Desalination of geothermal waters using a hybrid UF-RO process.Part II: Membrane scaling after pilot-scale tests

This paper presents the results of membrane scaling studies obtained after eight months of desalinating geothermal water in an integrated system that combines iron removal, ultrafiltration and two independent reverse osmosis systems fitted with low-pressure brackish water reverse osmosis membranes that are connected in series. Scaling prediction conducted using geochemical modelling showed that at a temperature of 30 °C and for neutral-pH water from the geothermal intake (GT-1 well) there were good conditions for the precipitation of aragonite, barite, calcite, chalcedony, dolomite, hydroxyapatite and quartz on the reverse osmosis membranes. A theoretically selected phosphonate scaling inhibitor failed to protect the membranes against sediment precipitation. Adjusting the pH of the water to ca. 5 by treating it with hydrochloric acid had a positive effect on the desalination process. On the basis of a morphological evaluation of sediments in the electron scanning microscope image and through the analysis of the chemical composition of the sediments crystallised on the membrane, silicates (SiO2) and aluminosilicates (Al2SiO5) as well as barite (BaSO4) and copper sulphate (CuSO4) crystalline structures were found. Calcium phosphate as well as celestite (SrSO4) and strontianite (SrCO3) sediments were also observed, which points to significant concentration polarisation. Most grains were a few to a maximum of 30 μm in size. Sediment clusters were usually heterogeneous in character. Desalination was performed with a high permeate recovery ratio (75–78%) with favourable key indicator retention ratios. © 2013 Elsevier B.V. All rights reserved.

During a pilot study of geothermal water desalination on a semiindustrial scale using DOW FILMTEC BW30HR-440i membranes, no decrease in the performance of the membrane system was observed and the efficiency of geothermal water desalination was high (75–78% at RO-1 and 75% at RO-2). However, the main factor that had a positive effect on the desalination process was the hydrochloric acid treatment of water fed to the first reverse osmosis stage, which decarbonised the water and also reduced the saturation index with respect to sulphate and silicate mineral forms. As far as the water studied is concerned, water whose carbonate hardness is on average 645 mg CaCO3/L, the use of a phosphonate antiscalant did not effectively prevent the precipitation of aragonite and calcite. The results of an examination of the membrane following eight months of water desalination showed a limited amount of absorbed sediment. This was mainly composed of sulphate (barite) mineral phases and silicate (chalcedony and quartz) forms. The prediction of the saturation index of acidified water fed to RO-1 membranes did not reveal solution oversaturation with respect to copper sulphates, calcium phosphate crystallised in the form of apatite, or celestite (SrSO4) and strontianite (SrCO3). This points to significant concentration polarisation. The thickness of the polarisation layer depends on the dimensions of the canals above the membrane surface and the nature of liquid flow. Adverse effects of concentration polarisation may be reduced by affecting either the permeate flux (Jv) or the mass transfer coefficient, which is a function of the diffusion coefficient and the rate of liquid flow above the membrane surface [8]. The limited amount of sediment found on the surface of the membranes demonstrates that water desalination in industrial installations may be conducted by treating water with hydrochloric acid in combination with: • a high permeate recovery ratio (75%) with a feed water pH ca. 5 before RO-1 and 10 before RO-2, in conditions where there is a risk of sparingly soluble substances crystallising (silica, barite, aluminosilicates and copper sulphates); • optimising module operating conditions or reducing process efficiency to obtain a permeate recovery ratio of ca. 50%.