MBRAOP Desalination Corrosion

Understanding and selecting an appropriate desalting process train for a specific application is complicated with an assortment of desalination membrane technologies offered in today's market. Designing a reliable desalting system requires a technical firm that applies fundamental principles of water quality and chemistry to properly engineer a cost effective design.

Desalination membranes have become an important process in water purification and water reuse; transforming previously unavailable waters into a valuable resource. In water treatment today, desalination membranes are being widely applied for treating brackish surface waters, groundwaters, reclaimed wastewater and seawater to create new water sources, or restore degenerated water sources. Today, there are three types of membrane processes used to remove dissolved solids from water: reverse osmosis (RO), nanofiltration (NF), and electrodialysis (ED).


Desalting membranes remove 70 to 99% of salts from saline water to produce a stream with fresh water, while a second stream known as reject or brine stream carries the removed salts to waste.

Two fundamental driving forces are used in these desalination membrane systems: pressure and electric potential. In pressure driven
systems (RO & NF), the feed water is pressurized to exceed the osmotic pressure and water passes through the semi-permeable
membrane; while dissolved solids are rejected and remain on the feedwater side, eventually concentrating into a brine solution. The
systems driven by electric potential (ED) apply voltage to opposite ends of a membrane stack; one positive terminal and one negative. The charged ions are attracted to the terminal with the opposite charge (i.e. Na+ migrates toward the negative terminal and Cl- migrates toward the positive terminal). So, while the pressure driven system selectively passes water and retains dissolved salts, the electrically driven system extracts the dissolved salts and retains the water. Either way, the water and salts are separated producing low salinity water.

Classes of pressure driven membranes
Pressure driven membranes have been classified into four categories based on the membrane rejection properties as follows:

  • Microfiltration (MF) membranes - have the largest pore size (0.1 - 3 micron), require low transmembrane pressure (1- 30 psi), and are used for turbidity reduction, removal of suspended solids, parasites (e.g. giardia and cryptosporidium), bacteria and some viruses
  • Ultrafiltration (UF) membranes - have a smaller range of pore sizes than MF membranes (0.01-0.1 micron) require low transmembrane pressure (1- 30 psi), and are capable of removing viruses as well as some color, odor, and organics removal, along with everything that the MF process can remove
  • Nanofiltration (NF) membranes - are relatively new porous membranes that have a pore size < 0.002 micron require moderate transmembrane pressure (75- 150 psi), and are primarily used for natural organic matter (NOM) removal for controlling disinfection byproduct precursor, water softening and sulfate removal
  • Reverse osmosis (RO) membranes - are effectively non-porous membranes that require high transmembrane pressure (150-500 psi) and are used for monovalent salt removal

MF/UF membranes are being increasingly employed in the desalination process to shield RO/NF/ED membranes from suspended solids and larger colloidal material that are detrimental to their performance. In other words, neither MF nor UF membranes are capable of salt rejection and, as a result, both are only relevant to desalination membrane systems as a pretreatment process. NF membranes can be visualized as tight UF membranes that not only reject materials (i.e. suspended solids, colloidal material, and bacteria) based upon a size exclusion, but also remove hardness (e.g. multivalent ions) based upon a charge repulsion mechanism. However, NF membranes poorly reject monovalent salts. Therefore, a majority of desalination is performed by non-porous RO membranes that provide physical barrier to a wide range of contaminants, including monovalents.

The EPA has designated RO as a best available technology (BAT) for removal of numerous inorganic contaminants, including antimony, arsenic, barium, fluoride, nitrate, nitrite, boron, selenium, radionuclides, and emerging contaminants, including endocrine disrupting compounds (synthetic and natural hormones), and several pharmaceutical compounds. Figure 1 summarizes various contaminants that would nominally be rejected by the various types of membranes.

 

Figure 1. Membrane Types.
Contaminants nominally removed by the four most commonly mentioned pressure-driven membrane processes.

The most common configuration of RO/NF element is the spiral-wound element, wherein a large number of flat sheet membranes are wrapped around a perforated PVC pipe (Figures 2a and b). Because this configuration results in a densely packed module, a significantly fewer number of membrane elements are required, which reduces the overall footprint of the desalination facility. Generally, four to eight elements are arranged in series in a pressure vessel (Figure 2c). During operation, when pressurized feed water enters the first element, a portion passes through the membrane material is collected as product water while the fraction remaining on feed side, now concentrated, becomes feed for the next element. Thus, the salt concentration on the feedwater side increases in each succeeding elements, with the last element receiving the most concentrated feed solution. The elevated feedwater concentration can cause increase in (1) concentration gradient across membrane that reduces the product water quality (e.g. higher TDS); (2) osmotic pressure that in turn reduces pure water flux or requires additional pressure to maintain this flux; and (3) inorganic scaling (precipitation) on the membrane surface.

Figure 2. The Spiral Wound Configuration.
Illustration of the geometry and function of a spiral wound RO module (a) & (b) and the installation of modules into a typical RO pressure vessel (c)

Concentration of reject streams leaving each element is directly related to the system recovery, i.e. fraction of feed water, which is recovered as product water. High recovery implies that a small amount of concentrated waste will be generated (which has economic and other advantages), but results in poor water quality and flux decline. On the other hand, low recovery operation translates into better membrane performance, but results in a large waste stream that is less economic and difficult to dispose of. Clearly, there is a trade-off between system recovery, overall membrane performance and project costs.

Electrodialysis
In contrast to the membrane types shown in Figures 1 and 2, the ED process applies an electric potential that moves dissolved salt ions (e.g. (+) ions are attracted to the (-) electrode) through an electrodialysis stack consisting of alternating layers of cationic and anionic ion exchange flat sheet membranes, creating alternate channels of desalted product water and concentrated reject water. A modification of the ED process; termed as EDR (electrodialysis reversal), reverses the polarity of the applied electrical potential on the stack, periodically, to minimize the effects of inorganic scaling and fouling by converting product channels into concentrate channels. EDR is most attractive for the desalination of brackish water and in cases where TOC removal and microbial control are not important. Below, a working diagram of ED process (Figure 3), and a photo of a full- scale EDR operating on tertiary effluent are shown (Figure 4).

Figure 3. Illustration of ED.
All cations attempt to migrate to the cathode and all anions attempt to migrate to the anode. Cation-exchange membranes allow only cations to pass and anion-exchange membranes allow only anions to pass. The effect is to remove the salt from every other cell. In EDR, the polarities are regularly reversed, dislodging deposits on the membrane surface.
Figure 4. EDR in Operation.
Photo of EDR plant with ion-exchange membranes in stack configuration.

Desalination Process Description
Membranes are typically applied to desalinate waters from three different water resources: seawater, brackish waters (groundwater and surface water), and tertiary treated wastewater. A schematic of a typical desalination facility is shown in Figure 5. Regardless of the membrane type or application, all desalination plants essentially consist of three major steps:

• Pretreatment
• Desalination Membrane Process
• Post-treatment
Figure 5. Illustration of desalination process.
Salt is removed from the pretreated seawater using desalination membranes.

Pretreatment is the cornerstone of any desalination process. Selecting appropriate pretreatment is imperative to the success of any desalination project. In the absence of adequate pretreatment, the membrane surface becomes rapidly covered by the rejected material present in the feed water and this results in membrane fouling or a drop in membrane productivity with time, increasing operational costs.

Fouling of desalination membranes can be broken into four distinct categories: fouling by particulate matter, organic fouling, biological fouling, and inorganic scaling.

Fouling with Particulate Matter and Suspended Solids. If not properly dealt with, particulate matter and fine suspended solids present in the feedwater are problematic and will reduce the water throughput of the desalination membranes with time. Although, ED membranes are thought to be traditionally more robust and can handle some influent particulate matter (i.e <2 NTU), excessive particulate matter and influent suspended solids will accumulate in the spacers for the NF/RO/ED membranes and increase the head loss through the system (no effect on transmembrane pressure). Depending on the feedwater quality, a filtration system needs to be designed to reduce the influent suspended solids and particulate matter before feeding the water to the desalting membranes.

Fouling with Organic Materials. Desalination membranes are susceptible to organic fouling depending on the source water quality. ED anion-exchange membranes are particularly susceptible to organic fouling due to the negative charge associated with natural organic matter. This can lead to process failures. Large organic anions cannot penetrate the anion-exchange membrane and will accumulate and adsorb to the membrane surface, increasing the stack resistance. Small organic molecules can also be problematic because they penetrate the membrane, but their electromobility is low and they remain inside the membrane. Fouling of this kind can make it quite difficult to clean and restore these membranes to their original electrical resistance. NF and RO membranes are fouled by organic adsorption as the membrane rejects these materials and membrane permeate is produced. It is difficult to determine the organic fouling potentials using aggregate organic measurements (i.e. TOC) in the feedwater. Some source waters with relatively low TOC (< 1 ppm) result in dramatic organic fouling while waters with high TOC (>8 mg/L) do not.

Biological Fouling. Biofilm control is important in virtually every unit process, which sets out to accomplish mass transfer in an aqueous system. Membrane manufacturers have come a long way in reducing the biodegradability of the membranes themselves, but most modern RO and NF membranes are sensitive to oxidants. Without the use of oxidizing disinfectants, it is unlikely that biofilm control will ever be adequately achieved.
Although turbulent cross-flow is maintained in all desalination membrane systems, bacteria are still capable of adhering to the membrane surface and excreting extracellular polymeric substances (EPS) to create a strong bond to the membrane surface. Once attached to the membrane, a complex community of microorganisms is created that is held together and fixed to the membrane with EPS. These consortia are generally referred to as a biofilm (Figure 6). Biofilms result in decreased membrane permeability in pressure driven applications and increased electrical resistance in potential driven processes, increasing the operational and maintenance costs of desalination membrane processes. The rejection of targeted contaminants can also be adversely affected.

Figure 6. Biological fouling of RO membranes.
A = scanning electron microscope (SEM) image of bacterial microcolonies formed on the surface of a cellulose acetate RO membrane after approximately 3 days of operation on pretreated municipal wastewater; B = a biologically fouled spiral-wound RO membrane element; C = transmission electron micrograph (TEM) of rod-shaped bacteria attached to an RO membrane surface; D = TEM of a mature membrane biofilm; E = nascent biofilm on permeate surface of a CA membrane. Courtesy of H. Ridgeway
Fouling From the Formation of Inorganic Scale. Rejection of dissolved solids with desalination membranes segregates salts into a waste stream commonly referred to as concentrate or brine. If the concentrations of these salts exceed their solution saturation, precipitates will form a scale of inorganic salts (e.g. Fe2O3, CaCO3, CaSO4, SiO2, CaF2, BaSO4, etc.) on the membrane surface (Figure 7). Scaling usually develops in the final stage of the RO or EDR process at the membrane surface because this is the active point of ion separation where concentrations are highest. By adjusting the feedwater recovery, the design engineer can estimate the concentrations of the dissolved solids and specify a system that does not suffer from inorganic scaling. Inorganic scaling, like other forms of fouling, will increase the operational costs and require operator attention to clean and restore the membrane system.
Figure 7. Inorganics fouling of RO membrane. Iron fouled feed side of a RO element.

Pretreatment for RO and EDR. The choice of a particular pretreatment process is based on a number of factors such as feed water quality characteristics, space availability, and RO/NF/ED membrane requirements. Typically, pH adjustments (between 5 and 6) and/or antiscalant additions are made to prevent scale formations on the membrane surface. As mentioned before, feedwater recovery is another key parameter that can be optimized to limit the scale formation. To control and minimize colloidal, organic, and biofouling either conventional pretreatment (e.g. coagulation/ flocculation/ sedimentation/ filtration), or modern pretreatment methods (membrane filtration using MF/UF) are employed prior to the desalination unit. Addition of mild oxidants is key to controlling biofouling of the membranes and generally, chloramines are typically used for this purpose.

Desalination membrane process

High Pressure Membrane System. A typical RO facility with cylindrical pressure vessels (each housing several membrane elements) is shown in Figure 8. A group of cylindrical pressure vessels are operated in parallel is called a stage. To increase system recovery, the concentrate from one stage can be fed to a subsequent stage (sometimes called as brine-staged or multistaged system); and to improve solute removal, permeate from one pass can be fed to a second pass (a two-pass system or permeate staged system). In multistage systems, the feedwater flow rate reduces in each succeeding stages due to permeate extraction from the feed water. To maintain sufficient crossflow velocity and to avoid solute accumulation on the membrane surface, multistage systems are designed with decreasing number of pressure vessels in each succeeding stages. Various configurations for multistage and two-pass systems exist and the specific design depends upon the feedwater and product water quality and treatment goals (see Figures 9a-c).

Figure 8. RO Skid.
Photo of RO pilot plant operating with 4” spiral wound membrane modules. The configuration of this skid is indicative of the configuration of full-scale systems as well.
Figure 9a. Array Configurations of RO facilities .
4-2-1 concentrate-staged array
Figure 9b. Array Configurations of RO facilities.
3-2-1 concentrate-staged array
Figure 9c. Array Configurations of RO facilities.
Two-pass system.

ED/EDR: The basic ED/EDR unit consists of several hundred cell-pairs bound together with electrodes on the outside and is referred to as a membrane stack. These stacks are then placed in series and a three-stage EDR process is shown in Figure 10. Unlike RO systems, concentrate and permeate both are fed to the subsequent stages and permeate water is taken only from the final stage (upon obtaining desired TDS concentration). Importantly, RO/NF energy cost is based on the volume of water treated; whereas for ED/EDR processes, it is proportional to the salts removed. Therefore, these processes are usually only suitable for brackish feedwaters with a salinity of up to 12,000 mg/L TDS. With higher salinities the ED/EDR process becomes more costly than other desalination processes. As a rule of thumb, approximately 1 kWh is required to extract 1kg additional salt using ED/EDR. Bacteria, non-ionic substances and residual turbidity are not affected by this process and can therefore remain in the product water and require further treatment before certain water quality standards are met.

Figure 10. EDR configuration.
Schematic of four parallel three-stage EDR process with 90% overall recovery and 87.5% solute removal.

Post treatment. Permeate water is typically low in pH, hardness and
alkalinity, requiring post treatment. Post treatment can be accomplished by a variety of treatment processes. Addition of disinfectant in product water is a must to ensure microbial safety in the distribution system. If the desalinated water is being combined with other sources of water supply, it is very important to ensure similar water quality characteristics for both water sources. If the desalinated water will serve as the sole water source for the distribution system, it is important that additional impacts on horticulture and taste and odor concerns are adequately addressed.

Links to key desalination membrane manufacturers in the USA:
GE OSMONICS
FILMTEC
TORAY MEMBRANE
HYDRAUNAUTICS
KOCH MEMBRANE SYSTEMS