Table of Contents

Return to previous page

View Article

CaCO3 deposits in reverse osmosis: Part II - Simulation model for hydrochemical predictions of reverse osmosis retentates and scaling propensity
S. Hager, M. Meinardus, T. Hofmann and K. Glas

Reverse osmosis (RO) technology is used in a wide variety of water treatment processes. In the brewing industry, mainly for the production of brewing and process water. The design and resulting efficiency of RO plants depends on hydro chemical predictions of the retentate flow. In order to avoid deposits in the membrane system, saturation limits of sparingly soluble salts are of decisive importance when selecting the maximum achievable fresh water yield. For the frequently present calcium carbonate (CaCO3), predictive indices of calcite saturation such as the Langelier saturation index (LSI) are widely used. However, due to simplifications of the water chemistry, the prediction quality must be critically questioned. In this context, we present a new simulation model for predicting RO retentate water chemistry. The calculations are performed in the freely available hydrochemical simulation software PHREEQC. Using a validation data set from German standard methods (DIN) for the determination of calcite saturations, a good agreement could be achieved with our calculation basis in PHREEQC software. The simplified description of LSI values from the ASTM International Standard and software from a membrane manufacturer provided greatly increased calcite saturations in comparison. In addition to the consideration of calcite saturation, we were able to extend our prediction of RO retentates to the saturations of further CaCO3 polymorphs as the monohydrate (MCC) and in the amorphous form (ACC). This allows us to reveal, for example, a metastable range of CaCO3 supersaturation as a function of the raw water recovery rate. Previously difficult-to-explain scaling behaviour of RO systems at high calcite supersaturation could be explained using this approach. The prediction of the retentate pH value and thus the membrane rejection of carbon in the membrane system showed good agreement with full scale data when simulating constant CO2 concentration independent of fresh water yield. Today, the hydrochemical design of reverse osmosis systems is usually based on unpublished calculation methods of membrane manufacturers, chemical suppliers and the ASTM guidelines. With our simulation approach, these predictions can be critically evaluated and comprehensive RO design calculations can be modularly coupled with our interface open simulation model in PHREEQC.

Descriptors: scaling, CaCO3, LSI, fouling, reverse osmosis, pH prediction, Langelier Saturation Index

BrewingScience, 75 (July/August 2022), pp. 54-68