Innovative strategies have been suggested to address the issue of brine disposal with inland desalination. One inland plant constructed a 30 km pipeline to send concentrate to the sea  while another proposed project will send, by gravity, concentrate from Red Sea desalination plants to the Dead Sea, producing energy, disposing of the brine, and returning much-needed volume of liquid to the Dead Sea . Such innovative solutions are more a rarity than the norm, however. As such, considerable effort must be spent toward engineering a cost-effective solution with the minimum environmental impact possible and potential sites must evaluate other options such as ZLD and/or beneficial uses and applications of the brine byproducts.
Many ZLD applications in operation today treat power plant cooling water with thermal desalination and/or evaporation ponds to gain maximum energy use of the heated water. Thermal desalination (distillation) has been practiced for over 30 years, particularly in the Middle East, and it is a mature technology unlikely to result in any major technological improvements. Although there have been a few design innovations over the years to optimize energy efficiency, thermal desalination remains an expensive, energy-intensive process. Therefore, membrane-based ZLD operations are gaining favor as the technology expands for improved membranes and energy recovery, though evaporative systems may be viable in regions with low humidity and large tracts of undeveloped land.
Enhanced evaporation systems use mechanical energy to increase the surface area of water in contact with air and thereby achieve evaporation rates several times that of conventional evaporation ponds. Such evaporation systems can be used alone or in combination with multistage membrane ZLD processes. Several enhanced evaporation systems have been tested, which may have the potential to significantly reduce the costs and area required for evaporation . Most rely on water lines with spray nozzles to disperse brine to the atmosphere as a mist, into a high-speed air stream, countercurrent air streams, or onto vertical surfaces to augment evaporation. A portion of the water is evaporated and residual salts fall to the ground. These systems may have the potential to achieve efficient evaporation in arid climates. A disadvantage to these systems is that windborne salt drift could potentially contaminate a sizeable surrounding area, which is especially of concern if the system is located near an urban area.
One possible membrane-based ZLD treatment process is illustrated in Fig. 1 . This process schematic shows a primary RO system for desalination of an inland source of water. Concentrate from the primary RO is treated to reduce its precipitation potential and then desalinated in a secondary RO system. The concentrate from the secondary RO, typically
Intermediate Concentrate Treatment
RO = reverse osmosis system BC = brine concentrator
Figure 1 Process schematic for desalination with zero-liquid discharge.
2-6% of the feed to the primary RO, is treated using a thermal desalination process, the brine concentrator. Brine concentrators typically recover about 95% of RO concentrate from brackish water desalination as a distillate with very low total dissolved solids (TDS). The residual concentrate, 0.1-0.3% of the feed to the primary RO, is discharged to an evaporation pond. Consequently, no liquid waste is discharged from the site.
Precipitation of solutes within membrane systems limits the amount of treated water that can be produced without some form of enhanced recovery. As recovery is increased, water on the feed side of the membrane becomes increasingly supersaturated with sparingly soluble salts such as calcium carbonate (CaCO3), calcium sulfate (CaSO4), and barium sulfate (BaSO4). Consequently, recovery in RO systems is limited by the precipitation potential of sparingly soluble salts. Crystallizers can be used to enhance recovery of water from the concentrated brine stream, which often is supersaturated .
An understanding of the factors that affect precipitation from a supersaturated solution is essential to development of a well-designed, functional ZLD process. The presence of foreign particles enhances precipitation kinetics by reducing the amount of free energy required for solids formation. Consequently, precipitation is made more energetically favorable. Induction time, the time required for precipitation to begin, decreases as the fit between the foreign particle and the crystal to be formed increases. The best fit between the two solid phases occurs when the particle is a seed crystal of the same salt (secondary nucleation). One group studied the effect of brushite (CaHPO4 • 2H2O) on gypsum (CaSO4 • 2H2O) precipitation . The researchers noted that the ability of one crystalline phase to grow on the surface of another is strongly dependent on the compatibility of their surface characteristics, and they observed a close fit in lattice structure between brushite and gypsum. Their results showed that brushite crystals served as effective nuclei for the growth of gypsum crystals. Another group studied the kinetics of seeded growth of gypsum in both the presence and the absence of antiscalant additives and concluded that seed concentration greatly influenced induction time . At the same antiscalant concentration and solution temperature, the induction time was 73 min for a seed concentration of 110 mg/mL and 31 min for a seed concentration of 193 mg/mL. Without the antiscalant, precipitation was immediate at each seed concentration. Without seed addition, the supersaturated solution was stable.
The schematic of a fluidized bed crystallization system optimized with these considerations in mind is illustrated in Fig. 2. The process is initiated,
Inland Desalination: Current Practices, EnvironmentalImplication 333 - -Effluent o °o ° °
o cfcf o
o ocp' fyp
Fluidized bed crystals: 0.2 - 2.0 mm
Periodic injection of sand grains (0.2 - 0.6 mm)
Periodic removal of crystals (1 - 2 mm)
Chemicals Influent Figure 2 Process Schematic for fluidized bed crystallization.
by introducing flow at the bottom of the reactor at a rate sufficient to fluidize the media without causing it to be washed out in the effluent. Sodium hydroxide or lime is fed at the bottom of the bed to achieve supersaturation with respect to CaCO3, and precipitation occurs as calcium and carbonate ions leave the solution and are adsorbed to the sand to form calcium carbonate pellets. Calcium removal continues with crystal growth of the pellets, and as the pellet diameter increases, the crystal surface area per unit volume of reactor decreases. The process is controlled by periodically removing larger crystals from the bottom of the column and adding new sand.
Crystallization was first applied to water softening in 1938  with the invention of the Spiractor®, a conical shaped upflow reactor. In 1971, a fluidized bed crystallizer, the Crystalactor®, was developed in the Netherlands with a cylindrical shape and water and chemical feed nozzles designed to enhance vertical plug flow and improve initial chemical mixing. This reactor was selected to provide water softening at the main water treatment plants in the Netherlands  and is currently in use at 25 treatment plants in Europe, Asia, and Australia. Fluidized bed crystallization has not been widely used for concentrate treatment, however . Several features of an optimized fluidized bed crystallization are expected to be advantageous compared to conventional softening for treatment of RO concentrate:
• Fluidized bed crystallization provides a large surface area of seed crystals for precipitation. Such precipitation occurs at lower supersaturation in the presence of crystals of the precipitate. Consequently, precipitation can be achieved with smaller chemical doses and at lower pH.
• The fluidized bed crystallizer produces near anhydrous pellets that are approximately 90% solid by weight, and these pellets drain rapidly under gravity to a solids content of 99%. Sludge produced in conventional softening has a solids content of 3-15%. Consequently, the fluidized bed crystallizer generates approximately 10% of the solids volume generated by conventional softening. Furthermore, calcium carbonate crystals have beneficial uses in agriculture and industry. All of the solids generated by fluidized bed crystallization softening in the Netherlands are reused. These calcite crystals have been used for treatment of aggressive groundwater, neutralization of acid wastewater, for road construction, cement manufacture, and in the metal industry .
• Crystallizers are designed with upflow velocities as high as 120 m/h (49 gallons per minute [gpm]/ft2) . Consequently, the fluidized bed crystallizer footprint is much smaller than that required for conventional softening.
For optimal performance in a fluidized bed reactor, it is important that calcium be removed by crystal growth rather than by spontaneous nucleation of calcium carbonate. According to Graveland et al. , good crystallizer design should include the following features:
• Proper water and chemical distribution to produce plug flow and avoid short-circuiting.
• Intensive mixing of the chemical to avoid locally high supersaturation and spontaneous nucleation of CaCO3 rather than crystal growth.
• Chemical mixing in the presence of a high seed surface area to promote crystal growth.
• Sufficient turbulence in the reactor to prevent scaling of nozzles and the reactor wall.
Was this article helpful?