Validation of a Method for Modelling Brine and Permeate pH in RO and NF Systems

Concentrate pH in RO and NF applications plays a significant role in the formation of certain scales and the rejection of various contaminants. Incorrectly predicting pH can lead to excessive chemical costs or operational issues. Likewise, permeate pH is of utmost importance for determining post-treatment requirements for the product water.

Most membrane and antiscalant projection software currently use a method similar to that of the ASTM standard D3739 for predicting concentrate and permeate pH. These calculations assume that the concentration of carbon dioxide in the feed, permeate and concentrate are equivalent. They determine the pH in each stream based on the equilibrium relationship according to the ratio of bicarbonate to carbon dioxide concentrations.

When comparing calculated values to measured pH values in real systems, it becomes very apparent that methods relying on HCO3/CO2 ratio predict concentrate and permeate pH very poorly outside a feed pH range of 5 – 6. They are overly conservative, always predicting excessively high concentrate pH values and excessively low permeate pH values. In real-world applications, there are cases where the measured concentrate pH is lower than the feed pH or where the permeate pH is higher. Such cases are never predicted correctly using the HCO3/CO2 ratio, which always determines that the concentrate pH must be higher than the feed pH, and that the permeate pH must be lower than the feed pH. Some have recently attempted to correct for this phenomenon by using the equilibrium relationship of carbonate to bicarbonate at higher feed pH, but we have found such calculated values to be inconsistent and unreliable. Others have looked at the interaction of hydronium and hydroxide ions with the membrane, but in the presence of any weak acids, the pH would only be impacted by free hydronium or hydroxide ion concentrations at pH extremes.

The approach described in this paper predicts concentrate and permeate pH very reliably in RO and NF systems. This model has been tested against real systems and verified to predict concentrate and permeate pH very accurately across the entire pH range. It has been found to be accurate in brackish water, seawater and industrial reuse applications.

This paper describes how the pH calculation method was derived and provides examples where the model was compared to onsite measurements for various types of feedwater.