is in the selection of a draw solute so that its presence in the product water is desirable, or so that it may be easily and economically removed. For example, if a combination of NH3 and CO2 gases is used as the draw solution, the energy requirements of the forward osmosis process are small quantities of electrical power (<0.25 kWh/m3) combined with low-quality heat (<50ËC), which could be provided as a reject stream from industrial or power production proc- esses (McGinnis and Elimelech, 2007). Dewvaporation In dewvaporation technology, a stream of air is humidified by a falling film of saline water along one side of a heat transfer surface. The air is partially heated by an external source (e.g., solar, waste heat). The heated air then is swept along the condensing side of heat transfer films, where the vapor condenses to a liquid, which is collected as product water. The condensation releases heat through the heat transfer surface to the evaporation side. A small prototype was built and operated, demonstrating the efficacy of the approach (Hamieh et al., 2001). Potential difficulties in the use of this process are the large heat transfer areas required, the impact of ambient weather, and the need for a low- temperature sink to permit condensation. Potential benefits include the efficient use of low-grade heat or solar energy, small footprint, and low capital costs com- pared to conventional thermal desalination methods. Membrane Distillation In membrane distillation, saltwater is warmed to enhance vapor production, and the vapor is exposed to a membrane that can pass water vapor but not liquid water. There are several different types of membrane distillation; the main four types are direct contact, air gap/sweeping gas, osmotic, and vacuum (Banat and Simandl, 1998; Celere and Gostoli, 2002; El-Bourawi et al, 2006; Srisurichan et continued
State of the Technology 83 al., 2006; Xu et al., 2006). Rejection of feedstream solutes is high and can be comparable to that of other distillation techniques. The possible advantages of the use of membrane distillation are that it has a small footprint relative to other thermal desalination technologies, lower capital costs, and the ability to use low- grade heat sources. Possible disadvantages include difficulty in maintaining the hydrophobicity of the membrane over long periods due to fouling and membrane degradation, the large enthalpy of vaporization required for the phase change of water transported across the membrane, and poor rejection of volatile feed- stream contaminants (Peng et al., 2005; USBR, 2004). Freeze Desalination The basis of freeze desalination technologies is to change the phase of wa- ter from liquid to solid. As ice crystals form, they exclude salt from their structure, enabling the possibility of washing the salt from the crystals. This approach seeks to take advantage of the relatively low enthalpy of phase changeâthe freezing of water at atmospheric conditions (334 kJ/kg)âwhereas evaporation would require 2,326 kJ/kg. The cooling required for the process must be supplied from a means of refrigeration, either mechanical or thermal (absorption cooling). Once crystallization of the water has occurred, the ice crystals need to be sepa- rated from the saline solution and washed to ensure final water quality. Potential benefits of the technology include improved energy efficiency compared to distil- lation processes because ambient seawater is always closer to its freezing point than its boiling point. Potential difficulties include effective separation and wash- ing of water crystals without prematurely melting them and redissolving the salt, maintenance of relatively complex system components, and achieving efficient operation in light of refrigeration requirements. While distillation processes can be cascaded in multiple stages or effects to reuse the latent heat of evaporation (reducing the 2,326 kJ/kg evaporation energy to less than 20 kJ/kg), it is chal- lenging to make a similar freezing configuration. Capacitive Deionization Capacitive deionization is an electrosorption process whereby ions are re- moved from water using an electric field gradient as the driving force. The saline feed flows through electrodes comprised of materials such as carbon-based aerogels. These aerogels have very high surface area (400-1,000 m2/g) and low electrical resistivity. The cations are attracted to the anodic electrode, while the anions are attracted to the cathodic electrode. A direct current is imparted, with a potential difference of 1-2 volts. Ions are held at the surface of the electrode in the electric double layer. Researchers at Lawrence Livermore National laboratory have determined that this technology can desalinate brackish water (2,000 ppm feed to 186 ppm permeate) using 0.14 kWh/m3, assuming 70 percent recovery of the stored electrical energy (Farmer et al., 1996). with tens of thousands of membrane elements required for large desalina- tion plants. By 2004, manufacturers began offering 16- and 18-inch- diameter elements in 40- and 60-inch lengths. Fewer, larger membrane elements may reduce overall capital costs through economies of scale
84 Desalination: A National Perspective and the need for fewer components (e.g., piping, connections) while re- ducing the operations and maintenance requirements. The large mem- brane surface area provided by large-diameter elements would also en- able the reduction of the channel length (or the number of membrane elements in the pressure vessel), thereby balancing the permeate through the membrane elements and eliminating operation under thermodynamic restriction where downstream membrane elements do not produce any water flux. Nanofiltration Similar to the RO process, the NF process also uses semipermeable membranes and a driving force of hydraulic pressure, in the range of about 50â250 psi. NF membranes are capable of rejecting divalent ions (such as hardness) and larger contaminants very well, while providing lower retention of monovalent ions (see Figure 4-9). NF can also remove synthetic organic chemicals and disinfection by-product precursors. Thus, NF is primarily used for softening and removal of organics. The NF process achieves removal via a combination of both the classic solu- tion/diffusion mechanism as well as steric (size) and charge exclusions (e.g., Childress and Elimelech, 2000; Timmer, 2001). Pilot testing of a two-pass NF system for seawater desalination is under way at the Long Beach Water Department. The first pass removes greater than 90 percent of the salinity, and the second pass removes greater than 93 percent, re- sulting in a total salt reduction of about 99.5 percent (Tseng et al., 2003). The presence of two passes of NF provides greater flexibility than con- ventional membrane processes. For example, the second pass can be op- erated at a higher pH by addition of a base, which allows very high rejec- tion of boron (see also Chapter 5). The overall recovery from the process is about 30 to 45 percent for seawater desalination, which is at the low end of the range observed with conventional RO desalination. For NF, the typical energy usage is lower than that for RO, depend- ing on the feedwater characteristics and the product water quality objec- tives. Similar to RO, energy recovery is possible using typical energy recovery devices. As with RO, fouling is a major challenge for efficient operation of NF plants, and pretreatment of feedwaters is needed (USBR, 2003). Developments in Energy Efficiency for RO and NF The reduction in energy use for RO in the past 20 years has been remarkable (see Figure 4-11) and has had a significant and direct effect
State of the Technology 85 on operating costs. Energy use of as low as 1.6 kWh/m3 is achievable using controlled, favorable conditions and commercially available state- of-the-art equipment, including energy recovery devices, feed pumps, and low-pressure membranes (Seacord, 2006a). Specific energy values would be larger using real-world conditions. Energy Recovery Devices. A key reason behind such improvements in the energy efficiency of seawater RO systems has been the develop- ment of highly efficient energy recovery devices that capture the energy resident in the concentrate stream of the RO process. Due to low net re- coveries of the highly pressurized feedwater, typically 40 to 60 percent of the applied energy in the process can be lost if the concentrate is dis- charged to atmosphere without any attempt to recover that energy. In general, energy recovery devices can recover from 75 to 96 percent of the input energy in the concentrate stream of a seawater RO plant (Sal- langos and Kantilaftis, 2003). Existing energy recovery systems can be divided into two categories. The first are devices that transfer the concentrate pressure directly to the feedstream (e.g., pressure exchanger, work exchanger), which have en- ergy recovery efficiencies of about 95 percent. The second category in- cludes devices that transfer concentrate pressure to mechanical power, which is then converted back to feed pressure (e.g., Pelton turbine, Fran- cis turbine, reverse-running pumps). The overall efficiency of energy 25 Minimum RO energy use (kWh/m ) 3 20 15 10 Affordable 5 Desalination Coalition 1.62 0 1975 1980 1985 1990 1995 2000 2005 2010 Year FIGURE 4-11. Seawater reverse osmosis energy use trend. SOURCES: Data from McHarg and Truby (2004).
86 Desalination: A National Perspective recovery here is about 74 percent (assuming a Pelton turbine efficiency of around 87 percent coupled with a pump efficiency of 85 percent). Specific efficiency values for several energy recovery devices are shown in Table 4-4. Although the pressure exchanger offers higher efficiency than the in- direct device group, the choice of energy recovery device for a specific plant design depends on a number of factors. For example, if energy is a critical issue to overall operating costs, the higher-efficiency pressure- work exchangers often will be the device of choice. However, if the pro- ject costs are dominated by the capital expenditures, todayâs pressure- work exchangers have a disadvantage due to their higher equipment costs and current component size limitations (that is, multiple units may be needed at large scales). Practical Limits to Energy Efficiency of RO. While desalination requires an intrinsic minimum available energy as described in Box 4-3, there are practical limits to approaching the minimum required energy of approximately 0.7 kWh/m3 for infinitesimally small recovery and 0.9 kWh/m3 for 40 percent recovery. The theoretical minimum energy de- scribed in Box 4-3 is pertinent to an ideal reversible process carried out extremely slowly with no energy loses (i.e., applied pressure is only in- finitesimally higher than the feed osmotic pressure). In actual desalina- tion processes, energy is lost because of the inherent irreversibility of the processes and a number of practical limitations. In practice, todayâs best available RO membranes are operated at pressures significantly above the osmotic pressure to produce practical product water fluxes through the membranes and thereby minimize the net capital expense of the desalination plant. This applied pressure re- quirement to overcome the osmotic pressure limitations in practice is further elevated to overcome the locally elevated osmotic pressures near the surface of the RO membrane that are caused by high concentrations of salts in the boundary layer near the membrane surface (AWWARF et TABLE 4-4 Typical Efficiency of Energy Recovery Devices Energy Recovery System Efficiency (percent) Francis turbine 76 Pelton turbine 87 Turbo charger 85 Work exchanger ~96 Pressure exchanger ~96 SOURCES: Geisler et al. (2001), Sallangos (2004), and Lieberman (2003).
State of the Technology 87 al., 1996; Brian, 1966; Sherwood et al., 1964). Further adding to this ef- fect is the increasing bulk salt concentration that results when the feed- water is progressively concentrated as it flows through the membrane pressure vessels (Song et al., 2003). Improvements in the permeability and salt rejection of properties of membranes (see Box 4-6) will also re- duce the pressure required to produce practical water fluxes. Through a simple mass and energy balance calculation (see Appen- BOX 4-6 Research to Improve Membrane Fouling Resistance, Flux, and Selectivity Current research efforts, including those in the rapidly growing field of nanotechnology, have the potential to advance technologies for water and wastewater treatment as well as desalination. Examples of efforts involving membrane modification, nanostructures, and nanomaterials for manufacturing new desalination membranes with improved water flux, permeability, fouling re- sistance, and selectivity are presented. Membrane Modification to Improve Fouling Resistance Organic and biological fouling are the result of interactions between solutes and the membrane surface. Thus, surface characteristics of membranes, such as hydrophilicity, surface charge, and roughness, will affect the rate and extent of fouling. Modification of commercially available membranes to alter surface char- acteristics to reduce fouling while maintaining or improving flux and selectivity is an established research area that shows promising results for RO and NF mem- branes (Abitoye et al., 2005; Belfer et al., 1998; Gilron et al., 2001). Although many types of modification methods exist, graft polymerization is the method most commonly utilized in RO and NF membranes. If a prudent choice of a monomer is utilized, graft polymerization can increase membrane hydrophilicity and, thus, resistance to fouling with little sacrifice in the flux or selectivity of the membrane. Carbon Nanotube-Based Desalination Membranes Theoretical studies and molecular dynamics simulations suggest that hydro- phobic channels, like carbon nanotubes, can have considerable water occupancy and that the flow of water in carbon nanotubes is frictionless, limited only by the barriers at the nanotube channelâs inlet and outlet. The observed flow rates in the molecular dynamics simulations were quite high, comparable to water flows ob- served in biological water channels (aquaporins) (Bolhuis and Chandler, 2000; Hummer et al., 2001; Kalra et al., 2003). Inspired by these studies, recent ex- perimental investigations demonstrated high water flows through carbon nano- tubes, with values exceeding those calculated from continuum hydrodynamic models by more than three orders of magnitude (Hinds et al., 2004; Holt et al., 2006). continued
88 Desalination: A National Perspective While these experimental observations are promising, no studies have been carried out so far that demonstrate rejection of salt by such nanotube mem- branes. It is also not clear at this time how such membranes will perform with seawater and brackish waters, where fouling can be an important factor. Finally, even if such nanotube membranes demonstrate desalination performance, scal- ing to membrane modules and cost of production will remain major obstacles. Nanocomposite Membranes Recent efforts aimed at improving the permeability and selectivity of dense polymeric membranes for gas separation demonstrated that dispersion of fumed silica nanoparticles in glassy amorphous polymers can enhance both the perme- ability and the selectivity of such membranes (Merkel et al., 2002). Mahajan et al. (2002) showed that composite materials comprised of molecular sieve domains (zeolites) embedded in polymer matrices can enhance membrane permeability and selectivity. While the materials and synthesis protocol of gas separation and RO membranes are different (Mulder, 2004), it is possible that nanocomposite RO membranes formed by dispersion of nanoparticles or molecular sieves in polymers would yield enhanced membrane performance. Recently, Jeong et al. (2007) reported the formation of novel thin-film nanocomposite RO membranes incorporating zeolite nanoparticles dispersed within the thin polyamide active layer. The results suggest that the nanoparticles in the active layer can play a role in water permeation and salt rejection. The reported performance data, how- ever, do not exceed the performance of current commercial RO membranes. Further research on refining the synthesis method of thin-film nanocomposite membranes may result in membranes with enhanced water flux. Biomimetic Membranes Aquaporins (water channels) and ion channels of biological cells are attract- ing great interest for their potential to overcome the limitations of polymeric dense membranes and increase water flux and selectivity (Miller and Mayer, 2005), even though there is no published work at this time on biomimetic mem- branes for engineered applications. Cell membranes of animals and plants are highly selective barriers that regulate the transport of water, ions, and uncharged solutes into the cell by means of specialized protein channelsâ¯aquaporins for transport of water and ion channels for regulating the transport of ions (Borgina et al., 1999). The most unique aspect of the water and ion channels is their very high selectivity. Aquaporins allow only water molecule transport through the pro- tein channel at flow rates several orders of magnitude larger than expected for a channel of only a few angstroms in diameter. Similarly, ion channels are highly selective structures in membrane cells, allowing for the selective transport of certain ions. No synthetic analogues have been developed with water permeabil- ity and ion selectivity as high as those found in aquaporins and ion channels, although Walz et al. (1994) incorporated aquaporin proteins into a lipid bilayer membrane that exhibited extremely high water permeability. Development of synthetic analogues for aquaporins or incorporation of aquaporins within a mem- brane matrix may lead to a major advancement in current desalination mem- brane technology.
State of the Technology 89 dix B), however, the committee concludes that the practical upper limit of energy savings that can be realized through advances in RO mem- branes is approximately 15 percent. This estimate was made by assuming a system operating at 40 percent recovery with a 95 percent energy re- covery device and a new advanced seawater RO membrane with twice the permeability of todayâs best available membranes while not sacrific- ing salt rejection characteristicsâa huge advancement above todayâs best available technology. This simple analysis implies that the RO proc- ess is approaching a state of diminishing returns as it relates to energy usage. Although these improvements would still provide a cost sav- ings to the desalination process, an improvement in energy savings be- yond 15 percent appears to be a significant challenge. Improvements in module design that enable operation at higher fluxes appear to have the greatest potential for reducing the overall operating costs of desalination because the capital costs and energy costs per cubic meter of permeate produced would simultaneously be reduced (see Box 4-6). Alternatively, a breakthrough in an alternate technology to RO may allow even greater energy savings (see Box 4-5). Electrodialysis and Electrodialysis Reversal The ED and EDR processes use ion-selective membranes and an electrical potential driving force to separate ionic species from water. Ionic species are driven through cation- and anion-specific membranes in response to the electrical potential gradient while the ion-depleted water passes between the membranes. The EDR process is similar to the ED process, except that it also uses periodic reversal of polarity to effectively reduce and minimize scaling and fouling, thus allowing the system to operate at comparatively higher recoveries. As of 2005, ED represented over 3 percent of the worldwide online desalination capacity and nearly 8 percent of online capacity in the United States (GWI, 2006b). EDR and ED processes are typically used to desalt brackish waterâ not seawaterâbecause the cost of these processes increases significantly with higher salinity or TDS (Figure 4-10). In general, municipal applica- tions for ED/EDR have been noted for brackish waters with TDS up to 7,500 mg/L (Mickley et al., 2006), although ED/EDR is typically cost- competitive with RO for TDS up to about 3,000 mg/L (Mallevialle et al, 1996). As with other membrane processes, ED membranes are subject to fouling and some pretreatment of the feedwater is necessary. Typically, prefiltration is required to remove suspended solids and CO2 is removed to improve energy efficiency (Kiernan and von Guttberg, 2005; Reahl, 2006; Weber, 1972).
90 Desalination: A National Perspective Even though the ED/EDR process is more energy intensive than RO with source water above approximately 3,500 mg/L TDS (Figure 4-10), the process still maintains an important niche in desalination technolo- gies. Unlike RO and thermal desalination processes, ED is only capable of removing ionic components from solution. As a result of this phe- nomenon, fouling by uncharged species like silica is less severe as com- pared to the RO process. Additionally, current ED/EDR membranes are resistant to chlorine, making them more robust for processing feedwaters with higher levels of organic matter that would typically foul RO mem- branes (e.g., water reuse applications). These features are important fac- tors that increase the practical application of ED/EDR over RO for such challenging applications. Because of the robust nature of EDR, it is also being applied in hybrid applications as a concentrate reduction method for RO processes (Kiernan and von Guttberg, 2005; Reahl, 2006). Thermal Desalination Processes Thermal distillation was the earliest method used to desalinate sea- water on a commercial basis, and thermal processes have been and con- tinue to be a logical regional choice for desalination in the Middle East for several reasons. First, the seas in the region are very saline, hot, and periodically have high concentrations of organics, which are challenging conditions for RO desalination technology. Second, RO plants are only now approaching the large production capacities required in these re- gions. Third, dual-purpose cogeneration facilities were constructed that integrated the thermal desalination process with available steam from power generation, improving the overall thermodynamic efficiency by 10-15 percent (Hamed et al., 2002; Hanafi, 2002). For these reasons combined with the locally low imputed cost of energy, thermal processes continue to dominate the Middle East. In other parts of the world, where integration of power and water generation is limited and where oil or other fossil fuels must be purchased at market prices, thermal processes are relatively expensive (GWI, 2006a). In the United States, thermal processes are primarily used as a reli- able means to produce high-quality product water (â¤ 25 ppm TDS) for industrial applications, because distillation processes are very successful at separating their targetâdissolved saltsâfrom the bulk feedwater. Dis- tillers almost completely reject dissolved species, such as boron, which can be problematic for RO. Distillers, however, are sensitive to volatile contaminants that may evaporate from the feedwater and carry over into the distilled water, where they may or may not condense. Three major thermal processes have been commercialized: MSF dis- tillation, MED, and MVC, and each is a mature and robust technology
State of the Technology 91 (see Box 4-7 and Table 4-2). MSF and MED processes demand both thermal energy (typically steam) and electrical energy. Thermal proc- esses are configured to use and reuse the energy required to evaporate water, known as the latent heat of evaporation (about 2,326 kJ/kg of wa- ter or 644 kWh/m3 at normal atmospheric conditions). How efficiently the latent energy is reused is a function of project-specific economics, considering capital and operating costs. BOX 4-7 Overview of Thermal Desalination Processes Three primary thermal desalination processes have been commercially de- veloped: â¢ Multistage flash (MSF) distillation, a forced circulation process, is by far the most robust of all desalination technologies and is capable of very large production capacities per unit. Globally, MSF is among the most commonly em- ployed desalination technologies. MSF uses a series of chambers, or stages, each with successively lower temperature and pressure, to rapidly vaporize (or âflashâ) water from the bulk liquid. The vapor is then condensed by tubes of the inflowing feedwater, thereby recovering energy from the heat of condensation (Figure 4-12). The number of stages used in the MSF process is directly related to how efficiently the system will use and reuse the heat with which it is pro- vided. FIGURE 4-12. Multistage flash evaporation. SOURCE: Buros et al (1980); Buros (2000). â¢ Multiple effect distillation (MED) is a thin-film evaporation approach, where the vapor produced by one chamber (or âeffectâ) subsequently con- denses in the next chamber, which exists at a lower temperature and pressure, providing additional heat for vaporization (Figure 4-13). MED technology is be- ing used with increasing frequency when thermal evaporation is preferred or re- quired, due to its reduced pumping requirements and thus its lower power use compared to MSF. MED plants were initially limited in size but MED technology 3 is planned for an 800,000 m /day desalination plant in Jubail, Saudi Arabia. continued
92 Desalination: A National Perspective Since the early 1990s, MED has been the process of choice for industrial low- grade-heat-driven desalination. The largest MED plants incorporate thermal va- por compression (TVC), where the pressure of the steam is used (in addition to the heat) to improve the efficiency of the process. FIGURE 4-13. Multiple effect distillation process. SOURCE: Buros et al (1980); Buros (2000). â¢ Vapor compression (VC) is an evaporative process where vapor from the evaporator is mechanically compressed and its heat used for subsequent evaporation of feedwater (Figure 4-14). VC units tend to be small plants of less 3 than 2,839 m /day that are used where cooling water and low-cost steam are not readily available. VC systems can operate at very high salt concentrations and the VC process is at the heart of many industrial zero liquid discharge sys- tems (Pankratz and Tonner, 2003). FIGURE 4-14. Vapor compression process. SOURCE: Buros et al (1980); Buros (2000). Other nonhybrid thermal desalination approaches, including solar stills and freezing, have been developed for desalination, although they have not been commercialized to date (Buros, 2000). In brief, a solar still uses the sunâs energy to evaporate water from a shallow basin, which then condenses along a sloping glass roof. Freezing technologies are described in Box 4-6.
State of the Technology 93 The combined energy requirements of thermal technologies are greater than that of membrane processes, but it is not simple to compare the total energy use of these diverse processes, because MSF and MED are capable of using low-grade and/or waste heat, which can significantly improve the economics of thermal desalination (see Box 4-8). Utilities in the United States have generally overlooked opportunities to couple thermal processes with sources of waste heat to produce desalinated wa- ter more economically. In the Middle East, the largest of the MSF and MED plants are built along with power plants and use the low- temperature steam exhausted from the power plant steam turbines. This âcogenerationâ approach combines water production with the generation of electric power using the same fuel and offers a method to improve the energy efficiency of desalination plants while sharing intake and outfall structures. Large MSF distillers are commonplace in the Middle East largely because of cogeneration. In another example, many of the largest modern cruise ships select the thermal MED desalination process be cause MED requires 20 to 33 percent of the electrical energy of RO and because the heat energy it requires can be obtained from the shipsâ pro- pulsion engines. MSF and, increasingly, MED units are also used in in- dustry to make water for liquid natural gas and methanol plants. These industrial processes have a relatively small demand for freshwater rela- tive to the massive quantities of waste heat generated by the petrochemi- cal process and can be designed to be quite inefficient. When the residual heat energy has little or no value, there is no economic justification to invest in more efficient designs. Scale deposition in thermal desalination units is a concern but is gen- erally mitigated by control of the operating temperatures and concentra- tions and use of polymer-scale inhibitors. The potential for mineral-scale deposition in a thermal desalination plant is an economic optimization issue, not a limitation of the process. Thermal technologies, including variations of MSFâs forced circulation configuration, can work with su- persaturated salt solutions and are used in brine concentrators for mini- mizing the volume of desalination concentrate. However, operating at extremely high recoveries is not usually economical for desalination ap- plications due to the boiling point elevation caused by the salt. In fact, economic considerations affected by boiling point elevation normally limit water recovery of thermal seawater desalination plant designs to about 35 to 50 percent, not considering cooling water. Although thermal desalination technologies are mature technologies, opportunities remain for additional cost savings. Thermal technologies are not optimized to the highest efficiencies due to current practical con- straints in materials and design and considerations of the source, condi- tion, and value of the thermal energy being utilized. All thermal proc-
94 Desalination: A National Perspective BOX 4-8 Low-Grade and Waste Heat for Desalination Low-grade heat and waste heat are two terms that are often used synony- mously but, depending upon the application, they may have completely differ- ent meanings. The term low-grade heat is often used to describe heat energy that is available at relatively low (near-ambient) temperatures that is of minimal value for industrial or commercial processes. In contrast, waste heat, which may or may not be low-grade heat, contains energy that is released to the environ- ment without being used. Both have potential value for desalination. Most of the largest desalination facilities in the world are dual- purpose facilities that produce both freshwater and electricity. In all of these facili- ties at least some of the electricity is generated by high-pressure steam when it is expanded through turbines. In the case of backpressure turbines, when the steam leaves the turbine, it can no longer produce electricity even though it is still slightly above atmospheric pressure. The waste energy from this exhaust steam is ideal for use by thermal desalination processes. In contrast, condensing tur- bines have a cool exhaust steam under vacuum conditions. Therefore, when condensing turbines are used in combination with thermal desalination, some low-pressure steam is extracted for use in the desalination process. Extracted low-grade steam could, in theory, be used to generate more electricity, but prac- tical circumstances (e.g., electricity demand, limited operating flexibility) influence whether this low-grade energy would, in fact, be used this way. Thus, low-grade heat might also be wasted under specific circumstances. Large slow-speed die- sel generators, such as those used to power large ships, also represent a source of low-grade heat that is often wasted. The cooling water can easily be used to heat both MED and MSF processes without affecting the efficiency of the power generation. Exhaust-gas boilers can also be added to capture otherwise wasted energy for use for desalination or to generate additional electricity. There are other potential sources for waste heat that are simpler to identify as waste, such as industrial stack emissions or cooling circuit heat that is re- jected to rivers, lakes, or the air via heat exchangers or cooling towers. Contrary to common belief, these heat plumes may contain useful energy, even though this energy may not be economical for use in the existing industrial processes. There are economic costs associated with the use of waste or low- grade heat, such as the cost of installing and operating the heat recovery system. The act of recovering the heat may also affect the efficiency of the main process. When a previously wasted energy stream is used, it may then be valued as a potential revenue stream by its owner. When these costs are considered, the energy is not free, but in many cases energy costs can be reduced to a small fraction of the total process cost of desalination. esses are affected by the cost of heat transfer surfacesâwhich are pri- marily copper or titanium alloysâand could benefit from development of new material options. Also, the methods of distributing feedwater over the heat-transfer surface of thin-film processes (e.g., MED, MED-TVC, VC) are proprietary and could benefit from development. There may be
State of the Technology 95 additional opportunities for improved efficiencies in new designs of thermocompressors for MED-TVC systems. There are also needs for additional research and development into improved configurations and applications to utilize low-grade and/or waste heat and into entirely new processes that optimize the use of low- grade heat (see Box 4-5). For example, there has been a recent review (Shih and Shih, 2007) of an industrial application that would utilize low- grade energy at sulfuric acid plants. Heat is produced when sulfur is burned and when concentrated acid is diluted. Thermal desalination plants incorporated into this process could therefore produce the water used to dilute the acid which in turn produces the heat required for the thermal desalination process. The location of low-grade and/or waste heat resources near saline water sources and large consumers of water, including industry, has not been investigated, and research on opportuni- ties to utilize low-grade and/or waste heat could yield economical appli- cations of existing thermal desalination technology in the United States. Ion Exchange Ion exchange is mainly used for water softening and demineraliza- tion, and applications of ion exchange at the municipal level are limited. In an ion-exchange process, water can be desalted by first passing it through a column of cation exchanger beads in the hydrogen (H+) form. Hydrogen ions replace the cations in the solution, which become bound to the exchanger. The water is then passed through a column of anion- exchange beads in the hydroxyl (OHâ) form where the anions replace the hydroxyl ions, which in turn react with the hydrogen ions in the water. This process can produce almost completely deionized water. When ex- hausted, the exchangers can be regeneratedâthe cation exchanger with acid and the anion exchanger with base. The problem is that removal of 1 pound of salt takes about 1.5 pounds of acid and 1.5 pounds of base to regenerate the exchangers (Xu, 2005). This process makes economic sense compared to other desalination processes only where there is a small amount of salt to be removed from the water. Therefore, the major application of ion exchange has been in the field of production of ultrapure water. Ion exchange is often used as a âpolishingâ step following another desalting process. Thus, ion exchang- ers alone cannot economically be used for desalination of seawater or brackish water.
96 Desalination: A National Perspective Hybrid Configurations Hybrid desalination configurations include combinations of proc- esses designed to improve process efficiency or reduce energy costs. Hy- brid thermal-membrane facilities incorporate both thermal and mem- brane desalting processes that are typically co-located with a power plant to improve overall process economics. One hybrid approach blends the product water from parallel RO and thermal desalination processes, which enables the RO membranes to operate with higher permeate TDS (Ludwig, 2004) and which can reduce the replacement costs of RO membranes by up to 40 percent (Hamed, 2005). Hybrid thermal- membrane facilities can also optimize water production and energy costs under seasonal variations in power loads, because operation can be switched from electrically driven RO to thermally driven distillation. In periods of high power demand there will also be associated abundant steam generation such that thermal desalination operations can be maxi- mized; conversely, when there is low power demand (and reduced quan- tities of available low-grade or waste heat), water production by RO is likely to become more economical (Ludwig, 2004). However, to utilize this operational flexibility and realize the cost benefits, the total installed capacity must be larger than the nominal demand for water. Fujairah in the United Arab Emirates is one facility of this type, with a total water production of 454,000 m3/day (120 MGD). Two-thirds of the production capacity is provided by MSF units, with the remaining capacity provided by seawater and second-pass brackish-water RO units. The facility also has the capacity to use warm MSF cooling water as part of the feedwater to increase the permeability of the RO process during winter months. Hybrid desalination facilities may also integrate multiple processes in series to increase the separation or concentration capabilities of the facility. These series hybrids are typically smaller in capacity. For exam- ple, zero liquid discharge (ZLD) systems (i.e., facilities with no offsite liquid-waste discharges) often concentrate the desalination waste stream by separating the process into logical steps and optimizing the entire sys- tem, using RO systems followed by distillation concentrators and crystal- lizers. Another hybrid example is the combination of ED and RO pro- posed by Davis (2006). The process uses ED to reduce the salinity of the reject stream from the RO so that the salt-depleted reject stream can be recycled to the RO to increase recovery. Hybrid configurations in series can also be used to create ultrapure water required by some industrial processes. The multitude of possible combinations of desalination proc- esses in hybrid configurations is limited only by ingenuity and the identi- fication of economically viable applications.
State of the Technology 97 Assessment of Desalination Process Technology The major desalination technologies currently in use are generally ef- ficient and reliable, but the cost and energy requirements remain high. Ongoing research efforts are motivated by the need to reduce cost or to overcome operational limits of a process, such as reducing membrane fouling or increasing energy efficiency. Although existing desalination technologies will continue to see incremental improvements, the current technologies are relatively mature, and the practical limits of further en- ergy savings through advances in RO membranes is approximately 15 percent. Thus, alternatives to the major desalination technologies con- tinue to be investigated to enhance or replace existing desalination proc- esses or fill niche applications where mainstream technologies are inap- plicable (see Boxes 4-5 and 4-6). As discussed earlier, no desalination process can overcome the thermodynamic limit of desalination in terms of energy use (see Box 4-3). Nevertheless, research on alternative desali- nation technologies is under way in hopes of more closely approaching the thermodynamic energy limit or finding ways to power the desalina- tion process with less-expensive energy sources, such as low-grade heat. POST-TREATMENT Desalinated water, produced directly from either thermal or mem- brane processes, is significantly stripped of dissolved solids, which re- sults in a water quality that has low hardness and alkalinity. Conse- quently, without proper post-treatment, this water would be corrosive to pipeline materials, including metals and concrete, and may introduce metals into drinking water and reduce the lifetime of water-system infra- structure. Current technology enables desalinated water to be made non- corrosive by adding chemicals such as calcium hydroxide (slaked lime) to increase the hardness and alkalinity and sodium hydroxide (caustic soda) to adjust the pH. Carbon dioxide is commonly used to normalize the pH. Post-treatment of water is a mature science. The water chemistry issues are generally well understood and methods of altering chemical conditions are feasible and generally available, although the exact proc- ess used will depend greatly on the particular chemistry of the desali- nated water and the existing infrastructure (see also Chapter 7). CONCENTRATE AND RESIDUALS MANAGEMENT All desalination processes leave behind a concentrated salt solution that may also contain some pretreatment and process residuals (see Ta-
98 Desalination: A National Perspective bles 4-1 and 5-1). Concentrate and residuals management involves waste minimization, treatment, beneficial reuse, and disposal, and conventional concentrate management approaches are described in this section. Each approach has its own set of costs, benefits, environmental impacts, and limitations (Sethi et al., 2006a). Further, state regulations may limit the concentrate management practices available at any individual site (Mick- ley, 2006). Because of the widely varying level of technology involved in concentrate management options and site-specific factors and regula- tory considerations that limit available alternatives, the cost of concen- trate management can range from a relatively small fraction of the cost of the main desalination system to as high as several times the cost of the desalination system. The state of the technology, including advantages and disadvantages, for each of the current methods of concentrate management is discussed in this section. A summary of the challenges and limitations in the cur- rent state of concentrate management methods is also provided in Table 4-5. The environmental impacts of concentrate management alternatives are discussed separately in Chapter 5. Surface Water Discharge Surface water discharge to a receiving body is the most common concentrate management practice in the United States, employed by ap- proximately 41 percent of municipal desalination facilities greater than 95 m3/day (Mickley, 2006; Figure 4-15) and at all seawater desalination facilities of significant capacity worldwide. Direct surface water dis- charge of concentrate is a relatively low-energy, low-technology solution to concentrate management. Costs are generally low assuming that the length of the pipeline is reasonable and the concentrate meets the permit requirements without the need for further treatment. However, it has the potential for negative impacts on aquatic organisms (see Chapter 5) and for complex permitting requirements. The salinity of the concentrate is typically higher than that of the ambient water, but good site location, engineering practice, diffuser design, and/or dilution with additional wa- ter (or treated wastewater) prior to discharge can likely minimize most potential negative environmental effects. Multiport discharge diffusers (Figure 4-16) are being employed at some seawater desalination plants to minimize environmental impacts. Studies show that concentrate, being denser than seawater, may sink and impact benthic communities. By employing multiple outlet ports, rather than a single open pipe, the mixing and dilution of the concentrate can be accelerated, lessening potential impacts in sensitive areas (EPA, 1991).
State of the Technology 99 Land disposal, 2% Unreported, 7% Evaporation ponds, 2% Surface waters , 41% Injection wells, 17% Sewers, 31% FIGURE 4-15. Identified methods of concentrate management, based on a sur- vey of the 234 municipal desalination plants in the United States with output greater than 95 m3/day (25,000 gallon per day). SOURCE: Data from Mickley (2006). FIGURE 4-16. Multiport diffuser for improved initial mixing of surface water con- centrate discharge. SOURCE: EPA (1991). Blending and diluting the concentrate with treated wastewater or power plant cooling water is also desirable to reduce environmental impacts but is not an option for all site locations. Shipboard desalination configura- tions can blend the concentrate with large volumes of ocean water, taken in specifically for that purpose.
100 Desalination: A National Perspective Sewer Discharge Discharge of concentrate into an existing sewage collection system is the second most common concentrate management practice in the United States, employed by approximately 31 percent of surveyed municipal desalting facilities (Mickley, 2006). This method is also relatively low in cost and in energy use but retains the potential for adverse environmental impacts from elevated concentrations of salt or trace elements in the treated effluent (Table 4-5). A permit from the local sewage agency is required to ensure that potential adverse impacts on the wastewater treatment processes, if any, are within acceptable limits. Large-volume discharges are typically not practical or suitable. Subsurface Discharge Concentrate discharge via a subsurface discharge structure, such as a deep well, can occur in both inland and coastal areas. Deep-well injec- tion is a mature technology that involves the disposal of concentrate into a deep geological formation, usually inland, that will serve to perma- nently isolate the concentrate from aquifers that may be used as a drink- ing water source. Appropriate geology with the presence of a structurally isolating and confining layer between the receiving aquifer and any over- lying source of drinking water is required. Suitable formations for injec- tion often contain water with TDS concentrations in excess of 10,000 mg/L. These conditions are determined through site-specific hydro- geologic assessments. Deep-well injection is commonly employed by desalination plants in certain parts of Florida and island applications where receiving aquifers can be found at relatively shallow depths, but it is less common elsewhere in the United States. Deep-well injection is used at 12 percent of municipal desalination facilities in the United States with output greater than 95 m3/day (Mickley, 2006). It is typically employed for larger desalination plants (e.g., >3,800 m3/day [>1 MGD]) because the costs for developing deep-injection wells are relatively high and are not largely reduced for smaller flows. For example, the typical capital cost of a 3,000-m-deep well is reported at $8.1 million for a con- centrate flow of 3,800 m3/day, which decreases to only about $5.1 mil- lion for a concentrate flow of either 380 or 38 m3/day (Malmrose et al., 2004). These costs exclude any pretreatment or standby disposal system. While capital costs for well injection are about average of typical inland concentrate management methods, the annual operating costs are rela- tively low as a percentage of total operating costs (Mickley, 2006). For
State of the Technology 101 example, the capital cost of well injection and several tens of miles of delivery pipes was 26 percent of total capital costs for the desalination project, but the operating costs are estimated to be 4 percent of the an- nual operating cost for the El Paso Water Utilities desalination facilities (Ed Archuleta, El Paso Water Utilities, personal communication, 2006; see also Box 5-2). If permitted by state regulations, existing wells from depleted oil and gas reservoirs could be used for concentrate injection, although their injection capacity would need to be evaluated, and the costs of transporting the concentrate may offset other cost savings. For seawater desalination, subsurface discharge involves using a beach well or percolation gallery beneath the beach or underneath the seafloor. Because mixing occurs beneath the surface and the discharge plume slowly dissipates into the surf zone, subsurface coastal discharge can be an effective way to minimize environmental impacts, although it requires specific hydrogeological conditions. Land Application Land application of brackish desalination concentrate can be used for lawns, parks, golf courses, or crop land; it is not practical for the large volume and highly saline concentrate from seawater desalination and is thus only considered for brackish water applications. Even for these ap- plications, a TDS greater than about 5,000 mg/L in the concentrate can typically preclude spray irrigation (Mickley, 2006); thus, there is typi- cally a need for addition of dilution water. Land application is usually practical and employed only for smaller concentrate flows and, because irrigation demands are seasonal, a second or backup disposal or storage method is also necessary for year-round operation (Malmrose et al., 2004). The key concerns with spray irrigation include the influence of concentrate on the soil and vegetation, potential contamination of groundwater, and runoff to surface water (see Chapter 5). The allowable salinity will depend on the tolerance of target vegetation, percolation rates, and the ability to meet the groundwater quality standards. In gen- eral the vegetation used is dependent on the site location; however, typi- cally grasses that are high in water and salinity tolerance are used with concentrate discharge. Research is now under way to develop genetically modified feed crops that would tolerate and take up more salt.
102 Desalination: A National Perspective TABLE 4-5. Concentrate Management Challenges and Limits Land Applicable Potential Capital O&M Area Permitting for Large Environmental a a Method Costs Costs Required Complexity Conc. Flows Impact Surface a a water L L -- H Yes M discharge Sewer a a L L -- M No M discharge Subsurface discharge M-H M L M Maybe L (deep well injection) Evaporation c H L H M No M Pond Land M L H M No M-H Application Thermal d d H H L L No L evaporation L = low; M = medium; H = high; dashes indicate not applicable. a Costs are highly site-specific; general trends in relative costs are indicated; cost for sur- face water or sewer discharge can be higher if the distance from desalination facility to the discharge water body or sewer is large, necessitating long pipelines and/or pumping facili- ties. b Energy use for surface water or sewer discharge or land application can possibly be higher if the distance from desalination facility to the discharge water body, sewer, or land application site is large, possibly necessitating pumping facilities. c O&M costs for evaporation ponds can possibly be higher if a significant amount of moni- toring wells and associated water quality analysis are required. d Permitting complexity and environmental impacts of thermal evaporation can possibly be higher if the feedwater-to-desalination process contains contaminants of concern that could be concentrated to toxic levels in the concentrated slurry or solids that are produced from this concentrate treatment process. e Low (L) pertains to Florida (where deep-well injection is commonly practiced) and moder- ate (M) pertains to other states in the United States. f Climate can indirectly influence surface water discharge by affecting the quantity of sur- face water available for dilution. Evaporation Ponds Evaporation ponds are a low-technology but high-cost approach to concentrate management, where the concentrate is pumped into a shal- low lined pond and allowed to evaporate naturally using solar energy. In evaporating environments, the thermodynamic activity of the concentrate decreases with increasing concentration and will approach the average relative humidity of the air. At that point, effective natural evaporation will ceaseâthat is, an evaporation pond that is not leaking will not
State of the Technology 103 Labor Needs Possible and Skill Public Special Pre-treatment Level (for Energy Perception Climate Geological Needs operation) Use Concerns Limitation Requirements b f M L L H Maybe N b L L L L N N e L L M L-M N Y L L L H Y Y b L L L H Y Y L H H L N N evaporate to dryness in most environments. Periodic removal and drying of accumulated solids is necessary for long-term physical sustainability of a site, although some evaporation ponds are closed and sealed once the pond is filled by solids. Evaporation ponds, under suitable climatic conditions, enable operation of the desalination plant under ZLD condi- tions, where no liquid waste leaves the plant boundary. Evaporation ponds can be a viable option in relatively warm, dry climates with high evaporation rates, level terrain, and low land costs. They are typically practical and employed only for smaller concentrate flows and are often coupled with high-recovery desalination processes. This disposal method has high capital costs due primarily to the land acquisition costs to ac- commodate the large surface areas required and also the costs of imper- meable liners, if needed. For example, assuming a relatively high evapo- ration rate of 0.1 L/h/m2, the typical capital cost of an evaporation pond is reported at $40 million for a concentrate flow of 3,800 m3/day (1 MGD), reducing to $4 million for a concentrate flow of 380 m3/day (Malmrose et al., 2004). The costs for a lower evaporation rate would be proportionally higher (e.g., four times as much for a rate of 0.025 L/h/m2). These costs exclude any solids disposal or seepage monitoring.
104 Desalination: A National Perspective Thermal Evaporation A thermal evaporator (also known as a brine concentrator) can re- duce desalination concentrate to a slurry of approximately 20 percent solids. Thermal evaporators are generally based on vapor compression technology (See Box 4-7). Most evaporators in operation are of the verti- cal-tube, falling-film type and employ a calcium sulfate seeded slurry process to prevent scaling (Mickley, 2006). Thermal evaporation uses a large amount of energyâmore than 18.5 kWh/m3 of feedwater (Bond and Veerapaneni, 2007). Thermal evaporators have been used in indus- trial RO applications and are known to be a viable and reliable technol- ogy. When an evaporator is followed with a crystallizer or spray dryer, the concentrated slurry can be further reduced to solids that are suitable for landfill disposal (a zero liquid discharge (ZLD) approach). Like thermal evaporators, crystallizers are also driven by vapor compression. In a crystallizer, the brine enters the vapor body at an angle and swirls into a vortex. As water evaporates, the salt crystals are separated using a centrifuge or filter. Spray drying is another means of producing dry product from concentrate. In this method the concentrated salt solution is reduced to a fine spray and mixed with a stream of hot gas, which pro- vides the heat for evaporation and carries off the moisture released from the concentrate. The resultant dry salt powder is collected in a bag filter. The crystallizer and spray drying processes are more capital-cost- and energy-intensive than the brine concentrator. The energy requirements of thermal evaporators combined with crystallizers can exceed 32 kWh/m3 (Pankratz, 2008). These thermal technologies may be coupled with other membrane- based high-recovery processing technologies, described earlier in the chapter, to reduce the overall energy requirements. In a pilot study of five inland brackish water sources, Bond and Veerapaneni (2007) were able to reduce the total energy use of desalination with ZLD to 0.45-1.9 kWh/m3 of product water by developing a process train involving two RO passes, intermediate concentrate treatment, and a brine concentrator, followed by an evaporation pond. In general, thermal evaporation-based processes are characterized with high capital costs and energy requirements. The capital and operat- ing costs of these thermal evaporation methods can sometimes exceed the cost of the desalting facility (see Chapter 6). Additionally, once most or all of the liquid is removed from the wastes, landfilling costs can be significant. The high costs, including high-energy requirements, are a large deterrent to application of this process, particularly for large mu- nicipal applications. Thus, at the present time, ZLD concentrate man- agement approaches are typically not considered for municipal drinking
State of the Technology 105 water applications. Nevertheless, thermal evaporation processes are be- ing considered for some water supply applications in inland regions where the concentrate flows are small and other methods of concentrate management are not feasible (e.g., the desalination facility at the Deuel Vocational Institution in Tracy, California, incorporates a brine concen- trator). For thermal evaporation applications to become more viable, im- provements are needed that reduce capital costs and/or energy usage. Development of beneficial salt reuse options and specific salt separa- tion methods are also important to cost reduction of the overall process (Drioli et al., 2004). Some examples of beneficial reuse of the solid product include extraction of gypsum and sodium chloride by means of selective precipitation. However, the economic viability of beneficial reuse of desalination by-product salts depends on finding local markets to avoid high transportation costs (Jordahl, 2006). Residuals Management In addition to the concentrate, there are other waste streams from de- salination that need to be managed. The state of the science here is rela- tively mature. The spent cleaning solutions are either disposed of with the concentrate or separately, usually into a sewer system. In the latter scenario, some pretreatment in the form of neutralization for pH adjust- ment may be required. In the pretreatment step, solids are removed. In some cases, these solids can be recombined with the concentrate discharge and disposed to the source water. However, more commonly the solids are separated with clarifiers and sent to a belt press for further dewatering. The resulting sludge is then hauled to a landfill. The use of a microfilter will reduce the volume of sludge to be settled in the clarifier as long as the mass of chemicals required to flocculate solids is reduced over conventional processes. The membranes and security filter cartridges also constitute a resid- ual when they reach the end of their effective life. These residuals are commonly disposed in landfills. A few companies recover used mem- branes and clean them for further use in a different application. CONCLUSIONS AND RECOMMENDATIONS Although RO and thermal-based processes are relatively mature, op- portunities exist to improve the energy efficiency and reduce costs. For membrane-based desalination, the most significant improvements can be realized through improved pretreatment and the creation of low-fouling,
106 Desalination: A National Perspective high-flux (i.e., high-permeability) membranes that can operate at lower pressures. For thermal-based desalination, costs can be reduced by more effective use of low-grade and/or waste heat. Improvements in concen- trate management would have both cost and environmental implications. These and other research findings are summarized in this section. RO technology is approaching the thermodynamic minimum en- ergy value. The membrane industry has made great strides in reducing energy use for the desalination process with the commercialization of high-efficiency energy recovery devices and improvements in membrane technology. Current energy use is within a factor of 2 of the theoretical minimum value for seawater desalination. Practical and economic con- straints are likely to inhibit RO energy use from decreasing more than approximately 15 percent from current values. This level of improvement would still be valuable to reducing cost and energy use, but greater re- turns on investment than this should not be expected. RO is the standard by which novel technologies should be assessed when specific energy use is the main consideration. Although the RO process is relatively mature, opportunities exist to further reduce the energy use by small but economically signifi- cant amounts. These opportunities include the following: â¢ Reduced fouling through pretreatment, â¢ Development of fouling-resistant membranes, â¢ Development of high-flux (i.e., high-permeability) mem- branes for operation at low pressure; â¢ Development of oxidant-resistant membranes; and â¢ Improving mechanical configuration of membrane modules and membrane system design. Operating the RO process at lower hydraulic pressure while main- taining high throughput is the key to reducing the specific energy for membrane-based desalination. To fully utilize the capacity of the high- permeability RO membranes and to accommodate even more permeable RO membranes in the future, it is imperative to reduce fouling and con- centration polarization effects and to develop new module configurations and system designs to avoid or overcome thermodynamic restriction. Fouling can be reduced by a more robust pretreatment and by the devel- opment of fouling-resistant membranes. Pretreatment for RO desalination can be improved by replacing conventional physicochemical processes with membrane-based (UF
State of the Technology 107 or MF) pretreatment. Conventional pretreatment technologies based on coagulation and sand filtration cannot always achieve sufficient removal of foulants. Membrane-based pretreatment, particularly UF, can produce water of superior quality with very low fouling potential. Such effective pretreatment is essential for efficient utilization of future high-flux mem- branes. Seawater desalination using thermal processes can be cost- effective when waste or low-grade heat is utilized effectively. Loca- tion of low-grade or waste heat resources near large water consumers may reduce the cost of heat energy and offset the higher specific energy requirements of thermal desalination when compared to RO. Hybrid membrane-thermal desalination approaches offer additional operational flexibility and opportunities for water production cost savings for facili- ties co-located with power plants. Thermal desalination technologies are themselves relatively mature; however, additional cost savings could be realized by improvements in materials, process configuration, and opti- mization of low-grade and waste heat resources. Few, if any, cost-effective environmentally sustainable concen- trate management technologies have been developed for inland de- salination facilities. Several methods are currently available for concen- trate management (e.g., surface water discharge, sewer discharge, deep- well injection, evaporation ponds, land application, and high-recovery/ thermal evaporation systems to minimize the volume of waste produced), and each method has its own set of site-specific costs, benefits, regula- tory requirements, environmental impacts, and limitations. Low- to mod- erate-cost inland disposal options can be limited by the salinity of the concentrate and by location and climate factors. Only evaporation ponds and high-recovery/thermal evaporation systems are ZLD solutions, but their costs are high for municipal application.