Date of Award

Spring 1-1-2017

Document Type


Degree Name

Master of Science (MS)

First Advisor

John Pellegrino

Second Advisor

Scott Summers

Third Advisor

Joseph Ryan


This research focuses on developing techniques to reliably analyze the crystallization kinetics of sparingly soluble salts in reject/concentrate streams from membrane-based, inland water supply processes. More usable water can be recovered (and lower disposal costs incurred) from these concentrate streams through efficient crystallization. Currently, pellet softening is a preferred mode of crystallizing the supersaturated salts from such streams. A conventional pellet softener in the ideal sense is a non-ideal plug flow reactor (PFR). In this work, we present the evolution of a steady state, continuously stirred tank reactors (CSTRs)-in-series approach, to approximate a PFR, to study crystallization kinetics of a model solution that is supersaturated in calcium carbonate. Furthermore, using this setup, we can systematically study crystallization, in general, and develop quantitative engineering-design, scale-up parameters. In the current work, we have used pH, conductivity and turbidity changes in the system to monitor crystallization in the CSTRs-in-series setup. We have used up to six CSTRs with individual residence times of approximately 2, 5 and 11 min. This system operates in a steady state mode with total treatment times between ~15-68 minutes. Our reactor setup was capable of handling up to 875 mL/min of hard water at the shortest residence time studied. The supersaturation was depleted ~25% with a total reactor residence time of 15 min and over 50% for overall times in the range of ~68 min. We have also been able to estimate induction times for crystallization based on the metric of 5 NTU as being the point of discernible crystal formation. Interestingly, crystallization always began in the second tank for all the residence times. Thus, the induction time appears to be influenced by other process variables beyond simply the time spent (by the inlet water) in the first tank of the CSTRs-in-series setup. Our initial hypothesis is that the induction time depends on the level of mixing in the first reactor stage and screening\studies of this variable are presented and discussed, but significant further work is needed.