Phosphorus recovery from urine and anaerobic digester filtrate: comparison of adsorption–precipitation with direct precipitation

Abstract

Hybrid anion exchange resin containing hydrous ferric oxide (HAIX-Fe) was used in column tests to remove phosphate (PO₄) from fresh urine, hydrolyzed urine, and anaerobic digester filtrate, and subsequently recover PO₄ as struvite (MgNH₄PO₄·6H₂O) or potassium struvite (MgKPO₄·6H₂O) via precipitation in the spent regenerant. The recovery of PO₄ using the two-step adsorption–precipitation process was compared with direct precipitation in urine and anaerobic digester filtrate considering chemical requirements for precipitation and mineral purity. Following the saturation of HAIX-Fe resin with PO₄ from urine and anaerobic digester filtrate, up to 95% of the PO₄ was desorbed using caustic brine during the regeneration phase. The spent regenerants were more concentrated in PO₄ than the original urine and anaerobic digester filtrate due to recycling of the regeneration solution. Precipitation in the spent regenerants and original wastewaters (urine and filtrate) yielded 96.7–99.8% PO₄ recovery as struvite or potassium struvite. Direct precipitation in fresh urine and hydrolyzed urine was more efficient than precipitation in the corresponding spent regenerants based on lower chemical requirements. Precipitation in the spent regenerant from HAIX-Fe resin saturated with PO₄ from anaerobic digester filtrate produced a higher purity mineral than direct precipitation in the anaerobic digester filtrate.

---

Water impact statement

Phosphorus (P) is a pollutant, finite resource, and critical element in agriculture. As such, there is growing interest in P recovery from waste streams for beneficial use as fertilizer. The key findings of this research are new experimental data on (1) P adsorption from urine and anaerobic digester filtrate to resin under continuous-flow operation and corresponding regeneration efficiency, and (2) chemical requirements for and mineral purity of P precipitation in the spent regenerants from resin adsorption/desorption and original waste waters. The results of this research are generalizable to resource recovery from domestic and industrial waste streams, and are significant because they advance source-separated and decentralized approaches to wastewater management that aim to increase recovery of water, nutrients, and energy.

Introduction

Recovery of phosphorus (P) from various types of wastewater has been proposed as a more sustainable alternative to phosphate rock mining due to concerns about the finite supply of phosphate rock, economic scarcity of phosphate, and the critical need for P in agriculture and food production.^1–4 Moreover, recovery of P from wastewater is attractive because wastewater treatment plants serve as a point source of P to receiving waters,⁵ which can harm aquatic ecosystems. In addition, excess P in wastewater can cause operational problems at wastewater treatment plants during sludge handling due to mineral scaling in pipes and pumps.⁶ Thus, P recovery from the most concentrated waste streams that comprise domestic wastewater, such as human urine and anaerobic digester sidestreams,^7,8 would provide an alternative to phosphate rock fertilizers while having the important benefits of decreasing nutrient loading to receiving waters and reducing operational problems caused by mineral scaling at wastewater treatment plants.

Phosphorus recovery by precipitation is a well-established process whereby the minerals struvite (MgNH₄PO₄·6H₂O),^7,9–12 and to a lesser extent potassium struvite (MgKPO₄·6H₂O),^13,14 have been investigated for various types of wastewater. For example, previous researchers have achieved 95% P recovery by precipitation of either struvite or potassium struvite in human urine,¹⁴ and 94% P recovery by precipitation of struvite in anaerobic digester sidestreams.⁷ In addition, struvite has the potential to be used as an effective slow-release fertilizer based on plant growth studies.¹⁵ One of the major challenges to P recovery by precipitation is achieving a high purity product. For example, the presence of calcium during struvite precipitation can decrease mineral purity by co-precipitating calcium carbonate and calcium phosphate minerals with struvite.^16,17 The extent to which co-precipitation is a problem depends of the final use of the struvite product. In some cases, potential contamination of struvite due to heavy metals^18,19 and pathogens²⁰ is possible, making the final product less desirable as fertilizer.

As an alternative to precipitation, adsorption processes can effectively and selectively remove phosphate (PO₄) from many types of wastewater.^21–23 For instance, hybrid anion exchange (HAIX) or ligand exchange resins combine polymer anion exchange resin with transition metals (e.g., Zr(IV),²⁴ Cu(II),²⁵ and Fe(III)²⁶) thereby enabling selective removal of PO₄ over competing anions such as sulfate, chloride, and bicarbonate.²⁷ This study used HAIX resin infused with hydrous ferric oxide (HFO); hereafter HAIX-Fe. The predominant mechanism of removal of PO₄ by HAIX-Fe resin is ligand exchange,^26,28 whereby the HFO immobilized in the resin serves as the Lewis acid which forms an inner-sphere complex with the Lewis base, PO₄.²⁹ HAIX-Fe resin has been used to remove PO₄ from synthetic and real secondary effluent (0.26–4.7 mg P L⁻¹),^26–28,30 wastewater-derived reverse osmosis concentrate (12 mg P L⁻¹),³¹ synthetic (324 mg P L⁻¹) and real (472 mg P L⁻¹) sludge liquor,³² and synthetic human urine (460–620 mg P L⁻¹).^21,33 The previous studies show greater removal of PO₄ by HAIX-Fe resin than anion exchange resins or granular ferric hydroxide, and selective removal of PO₄ by HAIX-Fe resin in the presence of competing anions. Upon saturation of HAIX-Fe resin with PO₄, regeneration of HAIX-Fe resin using caustic brine can desorb >90% of the adsorbed PO₄.^26–28,32 The drawback of PO₄ removal by adsorption is that an additional step is needed to recover the PO₄ in a useable form such as for fertilizer. For instance, PO₄ recovery has been demonstrated by adding magnesium and ammonia to the spent regenerant produced by HAIX-Fe regeneration thereby precipitating struvite.²⁸

The gap in the literature that motivated this research was the absence of data comparing PO₄ recovery by precipitation in waste streams (i.e., direct precipitation) with precipitation in the spent regenerant following regeneration of HAIX-Fe resin saturated with PO₄ (i.e., two-step adsorption–precipitation). Furthermore, there is only one previous study on struvite precipitation in the spent regenerant from HAIX-Fe resin²⁸ and no data on potassium struvite precipitation in spent regenerant. As such, there is not sufficient data available to compare the adsorption–precipitation process with direct precipitation in terms of chemical requirements for precipitation or purity of the final precipitated solid. In addition, there is limited data available on the continuous-flow column operation of HAIX-Fe resin to remove PO₄ from wastewater high in PO₄, total dissolved solids (TDS), and organics,³² and as such, it is not known how continuous-flow operation of HAIX-Fe resin will perform for source-separated urine. Accordingly, the goal of this research was to compare the two-step adsorption–precipitation process with direct precipitation for P recovery from source-separated human urine and anaerobic digester filtrate. The specific objectives of this research were to (1) evaluate PO₄ adsorption from urine and filtrate to virgin and regenerated HAIX-Fe resin under continuous-flow column operation; (2) evaluate the regeneration efficiency of HAIX-Fe resin saturated with PO₄ using fresh and recycled caustic brine; (3) quantify the chemical requirements for PO₄ precipitation as struvite or potassium struvite in the spent regenerants and wastewaters through chemical addition and batch precipitation; and (4) evaluate the P recovery and mineral purity achieved by batch precipitation in the spent regenerants and wastewaters. Source-separated urine and anaerobic digester filtrate were selected for this research because both have high PO₄ concentrations. In addition, there is increasing interest in urine diversion and anaerobic digestion as part of source-separated and decentralized approaches to wastewater management to increase recovery of water, nutrients, and energy.^34,35

Methods

Wastewaters

Three types of wastewater were used in this study: synthetic fresh urine, synthetic hydrolyzed urine, and real anaerobic digester filtrate (Table 1). Human urine represents wastewater high in PO₄ and TDS, whereas anaerobic digester filtrate represents wastewater high in PO₄, TDS, and organics. The composition of synthetic urine was based on previous studies.^14,36,37 Synthetic urine was used to ensure that the composition of urine was consistent throughout the study. Previous studies have shown similar precipitation behavior in synthetic and real urine,³⁶ and comparable adsorption results in synthetic urine³⁸ and real urine.³⁹ The composition of synthetic fresh urine assumed that no urea hydrolysis had occurred, whereas the synthetic hydrolyzed urine assumed complete urea hydrolysis with spontaneous precipitation of struvite and hydroxyapatite to eliminate the magnesium and calcium, respectively, and reduce the PO₄ concentration.^40,41 Theoretically, the ammonia would also be very slightly reduced by struvite precipitation. Potassium struvite does not precipitate during urea hydrolysis because it is not supersaturated. Therefore, the potassium concentration, as well as the concentrations of sodium, chloride, and sulfate, remained approximately constant in fresh and hydrolyzed urine.

Table 1 Chemical composition of synthetic fresh urine, synthetic hydrolyzed urine, and real anaerobic digester filtrate used in adsorption column tests and precipitation tests

Chemical^a	Fresh urine^b	Hydrolyzed urine^b	Anaerobic digester filtrate^c
a All concentrations mmol L^−1, except pH and phosphate distribution (unitless) and ionic strength (mol L^−1). b Prepared using NH4Cl, NH4HCO3, urea, NaCl, NaH2PO4, Na2SO4, KCl, CaCl2·2H2O, and MgCl2·6H2O in deionized water based on previous studies.^14,36,37 Concentrations are calculated not measured, except pH measured. c Real filtrate spiked with NaH2PO4; measured concentrations. d Not added, assumed 0. e From O'Neal and Boyer calculated using Visual MINTEQ.^21 f Measured as total nitrogen. Not measured (nm).
Urea (as N)	500	—^d	nm
Total ammonia	—^d	500	134^f
Cl^−	100	100	20
Total phosphate	20 (619 mg P L^−1)	13.6 (421 mg P L^−1)	2.5^f (77 mg P L^−1)
H2PO4^−/total PO4^3−	0.940^e	0.0078^e	0.284^e
HPO4^2−/total PO4^3−	0.0596^e	0.991^e	0.716^e
SO4^2−	15	15	2.6
Total carbonate	—^d	250	nm
Na^+	94	104	14
K^+	40	40	1.9
Ca^2+	4	—^d	4.3
Mg^2+	4	—^d	2
DOC	—^d	—^d	7.0
pH	5.8	9.3	7.0
Ionic strength	0.145^e	0.476^e	0.158^e

---

The anaerobic digester filtrate was collected from the belt press filtrate of anaerobically digested sludge from the Howard F. Curren Advanced Wastewater Treatment Plant in Tampa, FL, USA. The anaerobic digesters are operated at approximately 37 °C with a typical solids residence time of 15–25 d and a combination of mixed sludge and waste activated sludge. The anaerobic digester filtrate was filtered in series through filters with pore sizes of 5, 3, and 0.45 μm (Millipore). The composition of the anaerobic digester filtrate given in Table 1 is presented to allow for comparison with the synthetic urine used in this work. More detailed analysis of anaerobic digestion liquors can be found in the literature.^42,43 In particular, the nature of the dissolved organics could affect the adsorption and precipitation processes investigated in this work but was outside the scope of this project. Because the Curren plant does not practice enhanced biological P removal, the anaerobic digester filtrate was spiked with NaH₂PO₄ to increase the PO₄ concentration to approximately 80 mg P L⁻¹ to allow for comparison with previous data on anaerobic digestion sidestreams.²¹ In addition, the rationale for using synthetic urine and real filtrate was because the composition of urine is well documented in the literature, whereas there is not a standard recipe for anaerobic digester filtrate.

HAIX-Fe adsorption column tests and regeneration

The HAIX-Fe resin used in this work was a commercially available product (trade names PhosX^np or LayneRT, SolmeteX, Northborough, MA). Phosphate adsorption to HAIX-Fe resin using fresh urine, hydrolyzed urine, and anaerobic digester filtrate was investigated using continuous-flow column tests. The columns (Omnifit) were glass with a 10 mm inner diameter. All column tests were conducted using 8 mL of HAIX-Fe resin, which was defined as one bed volume (BV). The column tests were operated up-flow by pumping wastewater (Masterflex peristaltic pump) at a flow rate of 2.5 mL min⁻¹, which yielded an empty bed contact time (EBCT) of 3.2 min and a superficial linear velocity (SLV) of 3.1 cm min⁻¹. All column tests were performed at laboratory temperature of 24 °C. Effluent samples were collected at predetermined times and analyzed for pH and PO₄ concentration. Phosphate breakthrough was defined as C/C₀ = 0.1 and saturation was defined as C/C₀ = 1. An initial column experiment was performed in triplicate using fresh HAIX-Fe resin and fresh urine to ensure reproducible results (see Fig. S1 in ESI†). In subsequent experiments HAIX-Fe resin was used to treat each wastewater three times as follows: first with virgin HAIX-Fe resin, second with once-used HAIX-Fe resin that was regenerated with fresh regeneration solution, and third with twice-used HAIX-Fe resin that was regenerated with once-used regeneration solution. All column tests were run until HAIX-Fe resin was saturated with PO₄, at which point the test was stopped and the resin was regenerated as described in the next paragraph. The column containing HAIX-Fe resin was rinsed with 10 BVs of DI water between the adsorption and regeneration tests. The mass of PO₄ adsorbed to HAIX-Fe resin was calculated by taking the difference in influent and effluent PO₄ concentrations for each sample multiplied by the volume of wastewater treated per sample. The total mass of PO₄ adsorbed to HAIX-Fe resin was the summation of PO₄ adsorbed for all samples. The capacity (or column loading) of HAIX-Fe resin for PO₄ was calculated by dividing the total mass of PO₄ adsorbed by the mass of resin contained in the column.

Regeneration was performed in the same column as the adsorption tests using a solution containing 2.5% (m/m) NaCl and 2% (m/m) NaOH, pH 13.7, based on previous work.²⁸ Effluent samples were collected at predetermined times and analyzed for pH and PO₄ concentration. An initial regeneration test was performed in triplicate using HAIX-Fe resin that was previously saturated with PO₄ from column tests using fresh urine (see Fig. S2 in ESI†). After the initial regeneration test, the flow rate was decreased to 0.8 mL min⁻¹ for all subsequent regeneration tests, which corresponded to an EBCT of 10 min and a SLV of 1.0 cm min⁻¹. The HAIX-Fe resin was regenerated after each column adsorption test described in the previous paragraph. This resulted in three regeneration tests for each wastewater: first with fresh regeneration solution, second with recycled (i.e., once-used) regeneration solution, and third with recycled (i.e., twice-used) regeneration solution. No additional NaCl or NaOH was added to the recycled regeneration solution. The twice-used regeneration solution is referred to as spent regenerant, and was used in the precipitation experiments. The total mass of PO₄ desorbed from HAIX-Fe resin during regeneration was calculated by combining all effluent samples into a single sample, measuring the PO₄ concentration, and multiplying by the total volume of regeneration solution. The regeneration efficiency was calculated as the amount of PO₄ desorbed divided by the amount adsorbed.

Precipitation experiments

Precipitation experiments were conducted using each wastewater and spent regenerant, and either struvite or potassium struvite was precipitated. Potassium struvite was precipitated in fresh urine and spent regenerant from HAIX-Fe resin saturated with PO₄ from fresh urine and hydrolyzed urine. Struvite was precipitated in hydrolyzed urine, anaerobic digester filtrate, and spent regenerant from HAIX-Fe resin saturated with PO₄ from hydrolyzed urine and anaerobic digester filtrate. Potassium struvite precipitation was conducted by adding the appropriate chemicals (MgCl₂·6H₂O and KCl) such that the solution contained a K : Mg : P molar ratio of at least 1.5 : 1.5 : 1. (Fresh urine had a K : P molar ratio of >1.5 : 1.) Struvite precipitation was conducted by adding the appropriate chemicals (NH₄Cl and MgCl₂·6H₂O) such that the solution contained a Mg : N : P molar ratio of at least 1.5 : 1.5 : 1. (Hydrolyzed urine and anaerobic digester filtrate had an N : P model ratio of >1.5 : 1.) All precipitation experiments were conducted at ambient laboratory temperature of 24 °C. Precipitation was performed in 125 mL amber bottles containing 50 mL of solution. The pH was adjusted to 9.3 using 1 M NaOH or 12.1 M HCl to create favorable conditions for precipitation and avoid precipitation of Mg(OH)₂ at high pH. Samples were mixed on an Innova 2000 Platform Shaker at 200 rpm for 2 h. Precipitation experiments using wastewater were conducted in triplicate; precipitation experiments using spent regenerant were conducted singly because of the limited volume of solution. After the precipitation experiment was complete the solution was filtered through 1.5 μm glass fiber filter (Whatman 934-AH RTU), the solid precipitate was dried at 24 °C in a desiccator for a minimum of 48 h and weighed. The PO₄ concentration in solution was measured before and after precipitation to determine the P recovery by precipitation. The percent P by mass of the solid precipitate was calculated using the measured mass of solid and the difference in aqueous PO₄ concentrations. X-ray diffraction (XRD) was used to identify the crystal structure of the solid precipitate.

Chemical equilibrium calculations

Visual MINTEQ (ver. 3.0) was used to calculate the concentration of dissolved species, ion activity product (IAP), ionic strength, and saturation index (SI) of relevant solids in synthetic urine and spent regenerants used in the precipitation experiments.⁴⁴ Two different calculation approaches were used: supersaturated solids not allowed to precipitate and struvite or potassium struvite allowed to precipitate. The former approach was used to identify all possible solid phases that were supersaturated. The later approach was used to calculate the extent of PO₄ removal by precipitation of struvite or potassium struvite. The input values were taken from Table 1 or the initial composition of the regeneration solution, and were augmented by the addition of chemicals required for precipitation as described later in Table 3. The SI was calculated as log(IAP/K_s0) where K_s0 is the solubility product and SI > 0 indicates supersaturated conditions. The SI for anaerobic digester filtrate was not calculated because of incomplete information on its composition especially related to organics. Activity corrections were made using the Davies approximation with B = 0.3 instead of 0.2,²⁹ which was shown previously to be the most appropriate approximation to the activity coefficient in urine.³⁶ The K_s0 for potassium struvite (10^−10.62) was added to the database.⁴⁵ The K_s0 in the database for struvite (10^−13.26) was consistent with published data and was not changed.³⁶

Table 2 Phosphate adsorption, desorption, and regeneration efficiency during HAIX-Fe resin column tests

Column test	Phosphate adsorption/desorption	Fresh urine	Hydrolyzed urine	ADF
Anaerobic digester filtrate (ADF).a Adsorbed/(volume HAIX-Fe resin × density HAIX-Fe resin).b Freundlich parameters from O'Neal and Boyer.^21c Fouled = adsorbed − desorbed.d Total = adsorbed + fouled.e Total/(volume HAIX-Fe resin × density HAIX-Fe resin).f Synthetic anaerobic digester supernatant.
First use	Adsorbed, mg P	31.9	20.7	19.9
	Column loading,^a mg P g^−1	10.2	6.7	6.4
	Freundlich capacity,^b mg P g^−1	10.1	5.7	5.1^f
	Desorbed, mg P	30.2	17.4	18.9
	Fouled,^c mg P	1.7	3.4	1.0
	Regeneration efficiency	94.7%	83.8%	95.0%
Second use	Adsorbed, mg P	30.0	19.4	20.0
	Total,^d mg P	31.7	22.8	21.0
	Column loading,^e mg P g^−1	10.2	7.3	6.7
	Desorbed, mg P	15.9	15.3	14.4
	Fouled,^c mg P	15.8	7.4	6.6
	Regeneration efficiency	50.1%	67.3%	68.6%
Third use	Adsorbed, mg P	25.9	15.9	17.3
	Total,^d mg P	41.7	23.3	23.9
	Column loading,^e mg P g^−1	13.4	7.5	7.7
	Desorbed, mg P	7.2	10.1	12.1
	Fouled,^c mg P	34.5	13.2	11.8
	Regeneration efficiency	17.3%	43.4%	50.7%

---

Table 3 Phosphorus recovery by precipitation of struvite or potassium struvite directly in wastewater or in spent regenerant from HAIX-Fe adsorption

	Direct precipitation			Precipitation in spent regenerant
	FU	HU	ADF	FU	HU	HU	ADF
Fresh urine (FU). Hydrolyzed urine (HU). Anaerobic digester filtrate (ADF).a Total species.b NaOH/HCl added to reach pH 9.3.c Magnesium added to achieve Mg : P molar ratio of 1.5 : 1.d Potassium added to achieve K : P molar ratio of 1.5 : 1.e Ammonia added to achieve N : P molar ratio of 1.5 : 1.f Calculated from difference in aqueous concentrations.g Potassium struvite (MPP; MgKPO4·6H2O); struvite (MAP; MgNH4PO4·6H2O); halite (H).
Initial PO4^3−,^a mg P L^−1	674	487	78.2	783	652	652	693
Initial pH	5.8	9.3	7.0	12.8	12.8	12.8	13.0
1 M NaOH added,^b mL L^−1	28	—	36	—	—	—	—
12.1 M HCl added,^b mL L^−1	—	—	—	54	54	52	52
MgCl2·6H2O added,^c g g^−1 P	8	8.8	4.5	10	8.6	8.4	8.3
KCl added,^d g g^−1 P	—	—	—	3.7	3.1	—	—
NH4Cl added,^e g g^−1 P	—	—	—	—	—	2.2	2.2
Final PO4^3−,^a mg P L^−1	15.0	1.37	0.510	16.9	21.4	5.26	1.18
P removed,^f %	97.8%	99.7%	99.3%	97.8%	96.7%	99.2%	99.8%
P recovered,^f g L^−1	0.66	0.49	0.078	0.77	0.63	0.65	0.69
Solids collected, g L^−1	5.3	3.7	1.1	6.2	5.2	5.1	6.0
P in solid, % (m/m)	12.5	13.2	7.1	12.4	12.1	12.7	11.5
Mineral identified^g	MPP	MAP	MAP	MPP, H	MPP, H	MAP	MAP, H

---

Analytical methods

All stock solutions and synthetic urine were prepared using DI water and ACS reagent grade purity chemicals. Phosphate (as P) was measured following Standard Method 4500-P using a Hitachi U-2900 spectrophotometer at 880 nm and a 1 cm quartz cuvette.⁴⁶ Standard calibration checks and matrix spikes were performed to assess the accuracy of measurements; the relative difference between measured and known concentrations was <5% for standard calibration checks and matrix spike recoveries were 94–106%. The pH of each sample was measured with an Accumet AB 15 pH meter and Accumet combination pH/temperature electrode (Fisher Scientific), which was calibrated before use with pH 4, 7, and 10 buffer solutions (Fisher Scientific). Dissolved organic carbon (DOC) was measured using a Shimadazu TOC-V_CPH total organic carbon analyzer equipped with an ASI-V autosampler.⁴⁷ Inorganic anions and cations, excluding PO₄, were measured using a Dionex ICS-3000 ion chromatograph as described elsewhere.⁴⁸ X-ray diffraction of solid precipitates was performed with a computer-controlled Rigaku Ultima IV X-ray powder diffractometer with a stepping motor and graphite crystal monochrometer as described elsewhere.⁴⁹

Results and discussion

Phosphate adsorption column tests and regeneration

Fig. 1 shows the effluent concentration of PO₄ divided by the initial concentration of PO₄ as a function of the number of BVs treated by HAIX-Fe resin. The trends for increasing number of BVs of wastewater treated by HAIX-Fe resin until breakthrough (i.e., C/C₀ = 0.1) were hydrolyzed urine < fresh urine ≪ anaerobic digester filtrate and third use < second use ≤ first use (see Table S1 in ESI†). The trends for increasing number of BVs of wastewater treated by HAIX-Fe resin until saturation (i.e., C/C₀ = 1) were fresh urine < hydrolyzed urine ≪ anaerobic digester filtrate with no clear trend related to first, second, or third use (see Table S1 in ESI†). The results in Fig. 1, especially for anaerobic digester filtrate, share the greatest similarity with previous results on HAIX-Fe resin treatment of real and synthetic sludge liquor (472 and 324 mg P L⁻¹, respectively) in which saturation of HAIX-Fe resin using real and synthetic sludge liquor occurred at 31 and 120 BVs, respectively.³² All other previous column studies using HAIX-Fe resin has been conducted using wastewater with low PO₄ concentration (≤12 mg P L⁻¹) and low ionic strength (<1 × 10⁻³ M). Therefore, the results in Fig. 1 add new data to the limited data available on the continuous flow operation of HAIX-Fe resin in high PO₄ concentration (>400 mg P L⁻¹) and high ionic strength (>1 × 10⁻¹ M) wastewater.

Fig. 1 Phosphate adsorption to HAIX-Fe resin under continuous-flow column operation: (a) synthetic fresh urine (C₀ = 672 mg P L⁻¹, pH = 5.8), (b) synthetic hydrolyzed urine (C₀ = 491 mg P L⁻¹, pH = 9.3), and (c) real anaerobic digester filtrate (C₀ = 77 mg P L⁻¹, pH = 7.0). Column operating conditions: BV = 8 mL; EBCT = 3.2 min; SLV = 3.1 cm min⁻¹.

The pH and ionic strength of the wastewaters appears to explain some of the trends for number of BVs treated by HAIX-Fe resin until breakthrough and saturation (see Fig. S3 in ESI†). For instance, the lower number of BVs treated by HAIX-Fe resin in hydrolyzed urine than the other wastewaters was due to the higher pH and higher ionic strength of hydrolyzed urine than the other wastewaters. As the pH of the solution increases, the speciation of the resin-phase HFO shifts from positively charged (≡FeOH₂⁺) to neutral (≡FeOH) to negatively charged (≡FeO⁻), and the speciation of PO₄ shifts from monovalent to divalent (i.e., H₂PO₄⁻ ↔ HPO₄²⁻ + H⁺, pK_a = 7.20, I = 0 M).⁵⁰ Moreover, the charged HFO species become more prevalent as the ionic strength increases.⁵⁰ Hence, it is speculated that in hydrolyzed urine unfavorable electrostatic interactions occur between HFO and PO₄, whereas in fresh urine and anaerobic digester filtrate more favorable electrostatic interactions occur between HFO and PO₄. This is supported by previous research in which the maximum PO₄ adsorption to HAIX-Fe resin in secondary wastewater effluent was at pH 7–7.5.^26,27

Fig. 2 shows the effluent concentration of PO₄ as a function of the BVs of regeneration solution that passed through the HAIX-Fe resin bed. Phosphate desorption from HAIX-Fe resin showed very similar trends among fresh urine, hydrolyzed urine, and anaerobic digester filtrate, e.g., regeneration was complete within 5 BVs for the first, second, and third regeneration. This similarity was because the composition of the regeneration solution was initially the same for all three wastewaters. The concentration of PO₄ in the regeneration solution increased with repeated use of the regeneration solution (see Table S2 in ESI†), and this increase was most substantial for anaerobic digester filtrate. The concentration of PO₄ in the final, combined regeneration solution was 263 mg P L⁻¹ after the first regeneration, 502 mg P L⁻¹ after the second regeneration, and 693 mg P L⁻¹ after the third regeneration, which corresponded to 9.0× higher PO₄ concentration in spent regenerant than initial anaerobic digester filtrate. Fresh urine and hydrolyzed urine had 1.2× and 1.3× higher PO₄ concentration in spent regenerant than initial urine. Although not measured in this work, HAIX-Fe resin is able to bind sulfate to the strong-base anion exchange sites in the resin. For context, the phosphate–sulfate separation factor for HAIX-Fe resin was previously determined to be 46.²⁶ Although HAIX-Fe resin has a higher selectivity for PO₄ than sulfate, given the high concentration of sulfate in urine, HAIX-Fe resin likely accumulated sulfate and thereby the regeneration solution likely accumulated sulfate to some extent. The sulfate concentration in the regeneration solution was not measured, which is a limitation of the experimental results. Desorption of PO₄ and sulfate from HAIX-Fe resin during regeneration was reported by Sengupta and Pandit.²⁸ As a result, the accumulation of sulfate in the regeneration solution is expected to lead to decreased regeneration efficiency and possibly decrease the purity of the recovered PO₄ solid.

Fig. 2 Phosphate desorption from HAIX-Fe resin during column regeneration corresponding to adsorption tests using (a) synthetic fresh urine, (b) synthetic hydrolyzed urine, and (c) real anaerobic digester filtrate. Column regeneration conditions: 2.5% NaCl + 2% NaOH regeneration solution; BV = 8 mL; EBCT = 10 min; SLV = 1.0 cm min⁻¹.

It is possible that even higher PO₄ concentrations could be achieved in the spent regenerant if the regeneration volume was reduced from 9 to 5 BVs. However, the accumulation of PO₄ in the regeneration solution has the potential to decrease the regeneration efficiency as illustrated in Table 2. The regeneration efficiency decreased from 84–95% for the first regeneration to 17–51% for the third regeneration. It is not known why the first regeneration of HAIX-Fe resin saturated with PO₄ from hydrolyzed urine was 11 percentage points lower than the corresponding regeneration for fresh urine and anaerobic digester filtrate. Based on reported data (i.e., 90–98% regeneration efficiency of HAIX-Fe resin saturated with PO₄ from a variety of wastewaters excluding urine),^28,31,32 95% regeneration efficiency for the first regeneration is a more likely value than 84% regeneration efficiency. Overall, with each adsorption–regeneration cycle an increasing amount of PO₄ was retained (i.e., fouled) on HAIX-Fe resin, which resulted in a lower amount of PO₄ adsorbed in the subsequent adsorption step. HAIX-Fe resin saturated with PO₄ from fresh urine exhibited greater fouling than HAIX-Fe resin saturated with PO₄ from hydrolyzed urine or anaerobic digester filtrate because of the higher PO₄ concentration in the recycled regeneration solution. Fig. S4 in ESI† shows that the regeneration efficiency decreased linearly with increasing PO₄ concentration in the regeneration solution. As an alternative approach, Sengupta and Pandit were able to achieve 93% regeneration efficiency using recycled regeneration solution by precipitating PO₄ in the regeneration solution and then adding 1.5% NaOH to compensate for the loss of hydroxide during regeneration.²⁸

The maximum PO₄ loading on HAIX-Fe resin during the column tests was achieved using fresh urine, and increased from 10.2 mg P g⁻¹ (first use, adsorbed) to 13.4 mg P g⁻¹ (third use, total). The maximum PO₄ loading observed in this study was less than the maximum PO₄ loading reported by Sengupta and Pandit (23 mg P g⁻¹),²⁸ but greater than the PO₄ loading reported by Martin et al. (7.7 mg P g⁻¹).³⁰ The calculated PO₄ loading on HAIX-Fe resin using the Freundlich isotherm model (and model parameters corresponding to fresh urine, hydrolyzed urine, and anaerobic digester supernatant²¹) was in relatively good agreement (1–21% relative difference) with the experimental PO₄ loading on HAIX-Fe resin from the column tests (Table 2) and better than the Langmuir isotherm model (results not shown). Thus, batch equilibrium experiments and Freundlich isotherm modeling can predict the capacity of HAIX-Fe resin for PO₄ under continuous-flow column operation, and following the adsorption step, fresh regeneration solution can desorb up to 95% of PO₄ making it available for P recovery via precipitation.

Chemical requirements for phosphate precipitation

All chemical additions required for precipitation of struvite and potassium struvite in wastewaters and spent regenerants are listed in Table 3. Fresh urine and anaerobic digester filtrate required the addition of strong base to increase the pH to 9.3. The anaerobic digester filtrate required more NaOH than fresh urine because it had a higher buffer intensity than fresh urine due to the higher concentration of ammonia, carbonate, and DOC. Hydrolyzed urine did not require pH adjustment due to the increase in pH that occurs during urea hydrolysis, which yields a final pH of approximately 9.⁴¹

All wastewaters required the addition of magnesium to achieve an Mg : P molar ratio of 1.5 : 1. The anaerobic digester filtrate required the least amount of magnesium because the initial Mg : P molar ratio was 0.8 : 1 due to the initial magnesium present, whereas the hydrolyzed urine required the greatest amount of magnesium because magnesium was not initially present. No additional chemicals were required because fresh urine contained a stoichiometric excess of potassium for potassium struvite precipitation, and hydrolyzed urine and anaerobic digester filtrate contained a stoichiometric excess of ammonia for struvite precipitation. The chemical additions described herein are specific to the waste streams investigated and could vary for other wastewaters.

Although the comparison of saturation indexes for various solid phases does not allow for the determination of the least soluble phase, it is a useful approach for defining the universe of solid phases that could precipitate. Chemical equilibrium calculations (not allowing solid phases to precipitate) showed that potassium struvite, as well as nine other calcium and magnesium phosphate solid phases, were supersaturated in fresh urine at an Mg : K : P molar ratio of 1.5 : 2 : 1 and pH 9.3 (see Table S3 in ESI†). The presence of calcium in fresh urine increased the number of solid phases that were supersaturated. In hydrolyzed urine at an Mg : N : P molar ratio of 1.5 : 37 : 1 and pH 9.3, chemical equilibrium calculations showed that struvite, potassium struvite, magnesium carbonate, and magnesium phosphate solid phases were supersaturated (see Table S4 in ESI†). Of the supersaturated species, the solids that precipitate will depend on the nucleation kinetics. For example, previous precipitation studies investigating the NH₄H₂PO₄–CaCl₂–MgCl₂–H₂O system showed that brushite (CaHPO₄·2H₂O), monetite (CaHPO₄), amorphous calcium phosphate (Ca₃(PO₄)₂·xH₂O), newberyite (MgHPO₄·3H₂O), and struvite precipitated first, and that brushite, monetite, newberyite, and struvite were present as the final phases depending on concentration and pH.^51,52 The results in Tables S3 and S4† show that struvite had a higher SI in hydrolyzed urine (SI = 3.97) than potassium struvite in fresh urine (SI = 1.02). This was because struvite has a lower K_s0 than potassium struvite,⁴⁵ and the IAP of struvite in hydrolyzed urine was greater than the IAP of potassium struvite in fresh urine. Previous data are not available to compare the nucleation kinetics of struvite with potassium struvite; however, previous results for struvite show that crystallization induction time decreases with increasing supersaturation.⁵³ The amount and purity of the solids collected during the precipitation experiments, and the potential for co-precipitation of multiple solid phases are discussed in the following subsections.

The precipitation conditions in the spent regenerants were designed to be similar to the conditions in urine and filtrate, i.e., same pH and same Mg : P molar ratio. All of the spent regenerants required similar additions of strong acid to decrease the pH from 13 to 9.3 to avoid the precipitation of magnesium hydroxide at pH 13. The large amount of HCl required by the spent regenerants was due to the high buffer intensity of hydroxide at pH 13. The spent regenerants required the addition of two other chemicals: magnesium and potassium for precipitation of potassium struvite, or magnesium and ammonia for precipitation of struvite. The amount of magnesium, potassium, and ammonia required for precipitation was determined by the PO₄ concentration because none of these chemicals were initially present in the regeneration solution.

In the spent regenerant corresponding to fresh urine (Mg : K : P = 1.5 : 1.5 : 1 and pH 9.3), chemical equilibrium calculations showed that potassium struvite, Mg₃(PO₄)₂, and newberyite (MgHPO₄·3H₂O) were supersaturated (see Table S5 in ESI†). The SI for potassium struvite was lower in the spent regenerant (SI = 0.8) than in fresh urine (SI = 1.02) because the IAP was smaller in the spent regenerant than fresh urine. Because of the absence of calcium in the spent regenerant, only one solid phase had a higher SI than potassium struvite in the spent regenerant (i.e., Mg₃(PO₄)₂), whereas six other solid phases had a higher SI than potassium struvite in fresh urine (see Table S3 in ESI†). In the spent regenerant corresponding to hydrolyzed urine (Mg : N : P = 1.5 : 1.5 : 1 and pH 9.3), chemical equilibrium calculations showed that struvite, Mg₃(PO₄)₂, and newberyite were supersaturated (see Table S6 in ESI†). Ma et al. investigated struvite precipitation in MgCl₂–(NH₄)₂HPO₄–NaCl–H₂O system, which contained the same components as spent regenerant corresponding to hydrolyzed urine, and calculated that struvite (SI = 3.05), Mg₃(PO₄)₂ (SI = 3.98), and newberyite (SI = 0.74) were supersaturated at pH 9.¹² The SI for struvite was lower in the spent regenerant (SI = 3.08) than hydrolyzed urine (SI = 3.97) because the IAP was smaller in the spent regenerant than hydrolyzed urine. Struvite had the second highest SI in the spent regenerant and hydrolyzed urine; however, the solid with the highest SI changed from Mg₅(CO₃)₄(OH)₂·4H₂O in hydrolyzed urine to Mg₃(PO₄)₂ in spent regenerant. In comparing the PO₄ precipitation potential in urine and spent regenerant, fewer solid phases were supersaturated in the spent regenerant than urine.

Phosphate recovery by precipitation

The results on PO₄ recovery by direct precipitation and adsorption–precipitation are listed in Table 3. Struvite precipitation in hydrolyzed urine and anaerobic digester filtrate resulted in >99% PO₄ removal from solution, and chemical equilibrium calculations (setting struvite allowed to precipitate) predicted that essentially 100% of the PO₄ would precipitate as struvite in hydrolyzed urine. Additionally, the experimental results agree with previous literature, e.g., 99% PO₄ removal by precipitation of struvite in synthetic urine¹⁴ and 95% PO₄ removal by precipitation of struvite in real anaerobic digester supernatant.⁵⁴

Potassium struvite precipitation in fresh urine resulted in 98% PO₄ removal from solution. However, chemical equilibrium calculations (setting potassium struvite allowed to precipitate) predicted that only 75% of the PO₄ would precipitate as potassium struvite in fresh urine. Previous reports on PO₄ removal by potassium struvite precipitation differ, e.g., approximately 75% PO₄ removal by precipitation of potassium struvite in synthetic urine¹⁴ and 98% PO₄ removal by precipitation of potassium struvite and magnesium sodium phosphate in synthetic urine.⁵⁵ Thus, the high PO₄ removal observed for potassium struvite precipitation in this work is likely due to co-precipitation of potassium struvite and other phosphate-containing solid phases as discussed later.

Struvite precipitation in the spent regenerants corresponding to hydrolyzed urine and anaerobic digester filtrate resulted in >99% PO₄ removal from solution. Additionally, chemical equilibrium calculations (setting struvite allowed to precipitate) showed >99% PO₄ precipitation as struvite in spent regenerant. Previous reports show 99.9% PO₄ removal by struvite precipitation in synthetic spent regenerant³¹ and >90% PO₄ removal by struvite precipitation in real spent regenerant.²⁸ Hence, the experimental results, chemical equilibrium calculations, and previous literature all show high PO₄ removal by struvite precipitation in spent regenerant.

Potassium struvite precipitation in the spent regenerants corresponding to fresh urine and hydrolyzed urine showed 97–98% PO₄ removal from solution. However, chemical equilibrium calculations (setting potassium struvite allowed to precipitate) predicted only 55% PO₄ precipitation as potassium struvite in spent regenerant. The experimental results and chemical equilibrium calculations suggest that additional solid phases precipitated with potassium struvite, which is discussed in the next subsection. There are no published data on potassium struvite precipitation in spent regenerant to compare with, so the results of this work fill a gap in the literature on potassium struvite precipitation and illustrate the need for additional experimental work.

Identity and purity of precipitated solids

The percent P in the precipitated solid gives an indication of its mineral identity and purity, e.g., pure struvite is 12.62% (m/m) P and pure potassium struvite is 11.62% (m/m) P. Based on percent P, the solid precipitated in the spent regenerant corresponding to hydrolyzed urine showed the best agreement with pure struvite followed by the solid precipitated in hydrolyzed urine (see Table 3). The slightly higher P content in the solid precipitated in hydrolyzed urine than pure struvite suggests the presence of struvite and other magnesium phosphate solid phases. For example, newberyite (17.77% (m/m) P), bobierrite (Mg₃(PO₄)₂·8H₂O, 15.22% (m/m) P), and trimagnesium phosphate (Mg₃(PO₄)₂·22H₂O, 9.40% (m/m) P) are known to precipitate with struvite in solutions resembling urine.^12,45,52 Additionally, previous reports show that the percent P of struvite precipitated in urine can vary from 7.2–18%.⁵⁶ Based on XRD results, the solid precipitated in hydrolyzed urine matched well with the major d-spacings and relative intensities of pure struvite, whereas the solid precipitated in the spent regenerant corresponding to hydrolyzed urine matched the major d-spacings for pure struvite but the relative intensities deviated from pure struvite possibly due to orientation effects (see Fig. S4 in ESI†).^49,57

The solid precipitated in fresh urine and the spent regenerants corresponding to fresh urine and hydrolyzed urine showed higher P content than pure potassium struvite (see Table 3). This suggests that other calcium and magnesium phosphate solid phases precipitated with potassium struvite. For example, brushite (CaHPO₄·2H₂O, 18.00% (m/m) P), monetite (CaHPO₄, 22.77% (m/m) P), and hydroxyapatite (Ca₅(PO₄)₃OH), 18.50% (m/m) P) are known to precipitate in solutions resembling urine.^51,52 Additionally, trimagnesium phosphates (Mg₃(PO₄)₂·8H₂O, 15.22% (m/m) P; Mg₃(PO₄)₂·22H₂O, 9.40% (m/m) P) and magnesium sodium phosphate (MgNaPO₄·7H₂O, 11.54% (m/m) P) are known to precipitate with potassium struvite.^45,55 Furthermore, previous work investigating the co-precipitation of potassium struvite and magnesium sodium phosphate in synthetic urine reported 98% PO₄ removal,⁵⁵ which is in excellent agreement with 97–98% PO₄ removal by potassium struvite precipitation in this work. Based on XRD results, the solid precipitated in fresh urine and spent regenerants corresponding to fresh and hydrolyzed urine matched relatively well with the major d-spacings and relative intensity of pure potassium struvite (see Fig. S6 in ESI†).⁵⁷ In addition, the XRD results suggested the presence of halite in the solid collected from the spent regenerants even though the chemical equilibrium calculations showed halite was undersaturated. The XRD results were not quantitative in terms providing the amount of halite.

The solid precipitated in the anaerobic digester filtrate contained substantially lower P content than pure struvite (i.e., 5.5 percentage points lower), and the solid precipitated in the spent regenerant corresponding to anaerobic digester filtrate contained 1.1 percentage points lower P than pure struvite. The solid formed by direct precipitation in anaerobic digester filtrate was likely a combination of struvite, amorphous calcium phosphate (based on the Mg : Ca molar ratio of 1 : 2.1 in the filtrate¹⁶), and other unknown solids phases. Based on previous reports, the percent P of struvite precipitated in anaerobic digester sidestreams can vary from low to close match with pure struvite.^7,58 The XRD results for the solid precipitated in anaerobic digester filtrate matched most of the major d-spacings for pure struvite but the relative intensities deviated from pure struvite (see Fig. S5 in ESI†).⁵⁷ The XRD results for the solid precipitated in the spent regenerant corresponding to anaerobic digester filtrate showed a better match with the major d-spacings and relative intensities for pure struvite than the solid from direct precipitation. The XRD results for the solid precipitated in the spent regenerant corresponding to anaerobic digester filtrate also showed the presence of halite, which could account for the lower percent P than the solid precipitated in the spent regenerant corresponding to hydrolyzed urine in which halite was not detected.

Environmental implications

Phosphate adsorption to HAIX-Fe resin and desorption using caustic brine was more effective for anaerobic digester filtrate than fresh and hydrolyzed urine. Considering the waste streams investigated, the order of increasing chemical requirements for precipitation of struvite and potassium struvite was hydrolyzed urine < fresh urine ≈ anaerobic digester filtrate < spent regenerants. Based on the experimentally determined P content of the precipitated solids, the purity of struvite and potassium struvite precipitated in the spent regenerants was equal to or higher than the purity of the same mineral precipitated in urine and filtrate. Thus, direct precipitation of struvite in hydrolyzed urine is more efficient in terms of lower chemical requirements and similar mineral purity than adsorption–precipitation. However, adsorption–precipitation has potential benefits for PO₄ recovery from anaerobic digester filtrate and fresh urine as described below. Although outside the scope of this work, two possible scenarios in which PO₄ recovery via adsorption–precipitation could be more favorable than direct precipitation in source-separated urine are (1) to minimize contamination of the final mineral product^19,20 and (2) to improve process automation and control.^11,59 For example, the well-defined composition of the spent regenerant could allow for measurements such as pH, conductivity, and turbidity to control the precipitation process, whereas this type of monitoring and automation is challenging in urine because of its varying composition. Based on the data generated in this work, these scenarios are speculation; however, the scenarios represent the next logical step in the continuum of research on PO₄ recovery by adsorption–precipitation.

The adsorption–precipitation process is considered favorable for PO₄ recovery as struvite from anaerobic digester filtrate considering that direct precipitation produced a lower purity solid in this study and previous studies. If, however, the goal is to remove PO₄ to prevent fouling due to precipitation during sludge handling processes at the wastewater treatment plant then direct precipitation would be favored over adsorption–precipitation due to lower chemical addition requirements. For example, Lew et al. found that struvite precipitation of belt press filtrate from an anaerobic digester can be a solution to prevent clogging due to precipitation in pumps and pipes during sludge dewatering processes, but produces precipitates that could not be used for land application in agriculture due to impurities such as metals and low nutrient content.⁵⁸

The adsorption–precipitation process also has the potential to be more favorable than direct precipitation for PO₄ recovery as potassium struvite from fresh urine, especially if potassium struvite precipitation can be achieved at pH 13. Struvite precipitation is much less favorable at pH > 11 than pH 9–10 because of the shift in ammonia from NH₄⁺ to NH₃, and the subsequent decrease in the activity of NH₄⁺ and the corresponding decrease in the IAP and SI of struvite.¹² In contrast to struvite, Xu et al. showed nearly 100% PO₄ recovery as potassium struvite over the pH range 10–12.51.⁵⁵ Hence, if potassium struvite could be precipitated in spent regenerant with minimal pH adjustment, it would substantially decrease chemical requirements for the adsorption–precipitation process. To make the spent regenerant more conducive to potassium struvite precipitation, the regeneration solution could be prepared using KCl in place of NaCl so that magnesium would be the only external chemical needed.

Acknowledgements

This publication is based upon work supported by NSF CAREER grant number CBET-1150790. Any opinions, findings, conclusions or recommendations expressed in this publication are those of the authors and do not necessarily reflect the views of the National Science Foundation. The authors thank Willie Harris (University of Florida) for XRD analysis and Daniel Yeh (University of South Florida) for assistance obtaining anaerobic digester filtrate. This manuscript was improved by the comments of two anonymous reviewers.

References

1. J. S. Guest, S. J. Skerlos, J. L. Barnard, M. B. Beck, G. T. Daigger, H. Hilger, S. J. Jackson, K. Karvazy, L. Kelly, L. Macpherson, J. R. Mihelcic, A. Promanik, L. Raskin, M. C. M. Van Loosdrecht, D. Yeh and N. G. Love, A New Planning and Design Paradigm to Achieve Sustainable Resource Recovery from Wastewater, Environ. Sci. Technol., 2009, 43(16), 6126–6130.
2. D. P. Van Vuuren, A. F. Bouwman and A. H. W. Beusen, Phosphorus demand for the 1970–2100 period: A scenario analysis of resource depletion, NATO ASI Ser., Ser. I, 2010, 20(3), 428–439.
3. D. Cordell, J.-O. Drangert and S. White, The story of phosphorus: Global food security and food for thought, NATO ASI Ser., Ser. I, 2009, 19(2), 292–305.
4. R. W. Scholz, A. E. Ulrich, M. Eilittä and A. Roy, Sustainable use of phosphorus: A finite resource, Sci. Total Environ., 2013, 461–462(0), 799–803.
5. R. O. Carey and K. W. Migliaccio, Contribution of Wastewater Treatment Plant Effluents to Nutrient Dynamics in Aquatic Systems: A Review, Environ. Manage., 2009, 44(2), 205–217.
6. J. D. Doyle, K. Oldring, J. Churchley and S. A. Parsons, Struvite formation and the fouling propensity of different materials, Water Res., 2002, 36(16), 3971–3978.
7. E. V. Munch and K. Barr, Controlled struvite crystallisation for removing phosphorus from anaerobic digester sidestreams, Water Res., 2001, 35(1), 151–159.
8. J. A. Wilsenach and M. C. M. van Loosdrecht, Integration of processes to treat wastewater and source-separated urine, J. Environ. Eng., 2006, 132(3), 331–341.
9. L. E. de-Bashan and Y. Bashan, Recent advances in removing phosphorus from wastewater and its future use as fertilizer (1997–2003), Water Res., 2004, 38(19), 4222–4246.
10. K. S. Le Corre, E. Valsami-Jones, P. Hobbs and S. A. Parsons, Phosphorus Recovery from Wastewater by Struvite Crystallization: A Review, Crit. Rev. Environ. Sci. Technol., 2009, 39(6), 433–477.
11. B. Etter, E. Tilley, R. Khadka and K. M. Udert, Low-cost struvite production using source-separated urine in Nepal, Water Res., 2011, 45(2), 852–862.
12. N. Ma, A. A. Rouff and B. L. Phillips, A 31P NMR and TG/DSC-FTIR Investigation of the Influence of Initial pH on Phosphorus Recovery as Struvite, ACS Sustainable Chem. Eng., 2014, 2(4), 816–822.
13. K. Xu, C. Wang, X. Wang and Y. Qian, Laboratory experiments on simultaneous removal of K and P from synthetic and real urine for nutrient recycle by crystallization of magnesium–potassium–phosphate–hexahydrate in a draft tube and baffle reactor, Chemosphere, 2012, 88(2), 219–223.
14. J. A. Wilsenach, C. A. H. Schuurbiers and M. C. M. van Loosdrecht, Phosphate and potassium recovery from source separated urine through struvite precipitation, Water Res., 2007, 41(2), 458–466.
15. A. E. Johnston and I. R. Richards, Effectiveness of different precipitated phosphates as phosphorus sources for plants, Soil Use Manage., 2003, 19(1), 45–49.
16. K. S. Le Corre, E. Valsami-Jones, P. Hobbs and S. A. Parsons, Impact of calcium on struvite crystal size, shape and purity, J. Cryst. Growth, 2005, 283(3–4), 514–522.
17. X. D. Hao, C. C. Wang, L. Lan and M. C. M. van Loosdrecht, Struvite formation, analytical methods and effects of pH and Ca2+, Water Sci. Technol., 2008, 58(8), 1687–1692.
18. S. R. Sakthivel, E. Tilley and K. M. Udert, Wood ash as a magnesium source for phosphorus recovery from source-separated urine, Sci. Total Environ., 2012, 419(0), 68–75.
19. A. A. Rouff, Sorption of Chromium with Struvite During Phosphorus Recovery, Environ. Sci. Technol., 2012, 46(22), 12493–12501.
20. L. Decrey, K. M. Udert, E. Tilley, B. M. Pecson and T. Kohn, Fate of the pathogen indicators phage ΦX174 and Ascaris suum eggs during the production of struvite fertilizer from source-separated urine, Water Res., 2011, 45(16), 4960–4972.
21. J. A. O'Neal and T. H. Boyer, Phosphate recovery using hybrid anion exchange: Applications to source-separated urine and combined wastewater streams, Water Res., 2013, 47(14), 5003–5017.
22. M. R. Awual and A. Jyo, Assessing of phosphorus removal by polymeric anion exchangers, Desalination, 2011, 281(0), 111–117.
23. P. L. Sibrell, G. A. Montgomery, K. L. Ritenour and T. W. Tucker, Removal of phosphorus from agricultural wastewaters using adsorption media prepared from acid mine drainage sludge, Water Res., 2009, 43(8), 2240–2250.
24. X. P. Zhu and A. Jyo, Column-mode phosphate removal by a novel highly selective adsorbent, Water Res., 2005, 39(11), 2301–2308.
25. D. Zhao and A. K. SenGupta, Ligand Separation with a Copper(II)-Loaded Polymeric Ligand Exchanger, Ind. Eng. Chem. Res., 2000, 39, 455–462.
26. L. M. Blaney, S. Cinar and A. K. SenGupta, Hybrid anion exchanger for trace phosphate removal from water and wastewater, Water Res., 2007, 41(7), 1603–1613.
27. B. J. Pan, J. Wu, B. C. Pan, L. Lv, W. M. Zhang, L. L. Xiao, X. S. Wang, X. C. Tao and S. R. Zheng, Development of polymer-based nanosized hydrated ferric oxides (HFOs) for enhanced phosphate removal from waste effluents, Water Res., 2009, 43(17), 4421–4429.
28. S. Sengupta and A. Pandit, Selective removal of phosphorus from wastewater combined with its recovery as a solid-phase fertilizer, Water Res., 2011, 45(11), 3318–3330.
29. W. Stumm and J. J. Morgan, Aquatic Chemistry, John Wiley & Sons, Inc., New York, 1996.
30. B. D. Martin, S. A. Parsons and B. Jefferson, Removal and recovery of phosphate from municipal wastewaters using a polymeric anion exchanger bound with hydrated ferric oxide nanoparticles, Water Sci. Technol., 2009, 60(10), 2637–2645.
31. M. Kumar, M. Badruzzaman, S. Adham and J. Oppenheimer, Beneficial phosphate recovery from reverse osmosis (RO) concentrate of an integrated membrane system using polymeric ligand exchanger (PLE), Water Res., 2007, 41(10), 2211–2219.
32. A. Bottini and L. Rizzo, Phosphorus Recovery from Urban Wastewater Treatment Plant Sludge Liquor by Ion Exchange, Sep. Sci. Technol., 2012, 47(4), 613–620.
33. A. Sendrowski and T. H. Boyer, Phosphate removal from urine using hybrid anion exchange resin, Desalination, 2013, 322(0), 104–112.
34. T. Tervahauta, T. Hoang, L. Hernández, G. Zeeman and C. Buisman, Prospects of Source-Separation-Based Sanitation Concepts: A Model-Based Study, Water, 2013, 5(3), 1006–1035.
35. R. Rajagopal, J. W. Lim, Y. Mao, C.-L. Chen and J.-Y. Wang, Anaerobic co-digestion of source segregated brown water (feces-without-urine) and food waste: For Singapore context, Sci. Total Environ., 2013, 443(0), 877–886.
36. M. Ronteltap, M. Maurer and W. Gujer, Struvite precipitation thermodynamics in source-separated urine, Water Res., 2007, 41(5), 977–984.
37. K. M. Udert and M. Wächter, Complete nutrient recovery from source-separated urine by nitrification and distillation, Water Res., 2012, 46(2), 453–464.
38. B.-B. Lind, Z. Ban and S. Bydén, Nutrient recovery from human urine by struvite crystallization with ammonia adsorption on zeolite and wollastonite, Bioresour. Technol., 2000, 73(2), 169–174.
39. B. Beler-Baykal, N. P. Kocaturk, A. D. Allar and B. Sari, The effect of initial loading on the removal of ammonium and potassium from source-separated human urine via clinoptilolite, Water Sci. Technol., 2009, 60(10), 2515–2520.
40. H. L. Mobley and R. P. Hausinger, Microbial ureases: significance, regulation, and molecular characterization, Microbiol. Rev., 1989, 53(1), 85–108.
41. K. M. Udert, T. A. Larsen, M. Biebow and W. Gujer, Urea hydrolysis and precipitation dynamics in a urine-collecting system, Water Res., 2003, 37(11), 2571–2582.
42. L.-L. Wei, Q.-L. Zhao, K. Hu, D.-J. Lee, C.-M. Xie and J.-Q. Jiang, Extracellular biological organic matters in sewage sludge during mesophilic digestion at reduced hydraulic retention time, Water Res., 2011, 45(3), 1472–1480.
43. P. Battistoni, A. De Angelis, P. Pavan, M. Prisciandaro and F. Cecchi, Phosphorus removal from a real anaerobic supernatant by struvite crystallization, Water Res., 2001, 35(9), 2167–2178.
44. J. P. Gustafsson, Visual MINTEQ, version 3.0.http://www2.lwr.kth.se/English/OurSoftware/vminteq/ (accessed 24 June 2013).
45. A. W. Taylor, A. W. Frazier and E. L. Gurney, Solubility products of magnesium ammonium and magnesium potassium phosphates, Trans. Faraday Soc., 1963, 59(0), 1580–1584.
46. A. D. Eaton, L. S. Clesceri, E. W. Rice and A. E. Greenberg, Standard Methods for the Examination of Water and Wastewater, APHA, AWWA, WEF, 21st edn, 2005.
47. J. N. Apell and T. H. Boyer, Combined ion exchange treatment for removal of dissolved organic matter and hardness, Water Res., 2010, 44(8), 2419–2430.
48. K. Indarawis and T. H. Boyer, Alkaline Earth Metal Cation Exchange: Effect of Mobile Counterion and Dissolved Organic Matter, Environ. Sci. Technol., 2012, 46(8), 4591–4598.
49. W. G. Harris and G. N. White, X-ray diffraction techniques for soil mineral identification. In Methods of Soil Analysis: Part 5 – Mineralogical Methods, ed. A. L. Ulery and L. R. Drees, Soil Science Society of America, Madison, WI, 2008, pp. 81–115.
50. F. M. M. Morel and J. G. Hering, Principles and Applications of Aquatic Chemistry, John Wiley & Sons, New York, 1993.
51. F. Abbona, H. E. Lundager Madsen and R. Boistelle, The final phases of calcium and magnesium phosphates precipitated from solutions of high to medium concentration, J. Cryst. Growth, 1988, 89(4), 592–602.
52. F. Abbona, H. E. L. Madsen and R. Boistelle, The initial phases of calcium and magnesium phosphates precipitated from solutions of high to medium concentrations, J. Cryst. Growth, 1986, 74(3), 581–590.
53. N. C. Bouropoulos and P. G. Koutsoukos, Spontaneous precipitation of struvite from aqueous solutions, J. Cryst. Growth, 2000, 213(3–4), 381–388.
54. L. Pastor, D. Mangin, J. Ferrer and A. Seco, Struvite formation from the supernatants of an anaerobic digestion pilot plant, Bioresour. Technol., 2010, 101(1), 118–125.
55. K. Xu, C. Wang, H. Liu and Y. Qian, Simultaneous removal of phosphorus and potassium from synthetic urine through the precipitation of magnesium potassium phosphate hexahydrate, Chemosphere, 2011, 84(2), 207–212.
56. S. Antonini, M. A. Arias, T. Eichert and J. Clemens, Greenhouse evaluation and environmental impact assessment of different urine-derived struvite fertilizers as phosphorus sources for plants, Chemosphere, 2012, 89(10), 1202–1210.
57. H. Yang, H. J. Sun and R. T. Downs, Hazenite, KNaMg2(PO4)2·14H2O, a new biologically related phosphate mineral, from Mono Lake, California, U.S.A., Am. Mineral., 2011, 96(4), 675–681.
58. B. Lew, S. Phalah and M. Rebhum, Controlled struvite precipitation from belt press filtrate of anaerobic digester in a CSTR, Environ. Prog. Sustainable Energy, 2011, 30(4), 640–647.
59. A. Triger, J.-S. Pic and C. Cabassud, Determination of struvite crystallization mechanisms in urine using turbidity measurement, Water Res., 2012, 46(18), 6084–6094.

---

Footnote

† Electronic supplementary information (ESI) available: Additional figures and tables as referenced in the text. This material is available via the internet. See DOI: 10.1039/c5ew00009b

---