Advanced Membrane Technology and Applications

PART I

MEMBRANES AND APPLICATIONS IN WATER AND WASTEWATER

CHAPTER 1

Thin-Film Composite Membranes for Reverse Osmosis

TADAHIRO UEMURA and MASAHIRO HENMI

Global Environment Research Laboratories, Toray Industries Inc., Otsu Shiga, Japan

1.1 INTRODUCTION

Because of vastly expanding populations, increasing water demand, and the deterioration of water resource quality and quantity, water is going to be the most precious resource in the world. Thus, the 21st century is called the water century. In the 20th century, membrane technologies made great progress, and commercial markets have been spreading very rapidly and throughout the world. The key technologies fueling this progress are as follows:

1. Materials: Chemical design of high-performance materials suitable for each separation mode

2. Morphology: Morphological design of high-performance membranes

3. Element/Module: Element and module design for high-performance membranes

4. Membrane Process: Plant design and operation technology

In 21st century, to solve these water problems, membranes technology is going to be further expanded and new technology—further improvements of membrane performance, development of membrane systems, membranes stability such as antifouling membranes for wastewater treatment, and other highly qualified membranes—will be needed.

Among desalination technologies available today, reverse osmosis (RO) is regarded as the most economical desalination process. Therefore, RO membranes have played crucial roles in obtaining fresh water from nonconventional water resources such as seawater and wastewater.

1.2 APPLICATION OF RO MEMBRANES

Reverse osmosis membranes have been used widely for water treatment such as ultrapure water makeup, pure boiler water makeup in industrial fields, seawater and brackish water desalination in drinking water production, and wastewater treatment and reuse in industrial, agricultural, and indirect drinking water production as shown in Table 1.1.

TABLE 1.1 Application of Reverse Osmosis Membrane Process

The expansion of RO membrane applications promoted the redesign of suitable membrane material to take into consideration chemical structure, membranes configuration, chemical stability, and ease of fabrication. And along with the improvements of the membranes, the applications are further developed.

1.3 MAJOR PROGRESS IN RO MEMBRANES

1.3.1 Cellulose Acetate Membrane

Reverse osmosis systems were originally presented by Reid in 1953. The first membrane, which could be used at the industrial level in actual water production plants, was a cellulose-acetate-based RO membrane invented by Loeb and Sourirajan in 1960. This membrane has a so-called asymmetric or anisotropic membrane structure having a very thin solute-rejecting active layer on a coarse supporting layer, as shown in Figure 1.1. The membrane is made from only one polymeric material, such as cellulose acetate, and made by the nonsolvent-induced phase separation method. After the invention by Loeb and Sourirajan, spiral-wound membranes elements using the cellulose acetate asymmetric flat-sheet membranes were developed and manufactured by several U.S. and Japanese companies. RO technologies have been on the market since around 1964 (Kurihara et al., 1987). They were widely used from the 1960s through the 1980s mainly for pure water makeup for industrial processes and ultrapure water production in semiconductor industries; and some are still used in some of these applications.

Figure 1.1 SEM photograph of CA asymmetric membrane.

Figure 1.2 Representative chemical structure of linear polyamide membrane (B-10).

TABLE 1.2 Summary of Membrane Materials for RO

1.3.2 Aromatic Polyamide Hollow Fiber Membrane

Since then, there has been intensive and continuous R&D efforts mainly around the United States and Japan to meet the demands from commercial markets, and there exist many inventions and breakthroughs in membrane materials and configurations to improve the performances of membranes.

To overcome the problems of cellulose acetate membranes, many synthetic polymeric materials for reverse osmosis were proposed, but except for one material, none of them proved successful. The only one material, which could remain on the market, was the linear aromatic polyamide with pendant sulfonic acid groups, as shown in Figure 1.2. This material was proposed by DuPont, which fabricated very fine hollow fiber membranes; the modules of this membrane were designated B-9 and B-10. They have a high rejection performance, which can be used for single-stage seawater desalination. They were widely used for mainly seawater or brackish water desalination and recovery of valuable materials such as electric deposition paints, until DuPont withdrew them from the market in 2001.

1.3.3 Composite Membrane

Another approach to obtain a high-performance RO membrane was investigated by some research institutes and companies in the 1970s. Many methods to prepare composite membranes have been proposed, as shown in Table 1.2. In the early stage, very thin films of a cellulose acetate (CA) polymer coating on a substrate, such as a porous cellulose nitrate substrate, was tried. However, in spite of their efforts, this approach did not succeed in industrial membranes manufacturing.

The next approach used the interfacial polycondensation reaction to form a very thin polymeric layer onto a substrate. Morgan first proposed this approach (Morgan, 1965), and then Scala et al. (1973) and Van Heuven (1976) actually applied this approach to obtain an RO membrane. But it was Cadotte who invented the high-performance membrane using the in situ interfacial condensation method (Cadotte, 1985). In his method, interfacial condensation reactions between polymeric polyamine and monomeric polyfunctional acid halides or isocyanates takes place on a substrate material to deposit a thin film barrier onto a substrate. Some of the composite membranes were succeeded in industrial fabrication by another method, which was designated as PA-300 or RC-100.

Another preparation method for composite membrane is an in situ monomer condensation method using the monomeric amine and monomeric acid halide, which was also invented by Cadotte. Then, many companies succeeded in developing composite membranes using this method, and the membrane performance has been drastically improved up to now. Now, composite membrane of cross-linked fully aromatic polyamide is regarded as the most popular and reliable material in the world. Permeate flow rate and its quality have been improved 10 times greater than that of the beginning (Kurihara et al., 1987, 1994b).

1.4 TRENDS IN RO MEMBRANE TECHNOLOGY

Figure 1.3 shows recent trends in RO membrane technology with two obvious tendencies. One is a tendency toward low-pressure membranes for operating energy reduction in the field of brackish water desalination. The other is a tendency toward high rejection with high-pressure resistance in the large seawater desalination market.

Figure 1.3 Technology trends in RO membrane.

1.4.1 Progress of Low-Pressure Membrane Performance in Brackish Water Desalination

Figure 1.4 shows the progress of low-pressure membrane performance trends in RO membrane on brackish water desalination from the 1970s to the 1990s, including industrial water treatment such as ultrapure water production. In the 1970s much effort was devoted to developing high-performance membrane materials and improving the membrane performance. As a result performance was improved with a new developed material of cross-linked aromatic polyamide and by developing membrane morphology and fabrication technology. The cross-linked fully aromatic polyamide composite membrane developed in 1987 has four or five times larger water flux and five times higher product water quality than those of the CA membrane (Kurihara et al., 1987). Since 1987, membrane performance has been drastically developed. On the basis of the development of cross-linked fully aromatic polyamide composite membranes, RO membrane performance of brackish water desalination has improved very rapidly. Typical performances of the RO elements for brackish water desalination are shown in Table 1.3. The ultralow-pressure membrane, which can be used at ultralow pressures such as 0.75MPa, has been developed, which saves on the operating cost (Ikeda et al., 1996). And now the super-ultralow-pressure membrane elements, which can be used at super-ultralow-pressure, such as 0.5–0.3MPa, have been developed (Fusaoka, 1999). This membrane has three times the water permeability than the ordinary low-pressure RO membrane. This membrane can operate with onethird the pressure of a low-pressure membrane.

Figure 1.4 Performance trends in RO membrane for brackish water desalination.

TABLE 1.3 Typical Performance of Toray’s Brackish Water RO Elements

1.4.2 Progress of RO Membranes for Seawater Desalination

The progress of RO membranes for seawater desalination is shown in Figure 1.5 (Kurihara et al., 1994a). It is very important to increase the water recovery ratio on seawater desalination systems to achieve further cost reduction. Most seawater RO desalination systems in use today are confined to approximately 40% conversion of the feed water (salt concentration 3.5%), since most of commercially available RO membrane do not allow for high-pressure operation of more than around 7.0MPa.

Recent progress on high-pressure-high-rejection spiral wound (SW) RO elements, combined with proven and innovative energy recovery and pumping devices, has opened new possibilities to reduce investment and operating cost. The progress of RO seawater desalination from a point of view of water recovery is shown in Table 1.4 (Moch, 2000).

Toray has developed a new low-cost seawater desalination system called the Brine Conversion Two-Stage (BCS) system, as shown in Figure 1.6, which provides 60% water recovery of freshwater (Yamamura et al., 1996). Ohya et al. (1996) and Nakao (1996) also suggest that higher recovery of RO seawater desalination by the BCS system is most effective in saving energy yet keeping a low operating cost.

Figure 1.5 Performance trends in RO membranes for seawater desalination.

As for achieving the 60% RO seawater desalination system, it is absolutely necessary to make the RO membrane element, which can be operated under very severe operating conditions, with high pressure and high feed water concentration such as 9.0 MPa and 5.8%. Toray has developed a high-performance membrane (BCM element) that can be operated at high pressure and high concentration conditions, as shown in Table 1.5.

1.4.3 High Boron Rejection SWRO Membrane

The removal of boron is a significant problem in SWRO desalination processes (Fukunaga et al., 1997). Boron exists as boric acid in the natural water, and boric acid mainly shows the male reproductive tract when administered orally to laboratory animals. The World Health Organization (WHO) proposes that boron concentration in drinking water be below 0.5mg/L as a provisional guideline value (WHO, 2004). However, especially in SWRO desalination fields, this is not an easy goal to meet because boron concentration in seawater is comparatively high. Although conventional SWRO membrane elements have shown a little more than 90% of boron rejection, it is still inadequate. It is difficult for RO membranes to remove boric acid in water for the following reasons: First, the molecular size of boric acid is so small that it is difficult to remove by size exclusion. Second, since boric acid has a pKa of 9.14–9.25, it is not ionized in natural seawater with a pH of 7.0–8.0 and dissociates at pH 9 or more (Rodriguz et al., 2001). The boron rejection by the electric repulsive force between boric acid and the membrane cannot be expected in a neutral condition. Therefore, some posttreatment processes are necessary to meet the WHO proposal.

TABLE 1.4 Typical Performance of Toray’s Seawater RO Membranes

Figure 1.6 Typical flow diagram of BCS system.

The conventional SWRO membrane element TM820, which is typical with Toray, has exhibited 91–93% boron rejection, which was the highest level achieved by commercialized SWRO membrane elements (Toray, 2004; Redondo et al., 2003; Hiro and Hirose, 2000). This membrane element series has been installed in a large number of SWRO plants. And Toray has commercialized many types of SWRO membrane elements, which are for different pressure ranges due to total dissolved solids (TDS) concentration and temperature of the seawater, as shown in Table 1.6. However, the highest boron rejection in those membrane elements is 91–93%, which is the same as TM820. This means that the improvement of boron rejection by membrane material had been sluggish for a while. Meanwhile, the new membrane element TM820A was developed based on the following two concepts: (1) reduction of affinity with boric acid by control of hydrophobic property and functional groups may reject boric acid selectively, and (2) molecular structure design was considered as blocking the boric-acid-permeable large pore (Taniguchi et al., 2004). TM820A exhibited 94–96% boron rejection with high TDS rejection and high water productivity. The specification and typical performance of TM820A is shown in Table 1.7.

TABLE 1.5 Progress in RO Seawater Desalination Plant

Source: Moch (2000).

TABLE 1.6 SWRO Products Lineup Released from Toraya

aTesting conditions were: applied pressure, 800 psi (5.52MPa) for others; recovery ratio, 8%; feed solution: 32,000ppm, NaCl with 5.0mg/L boron for others; pH = 8; temperature, 25°C.

Seen from various viewpoints, a single SWRO system is the most ideal. Therefore, to evaluate the performance of TM820A, the amount of boron that TM820A could remove by a single-stage operation was estimated. Table 1.8 shows the permeate boron concentration that corresponds to the boron rejection performance used by membrane elements when each region of seawater is treated by a single-stage SWRO operation considering the aging factor. According to Table 1.8, TM820A meets the Japanese guidelines of below 1mg/L of boron concentration by a single-stage operation. But in severe conditions, for the WHO guideline grade and the Middle East seawater treatment, certain posttreatment processes are still needed. If 97% of boron rejection performance is gained, the WHO grade will be enabled until the Southeast Asia seawater treatment. Furthermore, at 99% boron rejection performance, the WHO guideline grade will be enabled even in the Middle East seawater treatment.

TABLE 1.7 Specifications and Typical Performances of TM820Aa

aTesting conditions were: applied pressure, 800 psi (5.52MPa); recovery ratio, 8%; feed solution: 32,000ppm, NaCl with 5.0mg/L boron; pH = 8; temperature, 25°C.

TABLE 1.8 Permeate Boron Concentration by a Single-Stage SWRO Operation (Calculated)a

aAssumed conditions: Plant: 7 elements/vessel, 14 lmh; operation: 25°C, pH 8.0, 800 psi, 3.5 L/min, recovery ratio 40%, after 3 years. Japanese law grade: <1.0mg/L, WHO guideline grade: <0.5mg/L.

Recently, Toray has been investigating SWRO membranes that focus on the removal of boron by the improvement of membrane performance. The history and future prospects of the boron removal at Toray and other companies are shown in Figure 1.7. Until 2000, although the boron rejection was also improved as various membranes were developed in each company, it was 90% at best. In the next period, from 2000 to 2003, the membranes in which a little more than 90% of boron rejection was shown were released, and these serve as main items for each company now.

From 2003 to 2005, Toray developed and released TM820A, whose performance was appreciably improved, and offered the membrane that showed around 95% boron rejection prior to other companies. However, the supportive systems are still required to meet the WHO proposal even by using TM820A as above. Thus, the next target is 97 or 99% boron rejection performance of renovative membrane. The further development of a new renovative membrane that can meet the WHO proposal for every seawater continues (Tomioka et al., 2005). Table 1.9 summarizes large seawater RO desalination plants around the world. TM820A is installed in a large seawater RO desalination plant in Singapore.

Figure 1.7 History and prospect of boron rejection performance of SWRO membrane element in Toray and comparable companies.

TABLE 1.9 Large Seawater Desalination Plants Utilizing RO Process (as of July, 2005)

1.5 REVERSE OSMOSIS/BIOFOULING PROTECTION

Biofouling has been regarded as the most serious problem in the operation of SWRO plants. The usual method to prevent biofouling is continuous chlorine dosing to intake seawater with sodium bisulfate (SBS) dosing at the RO portion. However, membranes performance deterioration occurred by oxidation in case of both polyamide and cellulose acetate membranes, and biofouling has not been solved yet. Toray has developed a new method that is effective to prevent biofouling on SWRO membranes and verified its effectiveness at actual plants.

First of all, by measuring the viable counts of bacteria at a plant, in case of the continuous chlorine/SBS dosing method, it was found that a number of bacteria drastically increased immediately after SBS dosing, as shown in Figure 1.8, and most of these bacteria were quite different from those found in raw seawater. Currently, the addition of SBS to feed water at relatively high concentration has been used for sterilization of RO membranes. However, when SBS was added to seawater, the pH was just dropped to 6, and most of the bacteria harbored in water were still alive. This result indicates that the sterilization ability of SBS is due to lower pH, and oxygen consumption with SBS only plays a role to repress the cell growth. Finally, Toray has developed a new agent, MT-901, which is effective in preventing biofouling on RO membranes. Adding MT-901 to seawater instead of SBS effectively killed bacteria in a few samples of seawater within a short time.

Finally, the effect of this method was verified at an SWRO plant. In this plant, when feed water was chlorinated and dechlorinated with SBS continuously and RO membrane modules were treated with SBS intermittently, differentiation pressure of the module increased gradually. MT-901 was used for membrane module treatment in place of SBS and the differential pressure decreased within 10 days. Moreover, using an intermittent chlorination method was effective to maintain the initial differential pressure with less concentration of MT-901 (Kallenberg et al., 1999).

Figure 1.8 Viable cell count assessment in RO plant.

1.6 LOW-FOULING RO MEMBRANE FOR WASTEWATER RECLAMATION

Wastewater reclamation and reuse plants have been constructed and operated around the world. Table 1.10 shows large wastewater reuse plants. RO membrane is necessary for wastewater reclamation to make the water quality reusable. The largest RO plant is operated in Kuwait since 2005.

For RO membrane modules, stable operation is very important. Many organizations, universities, and companies have reported many kinds of operation troubles. Fouling, membrane deterioration, and hardware problems have mainly caused these troubles, and the major troubles, which occupy 80%, are fouling problems.

As described above, it is important to consider the (1) proper RO membrane elements with low-fouling property, (2) proper pretreatment technology before the RO membrane, and (3) suitable sterilization methods and cleaning technology (Kurihara et al., 2003). The reasons for fouling of RO membrane are reported as consisting of chemical fouling, biological fouling, and scale precipitation.

TABLE 1.10 Large Wastewater Reuse Plants (as of July, 2005)

It is estimated that chemical fouling is caused by the adsorption of organic materials such as humic substances and surfactants in the feed water or on the membrane surface. Humic substances have various chemical structures depending on the water origin, such as land water or seawater, and regions in the world. However, it has both hydrophobic groups of aromatic and linear structure and ionic groups of amino acid and carboxylic acid. The material of RO membrane is polyamide with hydrophobic and ionic properties. As mentioned above, chemical fouling depends on hydrophobic interaction and electrostatic interaction between organic materials in the feed water and membrane surface.

On the other hand, in case of biological fouling, the following estimations are reported: (1) microbe adsorption by hydrophobic or electrostatic interaction, (2) propagation of microbe with nutrition in the feed water, and (3) deposition of exhaust material of biological metabolism. Case 1 is a reversible phenomenon; however, cases 2 and 3 are irreversible phenomena, which are difficult to remove by simple chemical cleaning.

As a result of R&D activities, Toray has developed low-fouling RO membrane for wastewater reclamation. The low-fouling RO membrane has the same level of pure water permeability as conventional RO membranes, SU-700 and SUL-G, and also has low-fouling property with keeping water permeability against chemical and biological fouling during the operation (Yamamura et al., 2002).

The low-fouling property of membranes is evaluated with a nonionic surfactant aqueous solution. Test result shows that, in operation, low-fouling RO membrane has a relatively small permeability declaration ratio of 27%, compared with initial pure water permeability and shows stable operation. On the other hand, conventional fully aromatic polyamide membranes, SU-700 and SUL-G, show 36–47% declaration ratio, even if they show stable operation. And concerning the chemical cleaning properties, low-fouling RO membrane shows better recovery of permeability after chemical cleaning.

To evaluate the fouling property against microbes, adsorption property of a certain hydrophobic microbe and other hydrophilic microbes were measured. The hydrophobic microbe was severely adsorbed to conventional RO membranes and caused biological fouling of RO membranes. In case of low-fouling RO membranes, the adsorption property of the hydrophobic microbe is quite low, which is less than one-tenth of conventional RO membranes. Initial performance of low-fouling RO membrane element TML-20 is described in Table 1.11.

A test of wastewater reclamation using low-fouling RO membranes and conventional low-pressure RO membrane SUL-G10 has been carried out in a wastewater treatment facility in Japan, as shown in Figure 1.9. In this test, secondary effluent was directly filtered by ultrafiltration (UF) membrane and permeate was fed to both of the RO membranes. In case of the SUL-G membrane, water permeability dropped to 60% of initial permeability in a day due to biological fouling; however, the permeability drop of low-fouling RO TML20 was smaller than that of SUL-G and the stable operation has been performed.

TABLE 1.11 Initial Performance of Low-Fouling RO Membrane Element TML-20a

aConditions: Pressure = 1.5 MPa, temperature = 25°C, feed water conc. = 1500 (NaCl mg/L), pH 6.5.

Figure 1.9 Water productivity of low-fouling RO compared with conventional RO.

Figure 1.10 Wastewater reclamation and reuse plant in Singapore (24,000 m³/day).

Figure 1.11 Process flow of wastewater reclamation and reuse plant in Kuwait (375,000 m³/day).

Figure 1.12 RO portion of wastewater reclamation and reuse plant in Kuwait.

The low-fouling RO membrane is strongly required for the stable operation of the wastewater reclamation plant. Two large water reclamation and reuse plants have been operated in Singapore and Kuwait, as shown in Figures 1.10, 1.11, and 1.12.

1.7 CHLORINE TOLERANCE OF CROSS-LINKED AROMATIC POLYAMIDE MEMBRANE

Chlorine tolerance is a very important characteristic to design an RO membrane process because chlorine dosing to water process is commonly used as a disinfection for microorganisms. Many authors have studied chlorine tolerance of RO membranes as listed in Table 1.12.

In our literature, we studied three kinds of RO membranes: a cellulose acetate asymmetric membrane (SC-3000), a cross-linked N-substituted polyamide composite membrane (UTC-60), and a cross-linked aromatic polyamide composite membrane (UTC-70) using immersion and operating test conditions (Uemura and Kurihara, 2003). The degradation was observed as the increase in both solute and water permeation coefficients, which can be expressed as functions of a quantity of chlorine concentration to the Xth power times the chlorine exposure time. It was found that the values of the X are in the range of 0.5–0.7 for cross-linked aromatic composite, 1.2 for cross-linked N-substituted composite, and 1.5 for cellulose acetate asymmetric. The value of X seems closely related to the degradation mechanism. In the case of rapid degradation, the degradation might be mainly caused by chlorination reaction, and the value of X should be close to 0.5. On the other hand, in the case of slow degradation, the degradation might be caused by oxidation reaction, and value of X should be close to 2. The morphological and structural changes due to the chlorination degradation were observed using electron microscopy and electron spectroscopy for chemical analysis (ESCA). It was clarified that, as the degradation reaction progresses, the membrane thickness is reduced and its looseness and fixed charge density are increased. The results are summarized in Table 1.13.

TABLE 1.12 Studies on Chlorine Tolerance of RO Membrane

TABLE 1.13 Chlorine Degradation Rate Obtained from Each Membranea

aC = chlorine conc. (ppm), t = chlorine exposure time (h).

Using the equations in Table 1.13, the membrane performance of both solute permeation coefficient (DAM /kδ) and water permeate coefficient (A) after certain chlorine concentration (C) and exposure time (t) can be roughly predicted for each membrane. But some catalytic acceleration reactions, which may be caused by the iron ion and other heavy-metal ions in water, must be taken into account in actual cases.

REFERENCES

Cadotte, J. E. (1985). Evolution of composite reverse osmosis membrane. In Materials Science of Synthetic Membranes. ACS Symposium Series 269. American Chemical Society, Washington, DC, p. 273.

Fukunaga, K., Matsukata, M., Ueyama, K., and Kimura, S. (1997). Reduction of boron concentration in water produced by a reverse osmosis sea water desalination unit. Membrane 22(4), 211.

Fusaoka, Y. (1999). Super ultra low pressure composite reverse osmosis membrane elements for brackish water desalination and ultrapure water production. Membrane 24(6), 319–323.

Hiro, A., and Hirose, M. (2000). Development of the high boron removal reverse osmosis membrane element for seawater desalination. Nitto Giho 40, 36.

Ikeda, T., Fusaoka, Y., Uemura, T., Tonouchi, T., and Fujino, H. (1996). Advanced ultra low pressure reverse osmosis membrane elements having a high water flux and a high solute rejection. In Preprints of International Congress on Membranes and Membrane Processes, Yokohama, Japan, Aug. 18–23, p. 182.

Kallenberg, K., Jose, P. V., Yamamura, H., and Kurihara, M. (1999). Brine conversion (BCS) enhances SWRO desalination case histories, operating data, novel design features. In Preprints of IDA World Congress, San Diego, USA, Vol. II, Aug. 29–Sept. 3, pp. 101–107.

Kurihara, M., Fusaoka, Y., Sasaki, T., Bairinji, R., and Uemura, T. (1994a). Development of cross-linked fully aromatic polyamide ultra-thin composite membranes for seawater desalination. Desalination 96, 133.

Kurihara, M., Himeshima, Y., and Uemura, T. (1987). In Preprints of ICOM, p. 428.

Kurihara, M., Matsuka, N., Fusaoka, Y., and Henmi, M. (2003). Newly developed wastewater treatment systems using separation membranes. In Proceedings Water Reuse & Desalination Conference, Suntec Singapore, Singapore, Feb. 25–27.

Kurihara, M., Uemura, T., Himeshima, Y., Ueno, K., and Bairinji, Y. (1994b). Development of crosslinked aromatic polyamide composite reverse osmosis membrane. Nippon Kagaku Kaishi 1994(2), 97–107.

Moch, I. (2000). The case for and feasibility of very high recovery sea water reverse osmosis plants. In Preprints of ADA Conference, Lake Tahoe, USA.

Morgan, P. W. (1965). Condensation polymers: By interfacial and solution methods. In Polymer Reviews, Vol. 10. Wiley, New York.

Nakao, S. (1996). Sea water desalination process for high recovery of fresh water by reverse osmosis. Bull. Soc. Sea Water Sci. Japn. 50(6), 406–412.

Ohya, H., Suzuki, T., and Nakao, S. (1996). Proposal and technological breakthrough of an integrated system for the complete usage of sea water. Bull. Soc. Sea Water Sci. Japn. 50(6), 389–395.

Redondo, J., Busch, M., and Witte, J. D. (2003). Boron removal from seawater using FILMTEC™ high rejection SWRO membranes. Desalination 156, 229.

Rodriguez, M., Ruiz, A. F., Chilon, M. F., and Rico, D. P. (2001). Influence of pH in the elimination of boron by means of reverse osmosis. Desalination 140, 145.

Scala, R. C., Ciliberti, D. F., and Berg, D. (1973). Interface condensation desalination membrane. U.S. Patent 3,744,642.

Taniguchi, M., Fusaoka, Y., Nishikawa, T., and Kurihara, M. (2004). Boron removal in RO seawater desalination. Desalination 167, 419.

Tomioka, H., Taniguchi, M., Okazaki, M., Goto, S., Uemura, T., and Kurihara, M. (2005). Milestone of high boron rejection seawater RO membrane. In Proceedings of IDA World Congress on Desalination and Water Reuse, Singapore, Sept. 11–14.

Toray (2004). Brochure of TM800. Tokyo, Japan.

Uemura, T., and Kurihara, M. (2003). Chlorine resistance of reverse osmosis membranes and changes in membrane structure and performance caused by chlorination degradation. Bull. Soc. Sea Water Sci. Jpn. 57, 498.

World Health Organization (WHO) (2004). Guidelines for Drinking Water Quality, 3rd ed. WHO, Geneva.

Van Heuven, J. W. (1976). Dynamic membrane. U.S. Patent 3,996,318.

Yamamura, H., Henmi, M., and Inoue, T. (2002). In Development innovative MF and RO membrane for wastewater treatment and reclamation. In Proceedings of the 2nd International Conference on Application of Membrane Technology, Beijing, China, Sept. 27–29.

Yamamura, H., Kurihara, M., and Maeda, K. (1994, 1996). Japanese Patent Applications H06-245184 and H08-108048.

CHAPTER 2

Cellulose Triacetate Membranes for Reverse Osmosis

A. KUMANO and N. FUJIWARA

Desalination Membrane Operating Department, Toyobo Co., Ltd., Osaka, Japan

2.1 INTRODUCTION

The reverse osmosis (RO) seawater desalination process has many advantages in the areas of saving energy, lower capital cost, short startup and shutdown time, short construction period, less installation space, and less total water cost. RO technology is becoming the key technology for obtaining freshwater from the sea, especially in the Middle East. Membrane manufacturers are working to develop membranes offering higher product water recovery, lower energy consumption, and lower installation costs in order to enable the RO process to be adopted as the most popular method for supplying freshwater around the world.

The commercialized RO modules consist of cellulose triacetate hollow-fiber type and polyamide spiral-wound type. The cellulose triacetate hollow-fiber RO modules are used around the world for seawater desalination mainly because of excellent features such as a chlorine tolerance and fouling resistance.

This summary describes the history of cellulose triacetate RO membrane, description of cellulose triacetate hollow-fiber RO modules of Toyobo for seawater desalination, actual operation results, and recent RO modules of cellulose triacetate including most recently developed advanced modules.

2.2 HISTORY OF CELLULOSE ACETATE MEMBRANE

2.2.1 Development of Loeb-Sourirajan Membrane

The Saline Water Act was enacted in the United States on July 3, 1952. Then, the Office of Saline Water (OSW) was installed in the Department of the Interior in order to study the method of obtaining freshwater from seawater and brackish water economically. This effort accelerated development of RO membranes. The RO process was proposed as one of the methods of desalination by Reid of Florida University in the beginning of 1953. Workers at Florida University studied various commercial polymer films in 1957 and announced a cellulose diacetate film as the outstanding semipermeable membrane in which salt rejection was 96% or more. However, the permeability of water was very small. After that, Loeb and Sourirajan of UCLA succeeded in developing a method of manufacturing a new asymmetric membrane in 1960. The obtained membrane had high permeability and consisted of a heat-treated cellulose diacetate asymmetric membrane. These improvements in advanced performance of a membrane led to a practical application of the RO membrane module that was promoted (Breton, 1957; Reid and Breton, 1959; Loeb and Sourirajan, 1964).

2.2.2 Development of Commercial Cellulose Acetate Membrane Modules

Gulf General Atomic was funded by OSW and developed a spiral–wound-type module. Moreover, the company applied for the patent of the basic structure in 1964. In the patent, the example that used the cellulose diacetate membrane of Loeb and Sourirajan was indicated (Westmoreland, 1968). The improvement was performed after that in each company, and the spiral-wound-type module using a cellulose acetate membrane was put on the market by many companies, such as UOP, Hydranautics, Envirogenics, Toray Industries, and Daicel.

As for tubular-type modules, from the 1960s, various models of tubular-type modules were developed and put on the market by many companies. The RO plant using the module that Loeb and others developed worked in 1965 (McCutchan and Johnson, 1970).

A hollow-fiber-type module was developed and a fundamental patent was filed by Dow Chemical in 1960. An RO module using the cellulose triacetate hollow fiber is indicated there (Mahon, 1966). A significant portion of research and development at Dow Chemical was carried out based on its research contract with OSW, and the development results of the RO module for brackish water was published in 1970 and for seawater in 1974. The RO module using cellulose triacetate hollow fiber for brackish water was marketed in 1974 (Dance et al., 1971; Ammons and Mahon, 1974).

Research and development of the hollow-fiber-type module using a cellulose acetate membrane was conducted by Monsanto, Toyobo, and others, in addition to Dow Chemical. Toyobo announced an RO module for one-pass desalination of seawater that used the cellulose triacetate hollow-fiber membrane module in 1979 (Orofino, 1970; Ukai et al., 1980).

2.3 TOYOBO RO MODULE FOR SEAWATER DESALINATION

2.3.1 Hollow-Fiber RO Membrane for Seawater Desalination

Seawater desalination by reverse osmosis is the most effective method for the production of freshwater among various desalination technologies. Hollow-fiber RO membranes and flat-sheet membranes have been developed for brackish water and seawater desalination by a two-pass process since 1976 (Ohya, 1976). In spite of a satisfactory result of two-pass seawater desalination processes, the one-pass process has the advantages of simple and compact plant, simple operation, easy maintenance, and the lowest energy consumption. Although several one-pass seawater desalination systems by reverse osmosis have been developed so far, the membrane performance, especially salt rejections, have not been satisfactory and not stable in long-term operation (Macgowan et al., 1976; Ohya, 1978). In 1979, Toyobo succeeded in developing the high-performance, especially high-salt rejection, hollow-fiber RO membrane of cellulose triacetate for seawater desalination by one-pass process for the first time at a practical level (Ukai et al., 1980). This hollow-fiber membrane had high-pressure resistance in addition to high-salt rejection. In general, the salt rejection of membranes varies inversely as permeate water flux. In flat-sheet membrane modules, such as spiral-wound modules, high membrane permeate water flux is required to obtain practical product water flux due to narrow membrane area in the modules. The spiral-wound module configuration may also result in insufficient salt rejection. To solve this conflict, thin-film composite membranes were developed. In contrast, the hollow-fiber membrane module usually has a large membrane surface area, which makes it possible to design the module performance without any special method such as thin-film composite membranes.

The cellulose triacetate (CTA) hollow-fiber RO membranes are prepared by spinning a doped solution of CTA polymer followed by soaking and annealing. An outline of hollow-fiber preparation is shown in Figure 2.1 (Aptel and Buckley, 1996).

Optimization of preparation conditions such as hollow-fiber spinning technology using high concentrations of polymers, micropore control technology in manufacture, and post-treatment by high-temperature treatment made it possible to increase the permeate water flux of the hollow-fiber membrane without additional steps. CTA–hollow-fiber RO membrane itself has outstanding pressure resistance retention, and the hollow-fiber design and selection of suitable dimensions provide high-pressure resistance. Outer diameter is about 165 μm and inner diameter is about 70 μm. Resistance of this hollow-fiber membrane against high pressure is a critical characteristic to achieve a practical performance level. A microscopic view of hollow-fiber membrane for seawater desalination is shown in Figure 2.2.

The membrane made from CTA, which give an improved membrane performance, continue to be used widely today because of high performance and long-term reliability. CTA–hollow-fiber features superior chlorine tolerance compared with polyamide membrane as shown in Figure 2.3.

In biologically active seawaters sterilization by chlorine is considered a very effective solution to prevent biological fouling from occurring in the RO process. Material properties of CTA chemistry make it possible that the CTA–hollow-fiber membrane RO process can be simply controlled by only chlorine injection, which prevents biological fouling. The sterilization operation mode can be optimized by continuous or intermittent chlorine injection.

Figure 2.1 Outline of hollow-fiber preparation.

Figure 2.2 Microscopic view of hollow-fiber membrane.

Figure 2.3 Comparison of chlorine resistance.

2.3.2 Seawater Desalination RO Module

The bundle of several tens of thousands of hollow fibers is fabricated into an element and installed in a module. Hollow fibers in an element are arranged in a mutually crossed configuration without any kinds of supporting materials between hollow-fiber layers.

The hollow-fiber membrane is wound into a bundle with a layered, cross arrangement, as shown in Figure 2.4, and moderately sized. Regular intervals are left between the hollow fibers to minimize pressure differentials and allow uniform flow and achieve small pressure drop.

These features minimize concentration polarization and extend the allowance of fouling index of feed water up to SDI = 4. This structure is also less prone to being clogged by fouling matter. From the microscopic photograph shown in Figure 2.5, the space of Toyobo’s hollow fiber is remarkably large, and an accordingly rapid increase of differential pressure is not observed in the RO process. This characteristic allows an RO plant to be operated with easy maintenance.

Figure 2.4 Cross arrangement of hollow fiber.

Figure 2.5 Cross section of hollow fibers.

Conventional hollow-fiber RO module consisted of a single element inserted in one pressure vessel. However, economical and technical considerations demanded modules with a number of elements inserted in one pressure vessel. Toyobo successfully produced double-element modules on a commercial basis for the first time in 1979 and was followed by various types of double-element modules based on their flow path design.

Toyobo double-element modules, with many advantages, were developed in 1981. This module structure and flow pattern is shown in Figure 2.6.

Feed water enters the feed center pipe of the feed-side element, feed water flows radially from the center pipe, past the hollow fibers in the bundle, and away from the element as concentrated brine of the feed-side element. Feed flow is uniformly distributed in the RO module throughout the cross-wound hollow-fiber bundle. The brine water of feed-side element flows on the periphery of the element, passes through the narrow space between the pressure vessel and element, and flows to the periphery of the brine-side element and enters and passes hollow fibers in the bundle radially from outside to inside. Brine water flows to and passes through the center tube of the brine-side element to brine port and away from the module. Permeate water is collected with supporting plates and passes through the permeate pipe of each side element.

Figure 2.6 Structure of the hollow-fiber RO module (HM type).

2.3.3 Features of Toyobo RO Module for Seawater Desalination

2.3.3.1 Hollow-Fiber Configuration

Hollow-fiber modules offer greater surface membrane area in the module than spiral-wound modules. This allows a high water production of the module and a smaller footprint of module banks in a desalination plant owing to high compactness. Figure 2.7 illustrates Toyobo’s hollow-fiber array fixed in epoxy resin at both ends.

Figure 2.7 Hollow fibers with epoxy resin.

This provides mechanical stability to the fiber array in the RO element. The fibers at one end of the element are precisely cut so that product water can be discharged from the bore of the fibers. Toyobo RO modules for seawater desalination have configurations in the doubleelement type as shown above. Because the permeate water of each element in a pressure vessel can be obtained directly, the quality of each permeate water can also be measured directly. This feature allows easier maintenance of an RO plant (Kannari, 1995).

2.3.3.2 Stable RO Performance

Membrane Physical Structure The physical structure of the hollow-fiber membrane in Toyobo RO modules is optimized to provide mechanical strength resistance to collapse in a hydrostatic pressure environment. This is achieved by spinning the hollow-fiber geometry at dimensions consistent with the mechanics of a tiny hollow tube or pipe and the material properties of a strong CTA polymer.

2.3.3.3 Superior Fouling Resistance

Chlorine Resistance Chlorine resistance is significantly better with the CTA membrane chemistry than with membrane chemistry based on polyamide. This allows the biological control and stable operation of the RO module by the intermittent chlorine injection (ICI) disinfectant method.

Optimum Permeability Continuous flux (permeability or flow per membrane surface area) across membranes can cause performance difficulties in a desalination process due to deposition of fouling materials on the RO membrane surface. Generally, the higher the flux is, the sooner the membrane surface will become coated with fouling material. Spiral-wound membrane elements offer relatively low membrane surface area, and hence high flow must be more restricted to prevent fouling. Hollow-fiber membrane elements offer about 10 times greater membrane surface area than spiral-wound membrane elements. Figure 2.8 shows a comparison of fouling tendency.

Figure 2.8 Comparison of fouling tendency in membrane type.

The larger membrane surface area advantage of the hollow-fiber element allows the same quantity of permeate water to be produced at a permeability about one-tenth that for spiral-wound elements. This results in much fewer restrictions in operation and less frequent cleaning.

Cross-Winding Style A hollow-fiber membrane manufactured via a multifilament spinning process allows element construction with parallel fiber arrays or filament winding, that is, cross-layered arrays. This unique fabrication allows regular open spacing between the hollow fibers, which minimizes element pressure drop and assures uniform flow throughout the element.

2.4 ACTUAL PERFORMANCE OF TOYOBO RO MODULE FOR SEAWATER DESALINATION

2.4.1 Actual Performance Around the World

The main features of a CTA membrane are summarized in following two points:

1. Chlorine is an effective disinfectant of CTA RO module and chlorine allows the RO membrane module to be directly sterilized.

2. A CTA membrane has low adhesion of the fouling matters due to its hydrophilic property to an organic matter.

These characteristics are most important for a seawater desalination plant because seawater has high potential for microbial growth in the module since there are many sources of nutrients. If the seawater RO module cannot be sterilized by chlorine, high microbe multiplication occurs and propagates on the membrane surface. Microbe propagation within an RO module will cause serious decrease of product water quantity and deterioration of water quality. Moreover, frequent membrane cleaning is required, plant downtime increases, and the amount of chemicals increases due to cleaning. However, such performance deterioration and trouble does not occur in the case of an RO module made from a CTA membrane, which can be sterilize directly by the chlorine as a disinfectant. Therefore, many RO modules made from a CTA membrane are adopted in the Arabian Gulf countries where the possibility of microbe multiplication is high due to surface intake unlike the beach-well intake and high temperature. The main desalination plants that have adopted RO modules made from CTA of Toyobo are listed in Table 2.1. Table 2.1 shows large-scale seawater desalination RO plants using Toyobo RO membrane modules.

2.4.2 Jeddah 1 RO Plant in Saudi Arabia

The Jeddah 1 RO seawater plant is introduced as an example of a Red Sea coast plant on the Arabian Peninsula. The Jeddah 1 RO Phase I plant, which has a capacity of 15 MGD (56,800 m³/day), went into operation in 1989. The same sized Phase II plant came on stream in 1994, giving the plant a total capacity of 30 MGD (113,600 m³/day) (Al-Badawi et al., 1995). The specifications of the Jeddah Phase I and Phase II RO plants are shown in Table 2.2.

The two plants are of almost the same construction, but the membrane at the Phase II plant was guaranteed for 5 years. Product water is blended with that from the MSF plant and distributed to Jeddah. The tight water supply in Jeddah just meets demand, and so product water from the plant is of crucial importance. The maintenance period must be kept to a minimum and the plant continuously operated at full capacity.

TABLE 2.1 Main Supply Record of Toyobo Hollow-Fiber RO Module

TABLE 2.2 Plant Specifications of Jeddah 1 RO Plants

Toyobo double-element-type hollow-fiber RO modules (HM10255) are used for both Jeddah Phase I and Phase II plants. Figure 2.9 illustrates the schematic flow of the Jeddah 1 Phase II plant.

Raw seawater is taken from the Red Sea and then disinfected by sodium hypochlorite produced by a chlorine generation plant using filtered seawater as feed. Then ferric chloride as coagulant is added to the seawater feed ahead of dual-media filter (DMF) to help in reducing SDI values to suit the recommended SDI required by the membrane manufacturer. The filtered feed water is then collected in a clear well, and particles greater than 10 μm will not pass a cartridge filter in order to avoid membrane plugging. Sulfuric acid is added to the feed water to adjust pH to about 6.5. Sodium bisulfite (SBS) is injected for 7 h every shift (8 h) ahead of the high-pressure pump to nullify residual chlorine and protect the membrane from oxidation by residual chlorine in the presence of heavy metals. To avoid biological fouling, 0.2 mg/L residual chlorine is allowed to pass through the membrane for 1 h every shift (8 h) intermittently by the ICI method instead of a conventional continuous chlorine injection (CCI) method.

Figure 2.9 Schematic flow of Jeddah Phase II plant.

Figure 2.10 Disinfecting power of chlorine.

The disinfecting power of chlorine was tested in Japan by cultivating bacteria taken from the Jeddah seawater. The results were as shown in Figure 2.10 (Fujiwara, 1999a).

It was confirmed that even when the bacteria count was high, disinfection with a chlorine concentration of 0.2 mg/L reduced it almost to zero in 30 min.

To determine whether the disinfection was sufficient, the concept of lead time was considered (Nada et al., 1994). Lead time is the time from when injection of SBS is halted and chlorine is fed into the module to when a certain chlorine concentration is detected in filtered water in the module. Except certain times when the SDI of feed seawater exceeded 4.5 and seawater was polluted, lead time remained stable at 10 min. This signifies that a 10-min injection of chlorine at a concentration of 0.2 mg/L was sufficient to disinfect the module. In fact, it was confirmed that differential pressure in the module was stable and no biofouling occurred during the test period.

2.4.2.1 Operation Results of Phase II Plant

Due to the successful site testing by the ICI method at Jeddah, the operating conditions were kept the same for the Phase II plant as shown in Table 2.3.

This plant is of crucial importance to Jeddah, which is prone to water shortages, and barring times when the feed seawater is polluted and the SDI exceeds 4.5, it operates continuously at 99% capacity. It is important to note that the membrane has not been replaced at all in the 5 years since the plant went into operation.

Product Flow Rate and Quality Changes in the permeate flow rate are shown in Figure 2.11. The permeate flow rate remained stable at the rated 56,800 m³/day throughout the entire period. As is evident from the diagram, all 10 systems at the plant were in virtually continuous operation. Permeate total dissolved solids (TDS) are shown in Figure 2.11. Permeate TDS stayed below 500 mg/L without the membrane being replaced at all in 5 years, wh ich is highly satisfactory (Fujiwara et al., 1999a).

TABLE 2.3 Plant Specification of Jeddah 1 RO Plants

Differential Pressure One advantage of ICI is the importance of preventing biofouling. While biofouling should not occur when the module is disinfected by CCI, intermittent disinfecting of the module necessitates caution. Disinfection tests were therefore conducted by cultivating bacteria taken from Jeddah seawater as described before, and extended testing was carried out using site test units to confirm that biofouling did not occur. As a result, differential pressure was stable at the Phase II plant as shown in Figure 2.12, and it was found that periodic cleaning with citric acid once a year reduced differential pressure to almost what it was at the start (Fujiwara et al., 1999a).

At the Jeddah Phase II plant, as a result of ICI adoption, product flow and product quality were excellent even with fact that the membrane was not replaced in 5 years.

Next we discuss the significant improvement of the performance of an existing plant by exchanging the RO membrane of the plant for a CTA membrane. Here, the operation test results at the AdDur plant in Bahrain is introduced as an example on the Arabian Gulf coast.

Figure 2.11 Permeate flow rate and permeate TDS of Jeddah Phase II plant.

Figure 2.12 Differential pressure of Jeddah Phase II plant.

2.4.3 The AdDur Plant in Bahrain

The AdDur RO desalination plant was designed to produce 10 MGD (45,500 m³/day) with polyamide RO hollow-fiber membrane modules from seawater containing up to 45,000 mg/L as TDS. This plant was commissioned in 1990. The AdDur plant process consisted of the dual-media filter (DMF) with coagulants, dechlorination by sodium bisulfate, high-pressure pumps, RO modules, and posttreatment by lime injection. The specifications of the AdDur plant are shown in Table 2.4 (Al-Badawi et al., 1995).

In 2000, rehabilitation work was conducted to obtain improved performance of the pretreatment by means of ultrafiltration (UF) membranes that proved most effective among the tests conducted. The work included modification of existing dual-media filter to singlemedia filter and additional installation of spiral-wound-type UF having a capacity of 130,000 m³/day after the media filter in order to reduce SDI values. Chemical dose was stopped together with modification of media filter. The filtered seawater from the UF process was fed to the existing first pass 8 RO trains of the polyamide membrane module. However, even after installation of the UF system, the RO membranes required frequent cleaning. It was concluded that additional action was required to achieve more stable operation. It was suspected that the UF membrane had poor removal of natural organic materials in the raw seawater, so that such organic matter would act as a nutrient for biological activities in the RO membranes.

2.4.3.1 Schematic Flow Diagram of CTA Membrane Test Unit

Reverse osmosis feed water to the test plant is taken from acidified UF filtrate and is exactly the same as the actual RO feed water to the plant. SBS is dosed into the feed water to remove residual chlorine contained in pretreated water. The SBS dose is stopped three times a day for 1 h to introduce residual chlorine into the RO membrane for sterilization purposes. Three Toyobo RO modules, HB9155 Model, were used in this test. A schematic flow diagram of the test unit is shown in Figure 2.13.

TABLE 2.4 Plant Specification of AdDur RO Plant

Figure 2.13 Schematic flow diagram of test unit.

2.4.3.2 Test Conditions

Test conditions and test requirements are shown in Table 2.5. The test was aimed at satisfying the following plant specification. For prevention of biological fouling, it was shown that the ICI method provided effective chlorine injection.

2.4.3.3 RO Performance at Site Test

Permeate Flow Rate and Permeate Quality Actual permeate flow rate is shown in Figure 2.14. The permeate flow rate was gradually increased to about 7.8L/min per module due to adjustment of operation conditions during about 10 days from startup. Then the permeate flow was set at about 7.5L/min per module (10.77 m³/day per module at design conditions). During the winter season the permeate flow rate was maintained by increasing feed pressure. The flow rate was stable and satisfied the plant requirements for 12 months. Actual permeate quality is shown in Figure 2.15.

TABLE 2.5 Test Conditions and Test Requirements

Figure 2.14 Permeate flow rate of test unit in AdDur plant.

The permeate quality was 764 μS/cm at the startup of operation and then gradually decreased to about 200 μS/cm due to natural membrane compaction. The permeate quality was stable at a level of 200 μS/cm (95 mg/L as TDS at design conditions) through the test period. This performance was much better than the plant requirement of 500 mg/L as TDS. RO performance (permeate flow rate and permeate quality) was very stable and satisfactory during the entire test period (Alawadhi et al., 2005).

Differential Pressure A differential pressure can be used as an indication of biological fouling growth in the RO module as shown in Figure 2.16. The differential pressure was stable at a low level of 20 kPa without the chemical cleaning. Therefore, it is considered that biological fouling had not occurred in RO modules for the 12-month test period. Also, it was concluded that the ICI method worked effectively to prevent biological fouling.

Figure 2.15 Permeate quality of test unit in AdDur plant.

Figure 2.16 Differential pressure of test unit in AdDur plant.

For a 12-month test period, RO performance of HB9155 based on CTA material was very stable and satisfied the test requirements. Therefore, it was demonstrated that RO plant performance could be recovered by replacement of existing polyamide membranes with CTA membranes. Differential pressure was stable at a low level of about 20 kPa without the need for chemical cleaning during the test period. Therefore, it is concluded that the ICI method worked effectively to prevent biological fouling.

In general, the Arabian Gulf region is a difficult area for an RO plant operation due to its biological activity. Control of biological activity in the RO membrane is key to a stable performance in the region. The result obtained through this test is helpful to the RO plant operation in this region.

2.5 MOST RECENT RO MODULE OF CELLULOSE TRIACETATE

2.5.1 Development of RO Module for Higher Recovery

The reverse osmosis seawater desalination process has many advantages from the viewpoints of saving energy, lower capital cost, short startup and shutdown time, short construction period, less installation space, and less total water cost. In a seawater RO process, the seawater is subjected to disinfection, coagulation-filtration, and acidification processes in the pretreatment section and forwarded as feed water to RO section. When the recovery fraction (the ratio of product flow rate to feed flow rate) is high, the amount of feed water required for a desired production is lower and hence, the pretreatment system, chemical cost, equipment sizing, and energy costs are significantly reduced (Ohya et al., 1996; Nakao, 1996).

Membrane manufacturers are working to develop membranes offering higher water recovery, lower energy, and lower installation cost in order to enable the RO process to be recognized as the most popular method for supplying freshwater around the world.

In areas such as the Middle East where seawater has high salinity, commercial seawater RO desalination plants were normally designed to operate at approximately 35% recovery. This relatively low recovery was due to the very high osmotic pressure of the seawater, and most commercially available RO membranes did not allow operating pressures above 7.0 MPa. Osmotic pressure of seawater from the Red Sea is approximately 3.2 MPa. In the case of operation with recovery at 35%, the osmotic pressure of brine is about 4.8 MPa in the RO module, as shown in Figure 2.17.

Figure 2.17 Osmotic pressure of brine water.

If the recovery is increase to about 50%, the osmotic pressure of the brine increases to about 6.1 MPa. Therefore, a higher feed pressure (i.e., driving force) is required for a high-recovery operation in comparison with a conventional low-recovery operation. In order to enable the high-recovery operation, the RO module must be designed and manufactured to withstand the higher pressure.

Toyobo recently developed a new type of RO module to achieve higher product water recovery in order to further reduce the cost of RO desalination. Toyobo’s hollow-fiber RO modules are widely used around the world in RO desalination plants. Based on the long operating experience and recent research efforts, Toyobo developed the high-pressure high-flux HB series modules. The HB series is an improved version of the conventional HM series type of module using the same materials (Ohnishi et al., 1997; Sekino, 1998; Fujiwara et al., 1999b). The hollow-fiber membrane in the HB series module is wound in a cross arrangement, designed to minimize pressure loss and allow uniform water flow in the module. The hollow fiber incorporated in the HB series has higher pressure resistance based on a change of the hollow-fiber outer diameter/inner diameter dimensions and optimization of manufacturing conditions. The specification of HB series modules are shown in Table 2.6.

The product flow rate of the new type improved by about 1.4 times compared with the conventional type. A high-pressure, single-pass desalination process of new HB series modules with high recovery was successfully conducted for the first time at an RO test plant on the Red Sea at the conditions of more than 52% recovery (Kumano et al., 2003).

2.5.2 Development of Both Open-Ended RO Module

Toyobo’s newest innovative RO module builds on the proven reliability of the HM-type module. The new technology is based on both open-ended (BOE) hollow-fiber membrane structures versus