(495d) Real-Time Mineral Scale Detection and Characterization Using a High Pressure Ro Ex-Situ Scale Observation Detector

In recent years there has been a growing interest in the desalination of inland brackish water in order to develop new water sources. Inland brackish water sources typically contain high levels of mineral salt ion precursors (e.g., calcium and barium cations and sulfate and carbonate anions) Economical desalting of brackish water requires operation at relatively high recovery (80% and higher). However, at such high recovery levels, the concentrations of ions near the membrane surface increase due to concentration polarization (CP) and can exceed the saturation level with respect to the solubility of sparingly soluble mineral salts (e.g., calcite, gypsum and barite). These mineral salts can precipitate and/or crystallize directly onto the membrane surface, thereby scaling the membrane. As a result permeate flux decline ensues and membrane productive life is shortened. In order to develop effective scale mitigation strategies, it is necessary to: (a) detect scale formation sufficiently early prior to any significant flux decline and (b) develop fundamental understanding of scale formation kinetics as impacted by process conditions.

In the present work an ex-situ scale observation detector (EXSOD) was developed with the capability of direct on-line observation of surface scaling as it occurs on the surfaces of NF/RO desalination membranes. The EXSOD system consists of a high pressure optical RO cell with the capability of recording magnified membrane surface images by a high resolution time-lapse digital photography. This system enables real-time observation of surface mineral salt crystallization while monitoring flux decline and other process parameters (i.e., pressure, flow rates, temperature, conductivity and pH). In the present study, gypsum and calcite crystallization were selected as the model scalants. While calcite scaling can be suppressed, to some extent by pH adjustment, gypsum solubility is essentially pH independent and its crystallization inhibition with antiscalants is limited. In order to demonstrate the level of fundamental information that can be evaluated by direct observation of surface crystallization, studies were conducted with various model solutions that mimic actual field compositions of brackish water over range of process conditions. Direct measurements were obtained for scaling induction time, the time evolution of number density of surface crystals, kinetics of single crystal growth, morphology of growing crystals, and the evolution of surface scale coverage. The results revealed that the time evolution of surface crystal number density followed a sigmoidal curve with faster nucleation in areas of higher supersaturation levels. Single surface crystal growth kinetics, which was described by a classical mass transfer growth model, demonstrated that the crystal growth mass transfer coefficient increased rapidly with crystal size during the initial period of growth but reaching an asymptotic value with crystal maturity. The crystal mass transfer coefficient was correlated effectively with a Sherwood-type mass transfer correlation. Crystallization kinetics was evaluated at various supersaturation indices and this information was utilized as input to a simple scaling model to simulate flux decline. The effectiveness of antiscalant treatment for suppression of scale formation was also evaluated via direct visual observation demonstrating the impact of antiscalant on both the ?observed? crystallization induction time and the kinetics and morphology of crystal growth. A series of studies were also carried out to evaluate the potential for cleaning a scaled membrane surface by forward surface washing with undersaturated feed solution. The effectiveness of surface cleaning was evaluated by direct visual observation of the dissolution of surface crystals and the accompanying change in flux and permeate conductivity. The present study demonstrated that the EXSOD system is capable of early detection of scale formation (i.e., significantly prior to the observation of any significant flux decline) and thus presents a unique capability for real-time monitoring of scaling and fouling in RO plant.