The solution pH directly affects the rejection of both the ammonium and the nitrate ions. A solution pH above 7 has a very negative impact on ammonium ion rejection, since the equilibrium between the monovalent ammonium ion and the uncharged ammonia molecule shifts toward ammonia as the pH approaches the basic range. With little ionic charge to enhance rejection, the very small ammonia molecule readily permeates RD membranes. This effect was demonstrated during the pilot testing, where as the pH was raised from 5.0 to 8.5 the ammonia rejection of the RD system dropped from 87% to 62%.
In contrast, nitrate ion rejection is improved by increasing the pH. During the lab test, adjustment of the solution pH from 2.0 to 4.1 improved nitrate rejection from 31% to 94%. The mechanism for this improvement is not as obvious, but probably also involves the effect on the apparent charge of the ion. Decreasing the pH shifts the equilibrium away from the monovalent nitrate ion toward an uncharged nitric acid molecule (HNO3). Again, such a small uncharged molecule more readily permeates the RD membrane.
Previous experience has demonstrated that, while the NH4+ rejection level is similar for the membranes of both an aromatic polyamide (PA) and a cellulose acetate (CA) blend (di- and triacetates), the PA membrane exhibits substantially better rejection of the N03- ion. To simplify, this can be attributed to the inherent difference in surface chemistry characteristics between the membrane polymers, which require smaller pores in the PA membrane - relative to CA - to achieve similar NaCl rejection levels (NaCl is the most common marker solute to commercially rate RD membranes). This pore size difference is demonstrated by the superior small organic molecule rejection capability of PA membranes relative to CA.
In this case, the 'fluffy' charge density of the N03- ion (due to resonance) reduces the effect of the dielectric repulsion mechanism (as defined by Sourirajan) and rejection becomes more dependent on pore size. Hence, lower rejection values are obtained for the CA membrane with its larger pores. In contrast, the NH4 ion, with a point charge density behavior more like typical salt ions, displays similar rejection levels for both membrane polymers.
Initial lab testing included a membrane scan and waste stream chemistry manipulation. An aromatic polyamide RO membrane was selected for field-site testing on the condensate from a nitric acid neutralization process. Varying the pH from 2.9 to 8.2 identified a technically acceptable range of 3.0 to 7.0. Economic considerations narrowed the optimum processing range to 3.0 to 4.0, thereby minimizing pH adjustment costs since the raw waste pH is about 1.5. Pilot testing on the actual wastewater further indicates that system operation at 280 psig (19.3 bar), 75% recovery and ambient temperature should produce permeate and concentrate streams suitable for reuse. Reuse of both streams will eliminate all discharge from the neutralizer process. Reuse of the concentrate stream and discharge of the permeate will reduce the nitrogen contribution to the plant outfall, presently discharged to the Mississippi River, by 78%. Based on current production levels this will result in an ammonium nitrate outfall reduction of over 200 kg (440 pounds) per day.