In the case of microfiltration, a more porous membrane is used than in the other membrane separation technologies, thus yielding a relatively higher flux. It is mainly useful in removing turbid causing materials and can replace conventional granular filtration processes. The most significant design parameters are:
• Transmembrane pressure
• Tangential velocity
• Size and geometry of modules
• Recirculation factor
The clean water flux across a membrane without any material being deposited follows Darcy's Law:
The net pressure differential across a membrane, taking into consideration the osmotic pressure is given by (AP - AH), and hence, the expression for the permeate flux is:
where Jw = permeate flux, m/s
AP = pressure difference, N/m2 R™ = internal membrane resistance, i/m Rc = resistance due to deposit on the surface, 1/m fi = dynamic viscosity, N-sec/m2
The resistance due to the deposit on the surface is given by the following relationship:
where e = porosity of deposit
6 = thickness of the deposit, m dp = average diameter of the particles, m
Commercial systems for wastewater treatment are designed to be submerged into built on-site rectangular concrete tanks. These are pre-engineered modular membrane systems that typically use a membrane with a 0.2-micron nominal pore size. A vacuum pump draws water through the membrane fibers of sub-modules submerged in the open top filter tanks. The fibers are the same polypropylene material as those used in the conventional filtration process. A typical system operates under vacuum, and utilizes improved filter cake characteristics at low pressures with a maximum driving pressure in the league of 85-100kPa.
There have been only a few studies have evaluated membrane microfiltration of secondary wastewater effluent. Microfiltration membranes might be used to achieve very low turbidy effluents with very little variance in treated water quality. Because bacteria and many other microorganisms are also removed, such membrane disinfection might avoid the need for chlorine and subsequent dechlorination. Metal salts of iron or aluminum may also be added to enhance membrane performance. For example, iron or aluminum coagulants may be added to precipitate otherwise soluble species such as phosphorus and arsenic as well as improving the removal of viral particles. Coagulation of colloidal materials may also increase the effective size of particles applied to membranes and increase permeate flux by 1) reducing foulant penetration into membrane pores, 2) forming a more porous cake on the membrane surface, 3) decreasing the accumulation of materials on the membrane due to particle size effects of particle transport, and 4) improving the backflushing characteristics of the membrane.
Membrane microfiltration at the pilot-scale has produced a permeate of similar or better quality than that produced by conventional filtration. Good removal of particulate contaminants, including coliform bacteria, have been observed. In this regard, the process appears to be as effective as chlorination for the removal of coliforms from secondary waste effluent. A key advantage is the ability to filter and disinfect in a single step without the need for subsequent dechlorination. Preliminary results indicate that coagulation pretreatment in conjunction with membrane microfiltration can be used to reduce phosphorus concentrations as well. There does not appear to be any advantage in running the microfiltration unit in a crossflow mode and there may even be some disadvantages. The permeate quality and evolution of pressure drop obtained from the membrane operated in the deadend mode is found similar or superior to that obtained under crossflow conditions.
Although membrane processes have been used successfully for many years in desalting brackish water and seawater, new kinds of membrane processes are now capable of treating water for a wide range of other uses. Some of these new, robust processes promise to do a better job meeting our current water treatment goals than such conventional processes as granular media filtration, carbon filtration and disinfection with chlorine. Engineers classify membranes in many different ways, including describing them by the driving forces used for separating materials (i.e., pressure, temperature, concentration and electrical potential), the mechanism of separation, the structure and chemical composition, and the construction geometry. In water treatment, the membranes most widely used are broadly described as pressure driven. Each membrane process is best suited for a particular water treatment function. For example, microfiltration (MF) and ultrafiltration (UF), which are very low pressure processes, most effectively remove particles and microorganisms. The reverse osmosis (RO) process most effectively desalts brackish water and seawater and removes natural organic matter and synthetic organic and inorganic chemicals. The nanofiltration (NF) process softens water by removing calcium and magnesium ions. These so-called nanofilters are also effective in removing the precursors to disinfection by-products that result from such oxidants as chlorine.
Nearly a decade ago, the use of low-pressure membranes such as MF and UF for disinfection and particle removal was only a concept being studied or was used only on a limited basis. The water community foresaw the possibility of providing primary disinfection without the use of chemicals. Moreover, Cryptosporidium, a waterborne enteric pathogen responsible for several disease outbreaks, was gradually showing resistance to traditional disinfectants such as chlorine. Thus, researchers believed that greater emphasis should be placed on removing organisms through physical means as opposed to chemical means.
Membrane processes also offer other advantages over conventional treatments. They reduce the number of unit processes in treatment systems for clarification and disinfection and increase the potential for process automation and plant compactness. Designers also thought membrane plants could be much smaller than conventional plants of the same capacity and, given their modular configuration, could be easily expanded. Additionally, these plants would produce less sludge than conventional plants because they wouldn't use such chemicals as coagulants or
Today many of the projected benefits of MF and UF have been realized. These technologies provide effective disinfection for potable water supplies as they reduce the levels of Giardia and Cryptosporidium, as well as a variety of bacteria, below detectable levels. MF and UF plants are now in operation throughout the world. In Europe there are several large UF plants. In the U.S., the San Jose Water Co. in Saratoga, Calif., was the first to construct a major MF plant (17,000 m3/day). A plant with 15,000 m3/day capacity followed in Rancho Cucamonga, Calif., and a 68,000 m3/day plant is under construction in Kenosha, Wisconsin. The largest plant for disinfection and particle removal, a 106,000 m3/day UF installation, is being planned in Del Rio, Tex. Today, there is more than 400,000 m3/day of MF and UF capacity in the U.S., either in operation or in the planning stage. The anticipated U.S. EPA Disinfectant/Disinfection By-Product Rule will lower the maximum contaminant level (MCL) for trihalomethanes (THM) from 100 to 80 fig/L, and set an MCL for haloacetic acids (HAA) at 60 ftg/L. Thus, NF and RO membrane processes are receiving considerably more attention since they are efficient at removing the precursors to these by-products. RO and NF membranes, in addition to desalting brackish water and seawater, are being used to remove inorganic chemicals such as nitrates, as well as synthetic organic chemicals. NF plants have been installed since the 1980s in Florida to remove color and hardness from groundwater. Today, there is 568,000 m3/day of installed NF and RO capacity
Membrane plant design begins with the selection of the membrane, which can be organic or inorganic in composition. Membrane manufacturers strive to formulate membranes that provide a desired permeate quality, are durable and resistant to fouling, and can be produced at a competitive cost. Most commercial water treatment NF and RO membranes are made up of organic polymers and are asymmetric. The active layer responsible for the separation process is typically a few micrometers thick and is supported on a highly permeable layer that adds mechanical strength to the membrane. One type of asymmetric membrane for NF and RO systems, the thin-film composite (TFC), shows great promise for potable water treatment. These membranes generally consist of an ultrathin active layer coated onto a microporous layer that, in turn, is supported on a mechanically strong base. TFC membranes typically have higher water permeability and chemical resistance than symmetric membranes. MF and UF membranes are constructed in either asymmetric or symmetric configurations. A number of hydrophilic and hydrophobic polymeric materials are used in manufacturing these membranes. These include cellulosic polymers, polypropylene, polysulfones and polyamides. The choice of material will influence contaminant rejection characteristics, durability and fouling potential.
Membrane systems consist of membrane elements or modules. For potable water treatment, NF and RO membrane modules are commonly fabricated in a spiral configuration. An important consideration of spiral elements is the design of the feed spacer, which promotes turbulence to reduce fouling. MF and UF membranes often use a hollow fiber geometry. This geometry does not require extensive pretreatment because the fibers can be periodically backwashed. Flow in these hollow fiber systems can be either from the inner lumen of the membrane fiber to the outside (inside-out flow) or from the outside to the inside of the fibers (outside-in flow). Tubular NF membranes are now just entering the marketplace. MF and UF systems can be designed to operate in various process configurations. A common configuration is one in which the feedwater is pumped with a cross-flow tangential to the membrane. The only pretreatment usually provided is a crude prescreening (usually 50 to 300 ¿im). The water that permeates the membrane is clean. The water that does not permeate is recirculated as concentrate and blended with additional feedwater just after the preliminary filter. To control the concentration of the solids in the recirculation loop, some of the concentrate is discharged at a specified rate.
MF and UF systems may also operate in a direct filtration configuration, with no cross-flow (or recirculation). This is often termed dead-end filtration. All of the prescreened feedwater passes through the membrane. Therefore, there is 100% recovery of this water, except for the small fraction of the water used to periodically backwash the system. MF and UF plants typically rely on either liquid or pneumatic backwashing systems. Most MF and UF water treatment plants use this direct flow configuration, since it saves considerably on energy by not requiring recirculation. There are also capital cost savings since there is no need to purchase recirculation pumps and associated piping.
RO and NF systems usually operate in a series of stages. In a three-stage system, the first stage consists of three pressure vessels, which usually contain four to eight membrane elements; the second stage has two pressure vessels; and the final stage has one. In full-scale plants, elements are approximately 1,000 mm long and have a diameter of 200 mm. Permeate is collected from each pressure vessel. The concentrate from the first stage serves as the feed to the second; concentrate from the second stage serves as the feed to the third. Consequently, each successive stage of the array increases the total system recovery.
For many groundwater applications, the pretreatment required for RO or NF
consists in adding acid or antiscalant and then passing the feed through a cartridge filter. However, for surface waters, more extensive pretreatment is necessary. This may involve conventional treatment, MF, UF, slow sand filtration or, in some cases, granular activated carbon adsorption.
One innovative process configuration for surface water and tertiary wastewater treatment involves the use of double-membrane systems, consisting of a low-pressure and a high-pressure membrane in series. This treatment is effective for both microbial and chemical contaminant control. The first membrane (MF or UF) is used to help prevent fouling of the second, higher-pressure membrane system (RO or NF). In the Netherlands, the Heemskirk water treatment plant, which will be completed late this year, will use a 53,000 m3/day double-membrane process (UF and TFC-RO) to treat water from the Ijssel River. San Diego, California, will also use a double-membrane system integrated into its treatment train as part of an approach for purifying tertiary wastewater for potable reuse. Another system uses membranes designed to be immersed in a process tank and suction instead of pressure to draw water through the membrane hollow fiber lumen.
Pilot testing is often a key aspect of successful membrane plant design. One of the most important reasons for conducting a pilot study is to evaluate the influence of water quality on membrane fouling. It is critical to determine if the process is feasible for a specific water source, particularly for those that exhibit significant water quality changes on a seasonal basis. Other reasons for pilot tests include demonstrating regulatory compliance, identifying the most effective and appropriate processes, evaluating new membrane products and establishing design criteria for a full-scale plant. Designers can also evaluate pretreatment options by conducting experiments using parallel treatment trains with varying pretreatment processes but identical operating conditions. This avoids the confounding factors in data interpretation, such as changing source water quality. Designers can also verify and fine-tune chemical cleaning procedures for site-specific conditions. In pilot studies, designers are able to optimize the operating conditions of each individual stage of a multistage installation.
As membranes filter out the impurities from the water, the membranes themselves become fouled (or clogged) and less effective. The fouling of membranes has been one the primary impediments to their more widespread application in water treatment. Membrane systems operate in one of two modes: constant transmembrane water flux (flow rate per unit membrane area) with variable pressure; or constant pressure with variable transmembrane water flux. The former is the more common. Membrane fouling occurs during an increase in transmembrane pressure to maintain a particular water flux or during a decrease in water flux when the system is operated at constant pressure. In general, membranes can be fouled by an accumulation of inorganic particles (for example, clays, iron, manganese and silica) and organic compounds (such as humic and fulvic acids, hydrophilic and hydrophobic materials, and proteins). Bacteria can also adhere to the membranes and create a biofilm. Accurate tests to predict fouling still do not exist. However, researchers can conduct "autopsies" on fouled membranes prior to chemical cleaning to analyze the nature and composition of the contamination. They can then adjust pretreatment and chemical cleaning procedures. Fouling can be controlled by hydrodynamic and chemical methods, periodic backwashing, and chemical cleaning. Other methods include improving pretreatment and changing operating conditions. Membrane fouling rates are functions of the operating conditions such as water flux and recovery. Typically, reducing the operating flux and recovery will reduce fouling. Research shows that, for MF and UF, increasing the frequency of backwashing also decreases the rate of fouling. Because filtered or raw water is used for backwashing, the net recovery of direct-flow MF and UF systems decreases as the frequency of backwashing increases. Improved pretreatment also reduces membrane fouling. Water treatment operators have decreased MF and UF fouling rates by using coagulation/flocculation/sedimentation and dissolved air flotation pretreatment.
Chemical scaling is another form of fouling that occurs in NF and RO plants. The thermodynamic solubility of salts such as calcium carbonate and calcium and barium sulfate imposes an upper boundary on the system recovery. Thus, it is essential to operate systems at recoveries lower than this critical value to avoid chemical scaling, unless the water chemistry is adjusted to prevent precipitation. It is possible to increase system recovery by either adjusting the pH or adding an antiscalant, or both.
Was this article helpful?