Urine separation is a very interesting process to recover phosphate from wastewater because it contains 550 mg PO₄-P l^-1. The preferred technology at the moment is struvite precipitation by the addition of MgCl₂ or MgO. This is not ideal at the household level, as the particularly dosage of the correct amount of magnesium salts is not very easy. Thus the technical feasibility of two alternative technologies was estimated in batch experiments at the lab-scale: adsorption of phosphate on GFH and calcite, or phosphate precipitation by electrochemical dissolution of magnesium.
The adsorption of phosphate out of untreated synthetic urine (pH=8.9) on a typical GFH/calcite mix (7:3) obtained maximally a load of 9 mg PO₄-P g^-1 adsorbent which was considered to be insufficient.
Further tests were conducted on synthetic urine after nitrogen removal (pH=6.3). GFH showed very good adsorption capacity of 47.5 mg PO4-P g^-1 GFH according to a Langmuir isotherm and high affinity to phosphate (32 mg PO₄-P g^-1 can be reached at 25 mg PO₄-P l^-1 in solution). However, the adsorption kinetics was rather slow, as 7 days were needed with grain size 0.8-1.25 mm to get in equilibrium. The application of a surface diffusion model could not well approximate the equilibrium time due to the high solute concentrations. 75% of the adsorbed phosphate could be recovered by 7 days desorption in 0.1 M Na(OH).
The addition of calcite was shown not to decrease the adsorption/desorption properties of GFH at pH=6.3 while its maximal phosphate load at that pH was maximally 7.4 mg PO₄-P g^-1, thus too small.
The process of electrochemical magnesium dissolution is shown to be a technically feasible process. Very good magnesium (99%) and current (over 100 %) efficiencies were obtained by applying a constant current (1.95/3.9 mA cm^-2] on two magnesium electrodes plunged in untreated synthetic urine when the polarity was changing every 30 minutes until phosphate was depleted. Current efficiencies above 100% are thought to be due to Local Cell Action which leads to magnesium electrode self-dissolution. A reduction of current efficiency was observed with time, as a precipitation layer formed on the electrodes. This layer had an impact on the energy needed to resolve the magnesium (up to 4 Wh g^-1 P) but also on the current efficiency which decreased with time in otherwise equal experiments.
The same passivation behaviour of the sacrificial magnesium was observed when stainless steel was used as cathode. By their intrinsic potential difference (measured as 1.2 V) , elimination rates of 0.29 mg PO₄-P l^-1 h^-1cm^-2 were observed. The current efficiencies were lower than in the experiments with two magnesium electrodes, ranging around 50%, depending negatively on the period the magnesium electrode was used and on the applied current intensity. This is thought to be due to the passivation of the electrode and to the fact that more self-dissolution could occur if the electrodes are longer in contact with urine. One series of measurements furthermore suggests that the oxidation of chloride to chlorine gas could account for a substantial part of the inefficiencies.
No of pH, conductivity and voltage measurement could be identified as process control master variable, as it a difference in absolute value cannot be estimated,
Unfortunately, neither of pH, conductivity and voltage measurement could be identified as potential process control because their behaviour was not observed to change when phosphate in solution was depleted.