When does commercial software fail in predicting scaling tendency in reverse osmosis and what can we do better?

Avoiding chemical scaling by sparingly soluble salts is a critical design component in desalination applications based on reverse osmosis. If not prevented, scaling can rapidly reduce membrane lifespan and performance while considerably increasing desalinated water production costs. Therefore, an important part of reverse osmosis process design is the calculation of scaling tendency of the concentrated rejected solution. For this purpose, commercial software packages developed by membrane manufacturers or suppliers of antiscalant chemicals are commonly used. Nevertheless, the ability of commercial software to produce realistic predictions of scaling tendency was scarcely studied. Here the performances of three prominent computer programs, one from a large membrane manufacturer and two from prominent antiscalant suppliers, are systematically tested by comparing their predictions to simulation results made by WATRO (weak acid transport in reverse osmosis)-an experimentally validated mechanistic model recently developed by the author. Three different cases were examined: 1^st pass seawater reverse osmosis, brackish water reverse osmosis and 2^nd pass of reverse-osmosis, treating seawater reverse-osmosis permeate. In many cases, it was found that the predictions made by the commercial programs were unrealistic, resulting in unnecessarily high recommended doses of antiscalants chemicals to the feed. The unrealistic predictions are mainly due to the incompetence of the tested commercial software in modeling the change in pH, developing in the concentrate upon increasing the permeate recovery ratio. Using the WATRO code, it was revealed that retentate pH calculations made by the commercial software is likely based on CO₂ permeation as the sole mechanism, together with ad-hoc modifications. However, this mechanism is often not the dominant one, while other dominant phenomena controlling acid-base dynamics in the retentate during reverse osmosis are unaccounted for or improperly treated, giving rise to unrealistic predictions. Specifically, the influence of background ions on proton dissociation constants and the cross-membrane transport of hydronium and hydroxide ions should be accounted for in order to improve commercial software predictions. Fortunately, the WATRO code provides a suitable framework for expanding the predicting capacity of currently available design tools, thus promoting a more optimal design.