Phosphate Recovery 2.0

As struvite, a fertilizer, is not the solution for phosphate recovery in countries with already enlarged concentrations of P in their soils, the recovery of phosphate as calcium phosphate was investigated as the latter can we a substitute for phosphate ore, which itself largely consists of calcium phosphate. The recovery of phosphate as calcium phosphate is however in competition with calcium carbonate formation from dissolved inorganic carbon (DIC).

Nitrification of the ammonium present in waste water results in a bi-equimolar removal of TIC, according to reaction 1. A further denitrification leads according to reaction 2 to an increase of the pH, a condition favourable for calcium phosphate precipitation, but also leads to an increase of the TIC concentration; reaction 2.

4 NH4+ + 8 O2 + 8 HCO3- → 4 NO3- + 8 CO2↑ + 12 H2O                  Reaction 1

5 H-C°-OH + 4 NO3-  → 2 N2 + CO2 + 4 HCO3- + 3 H2O                 Reaction 2

Simulation of the SI at different pH values and at different Ca/P ratios of ionic combinations present in anaerobic effluent and modifications obtained after nitrification, partial or total removal of DIC and denitrification, showed that different carbonates and magnesium phosphate easily can coprecipitate with calcium phosphate. Simulation of the step-by-step precipitation of the different phosphates in batch tests was used to predict the optimal pH interval for phosphate recovery as calcium phosphate when operating a CSTR. The results of the different continuous lab experiments agreed with the simulations: the remaining phosphate concentration was well predicted and the different contaminations of the isolated calcium phosphate were expected based on their SI values The most prominent conclusions are i) that a nitrification is a good pre-treatment; ii) that a supplementary denitrification resulted in phosphate loss during denitrification and an increased contamination of the calcium phosphate with magnesium phosphate and calcium carbonate; iii) that, at neutral pH, to obtain a good phosphate recovery, the added Ca/P at least should be 4, independent of the endogenous amount of calcium ions.

At pilot scale, after nitrification and at a CaADDED/P ratio of 4.68 ± 050 (Pout = 7.30 ± 3.65 mg.L-1), a first small (1.5 kg; 21.0 % P2O5; 6.51 % TOC; 25.64 % CaCO3) and a second larger amount (10,12 kg; 34.9 P2O5; 4.86 % TOC; 14.21 % CaCO3) of precipitate were obtained. Filtration over 25 µm and ultrafiltration as additional pre-treatment to remove organic material were tested (CaADDED/P= 9.39 ±5.4’; Pout = 6.45 ±6.70 mg.L-1)); results (resp. 32.4 % P2O5; 6.33 % TOC and 35.5 % P2O5; 2.26 % TOC) showed a clear effect of ultrafiltration. It was also observed that aeration of the reactor enhanced selective phosphate precipitation (the % CaO / % P2O5 decreased from 1.28 to 1.21). In an experiment of 3 months on nitrified, ultra-filtered and aerated effluent (CaADDED/P = 6.79 ± 2.33; Pout = 5.26 ± 2.92 mg.L-1), a black colour developed (tannins?). The dark gray precipitate was isolated, washed and dried: 17 kg of a slightly gray product was obtained (37.5 % P2O5; 2.80 % TOC; 15.36 % CaCO3). The following experiment (CaADDED/P = 3.93 ± 0.52; Pout = 11.65 ± 4.22 mg.L-1) showed that discontinuous aeration was possible, but that the colour remained. The phosphoric acid produced from the 2 largest preparations was black coloured but fulfilled all the chemical criteria.

In view of the mentioned depletion phosphate rock, the fractionation and concentration of phosphate from wastewater prior to the precipitation step can be very promising. For this purpose, anionic selectrodialysis (aSED), which uses a conventional electrodialysis stack that is supplementary equipped with monovalent selective anion-exchange membranes, was used. First it was demonstrated by batch experiments on pilot scale that the PC-Acid 100 OT standard anion exchange membrane allowed the best phosphate migration from nitrified effluent and was therefore selected to perform a long-term pilot experiment in the feed and bleed mode. Next, the long-term pilot experiment demonstrated that the recovery of nutrients and water from nitrified effluent is technologically feasible by means of the combination of ultrafiltration and aSED. When sufficient membrane area is used, the nitrified effluent is completely desalinated and thus also dephosphated (in the context of water reclamation), phosphate accumulates in the product stream and KNO3 accumulates in the brine stream (in the context of nutrient recovery). The product effluent contained on average about 162 mg.L-1 P and was fed to a lamella separator to which CaCl2.2H2O was dosed (Ca/P ratio of 1.93). This resulted in a phosphate removal of 98%. From the XRD analysis, the precipitate formed was found to be crystalline hydroxyapatite.

It was thus demonstrated on pilot scale that it is possible to recover phosphate as calcium phosphate from anaerobic effluent of a potato processor after a preceding nitrification, UF filtration and under aeration. Contamination of the precipitate with organic carbon and calcium carbonate and discoloration (tannins?) were problems, but can be resolved by calcination. The finally obtained phosphoric acid was dark in colour, but fulfilled all other chemical criteria. Further refinement of the procedure is necessary e.g. by the suggested calcination since the recovered phosphate is a valuable recycled product that can be added to mined phosphate ore. The selectrodialysis is very promising as it  can lead to a technology that both recovers water and various nutrients  Phosphate was recoverd from the product phase of the system in a yield of 98 %.