Search

Reverse osmosis is becoming popular for the purification of water resources e.g. sea and brackish water. However, one of the major limitations in efficient operation are the fouling of membranes and relevant technique used to inhibit scaling due to the presence of sparingly soluble salts. The occurrence of this type of fouling results in a significant increase in both operation and maintenance. This review begins by providing a general overview of membrane fouling by minerals usually found in sea and inland water resources. The paper then discusses some important aspects of the phenomena itself: different mechanisms; concentration polarization; the major types of mineral scales including their prevalence and characteristics; and consequences of fouling on membrane performance. This is followed by a discussion of the major strategies currently considered for the control and prevention of scale formation. The penultimate section discusses the drawbacks and deficiencies of current systems in operation and highlights potential future directions that need to be pursued in order to achieve overall process improvements in terms of sustainability. In particular, the need for evaluation under more realistic conditions, studying the interactions between membrane and spacer surface and foulants, and the development of environmentally benign antiscalants are stressed.

The surface of commercial reverse osmosis (RO) membranes was modified by the deposition of 2-hydroxyethyl methacrylate-co-perfluorodecyl acrylate (HEMA-co-PFDA) copolymer films using an initiated chemical vapor deposition ( i CVD) technique. Antifouling characteristics of the modified and unmodified membranes were evaluated under cross-flow permeation conditions using sodium alginate as a model organic foulant. The permeate flux decline was lower in the surface-modified membranes as compared to the virgin ones while salt rejection remained almost unaltered in both virgin and coated membranes. The presence of the HEMA-co-PFDA copolymer on the fouled membrane surface is confirmed by simultaneous analyses with Attenuated Total Reflectance–Fourier Transform Infra-Red (ATR-FTIR) and X-ray Photoelectron Spectroscopy. Field Emission Scanning Electron Microscopy (FESEM) images showed a dense and continuous foulant layer on the virgin samples as opposed to a porous and discontinuous one for the coated membranes. The water contact angle values for the coated membranes registered a decrease indicating the more hydrophilic nature of the adsorbed alginate. To conclude, although fouling appears inevitable, it is considerably slowed down by this surface modification strategy.

Polysuccinimide (PSI) of various molecular weights (MW) was synthesized from aspartic acid. The PSIs were hydrolyzed with NaOH to obtain sodium polyaspartates. Partial hydrolysis of a PSI gave the copolymer poly(SI- co -sodium aspartate). PSIs were treated with one equivalent of cyclic amines: pyrrolidine, piperidine, and azepane to afford the corresponding polyaspartamides in excellent yields. Treatment of PSI with 0.5 equivalents of the cyclic amines led to the formation of 1:1 copolymer poly(SI- co -aspartamide) which on basic hydrolysis gave the corresponding 1:1 copolymer poly(sodium aspartate- co -aspartamide). The antiscalant behavior of the newly synthesized compounds was evaluated at various concentrations (2.5 ppm–10.0 ppm) for the inhibition of gypsum scale formation. The antiscalant evaluation experiments were performed by measuring conductivity and using concentrated brines. The study revealed the performance of low MW antiscalants to be superior to their high MW counterparts. Thus, a polyaspartate is found to be a better antiscalant than its corresponding copolymer of SI/aspartate, i.e. a partially hydrolyzed PSI. Derivatives of 7-membered cyclic amine azepane performed better than their 5- and 6-membered counterparts. At the same degree of polymerization, performance with decreasing order of effectiveness was found to be: polyaspartate > poly(aspartamide- co -aspartate) > (SI- co -aspartate) > poly(SI- co -aspartamide) > polyaspartamide.

This study is focused on precipitation and inhibition studies of calcium sulfate (gypsum) scale in reverse osmosis (RO) desalination plants. The temperature and concentration range encountered in brackish water RO desalination process has been considered in conducting precipitation and inhibition experiments. Initial screening of a number of additives has been carried out. Formation of gypsum scale has been verified by XRD and SEM analysis. Successful additives were evaluated by measuring induction times using conductivity of solutions and their performance was compared with that of blank supersaturated calcium sulfate solution. The induction time of STD100, SHMP, STD101 was found to be 70, 690 and 1200 min respectively at dose level of 10 ppm, whereas a blended formulation FM101 (10 ppm) did not show any precipitation till 3000 min. The SEM micrographs of precipitate collected when FM101 was added as a scale inhibitor with that of a blank solution and other inhibitors indicate that, the formulation FM101 has significantly modified the crystal morphology and is probably responsible for increased induction time.

Thin film composite membranes were fabricated via spin assisted layer by layer (SA-LbL) assembly by depositing alternate layers of poly(allyl amine hydrochloride) (PAH) and poly(acrylic acid) (PAA) on a polysulfone (PSF) ultrafiltration membrane as support. The suitability of these membranes for potential water purification applications was explored by testing the stability of the deposited thin films and their permeation characteristic using cross-flow permeation cell. Permeation test conducted at a pressure of 40 bar, temperature of 25 °C, pH of 6 and feed water concentration of 2000 ppm NaCl demonstrated that the PAH/PAA multilayer film deposited on polysulfone support remained stable and intact under long-term test conditions. The 120 bilayers of PAH/PAA membrane tested at the above condition showed flux of 15 L/m2.h and salt rejection of 65%. The membrane performance evaluation also revealed that SA-LbL PAH/PAA membrane follows the characteristics of the solution diffusion membrane.

A study was undertaken to identify various types of scales that were responsible for shortening the useful life span of the membrane permeators in a commercial reverse osmosis (RO) desalination plant. Compositions of the raw and treated feed water and of the reject brine were determined using the inductively coupled plasma (ICP) spectrometry and ion chromatography (IC). Various scaling index calculations showed that the feed and brine were non-scale forming with respect to CaCO 3 (calcite), SrSO 4 , CaSO 4 .2H 2 O (gypsum), and silica (SiO 2 ). Two completely fouled membrane permeators, retired from stage 1 and stage 2 of a commercial plant, were subjected to membrane autopsy using scanning electron microscopy (SEM), X-ray diffractometry (XRD), optical microscopy (OM), and energy dispersive x-ray florescence (XRF). The deposits were predominantly amorphous in nature. The membrane autopsy showed that CaCO 3 , SrSO 4 , and CaSO 4 .2H 2 O (gypsum) scales did not constitute a serious problem in the plant. The advanced phosphonate+polyacrylate based scale inhibitor had itself formed Ca phosphonate sludge, but the amount was quite small. Though below saturation, silica is believed to have been precipitated due to the catalyzing effect of trivalent Al 3+ and Fe 3+ ions. Iron fouling was the major cause of reduced life span of the membranes and, to a lesser extent, calcium-alumino-silicates.

Sodium hexa-meta-phosphate (SHMP), inhibitor-A (acrylic polymer), and inhibitor-B (polyacrylate+phosphonate), each at a dose level of 1 ppm (on active basis) were evaluated at 25°C for their efficacy against CaSO4, SrSO4, and CaCO3 scales in a synthesized brine which corresponded to 80% recovery from the locally available aquifer brackish water. At periodic time intervals brine pH was measured and samples were withdrawn and analyzed for Na+, Ca2+, Sr2+, SO42−, and total alkalinity (TA). The total time of equilibration was 168 h (7 days). The criterion adopted in comparing efficacy of the inhibitors was that if an inhibitor could achieve an induction period of at least 15 min, it would be considered an effective inhibitor at that dose level against that particular scalant. In the control experiment (no inhibitor added), there was an immediate drop in Ca2+, Sr2+, and TA levels and the solution became turbid much earlier than 15 min, suggesting that the brine was highly supersaturated and scale-forming with respect to CaCO3 and SrCO3 scales and would require injection of an inhibitor. In 168 h, there was no drop in SO42− concentration suggesting that the brine possessed no potential to form CaSO2.2H2O and SrSO4 scales. SHMP was able to keep the scale-forming constituents (Ca2+, Sr2+, and TA) in solution for 91 h, inhibitor-A for 50 h, and inhibitor-B for all of 168 h, suggesting that the inhibitors at 1 ppm gave completely scale-free operation up to 80% product recovery and that performance of inhibitor-B was the best.

A comparative evaluation of a holloe fine fiber (HFF) and a spiral-wound (SW) membrane was carried out by operating them in parallel units of an RO pilot plant. The two units received the same brackish feed water, identical scale control treatment (H2O+SHMP) and operated at the same recovery (70%). Unit #1, using the HFF membrane, was operated at a feed pressure of 27.6 bar (400 psig); unit #2, using the SW membrane, was operated at 15.2 bar (220 psig). The trial lasted 7500 h. The SW membrane produced water of excellent quality (86 ppm) vs. 470 ppm by the HFF membrane, and it consumed only half as much pumping energy, thereby outperforming the HFF membrane. Following the 7500 h trial, an autopsy of the two membranes was carried out to identify the scale deposits. Visual observation, scanning electron microscopy (SEM), X-ray diffractomery (XRD), inductively coupled plasma (ICP) spectroscopy, coulometry, photometric and gravimetric techniques were used. Both membranes were found in fouled condition. However, the HFF membrane was much more fouled. SEM and XRD analyses showed that 90–100% of the deposits were amorphous in nature. Results of the coulometric and photometric methods suggested that sodium hexa-meta-phosphate (SHMP) was effective in controlling the CaCO3 scale but was itself responsible for the fouling of both membranes by reversion to orthophosphate (PO43−). The ICP results showed that iron contributed to the fouling. Among the inorganic scales, calcium alumino silicate clay was the major scale; but the great bulk of the deposit on both membranes was organic matter or biomass.

The HFF and SW membrane permeators were installed in parallel units of a reverse osmosis (RO) pilot plant. Brackish feed water to the two units was treated with identical scale control treatment consisting of 6 ppm of sodium-hexa-meta phosphate (SHMP) and about 46 ppm (98% basis) of sulfuric acid. The pH of the acidified feed was about 6.4. Unit #1, employing the HFF membrane, was operated at feed pressure of 27.6 bar (400 psig) and unit #2, employing the SW membrane, was operated at 15.2 bar (220 psig). The water recovery in both units was 70%. The pilot plant trial lasted for 7,500 h. The techno-economic evaluation showed that performance of the SW membrane was far superior; its product TDS in 7500 h was 86–98 ppm whereas that of the HFF membrane increased from 80 ppm to 468 ppm. Water output of both membranes declined appreciably, but output of the SW membrane declined 12% less. The average electrical energy consumed by high pressure pump of the SW membrane was nearly half the energy consumed by the HFF membrane (0.70 kWh/m3 product vs. 1.51 for HFF). Variations in the pressure drops across the HFF and SW permeators were modest (5–10 psi) due to amorphous nature of the scale deposits found on fibers of both membranes and due to significant decrease in feed flows experienced by the permeators. The pretreatment, the scale control treatment cost and water recovery being the same for the two units, the SW membrane produced excellent quality water (86 ppm) at half the electrical energy cost compared with the HFF membrane which produced 468 ppm water. The SW membrane out-performed the HFF membrane.

An advanced anti-scalant, consisting of a polyacrylate and hydroxyethylidene diphosphonate (HEDP), was tested against the conventional H2SO4 and sodium hexa-meta-phosphate (SHMP) inhibitors in a pilot plant which had parallel RO units. The two units were fitted with identical brackish hollow fine fiber (HFF) permeators arranged in a two-stage mode which received the same brackish feed water, and were operated at the same feed pressure (27.6 bar/400 psig) and product water recovery (70%). Unit #1 was treated 9 ppm of as-received advanced anti-scalant, and unit #2 was treated with 6 ppm of SHMP and about 130 ppm of 98% H2SO4. The pilot plant trial lasted for 3,000 h. The techno-economic evaluation of the two scale control treatments was carried out in terms of pumping energy, the anti-scalants consumed, and the quality and total output of the product water produced in 3,000 h. The pumping energy and scale control treatment costs (as $/m3 of water produced) were same for the two treatments; but in terms of salt rejection, product TDS and output, performance of the advanced anti-scalant was inferior; after 3,000 h the salt rejection of unit #1 (which operated on the advanced anti-scalant) was about 81% vs. 94% for unit #2 (H2SO4+SHMP). Similarly, in 3,000 h unit #1 produced 20% less water of unacceptable quality (620 ppm) whereas unit #2 produced more water of excellent quality (180 ppm), providing a clear testimony to the superior performance of the conventional H2SO4+SHMP treatment over the advanced anti-scalant.

A polyacrylate and hydroxyethylidene diphosphonate (HEDP) based advanced anti-scalant was tested against the conventional H2SO4 and sodium hexa-meta-phosphate (SHMP) inhibitors in a pilot plant which had parallel RO units, each fitted with identical hollow fine fiber (HFF) permeators which received same brackish feed water. At the end of the 3000 h trial, product TDS of unit #2 operating on the conventional treatment was only 180 ppm vs 620 ppm for unit #1 operating on the advanced anti-scalant, and it produced an average of 20% more product water. To identify the scale deposits, autopsy of the first and second stage permeators was carried out using visual observation, scanning electron microscopy (SEM), x-ray diffractometry (XRD), x-ray fluorescence (XRF), inductively coupled plasma (ICP) spectroscopy, ion chromatography (IC), and coulometry techniques. Whereas thick deposits of soft and non-adherent types of scales were found on the permeator fibers of unit #1, no scale was evident on unit #2 permeator fibers and no samples could be collected from them. Analyses of the solids removed from the unit #1 fibers revealed that bulk of the deposited scale was amorphous in nature which the SEM and XRD techniques were unable to charaterize. The XRF, ICP, and IC results suggested that the amorphous sludge was predominantly alumino-silicate type. In addition, membrane fibers of stage 1 had calcium/magnesium phosphonate, formed by the advanced anti-scalant itself by reacting with Ca2+ and Mg2+ of the brine. No CaCO3, SrCO3, CaSO4, SrSO4, or iron-based scales were detected, indicating that the advanced anti-scalant was effective against these scales.

Field trials of the additive-only (Albrivap-G) treatment and the hybrid (acid+Albrivap DSB) treatment were carried out on an MSF unit at 112.8°C top brine temperature. The main objectives were to reduce the abnormally high frequency of acid cleaning with the existing additive treatment, and to test the techno-economic viability of the hybrid treatment. The economics and the corrosion and thermal performance of the hybrid treatment were outstanding: the period of operation between consecutive acid cleanings was increased from 19 days to at least 164 days (5 1 2 months) and the frequency of acid cleaning was thereby reduce from seven to at most one per year of 5 1 2 operating months. The cost of scale control treatment was reduced by 50% while energy cost of the extraction steam was reduced by 11% resulting in considerable savings in the operating cost of the 3-unit MSF plant.