By Dr. Mo Mukiibi
Reverse Osmosis (RO) was developed in the 1950s as a technique, initially governmentally funded, for desalting seawater. Over the years, RO has evolved in terms of scale of versatility, applications, reduced cost and increased efficiency.
RO has become effective in treating brackish, surface and ground water. Besides treating large water volumes, RO has been modified to treat small water flows. Commercially available RO units are readily available to treat two to 100 gpd of household drinking water. On a larger industrial scale, an RO plant can treat up to 20 million gallons per day (mgd) of ground water. In addition, such plants are often used to process secondary effluent and treat highly saline water for reuse.
Other industrial applications where RO is commonly used include:
- Pharmaceutical: Reverse osmosis is an approved technology for the production of United States Pharmacopeia (USP) grade water meant for pharmaceutical applications.
- Boiler feed water treatment: RO produces high-purity water, which when injected into a gas turbine, can improve efficiency and increase energy output by over 10 percent.
- Food and beverage: RO is often used to treat water used for food and beverage processing.
- Semiconductor: Companies like Intel use reverse osmosis to process and produce ultrapure water (UPW) required for semiconductor processing.
- Metal finishing: Several types of metal processing, finishing and plating industries use RO systems in their operations.
RO systems are typically designed to recover as much water as possible, reduce the levels of TDS and provide public health benefits. These benefits include removing micro organisms such as Cryptosporidium and Giardia, which may cause adverse health effects.
RO is also very effective in reducing total organic carbon (TOC), precursors to disinfection byproducts (DPB). Other economic benefits of implementing RO include provision of low-salinity product water with low corrosion rates toward home appliances such as water heaters, swamp coolers and plumbing fixtures.
How RO works
Pretreatment: This involves manipulation of the feed water chemistry (using chemicals/coagulants) and/or using pre-filtration systems, aimed at controlling scaling and fouling as well as enhancing the performance of the RO treatment systems. Prefilters are often used to remove large particles that may plug the RO membranes. The various types of prefilters that are available include rotating disc (drum) or bag filters (for small systems).
Filtration: RO membranes are semi-permeable. They are designed to separate water from a solution of dissolved solids. RO membranes have a pore size of about 0.0001 µm, which is about 500, 000 smaller than a human hair (60-75 µm). One theory is that when pressure is applied, it forces the lighter molecules (e.g. ,water) to filter through but blocks passage of larger compounds (dissolved salts).
RO recovery efficiencies (product water flow divided by feed water flow) are based on design factors such as the type of contaminant, its initial concentration levels, the type of membrane selected and the water pressure. Generally, a large brackish water RO treatment system for treating secondary treated wastewater is expected to operate at 70 to 80 percent recovery.
Recoveries of about 90 percent can be achieved by returning a portion of the reject water back into the modules for additional treatment. Unfortunately, not all the water is recovered by RO membranes. The remainder of the initial feed volume contains large amounts of TDS and is generally wasted (disposed of).
Post-treatment stabilization: Depending on feed water chemistry, more often than not, RO treatment processes require manipulation of feed water chemistry, which in turn affects the treated water quality. In the event that chemical manipulation is required, product water (RO permeate) must be stabilized to meet drinking water or reuse standards. The stabilization process may involve pH adjustment (to ~pH 7) and/or degasification before water is supplied for potable use. In most cases, the degree of stabilization depends on intended use as well the product water characteristics.
Types of RO membranes
Recent advances in technology have resulted in development and production of different configurations of RO membranes. The most popular and widely used configuration for RO is spiral wound. Most RO membranes are either cellulose acetate or aromatic polyamides.
For water treatment applications, cross linked polyamide (PA) type of membranes are preferred to cellulose acetate (CA) membranes, because of their increased chemical and physical stability, high water recoveries and high resistance to bacterial degradation. The upper portion of Figure 1 shows a synthetic pathway used to prepare PA thin film membranes with good desalination properties.1
Unfortunately, widely desired PA membranes are susceptible to oxidation and are often impacted by the side chain reaction between the free chlorine (a disinfectant) and PA amide groups. This reaction disrupts their stable linkages and conformational structure. Consequently, this renders them ineffective for their intended filtration functions.
There have been several attempts to create chlorine-resistant membranes for more than twenty-five years, but without success.2
In general, RO systems require little maintenance and are easy to use. They are highly automated, requiring minimal operator attention. The longevity of RO elements, however, is strongly influenced by membrane fouling and scaling, both of which cause permeate flux decline (reduced efficiency).
Scaling: This is a main cause of permeate flux decline. Scaling is normally caused by precipitation of super-saturated salts on membrane surfaces. The problematic salts responsible for scaling include calcium carbonate (CaCO3), barium sulfate (BaSO4), silica (SiO2), calcium sulfate (CaSO4), calcium fluoride (CaF2), strontium sulfate, (SrSO4) and magnesium hydroxide (Mg (OH)2).
For most of the problematic salts mentioned above, surface precipitation is a gradual process, which mechanistically occurs in three steps.
The first step is concentration polarization (resulting in flux reduction). The second is nucleation (resulting in crystal formation) and the third is cake formation (resulting in scaling). It is therefore not surprising that cake formation (the last step in forming precipitates), normally occurs in the back-end stages of an RO treatment train where reject water is highly concentrated.
Cake formation (precipitation) of any of the above salts on the membrane surface occurs when their corresponding solubility product is exceeded. Table 1 lists the solubility constant, known as Ksp, for common scaling compounds.
In some RO operations, forecasting of the scaling potential on membrane surfaces is often achieved by computing the saturation index (SI), which is a ratio between ion products of the respective salt (Ip) to their corresponding solubility product (Ksp) under the same conditions. A saturation index greater than 1.0 means that there is a high probability of the occurrence of precipitation and scaling.
The Langelier saturation index (LSI) is often used to determine specifically if the main cause of scale formation is calcium carbonate (CaCO3). This evaluation is normally used if source water TDS levels are less than 10,000 mg/L.
A positive LSI is indicative of scale formation potential. The advantage of using LSI is that other than calcium concentration, it takes other factors like alkalinity concentration, pH and temperature into consideration.
Chemical fouling: Fouling is a phenomenon caused by accumulation of particles (organic and/or inorganic), that are present in the feed water and are later deposited on the membrane surface. At the membrane-liquid interface, several parameters can cause particles to accumulate on polymer membranes, the principle of which is electrostatic interactions.
Influent water usually contains particulates that could be greater than 0.45µm in diameter. It is not surprising that fouling is usually most pronounced on membranes that are located in the front-end stages of an RO treatment system.
Biofouling: Biological fouling is normally caused by colonization and attachment of bacterial communities on the hydrophobic membrane surface. Evidence of RO fouling micro organisms exhibiting very different affinities for surfaces of different membranes has been investigated.3 Researchers concluded that membrane surface chemistry and structure play an important role in the bacterial adhesion process and that PA membranes exhibit greater biofouling than other types of RO membranes.
Other than membrane structure, additional factors determine the attachment of bacteria to a membrane surface. These include: the hydrophilic/hydrophobic nature of the membrane surface; the ionic and nutrient composition of the feed water, and the nature of the bacterial cell surface, membrane charge, surface roughness and available surface area.
Scaling, chemical fouling and biofouling of RO membranes result in a significant decline in performance and cost-effectiveness. Most industrial systems utilize pretreatment processes consisting of a series of physicochemical application processes designed to control membrane fouling and scaling.
These processes are geared towards enhancing the performance of the RO treatment system. Even with pretreatment in place, RO units require both regular monitoring and maintenance to retain effectiveness over an extended period of time. Most commercial or municipal RO systems employ regular clean-in-place (CIP) applications to extend membrane performance.
Pretreatment of RO feed
Pretreatment involves processes that are designed to reduce membrane fouling and scaling while increasing the recovery of water. These processes can be classified into two major categories: those technologies that can remove and limit the availability and solubility of scalants and/or foulants present in the RO feed water and those processes which increase solubility beyond solubility limits (super saturation).
The former is normally accomplished by using precipitative softening, ultrafiltration (UF) or microfiltration (MF). Super saturation is normally accomplished by adding antiscalants/ inhibitors. The choice of the best method for RO applications depends on source water composition and chemistry to determine the best option.
Microfiltration versus ultrafiltration
For pretreatment, some RO systems rely on either UF (membranes with a pore size of 0.001-0.1 µm) or MF (pore size range of 0.1-10 µm). This is because ultrafiltration is effective in removing viruses, colloids and high-molecular weight organic compounds that may be present in the source water whereas microfiltration is effective in removing bacteria and colloidal material.
Many water sources have macromolecules such as proteins and polysaccharides that typically have molecular weights that range from 104 to 106 Daltons (Da) and molecular diameters that range from 0.002 to 0.01 μm. Specifically, enzyme (protein) molecular weights normally range from 104 to 105 Da and have molecular diameters that range from 0.002 to 0.005 μm.
If such macromolecules are present in the source water, they should be able to pass through MF and may not be completely captured by UF. A greater fraction of the dissolved (less than 0.45 μm) organic or inorganic matter cannot be removed by MF or UF membranes without advanced pretreatment. This demonstrates that the effectiveness of either UF or MF in reducing the fouling rate of RO membranes requires a full understanding of the relative fouling potential of large versus small-size organic matter.
Precipitative softening (PS): This technology involves removing problematic salts or their precursors prior to membrane application. Process control normally involves the addition of caustic (NaOH), lime (CaO) or sodium bicarbonate (Na2CO3) to the feed water.
However, PS/RO systems have a large footprint and require additional chemical and dewatering services. Environmental impacts of PS/RO systems include high chemical usage and disposal of large quantities of sludge generated during the process.
Attempts are currently underway to improve the efficiency of precipitative processes by using polymers (to increase floc settling rates and sludge dewatering), specific ion additions (to target early precipitation of certain cations) and designer particle additions (to provide preferential nucleation sites and increase settling rates).4
Acidification: This process involves increasing the solubility of sparingly soluble salts by adding a strong acid (such as sulfuric or hydrochloric acids) to the feed water. In most natural water sources, calcium and carbonate are in equilibrium, as shown in the following equation: Ca2+ + HCO3– ↔ H+ + CaCO3 (s)
Acidification of the feed water also has direct effects on RO-treated water quality. To illustrate this impact, consider a case of most natural waters, where carbon chemistry and speciation is controlled by pH, as shown in the following equation: CO2 + H2O ↔ H2CO3 ↔ H+ + HCO3–
If the pH of feed water is reduced to below 8.2 by adding acid, equilibrium starts to shift to the left and as the pH is lowered more bicarbonate (HCO3–) is converted to carbonic acid (H2CO3), which in turn dissociates into carbon dioxide (CO2). The dissolved carbon dioxide then permeates through RO membranes and adversely impacts treated water quality.
This effect is worsened if the treated water lacks sufficient alkalinity to mitigate immediate fluctuations in pH. This series of events may trigger the need for post-treatment stabilization process control involving degasification of the treated water.
Antiscalants/scale inhibitors: Another way of increasing solubility of the problematic compound is by adding antiscalant/inhibitors to feed water. Inhibitors/antiscalants are categorized according to their mode of operation, but they all have a common feature in that they work by interfering with crystal growth patterns and reduce scale formation.
Threshold inhibitors such as polyphosphates and organo-phosphorous compounds function by adsorbing onto the surface of the scale, hence causing a delay in precipitation. Dispersant inhibitors such as polyelectrolytes function by forming an electrostatic charge on the growing crystal, causing a mutual repulsion. Crystal distortion inhibitors work by tempering and modifying the normal crystal growth patterns and this forces the scale to form an irregular crystal structure with poor scaling abilities.
There are several types of antiscalants commercially available. The choice of additives for RO application depends on feed water/brine stream chemistry, composition, recovery condition and target scaling compound.
As mentioned earlier, RO systems generate two products (permeate and concentrate). The big question that still remains is what to do with a concentrate stream that contains a TDS concentration greater than 50,000 mg/L, possibly laden with huge quantities of RO cleaning chemicals?
To illustrate the challenges involving disposal of RO brines, consider the Central Arizona Project (CAP) source water, which draws from the Colorado River. This major source of potable and irrigation water brings over 1,000,000 tons of salt each year into central and southern Arizona.4
The TDS level in this source water is 700 mg/L, which exceeds the US EPA Secondary Standard of 500 mg/L. To remove this salt from CAP water, a 107.6 mgd RO plant at 85 percent water recovery will produce 91.5 mgd of product water (average TDS~56mg/L) and a brine flow of 16.1 mgd with a concentrate TDS of 4,400 mg/L. This translates to a production of 108,600 tons/year of salts, which need to be disposed of in an environmentally safe manner.
If not properly disposed of, brine streams with such high TDS levels pose a major environmental concern because high salinity causes pipe corrosion and harms salt intolerant plants and animals. In addition, it can also affect the ability to reuse the water for agricultural, industrial and ground water recharge.
Therefore, proper management and disposal of highly saline waters is inevitable. Some RO applications have taken additional process steps to completely solidify the brine product for final disposal (zero liquid discharge [ZLD]).
Even with additional processing in place, the final product must be disposed of in an environmentally safe manner. Options for final disposal include deep well injection, landfills, down stream discharge to sewers, ocean discharge and evaporation ponds. Each of these options has limitations.
The selection of the best option for managing a given RO concentrate waste stream depends on a number of factors, including cost. More often, the costs associated with a given option are much higher than standard treated water costs.
As approximately 48 percent of RO facilities in the US are located inland, disposal of RO concentrate presents a serious challenge and limits the implementation of RO systems. Concentrate management and disposal options will be discussed in details in a follow-up article.
- “Polyamide Reverse Osmosis Membrane Fouling and Its Prevention: Oxidation-Resistant Membrane Development, Membrane Surface Smoothing and Enhanced Membrane Hydrophilicity.” Bureau of Reclamation final report 2000.
- Glater, J., Seung-Hong and M. Elimelech, “The Search for a Chlorine-Resistant Reverse Osmosis Membrane,” Desalination, 95, (1994) 325-345.
- Ridgway, H. F. and H. C. Flemming, “Membrane Biofouling”, pp 6.1-6.62, in: Water Treatment Membrane Processes, J. Mallevialle, P. E. Odendaal, and M. R. Weisner (Eds.), McGraw-Hill Publishers, New York (1996).
- Salinity Management and Desalination Technology for Brackish Water Resources in the Arid West. Summary report of workshop on “Improving Salinity Management and Desalination Technology for Brackish Resources in the Arid West” August 6, 2007. Wendell Ela, et al. 2008.
About the author
Dr. Mo Mukiibi is a Process and Design Engineer with CH2M Hill, Inc.’s Water Business Group, based in Phoenix, AZ. His expertise ranges from advanced water and wastewater treatment, process control and water systems design. Dr. Mukiibi is a technical expert (water and sanitation) for the World Health Organization and is skilled in desalination technologies, concentrate management and disposal. Well published and with presentation at many conferences locally and internationally, Dr. Mukiibi is a member of the WC&P Technical Review Committee and is an active member of WEF, NGWA and AWWA. He can be reached via phone (480) 966-8188 ext. 38228 or by email at firstname.lastname@example.org.