Assessment of the scaling potential for sparingly soluble salts in RO and NF units

This paper briefly discusses the current industrial practice for assessing the scaling potential of sparingly soluble salts and suggests a more theoretical approach which is expected to provide more accurate results. This theoretical approach is based on the use of Gibbs free energies of reaction for prediction of dissociation and solubility constants which have usually been estimated crudely through empirical means. In addition, this systematic approach incorporates determination of the activity coefficients by considering short and long ion-interactive forces through incorporation of thermodynamic electrolyte models. Furthermore, since the basic principles and individual ioninteractive forces are incorporated. The results demonstrate that total dissolved salts cannot be used to incorporate the effect of salinity on scaling potential. This has further implications in membrane modules as the individual ionspecific passages through the membranes are different for different salts and the relative composition of the feed stream changes from the inlet to the outlet. As such the specific ion composition within modules should be used for assessing the scaling potential along the modules. Advances in computational techniques enables us to asses the scaling potential of sparingly soluble salts from a more fundamental and theoretical basis as discussed in this paper. Further incorporation of computational techniques and basics of thermodynamics would allow incorporation of other factors and even interactive effects.

In practice various indices and relationships are used for assessing the scaling potential of sparingly soluble salts. These indices/methods are at least partially based on empirical relationships and were obtained originally for conditions different from those encountered in membrane systems which have high recoveries and specific ion passages. These methods in addition to being partly empirical, do not take into account the specific ion electrostatic and interaction forces; they consider TDS as the only measure of ion activity. Practice has shown that interactive effects even by trace species cannot be ignored and that they have significant effect on scaling potential of given water. The advances in computation allow us to revert to fundamentals and formulate a general scaling potential index for all sparingly soluble salts based on thermodynamics of electrolyte systems. This paper has introduced the Scaling Potential Index (SPI) which can be used for all sparingly soluble salts and at salinities up to 6 M; SPI is formulated based on the Gibbs free energy of reaction for calculation of thermodynamic solubility product and by incorporation of the electrostatic and ion interactive forces using the comprehensive Pitzer model for calculation of ion activities. The validity of Pitzer model for salinities up to 6 M is already assessed in the literature. When SPI is negative, certainly no scale will form; when zero the system is at equilibrium; when the SPI is positive, the salt has a “potential” to form scale. It should be noted that any water having a positive SPI would not have the capacity to form the scale or the kinetics might be too slow. The SPI is a conservative estimate for assessing scaling potential. The effect of system kinetics and hydrodynamics on formation of the scale is discussed and should not be ignored. The calculated values of thermodynamic solubility constants are compared with experimental ones and they are in excellent agreements. In addition, the importance of other factors such as kinetics and hydrodynamics of the system has been discussed and SPI calculations for CaSO4 solutions have shown that even at a given SPI value when the scaling potential for solutions are identical, higher salinities can significantly increase the kinetics and reduce the induction times and as such promote scale formation. The future directions that theoretically could incorporate the interactive effects were introduced. Furthermore, the importance of the residence time distribution within modules has been discussed. Finally, the calculations have shown that the specific ion concentration contributing to the value of TDS is of importance and just a global value of TDS is not sufficient to be used to incorporate ionic effects; solutions having the same increase in their TDS value due to increases in NaCl, MgCl2 or NaSO4 do not have the same SPI and as such the same scaling potential with respect to CaCO3. These are very important in case of NF and RO membrane systems that have high recoveries in comparison to thermal units and specific ion passages. Time has come with the advances in computation to rely on theoretical assessment of scaling potential and obtaining a unified approach which could be improved to include all interactive effects from a fundamental points of view.