Interaction of divalent metals with struvite: sorption, reversibility, and implications for mineral recovery from wastes

ABSTRACT Phosphorus (P) recovered from wastewater as struvite (MgNH4PO4·6H2O) can meet high P demands in the agricultural sector by reuse as a P fertiliser. Heavy metals are prevalent in wastewaters and are common fertiliser contaminants, therefore struvite as a sorbent for metals requires evaluation. Struvite sorption experiments were conducted in model solutions with cadmium (Cd), cobalt (Co), copper (Cu), nickel (Ni), lead (Pb), and zinc (Zn) at 1–5 μM concentrations from pH 7–10. The struvite metal loading increased with dissolved metal concentration and pH, ranging from 2 to 493 mg kg−1. Highest loadings were observed for 5 μM Pb, which exceeded the 120 mg kg−1 European Union (EU) struvite fertiliser limit at all pH values. At 5 μM concentrations, Ni and Cd loadings exceeded EU limits of 100 mg kg−1 at pH 10, and 60 mg kg−1 at pH 8–10, respectively. In desorption experiments, 10–85% metal was released after resuspension in metal-free solutions, with a positive correlation between initial loading and amount desorbed. Distortions of the struvite phosphate band, by Fourier transformation infrared (FTIR) spectroscopy, indicated lowered symmetry of phosphate vibrations with metal sorption. X-ray absorption fine structure spectroscopy (XAFS) analysis of pH 9 solids indicated tetrahedral coordination for Cu and Zn, octahedral coordination for Co and Ni, and Pb in 9-fold coordination. Precipitation of Pb-phosphate minerals was a primary mechanism for Pb sorption. The results provide insight into metal contaminant sorption with struvite in wastewaters, and the potential for metal desorption after recovery. GRAPHICAL ABSTRACT


Introduction
Phosphorus (P) is an essential element for all living beings, serving as a source of energy at the cellular level [1], and playing a critical role in cell multiplication, biological activities, and plant growth and sustenance [2]. As a result, modern agriculture is heavily dependent on P fertiliser, the application of which has increased over the last century, to support increased crop production and meet the needs of the rising world population [3]. However, the quantity of fertiliser applied in crop production exceeds its demand with only 16% of the fertiliser input sequestered by the plant [4], and any excess P in the soil eventually washed out. Runoff from fertilised fields and high-density livestock operations are the two major contributors of P to water bodies [5], causing nutrient enrichment or eutrophication. Enrichment results in mass multiplication of algal and other microbial matter in surface waters, causing algal blooms, and leading to anoxic dead zones [6]. This constitutes a net loss of P from agricultural systems, with detrimental environmental impacts. Furthermore, commercial P fertilisers are usually manufactured from phosphate rock, which is a finite and nonrenewable source. Approximately, 80-90% of the phosphate mined today is used in fertiliser production [7], and with current rates of consumption, reserves are projected to become depleted in the next 50-100 years [8].
To mitigate global P demand in the face of limited phosphate reserves, it is necessary to adopt sustainable, non-traditional solutions for P recovery and reuse from wastewater and agricultural runoff streams. Crystallization of struvite (MgNH 4 PO 4 ·6H 2 O) from waste streams has generated interest as a viable economical approach for P fertiliser production. As a fertiliser product, the low solubility of struvite prolongs the duration of nutrient availability [9], eliminating frequent reapplication and spraying of fertiliser [6], and creating sustainable crop production systems [10]. Struvite can therefore be marketed as a slow-release fertiliser compared to other commercial fertilisers [10,11]. The geochemical parameters controlling struvite precipitation such as pH, the molar ratio of aqueous constituents, and saturation index, have been extensively studied [12][13][14][15][16][17][18]. However, wastewater streams targeted for struvite recovery contain various organic and inorganic contaminants in trace and elevated concentrations. These contaminants can in turn be recovered with struvite during and after crystallization [19,20]. Though the impurities in recovered struvite can be lower than that of commercial phosphate fertilisers [21], concentrations can still exceed safe environmental limits, and as a result there are regulatory limits for common contaminants [22].
Heavy metals like cadmium (Cd), chromium (Cr), copper (Cu), iron (Fe), manganese (Mn), nickel (Ni), lead (Pb), and zinc (Zn), are regulated struvite contaminants that are prevalent in wastewater systems. For example, dairy, swine, and greenhouse wastewaters collected to recover struvite contained Zn, Cu, and Ni at concentrations ranging from 1 to 4 µM [23,24]. In raw swine manure, aluminum (Al), Fe, Cu, and Zn concentrations were found to range from 3.1 to113 mg L −1 , in exceedance of the United States Environmental Protection Agency (U.S. EPA) recommendation limits for agricultural use in the long (> 20 years) and short (< 20 years) term scenario [25,26]. The presence of heavy metals can impact struvite recovery, reaction kinetics, and morphology [27,28], with for example increasing Cu, Pb, and Zn concentration in solution reducing the rate of struvite precipitation by ∼65-76% [27]. Metals can also be recovered with struvite both during and after precipitation [20,23,[29][30][31][32][33][34][35]. When precipitated from model solutions, struvite contained ∼13-3100 mg kg −1 Zn [32] and ∼22-3030 mg kg −1 Cr [30], with elevated loadings observed at higher initial concentration in solution. Concentrations of metals in recovered struvite ranged from 64 to 247 mg kg −1 Zn and 12 to 54 mg kg −1 Cu for greenhouse wastewater [23] and 0.42 to 0.43 mg kg −1 Pb and 1.12 to 1.16 mg kg −1 Zn [20] for poultry wastewater. The overall metal content in recovered struvite ranged from ∼3857 to 5884 mg kg −1 for dairy and ∼653 to 1111 mg kg −1 for swine wastewater [24]. In addition to concerns about contaminant content, metal impurities impacted the struvite structure and decomposition pathways, with implications for nutrient release properties [23,36]. Metal association with struvite after mineral precipitation, has not been as widely addressed [23,29,32], but is of importance when considering the role of struvite as a sorbent for dissolved metals remaining in solution [37]. There may also be an influence on the metal binding configuration which can in turn impact mobility. Zinc, when present during struvite precipitation formed octahedrally coordinated complexes however, tetrahedral monodentate surface complexes dominated in the presence of pre-existing struvite [32]. Though metal mobility was not directly evaluated, it was expected that surface complexes may be more susceptible to remobilisation.
The aim of the present research is to provide a complete investigation of the extent of divalent metal interaction with struvite, and the governing geochemical processes in simulated matrices reflective of systems feasible for nutrient recovery. Struvite sorption with Cd, Co, Cu, Ni, Pb, and Zn, a suite of common heavy metals found in agricultural wastewaters, was evaluated. The influence of initial metal concentration and pH on sorption was assessed using simplified electrolytic background solutions to isolate the metal-struvite sorption processes, without interference from the range of organic and inorganic contaminants in true wastewater matrices. As there are no previous metal desorption studies, experiments were also conducted to determine the extent and reversibility of metal sorption, and the role of pH in this process. Recovered solids were evaluated to determine the extent of metal sorption and desorption, the impact on the struvite structure, and the dominant sorption configuration. In the absence of a single comprehensive study that evaluates divalent metal sorption with struvite, this research will provide fundamental information concerning the macroscopic and molecular-scale sorption processes as controlled by pH and dissolved metal concentration. Results will aid the understanding of interactions of divalent metals with struvite, elucidating sorption processes at the mineral-water interface in struvite recovery systems.

Metal sorption
A 0.1 M sodium nitrate (NaNO 3 ) solution was prepared for use as a background electrolyte. This background was chosen as the constituent ions are commonly found in many types of wastewaters [23,30,32], and to be consistent with conditions used in previous studies [23,32]. For pH-dependent sorption experiments, 1.25-2.88 g of commercial struvite (Alfa Aesar, 98% purity, ∼100 µm crystal size, 40 m 2 g −1 surface area, [38]) was added to 1 L of background solution. Struvite was added in excess to account for dissolution and release of magnesium (Mg) and P at the desired pH ( Figure  S1.1) for a final loading of 1 g L −1 . Solution pH was adjusted to pH 7, 8, 9, or 10 (±0.1) using 5 N nitric acid (HNO 3 ) and 0.25 N sodium hydroxide (NaOH). A corresponding set of blank solutions with 0.1 M NaNO 3 background, containing no struvite, were also prepared at pH 7-10. After 72 h of equilibration, solutions were adjusted to the final pH, and spiked with 1 or 5 μM M 2+ (where M = Cd, Co, Cu, Ni, Pb, Zn). The experimental pH of sorption solutions was monitored throughout the experiment ( Figure S1.2). All experiments were performed in duplicate. The speciation of metal ions, and saturation indices of solid phases were calculated using Visual MINTEQ software using the thermo.vdb database [39].
Metal sorption solutions and corresponding blank solutions were reacted for 7 d. Aqueous samples (10 mL) were withdrawn at regular intervals, filtered through a 0.45 µm PTFE syringe filter, acidified with 5 N HNO 3 , and stored for further analysis. While the entire reaction was not performed under continuous agitation, samples were stirred at ∼1000 rpm for 1 min to create a homogenous suspension prior to sample collection. At the end of the 7 d reaction, solids were recovered by filtration using a 0.45 μM nitrocellulose membrane filter, air-dried, and stored for further analysis. Aqueous samples and 2% HNO 3 digested solids were analyzed for metal concentrations by flame atomic absorption spectrometry (AAS, ThermoFisher Scientific ICE 3000-series) and inductively coupled plasma-optical emission spectroscopy (ICP-OES, Agilent 5110 SVDV). Batches of samples analyzed by AAS were reanalysed by ICP-OES to ensure that obtained analytical results were consistent across both analytical instruments.

Batch desorption
Desorption experiments were conducted for solids from 5 μM sorption experiments to determine the lability of the sorbed metal and the reversibility of sorption. Solutions for resuspension were prepared by adding the mass of struvite required for a final 1 g L −1 loading at pH 7-10 to 0.1 M NaNO 3 background solutions, as described for batch sorption. Pre-equilibration of solutions with struvite at the desired pH was done to minimise dissolution of the metal-sorbed struvite upon resuspension. After 72 h equilibration, the solutions were filtered using a 0.45 μm nitrocellulose membrane filter to remove excess undissolved struvite and adjusted to the final pH using 5N HNO 3. Solids for desorption experiments were prepared as in sorption experiments, with 5 μM M 2+ and 1 g L −1 struvite at pH 7-10. After 7 d reaction, the metal-sorbed struvite was recovered from the sorption experiment, re-suspended in the relevant background solution at a loading of 1 g L −1 , and allowed to react for 72 h. All experiments were conducted in duplicate. Aqueous and solid samples were recovered and analyzed as described for sorption experiments.

Spectroscopic analysis
Solids recovered from 5 μM sorption experiments were analyzed by attenuated total reflectance Fourier transform infrared spectroscopy (ATR-FTIR, Perkin Elmer Spectrum 100). Spectra were recorded in the region 650-4000 cm −1 at a resolution of 4 cm −1 . The final spectrum was the average of 20 scans. Select solids collected from batch sorption and desorption experiments were characterised by X-ray absorption fine structure spectroscopy (XAFS). Data was collected at the 6-BMM Beamline at the National Synchrotron Light Source-II (NSLS-II), at Brookhaven National Laboratory (BNL), Upton NY. Data was collected in fluorescence mode using a Si (111) monochromator and a 4-element vortex silicondrift detector with metal filters to reduce the impact of elastic scattering. Multiple scans (4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20) were recorded per sample to improve the signal-to-noise ratio. The IFEFFIT software package was used to perform data reduction, analysis and fitting [40]. Data was collected for all metals, with the exception of Cd as the K-edge energy (26.71 keV) exceeded the maximum permissible energy of the beamline (23 keV).

Metal sorption from solution
The percent metal remaining in solution over the 7 d sorption period is reported for 1 and 5 µM initial metal concentrations ( Figure 1). The metal solution concentration decreases rapidly within the first 72 h, due to sorption to readily available sites, followed by a slower sorption step as sites are filled [41,42]. For all solutions, except for 5 µM Pb, metal sorption increases with pH. The trend in sorption can be compared to the pH-dependent metal speciation in solution as modelled using Visual Minteq ( Figure S2 and S3). At pH 7-8 Ni, Pb, and Co form free metal ions (M 2+ ), and Cd and Cu are speciated as metal hydrogen phosphates (MHPO 4 0 ). At pH > 9 free Cd 2+ , Pb-hydroxide (PbOH + ), and Co, Cu, Ni, and Zn metal-hydroxides (M(OH) 2 0 ) dominate. There is no correlation between modelled metal speciation and the extent of sorption [41], indicating that sorption is not species dependent. Rather, struvite surface charge may influence the extent of uptake. The zeta potential (ζ) of struvite has been measured at −17.5 to −27.6 mV from pH 8.5-10.5 [13,15]. An increase in net negative surface charge with pH facilitates electrostatic interaction between positively charged metal species, promoting surface adsorption. The percent metal associated with the solid after 7 d, as calculated from initial and final aqueous concentrations, is reported as a function of pH ( Figure 2). Except for Pb, the percent metal-sorbed increases with pH, with higher sorption at 5 µM initial concentration, and consistent with the metal removal trends observed in solution. Results are in agreement with previous struvite sorption studies, where sorption of Zn and Cu increased with initial metal concentration [23,32]. Though adsorption may occur at lower initial concentration and pH, at higher values metals are thermodynamically oversaturated with respect to several phases (Table S1). The exception is Pb for which Pb-bearing mineral phases are oversaturated under all studied conditions (Table S1). Precipitation is likely the dominant mechanism contributing to increased metal removal at higher pH and initial concentration.
Divalent metal sorption for 1 μM sorption experiments at pH 7 followed the selectivity sequence Pb > Ni > Cu > Cd > Zn > Co (Figure 2(a)). At this concentration and pH, all metals except for Pb are undersaturated with respect to potential precipitate phases (Table S1), therefore adsorption may be a dominant mechanism. The metal sorption trend is different from the Irving Williams order for divalent metal cations Zn > Cu > Ni > Co [43], but is somewhat consistent with observations for these metals sorbed with phosphate minerals. For hydroxyapatite, metal sorption increased with pH in the order Pb > Cd > Zn, with rapid Pb sorption, whereas other metals reached sorption maxima at ∼24 h [44]. Sorption selectivity was Pb > Cd > Zn for apatite and phosphatic clay [45,46] and Pb > Cu > Zn on phosphate rock [47]. However, for hydroxyapatite at pH 9 Co > Ni sorption with >90% metal-sorbed [48]. Overall consistency in metal sorption trends with struvite compared to other phosphate substrates may be due to similarities in sorption mechanisms. Assuming that metals primarily bind to phosphate groups, the selectivity sequence can be compared to the metal phosphate (M 3 (PO 4 ) 2 ) solubility (K sp ) (Table S2) for which the trend is Pb < Cu < Co < Zn < Cd < Ni. That phosphate solubility does not completely predict the relative affinity of metals for struvite suggests that sorption is not limited to free metal aqueous species and/or binding to phosphate, as expressed in the solubility equations (Table S2).

Metal loadings on recovered struvite
The sorbed metal loadings range from 2 to 49 mg kg −1 and 28 to 493 mg kg −1 for struvite recovered from 1 and 5 µM sorption experiments, respectively (Table 1). Higher sorbed metal with higher initial concentration correlates with the observed trend in metal removal from solution ( Figure 1) and is consistent with previous struvite sorption studies conducted in the same background, but with select metals at a single pH [23,30,32]. The metal loadings were compared to the European Union (EU) Products Regulation limits for inorganic contaminants in struvite ( [22], Table 1). The EU 2019 limits were used as the United States (U.S.) has   (Table S1) may account for the observed solid loadings. The thermodynamic calculations for Pb predict oversaturation of hydroxylpyromorphite (Pb 5 (PO 4 ) 3 OH), lead phosphate (Pb 3 (PO 4 ) 2 (s)), lead hydrogen phosphate (PbHPO 4 (s)) at pH 7-10, and lead hydroxide (Pb(OH) 2 (s)) at pH 8-10 (Table S1). For pH-dependent sorption of Pb on apatite minerals, precipitation of oversaturated Pb-phases such as pyromorphite and Pb oxides were also determined to be a major mechanism of Pb removal from solution [45]. Therefore, consistently higher loadings of Pb in comparison to other metals can be attributed to the precipitation of low solubility Pb minerals. In a study of Zn and Cu sorption at pH 7, Rouff et al. [23] delineated that synergistic co-sorption of the metals occurs, whereby Cu enhances Zn uptake. The Zn concentrations were 142.9 mg kg −1 , increasing to 377 mg kg −1 in the presence of Cu, while Cu sorption was 384 mg kg −1 both in the presence and absence of Zn. These solid loadings are different from those reported here due to higher initial concentrations and varying ratios of Cu and Zn used in the previous study. However, these results indicate that in solutions with more than one metal, initial concentration is not the sole factor impacting metal sorption, with both competitive and synergistic effects between metals ultimately dictating their final concentrations in struvite. Experiments monitoring Cu and Zn removal from greenhouse wastewater found less sorption of both metals (54 and 247 mg kg −1 , respectively) when compared to removal from 0.1 M NaNO 3 model solutions (99 and 431 mg kg −1 , respectively) [23]. This was attributed to the presence of inorganic and organic ligands in in the wastewater that complex the metal, reducing sorption. Therefore, in wastewaters, higher dissolved metal concentrations may not necessarily result in metal loadings exceeding regulatory limits, as the concentrations and types of metals and overall wastewater composition will dictate competitive and synergistic effects between metals, as well as metal complexation.

Sorption reversibility
To determine the reversibility of sorption, metal-sorbed struvite recovered from 5 µM sorption experiments was re-suspended in metal-free pre-equilibrated solutions. These solids were chosen due to higher metal loadings, in some cases exceeding regulatory limits. For all metals, a detectable concentration in solution is observed within the first few hours of resuspension ( Figure S4). Except for Pb, which shows little change in release over time, the desorption curves at all pH values increase up to 24 h, after which the reaction slows, approaching equilibrium ( Figure S4). Compared to sorption (Figure 1), the desorption rate is slower, likely due to higher activation energy required to break metal-sorbate bonds [49][50][51]. In the case of Pb, minimal aqueous concentrations are observed with no impact of time or pH.
The metal loading remaining on the solids after desorption was directly measured and compared to the loading before desorption. The percent metal desorbed ranged from 10% to 85% for all metals (Table S3). Though measured Pb aqueous concentrations were low, 37-84% was desorbed indicating release as secondary precipitates with particles >0.45 µm which are removed from solution by filtration prior to aqueous analysis. Release of metals from struvite is an indication that a significant portion of sorbed and/or secondary metal precipitates are surface-bound, which resolubilize upon resuspension. The correlation between metal loading and the amount of metal desorbed was determined by plotting these values for each pH (Figure 3). In general, the amount of desorption increases with increasing sorption, with highest values of both sorption and desorption observed for Pb and Zn. The order of desorption also varied with pH (Table S4). The entire suite of metals has not been previously studied for other phosphates, however, Peld et al. [52] observed that for apatite desorption of Zn > Cd, whereas for phosphate rock Zn > Cu > Pb [47]. The order of desorption among metals is consistent with what is observed here, with the exception of Pb, which forms precipitates that are subsequently released to solution as particulates, as discussed above. In general, the extent of desorption is likely dependent on the sorption mechanism, with exchange of surface-bound complexes and/or solubility of metal precipitates dictating the process.
Metal desorption is an important phenomenon as this controls the availability of metals associated with struvite and can inform strategies for reuse of the mineral as a fertiliser. For example, metal release, particularly with increasing loadings can introduce harmful contaminants to soil solution and agroecosystems. In this case, metals which are readily desorbed can be more easily removed prior to application. Alternately, metal-doped struvite containing ions like Cu and Zn can serve as a source of slow-release fertiliser supplying micronutrients [37], with the extent of desorption influencing initial metal release to soils.

Fourier transform infrared spectroscopy (FTIR)
The FTIR spectra for metal-sorbed struvite recovered from 5 µM sorption experiments were evaluated and compared to metal-free (blank) struvite equilibrated at the relevant pH ( Figure 4). For unreacted struvite, bands centred at 1000 cm −1 are associated with the ν 3 PO 4 3− anti-symmetric stretching and bands at 1430 and 750 cm −1 are due to NH 4 + v 4 anti-symmetric bending, and hydrogen bonding of water (H 2 O), respectively [23,24,29,30,32,53,54]. For all solids, there was no direct effect on the NH 4 + and H 2 O bands, regardless of sorption pH and metal loading.
The shape of the PO 4 3− band for blanks did not change with pH, with a single peak at 1000 cm −1 (Figure 4(a-d)), indicating no effect of pH alone on this band. The impact of metal sorption on this band can be broadly classified into two categories, either peaks splitting to form two distinct bands, or peaks merging to form a single broad band centred at 1000 cm −1 . For Cu, Cd, Ni, and Zn, as pH increases the PO 4 3− peak splits to form two distinct peaks at 970 and 1020 cm −1 . Splitting of the peaks is associated with distortion of the PO 4 3− tetrahedron, reflective of possible lowering in the symmetry of the crystal lattice due to metal sorption [32,53]. For Cd, Ni, and Zn, with increasing pH and a corresponding increase in metal loading, the peak at 970 cm −1 becomes less-significant, whereas the peak at 1020 cm −1 is increasingly prominent. For Cu, higher metal loadings resulted in a more prominent peak at 970 cm −1 with a smaller shoulder at 1020 cm −1 . For Zn, this is similar to observations made in previous studies where Zn sorption on struvite at pH 9 affected the PO 4 3− band considerably, with emergence of a single broad peak as Zn concentration on the solid increased [32].  [57]. At pH 10 where multiple Pb solid phases are oversaturated (Table S1), this confirms that the presence of mineral precipitates affects the PO 4 3− band considerably.

X-ray absorption fine structure spectroscopy (XAFS)
To determine mechanisms of sorption, select solids recovered from 5 µM sorption experiments were analyzed by XAFS (Table 2, Figure 5, Table S5, Figure S5). Solids reacted at pH 9 were selected as this is the optimal pH for recovery of struvite from wastewater sources [58], and metal loadings were suitable for XAFS analysis. For Pb, all pH-dependent samples and the pH 10 desorption sample were analyzed. Cadmium was not analyzed as the Cd K-edge energy (26.71 keV) exceeded the maximum permissible energy of the beamline (23 keV).  (Table S5) and are commonly observed configurations for sorption of these metals on other substrates such as oxides (gibbsite, pyrolusite, and amorphous silica), pyromorphite, and pyrophyllite [59][60][61]. For Cu and Zn, tetrahedral coordination through struvite PO 4 3− groups has been previously observed [23,32]. A P shell is detected for all samples, due to metal complexation with PO 4 3− groups.  [64]. At pH 10, the fits to both the sorption and desorption samples are similar, indicating no change in sorption configuration after desorption. Therefore, the Pb particulates released, as observed in macroscopic results, are Pb precipitates that are dislodged from the surface during desorption.

Conclusions
Sorption of the divalent heavy metals, Cd, Co, Cu, Ni, Pb, Zn with struvite was evaluated at 1-5 μM concentrations at pH 7-10. All metals were found to increasingly sorb with the mineral at higher initial concentration and pH, with highest removal of Pb under most conditions. The final metal loadings ranged from 2 to 493 mg kg −1 , with select Cd, Ni, and Pb loadings from 5 μM solutions exceeding regulatory limits for struvite fertiliser [22]. The lability of sorbed metals ranged from 10% to 85% when in contact with metal-free solutions at the same pH. Of the metals, Cd was least susceptible to desorption at higher loadings, with Pb exhibiting the most desorption as loading on struvite increased. When sorbed with struvite at pH 9, Co and Ni were in octahedral coordination, whereas Cu and Zn were in tetrahedral coordination. Sorbed Pb was a mix of adsorbates and precipitates at pH 7, with precipitation as pyromorphite dominating from pH 8-10, and after desorption at pH 10. As with Pb, it is likely that for the other metals adsorption occurs, at lower pH and concentration, and is accompanied by oversaturation and precipitation of secondary phases as pH and initial concentration is increased.
The objective of this study was to develop an understanding of metal sorption with struvite using simplified model solutions to discern fundamental processes at the metal-mineral interface. Evaluation of the metals under the same conditions is important to compare the extent and mechanism of sorption, the effect on the mineral properties, and the potential for desorption. Wastewater systems viable for recovering struvite may contain metal concentrations in exceedance of those used in this study, resulting in higher struvite metal loadings. Conversely, in actual wastewaters, metal sorption may be mitigated by competitive and synergistic effects between metals, and aqueous complexation with inorganic and organic ligands. Once sorbed, metals that are readily desorbed can be removed prior to application as fertiliser, limiting metal contamination to soils and potential toxicity to plants. Overall, the potential of struvite as a fertiliser is dependent on multiple factors that impact the quality of recovered product, including the concentration of heavy metal contaminants. Therefore, elucidating fundamental metal-struvite sorption processes is a critical first-step in development of a holistic strategy of reusing struvite as fertiliser. Results of this study can help to reduce dependency on existing phosphate sources by evaluating the environmental risk associated with using recovered struvite, ultimately improving the understanding of processes that are often considered as barriers to struvite use as fertiliser.