**2. Materials and Methods**

The crystallization of struvite (MgPO4NH4·6H2O) was carried out by magnetic stirring (with an HI 190 stirrer, Hanna Instruments, Woonsocket, RI, USA) in a cylindrical 2 L Pyrex cell. The working solutions were prepared by mixing two solutions of magnesium chloride hexahydrate (MgCl2·6H2O) and ammonium dihydrogen phosphate (NH4H2PO4) in distilled water. The solutions of NH4H2PO4 and MgCl2·6H2O were prepared by dissolving the corresponding solids NH4H2PO4 (purity > 99%, Sigma Aldrich, Burlington, MA, USA) and MgCl2·6H2O (purity > 99%, Fluka, Charlotte, NC, USA) in distilled water. Tablets of NaOH (>99% purity, Sigma Aldrich) were dissolved in distilled water to obtain 1 M NaOH solution. This alkaline solution was used to adjust the initial pH of the working solution to the desired value. For all experiments, the molar ratio of Mg:P:N was set at 1:1:1 corresponding to struvite (MgNH4PO4·6H2O) solid. The solution temperature was fixed at 25 ◦C using a thermostat with circulating water. The solution pH was monitored using a pH meter (pH 213, Hanna Instruments, USA) with a combination glass/Ag/AgCl electrode after calibration with commercially available standard buffer solutions from Biopharm at pH 4, 7, and 10. Each experiment was performed at least four times for reproducibility, and the mean values are reported in the present work. The error on the reported values was less than 5%.

To study the effect of stirring on struvite precipitation, four series of experiments were performed. All the experiments performed in the present work aimed to optimize the crystallization of struvite at high stirring rates. In the first series of experiments, the stirring rate varied between 100 and 700 rpm (100, 300, 500, and 700 rpm), the initial solution pH was adjusted to 8, the initial phosphorus concentration was 200 mg·L−<sup>1</sup> (6.45 mM) and the solution volumes were 600 and 1200 mL. In the second series of experiments, the initial phosphorus concentration was assessed at 600 mg·L−<sup>1</sup> (19.35 mM) and the initial pH, the stirring rate, and the solution volume were assessed as those of the first series of experiments. The third series of experiments was performed at different solution volumes between 600 and 1200 mL, a stirring rate of 700 rpm, an initial solution pH of 8, and an initial phosphorus concentration of 200 mg·L−1. The increase in the solution volume was expected to lower the flow turbulence caused by stirring and favor struvite precipitation. Finally, a series of experiments was set at initial phosphorus concentrations ranging from 200 to 800 mg·L−1. The increase in the constituent ion concentrations may overcome the effect of turbulence on struvite crystallization. For this series of experiments, the stirring rate was set at 700 rpm, the initial pH at 8, and the volume of the solution at 600 mL.

During each experiment performed in this work, periodic samples of 1 mL were withdrawn from the solution and then filtered through a 0.45 μm membrane filter. Next, the concentration of phosphorus was determined photochemically by a HACH DR/4000 spectrophotometer. At the end of each experiment, the solution was filtered using a 0.45 μm membrane filter. The recovered precipitates were dried at room temperature and analyzed by XRD, SEM, and FTIR spectroscopy. XRD was carried out at room temperature using a Philips X'PERT PRO diffractometer in step scanning mode using Cu-Kα radiation. The

XRD patterns were recorded in the scanning range 2θ = 10–70◦ using a small angular step of 2θ = 0.017◦ and a fixed counting time of 4 s. The positions of the XRD reflections were determined using 'X-Pert HighScore Plus' software, Version 2.1. The XRD patterns of the collected precipitates were compared with the Joint Committee on Powder Diffraction Standards data. The FTIR spectra were recorded between 400 and 4000 cm−<sup>1</sup> (Shimadzu IRAffinity-1) using pressed powder samples in KBr medium. The size of particles was measured using a Mastersizer 2000 (Version 6.01, Malvern Instruments, Malvern, UK) combined with a Hydro 2000 MU.

### **3. Results**

#### *3.1. Effect of Stirring*

For all the experiments carried out in the present work, the precipitated phase observed was struvite. A typical X-ray diffractogram and FTIR spectrum of the precipitates obtained are given in Figure 1a,b, respectively. The XRD pattern (Figure 1a) showed that the characteristic reflections of the obtained precipitates were comparable to those of the pattern for the struvite standard (JCPDS 15-0762) [32]. The FTIR spectrum (Figure 1b) presented the characteristic band of PO4 <sup>3</sup><sup>−</sup> located at 1005 cm−<sup>1</sup> and the characteristic band of NH4 + located at 2933 cm<sup>−</sup>1. Comparable struvite spectra were reported by Korchef et al. [13] and Zhang et al. [33]. The following absorption bands were also observed: a large band between 3600 and 2270 cm−<sup>1</sup> corresponding to O-H and N-H stretching vibrations, a low-intensity band located at 1660 cm−<sup>1</sup> attributed to H-O-H bending modes, which indicates water of crystallization, and a band at 1432 cm−<sup>1</sup> attributed to the asymmetric bending vibration of N-H in NH4 +. The bands detected between 1005 and 456 cm−<sup>1</sup> are characteristics of the ion PO4 <sup>3</sup>−, where the wide low-intensity band at 1005 cm−<sup>1</sup> was associated with the asymmetric bending vibration of PO4 <sup>3</sup>−, the weak band at 870 cm−<sup>1</sup> was attributed to the vibration of coordinated water−, the band at 566 cm−<sup>1</sup> was associated with the asymmetric bending modes of PO4 <sup>3</sup>−, and the band at 456 cm−<sup>1</sup> was attributed to the symmetric bending vibration of PO4 <sup>3</sup> units. SEM analysis confirmed the results obtained by XRD and FTIR spectroscopy and showed precipitates with the typical prismatic pattern characteristic of struvite crystals (Figure 2).

**Figure 1.** Typical examples of the (**a**) XRD pattern and (**b**) FTIR spectrum of the obtained precipitates.

**Figure 2.** SEM micrograph of the precipitates obtained with the typical prismatic pattern characteristic of struvite crystals.

Figure 3 illustrates the changes in pH and phosphorus concentration over time at varied stirring rates between 100 and 700 rpm for an initial phosphorus concentration of 200 mg·L−<sup>1</sup> and solution volumes of 600 and 1200 mL. For all experiments performed, the pH decreased over time, and then from a certain time (depicted herein as the end of precipitation time), it remained practically constant (Figure 3a). Concomitantly, a decrease in phosphorus concentration was observed, followed by a plateau (Figure 3b). The higher the stirring rate was, the lower the concentration obtained at the end of the precipitation. The decrease in pH and phosphorus concentration over time was due to struvite precipitation according to the following reaction [34]:

$$\text{Mg}^{2+} + \text{NH}\_4^+ + \text{H}\_n\text{PO}\_4^{n-3} + 6\text{H}\_2\text{O} \rightarrow \text{MgNH}\_4\text{PO}\_4\cdot6\text{H}\_2\text{O} + n\text{H}^+\tag{1}$$

where *n* = 0, 1, and 2 depending on the solution pH and the initial phosphorus concentration.

The constant values of pH and phosphorus concentration for advanced reaction times indicated the end of precipitation. The precipitation of struvite occurred as soon as the solution pH was adjusted to 8, and it was manifested by a significant decrease in phosphorus concentration and pH in the first minutes. Comparable results were found when struvite precipitated from pig manure in an anaerobic digester [35]. At the end of precipitation, the solution pH reached values between 7.2 and 7.4 for all experiments. This agreed with the results of Stratful et al. [36], who observed that at a pH equal to 7, no precipitation of struvite occurred, and at a slightly higher pH of 7.5, only a small amount was recovered through very small crystals. Several studies on the effect of pH on struvite crystallization were reported in the literature, and it was depicted that pH affected the solubility of struvite. Indeed, it was found that for pH values between 8 and 10 the solubility of struvite significantly decreased, and the struvite crystallization rate increased [23,37–39].

The pH and phosphorus concentration values obtained over time for the stirring rates 100 and 300 rpm and a solution volume of 600 mL were comparable but slightly higher than those obtained for 500 and 700 rpm (Figure 3a,b). When the volume increased to 1200 mL, the curves showing the changes in pH (Figure 3c) and phosphorus concentration (Figure 3d) over time presented the same evolution trends as those for a volume of 600 mL.

**Figure 3.** Variations of pH and phosphorus concentration over time for different stirring rates, an initial phosphorus concentration of 200 mg·L−<sup>1</sup> and solution volumes of (**a**,**b**) 600 mL and (**c**,**d**) 1200 mL.

Table 1 gives the precipitation end time (tf), the initial precipitation rate (Vi), and the phosphorus removal efficiency, *R* (%), obtained at different stirring rates and solution volumes of 600 and 1200 mL. The precipitation end time (tf) corresponded to the first point of stabilization of the phosphorus concentration after precipitation. The initial precipitation rate (Vi) was determined from the slope of the linear part of the time curve of phosphorus concentration for times less than tf. The phosphorus removal efficiency, *R* (%), was calculated using the following equation:

$$R(\%) = \frac{\left(\mathbb{C}\_i - \mathbb{C}\_f\right)}{\mathbb{C}\_i} \times 100\tag{2}$$

where (*Ci*) and (*Cf*) are the initial and final phosphorus concentrations, respectively.

As the stirring rate increased, the precipitation end time (tf) decreased for both volumes used. For example, tf decreased from 75 to 45 min when the stirring rate increased from 100 to 700 rpm, respectively, for a solution volume of 600 mL (Table 1). For a given stirring rate, the precipitation end time obtained for a solution volume of 1200 mL was greater than that obtained for a solution volume of 600 mL. The highest precipitation end time (90 min) was obtained for a stirring rate of 500 rpm and a solution volume of 1200 mL (Table 1) where, in addition to precipitation in the bulk solution, a significant amount of struvite precipitated on the cell walls. The change in the precipitation nature (from precipitation in the bulk solution to precipitation on the cell walls) could be the cause of the increase in the precipitation end time at the stirring rate of 500 rpm and the solution volume of 1200 mL. By increasing the stirring rate to 700 rpm, precipitation occurred mainly in the bulk solution and the precipitation end time decreased to 45 min for both volumes used.

**Table 1.** End of precipitation time (tf), initial precipitation rate (Vi), and phosphorus removal efficiency (*R*%) for different stirring rates, solution volumes of 600 and 1200 mL, and an initial phosphorus concentration of 200 mg·L<sup>−</sup>1.


The initial precipitation rate (Vi) increased with the increase in the stirring rate for both solution volumes of 600 and 1200 mL. At 100 rpm and a solution volume of 600 mL, the initial precipitation rate Vi was equal to 0.22 mmol·L−1.min−1. The increase in the stirring rate to 500 rpm resulted in a slight increase in Vi to 0.36 mmol·L−1·min−<sup>1</sup> (Table 1). A more pronounced increase in the initial precipitation rate was observed when the stirring rate increased to 700 rpm., i.e., it became approximately five times higher (Vi =1.13 mmol·L−1.min−1). Comparable results were found for a solution volume of 1200 mL. However, the acceleration of struvite precipitation was accompanied by a slight decrease in the phosphorus removal efficiency. Indeed, increasing the stirring rate from 100 to 500 rpm significantly improved the phosphorus removal efficiency, and comparable values were obtained for both solution volumes. For example, the phosphorus removal efficiency, at a solution volume of 600 mL, increased from ~90% to ~95% when the stirring rate increased from 100 to 500 rpm, respectively, and it decreased slightly to ~94% at a higher stirring rate of 700 rpm (Table 1). This decrease was more pronounced for a solution volume of 1200 mL where the removal efficiency of phosphorus reached ~89% at a stirring rate of 700 rpm. A detailed study of the effect of the solution volume on struvite precipitation is given herein.

To conclude, for an initial phosphorus concentration of 200 mg·L−1, the optimal stirring rate (which gave the highest efficiency) was 500 rpm for both solution volumes of 600 and 1200 mL. Comparable results were depicted by Perera et al. [29], who found that magnetic stirring at 500 rpm gave the highest phosphorus removal through struvite crystallization from swine waste biogas digester effluent where the molar ratio of N:Mg:P was 1:1.2:1.2, the working volume was 3 L, and pH was equal to 9.

Figure 4 shows the variation in phosphorus concentration over time at stirring rates between 100 and 700 rpm, an initial phosphorus concentration of 600 mg·L−1, and a solution volume of 600 mL. As soon as the initial pH was adjusted to 8, the concentration of phosphorus in the solution decreased. This decrease became more pronounced as the stirring rate increased. In addition, the pH attained at the end of precipitation reached ~6.9 for a stirring rate of 100 rpm and ~6.7 for stirring rates of 500 and 700 rpm. The initial rate of struvite precipitation increased with the stirring rate, and the end of precipitation time decreased (except at 700 rpm where a slight increase in tf was observed). In addition, the phosphorus removal efficiency increased with the stirring rate. Indeed, it increased from ~97.5% for a stirring rate of 100 rpm to ~99.4% for a stirring rate of 700 rpm (Table 2). Therefore, at the high initial phosphorus concentration of 600 mg·L−<sup>1</sup> and a solution volume of 600 mL, increasing the stirring rate accelerated the precipitation of struvite, and higher struvite amounts were obtained at relatively shorter times. Comparing the values of tf, Vi, and *<sup>R</sup>* (%) obtained for the initial phosphorus concentrations of 200 and 600 mg·L−<sup>1</sup>

and the solution volume of 600 mL (Tables 1 and 2) indicated that the values obtained for 600 mg·L−<sup>1</sup> were significantly higher than those obtained for 200 mg·L−1. This can be explained by the flow turbulence of the solution. Indeed, when the turbulence of the solution became stronger (by increasing the stirring rate at a fixed solution volume), the transition of the constituent ions of struvite from the liquid phase to the solid phase became more difficult, and consequently, the reaction with the struvite crystal surface was inhibited. This resulted in a decrease in struvite crystal growth and fewer amounts were obtained. The limiting effect of turbulence became less consequent to the precipitation of struvite when the phosphorus concentration increased to 600 mg·L−<sup>1</sup> at a fixed solution volume of 600 mL.

**Figure 4.** Variations of (**a**) pH and (**b**) phosphorus concentration over time for different stirring rates, an initial phosphorus concentration of 600 mg·L<sup>−</sup>1, and a solution volume of 600 mL.

**Table 2.** End of precipitation time (tf), initial precipitation rate (Vi), and phosphorus removal efficiency (*R*%) for different stirring rates, an initial phosphorus concentration of 600 mg·L−1, and a solution volume of 600 mL.


Changes in pH and phosphorus concentration over time for different solution volumes between 600 and 1400 mL, an initial phosphorus concentration of 200 mg·L<sup>−</sup>1, and a stirring rate of 700 rpm are shown in Figure 5a,b. The obtained curves showed the same evolution trends as those obtained for the different stirring rates presented in Figure 3. The increase in the solution volume resulted in an increase in the phosphorus concentration obtained at the end of precipitation. That is, fewer amounts of struvite were obtained when the solution volume increased. This explains the slight increase in the solution pH obtained at the end of struvite precipitation. Indeed, struvite precipitation is accompanied by the release of protons (see Equation (1)). The initial precipitation rate Vi, the end of precipitation time tf, and the phosphorus removal efficiency *R* (%) for different solution volumes, an initial phosphorus concentration of 200 mg·L<sup>−</sup>1, and a stirring rate of 700 rpm are given in Figure 5c,d. The initial precipitation rate decreased from ~1.13 to ~1 mmol·L−1·min−<sup>1</sup> when the volume of the solution increased from 600 to 800 mL, respectively. For higher volumes, it remained practically constant. The end of precipitation time remained practically constant up to a volume of 800 mL, then it increased significantly (Figure 5c). The phosphorus removal efficiency decreased as the volume of the solution increased. For example, it decreased from ~94% to ~85% as the volume increased from 600 to 1400 mL, respectively (Figure 5d). At a constant stirring rate and initial phosphorus concentration, the increase in the solution volume affected the transfer of struvite constituent ions from the liquid phase to the solid phase and inhibited their reaction with the crystal surfaces, which limited the rate of struvite crystal growth. Consequently, the amount of precipitated struvite decreased. This was confirmed by the decrease in the initial precipitation rate and the increase in the end of precipitation time. This agreed with the fact that increasing the stirring rate up to 500 rpm at a constant volume of 600 mL increased the phosphorus removal efficiency by increasing the transfer rate of struvite constituent ions. At 700 rpm, the solution became strongly turbulent, and the opposite effect (limitation of the precipitation reaction) occurred. In conclusion, under the experimental conditions of the present work, to obtain a high phosphorus removal efficiency through struvite precipitation, the stirring rate should be set at 500 rpm when working at constant volume. For higher stirring rates, the volume of the solution should be reduced.

**Figure 5.** Variations of (**a**) pH, and (**b**) phosphorus concentration over time for different solution volumes, and variations of the (**c**) end of precipitation time tf and initial precipitation rate Vi, and (**d**) phosphorus removal efficiency with the solution volume for an initial phosphorus concentration of 200 mg·L−<sup>1</sup> and a stirring rate of 700 rpm.
