High Recovery Desalination of Brackish Water by Chemically-Enhanced Seeded Precipitation
- Author(s): McCool, Brian Carey
- Advisor(s): Cohen, Yoram
- et al.
Various regions around the world are confronted with dwindling water supplies and thus the need for exploiting non-traditional inland brackish water resource, as well as reclamation and reuse of municipal wastewater and agricultural drainage (AD) water. Reverse osmosis (RO) membrane desalination is the primary technology for inland brackish water desalting. However, successful implementation of RO technology requires operation at high product water recovery (>85%) in order to minimize the volume of generated concentrate (i.e., brine). Brine management is a key factor governing the economics of inland water desalination. Therefore, brine volume reduction is critical to enabling various brine residual management options. At high water recovery, dissolved mineral salts (e.g., CaSO4, BaSO4, CaCO3) may become concentrated above their solubility limits and may crystallize in the bulk and onto the surface of the RO membranes. Mineral crystallization leads to membrane scaling and hence leads to flux decline, increased process costs, and shortening of membrane life. Therefore, the attainable desalination water recovery is limited by mineral scaling. Many inland brackish water sources contain high concentrations of sparingly soluble mineral salts. In certain areas, such as in California's San Joaquin Valley (SJV), brackish water is near saturation with respect to calcium sulfate and barium sulfate. Based on the current work, single-stage RO desalination in SJV would generally be limited to ~50-70%.
In order to desalt brackish water of high mineral scaling propensity at a high recovery level (>85%), the feasibility of intermediate concentrate demineralization (ICD) of primary RO (PRO) concentrate, as a means of enabling secondary RO (SRO) desalting, was investigated with a focus on brackish water having high concentrations of gypsum salt precursor ions (i.e., calcium and sulfate). Accordingly, a two-step chemically-enhanced seeded precipitation (CESP) ICD process was developed in which the PRO concentrate is treated prior to further SRO desalting. The first step is lime precipitation softening (PS) which serves to induce sufficient CaCO3 crystallization in order to remove residual antiscalant (AS), a PRO feed treatment additive (generally polymeric) used for scale control, that would otherwise inhibit precipitation (in the ICD) of the target mineral salt scalants. Subsequently, gypsum seeded precipitation (GSP) is carried out to reduce the level of calcium sulfate saturation.
The CESP process was evaluated experimentally, in a batch crystallizer, using synthetic PRO concentrate and also PRO concentrate generated in the field, from AD water, using a spiral-wound RO pilot plant. The effect of residual AS (from the PRO stage) on retardation of mineral salt precipitation (in the ICD) was evaluated using both a generic (polyacrylic acid) and a commercial AS. Laboratory batch CESP studies were carried out in which the CESP process conditions were first optimized with respect to the required lime and gypsum seed doses. For raw brackish water that was about 98% saturated with respect to gypsum, PRO desalination at 52%-62% recovery yielded a brine stream 70-150% above saturation. CESP treatment, at lime doses of 0.25-0.35 mg/L and gypsum seeding of 4-5 g/L, enabled reduction of gypsum concentration to only 10-15% above its saturation. In general, the sequential processes of lime treatment for 10-20 minutes followed by ~1 hr of GSP were sufficient to achieve the above level of gypsum desupersaturation. GSP alone reduced gypsum saturation by only ~5%. PRO brine desupersaturation via CESP was feasible due to the effectiveness of AS removal (up to 90% for AS content of up to 10 mg/L in the PRO brine). Analysis of AS removal using a fundamental AS adsorption model, along with measurements of the size distribution of precipitating CaCO3 crystals, indicated that the area for AS adsorption provided by lime-induced nucleation of CaCO3 crystals is the key factor governing AS removal. In order to establish the feasibility of deploying CESP as a continuous process, a numerical model was developed for a fluidized bed reactor for the GSP stage. Model simulations indicated that the required level of calcium sulfate desupersaturation could be maintained by solids recycling leading to a steady-state particle size distribution.
Process simulations and economic analysis were carried out for the integrated process of PRO, CESP and SRO (PRO-CESP-SRO) demonstrating the existence of an optimal recovery (with respect to product water treatment cost). For the evaluated SJV brackish AD water source, the optimal recovery was about 93%. Overall brackish water treatment cost, when considering the disposal cost of high salinity AD water, was lower for PRO-CESP-SRO relative to a similar process based on conventional PS or utilizing a single stage RO which would be of limited recovery (<62%). The current work suggests that, at inland areas with limited options for management of high salinity water, high recovery desalination can be economically attractive when considering the reduction in brine disposal cost and the value of the product water produced by the PRO-CESP-SRO process.