Advanced Phosphorus Recovery from Municipal Wastewater using Anoxic/Aerobic Membrane Bioreactors and Magnesium Carbonate-Based Pellets

Abstract

Effective recovery of phosphorus from municipal wastewater could be one of the best practical alternatives to protect aquatic environments from eutrophication and save natural phosphorus resources. This paper focuses on validating magnesium carbonate (MgCO₃)-based pellets combined with a bench-scale anoxic/aerobic membrane bioreactor (MBR) system for advanced phosphorus recovery from municipal wastewater. As the flow rate of wastewater into the MgCO₃ column decreased from 10 L/d to 2.5 L/d, the phosphorus recovery rate of the MgCO₃-based pellets increased from 54.3 to 93.5%. However, the column’s severe clogging was found after a 13-days operation due to the high removal of total suspended solids (TSS) (~82%) through the MgCO₃ column. The anoxic/aerobic MBR introduction provided efficient removal of TSS, organic matter, and ammonia nitrogen before the MgCO₃ column. The combination of MBR with the MgCO₃ column achieved 73.1% phosphorus recovery from municipal wastewater without physical clogging. The P recovery capacity of the MgCO₃-based pellets was maintained at 0.47 mg ortho-P/g MgCO₃-based pellet during the continuous operation. Physical and chemical properties of MgCO₃-based pellets before and after the experiment were characterized using scanning electron microscopy (SEM) with energy-dispersive X-ray spectroscopy (EDX), X-ray diffraction (XRD), and Brunauer–Emmett–Teller (BET) surface area analyzer.

Graphical Abstract

1. INTRODUCTION

Phosphorus (P) is an essential element for life and plays a crucial role in farming practices,¹ derived from the depleting mineral phosphate rocks, which mandates the need to recover and reuse it. On the other hand, municipal wastewater contains a large amount of phosphorus from human feces and urine, washing machine and dishwasher detergents, and personal care products,² leading to harsh environmental impacts such as eutrophication in the aquatic environments if not controlled before discharge.^3,4 Indeed, the increasing occurrence of harmful algal blooms in lakes and rivers has become an emerging concern threatening human and environmental health because certain cyanobacteria can produce and release potent toxic compounds in sources of water supply.^5–8

Among various phosphorus removal and recovery methods, including adsorption, chemical precipitation, biological removal from wastewater, chemical precipitation methods like struvite precipitation are the most used method, done by adding metallic salts such as iron, magnesium, calcium, and aluminum. Nevertheless, phosphates precipitated as iron and aluminum salts are not suitable for further applications as fertilizers.^9,10

The biological phosphorus removal method, often referred to as enhanced biological phosphorus removal (EBPR), occurs when the sludge enriched with polyphosphate-accumulating organisms (PAOs) are exposed under alternating anaerobic and aerobic conditions. Under the anaerobic condition, the PAOs uptake carbon sources, store polyhydroxyalkanoates (PHAs) with the energy produced from polyphosphate hydrolysis, and release P into water.^11,12 Under the aerobic condition, PAOs utilize the PHAs stored in the anaerobic phase for biomass growth and uptake phosphorus from water for polyphosphate synthesis.¹³ It is a cost-effective method and provides sludge free of metal ions and coagulants.¹⁴ However, the EBPR process has certain drawbacks, such as ample space requirement for secondary clarifier, solid (biomass)–liquid (water) separation issues, and excess production of waste sludge.¹⁴

Adsorptive removal of phosphorus from wastewater effluents has attracted attention due to its potential for sludge reduction and phosphorus recovery for further applications. Physical adsorption of phosphorus in the membrane permeates using solid adsorbents can be a feasible technology to efficiently recover phosphorus from wastewater without further hindering particles and microorganisms, which reduce the adsorption efficiency of phosphorus from water.^15–17

Solid adsorbents are more straightforward in design, less initial cost, easy to separate, and insensitive to toxic substances. It helps to complete the contaminant’s removal even under lower concentrations,¹⁸ and is an attractive method as it allows to desorb the phosphorus back from the adsorbent material for reuse. For this application, adsorbents should be durable, low cost, highly selective for phosphorus recovery from wastewater, and easy to recycle and reuse.

Among various adsorbents, magnesium carbonate (MgCO₃)-based materials (particularly beads, tablets, or pellet types) showed great potential for the efficient and sustainable recovery of phosphorus from wastewater due to its physical and chemical stability in water and eco-friendly properties for agricultural applications.^19,20

Our previous study²⁰ showed that magnesium-based pellets made of MgCO₃ powder and a binder (i.e., cellulose) achieved over 80% recovery of 160 mg/L ortho-phosphate in deionized water. The pseudo-second-order kinetics model fits best, suggesting that the adsorption occurring was chemisorption.

The average adsorption capacity achieved in the isotherm study was 96.4 mg P per g MgCO₃ at equilibrium through batch experiments for 25 days. However, this study was conducted under the ideal condition (i.e., at a high soluble phosphorus concentration in synthetic wastewater without particles and other organic and inorganic contaminants).

From desorption tests in deionized water, 0.1 M HCl, and 0.1 M NaOH suspensions, it was found that the leaching of P into various aqueous suspensions (pH 4–10) was below 4.5% of the adsorbed-P after 25 days, indicating a great application potential as a slow-release material.²⁰

Membrane bioreactor (MBR) has advantages over the conventional activated sludge process for wastewater treatment, such as smaller footprint, less sludge production.²¹ MBR can produce high-quality effluent due to the membrane’s adequate pore size (typically 0.01–0.1 μm) that efficiently retains suspended solids (SS) and microorganisms.²² However, its typical long solid retention time (SRT) has a negative effect on phosphorus removal.²³

Therefore, the present study evaluates the effects of real wastewater matrices on phosphorus recovery from MBR effluent using the MgCO₃-based pellets. The effects of SS and various feed flow rates on phosphorus recovery from municipal wastewater were studied using a continuous-flow column packed with the MgCO₃-based pellets.

2. MATERIALS AND METHODS

2.1. Synthesis of MgCO₃-Based Pellets.

Analytical grade MgCO₃ powder (Fisher Scientific) was made into pellets (6 mm in diameter and 17 mm in length, Supporting Information (SI) Figure S1) without using any binding material using the MZL Flat Die Pellet Mill (Xuzhou Orient Industry). The adsorbent mass of each pellet was 0.4 ± 0.05 g. The pelletizer instrument allowed for variation in the pellets’ length, but the die hole size controlled the diameter. It meant the standard deviation of diameter was small, while the deviation in length was large.

2.2. Characterization of MgCO₃-Based Pellets.

Surface morphology of MgCO₃-based pellets before and after phosphorus recovery from municipal wastewater was examined using scanning electron microscopy (SEM) with energy dispersive X-ray (EDX) (JEOL JSM-6490LV) for both a line scan and a cross-section scan of the pellets. The uniformity of MgCO₃-based pellets was determined by comparing the percentage of carbon (C), oxygen (O), magnesium (Mg), and phosphorus (P) present across each SEM-EDX scan. The crystal structure was determined using X-ray powder diffraction (XRD) (PANalytical X’Pert) with the 2-theta diffractometer under CuK_αradiation and a wavelength of 1.54 μm, and the XRD patterns were analyzed using JADE software (MDI, Inc.). The BET surface area was determined using the NOVA 2000e Surface Area and Pore Size Analyzer (Quantachrome). Samples were first purged with nitrogen gas at 150 °C overnight before analysis.

2.3. A Bench-Scale Continuous-flow Column Reactor with MgCO₃-Based Pellets.

Recovery of phosphorus from the MBR effluent using MgCO₃-based pellets was studied in a bench-scale continuous flow column (75 cm in length and 3.5 cm in diameter) at different flow rates (SI Figure S2). The volume of the column and the packing density were 0.72 L and 153.1 g/L, respectively. The MgCO₃-based pellets were manually broken into 2–3 parts before being packed into the column. The pellets’ weight packed into the column was 85.5 g for Run 1 and 110.5 g for Runs 2–4. The column was filled with 80 g of gravel in the top and bottom to serve as support layers.

The column study was conducted at room temperature. Table 1 shows the experimental conditions of the bench-scale column with MgCO₃-based pellets. In Run 1, raw wastewater (i.e., before primary settling) was pumped at a flow rate of 10 L/d to study the effects of total suspended solids (TSS) on phosphorus recovery using these pellets. From Run 2, the effect of flow rates (i.e., 2.5, 5, 10 L/d) on phosphorus recovery was studied using primary effluent (after primary settling). Table 2 shows the typical characteristics of raw wastewater and the primary effluent used in this study.

Table 1.

Experimental Conditions of the Bench-Scale Column with MgCO₃ pellets

	run 1	run 2	run 3	run 4
feed	raw wastewater	primary effluent
flow rate, L/d	10		5	2.5
empty bed contact time (EBCT), min	104		208	416
weight of MgCO3 pellets, g	85.5	110.5

Table 2.

Characteristics of Municipal Wastewater from the Mill Creek Wastewater Treatment Plant (Cincinnati, Ohio)

parameter	unit	raw wastewater (run 1)	primary effluent (runs 2–6)
pH	-	7.2 ± 0.2	7.3 ± 0.2
total suspended solids (TSS)	mg/L	306 ± 45	196 ± 116
chemical oxygen demand (COD)	mg/L	403 ± 60	347 ± 28
ammonia-nitrogen (NH3–N)	mg/L	11.3 ± 1.8	12.4 ± 6.2
total phosphorus (TP)	mg/L	2.8 ± 0.7	3.4 ± 0.2
orthophosphate (Ortho-P)	mg/L	1.0 ± 0.3	1.1 ± 0.4
turbidity	NTU	87.7 ± 6.5	10.1 ± 2.7

The adsorption capacity and removal efficiency of phosphorus using the MgCO₃-based pellets were calculated using the following equations:

$$\text{absorption}\text{capacity}(\text{mg}\text{P}/\text{g}\text{MgC}{\text{O}}_3-\text{based}\text{pellets})=\frac{\sum \left(C_{\text{in}}-C_{\text{out}}\right)Q}{W}$$ (1)

where C_in and C_out are the daily average concentration of phosphorus in the influent and effluent (mg/L), respectively, Q is the flow rate (L/d), and W is the weight of MgCO₃-based pellets (g).

$$\text{removal}\text{efficiency}(\text{\%})=\left[\left(C_{\text{in}}-C_{\text{out}}/C_{\text{in}}\right)\right]\times 100$$ (2)

where C_in and C_out are the daily average concentration of various constituents in the influent and effluent (mg/L), respectively.

2.4. Anoxic/Aerobic MBR System.

Figure 1 shows a process flow diagram of the anoxic/aerobic MBR system and a column reactor packed with MgCO₃-based pellets used in this study. The anoxic (working volume = 10 L) and aerobic (working volume = 10 L) reactors were inoculated with return activated sludge (RAS) from the Mill Creek Wastewater Treatment Plant (Cincinnati, Ohio). The anoxic reactor has a mixer impeller, which provides complete mixing of mixed liquor suspended solids (MLSS).

Figure 1.

Schematic diagram of the anoxic/aerobic MBR system and a column reactor packed with MgCO₃-based pellets for enhanced phosphorus recovery from municipal wastewater.

A polyvinylidene fluoride (PVDF) hollow fiber membrane (nominal pore size = 0.03 μm, LOTTE Chemical, South Korea) (SI Table S1, Figure S3) module was submerged into the aerobic reactor. Air bubbles are sparged into the reactor to provide aeration (dissolved oxygen concentration = 3–5 mg/L) and simultaneously complete mixing MLSS. MLSS was recycled from the aerobic reactor to the anoxic reactor at 300% of the feed flow rate.

Hydraulic retention time (HRT) and SRT of the MBR system were maintained 24 h and 60 days, respectively. Excess sludge was taken from the aerobic reactor to control SRT. The MBR system was operated at room temperature (20.8 ± 0.5 °C). The system was fed with primary effluent (i.e., after primary settling) from the Mill Creek Wastewater Treatment Plant. Samples were collected from the influent storage tank, anoxic reactor, and MBR effluent storage tank (Figure 1).

2.5. Integration of the MBR System and the MgCO₃ Column for the Enhanced Recovery of Phosphorus from Municipal Wastewater.

First, the MBR system was operated over 30 days to stable the removal of organic matter, TSS, and NH₃–N from municipal wastewater (Run 5, Table 3). Primary effluent was fed into the anoxic reactor at a flow rate of 10 L/d. The MBR system was then connected with a column reactor packed with MgCO₃-based pellets in series (Figure 1) and operated for 10 days (Run 6, Table 3). In Run 6, the MBR effluent through the PVDF membrane was fed into the column at a flow rate of 2.5 L/d. The MBR reactors and the MgCO₃ column were operated at room temperature (21.3 ± 0.8). Samples were collected from the influent storage tank, anoxic reactor, MBR effluent (i.e., MgCO₃ in), and MgCO₃ out.

Table 3.

Experimental Conditions of the MBR System Treating Primary Effluent

operating condition		run 5	run 6
flow rate, L/d	MBR	10	10
	MgCO3 column	not applicable	2.5
weight of MgCO3-based pellets, g		not available	110.5
operating period, day		31	25

2.6. Characterization of Municipal Wastewater.

Concentrations of COD, TP, ortho-P, and NH₃–N were measured using a UV/vis spectrophotometer (DR6000, HACH). TSS and volatile suspended solids (VSS) were measured according to Standard Methods.^24,25 Temperature and pH were measured using temperature and pH electrodes connected with a benchtop pH meter (ThermoScientific). Turbidity was measured using a turbidimeter (TL2300, HACH). Particle size in raw wastewater and primary effluent was measured using a professional infinity plan phase contrast Kohler trinocular microscope with a 14 MP high-speed digital camera (detection limit = 1 μm, AmScope, U.S.). Image analysis of the microscopic images obtained was performed using Image-Pro version 4.5 (Media Cybernetics, Inc. Bethesda, MD).

3. RESULTS AND DISCUSSION

3.1. Effect of TSS on Phosphorus Recovery from Raw Wastewater using MgCO₃-Based Pellets (Run 1).

The physical and chemical properties of pristine MgCO₃-based pellets were characterized. The BET surface area of MgCO₃-based pellets was 27.5 m²/g. The XRD patterns of MgCO₃-based pellets showed hydromagnesite (Mg₅(CO₃)₄(OH)₂· 4H₂O) was a major form with some remaining brucite (MgOH) and magnesium phosphate (SI Figure S4).

Raw wastewater (turbidity = 87.7 NTU) from the Mill Creek Wastewater Treatment Plant contains large particles and microorganisms (SI Figure S5). When the raw wastewater (before primary settling) was fed into the continuous-flow MgCO₃ column at a flow rate of 10 L/d, it was observed that the column was clogged within 3 days as particles physically clogged the pores. Average TSS and COD concentrations in raw wastewater were 306 mg/L and 403 mg/L, respectively (Table 2). Phosphorus recovery efficiency from raw wastewater was 45–46% using MgCO₃-based pellets during the first 2 days, but there was no significant change in the influent and the effluent on the third day (SI Figure S6). Turbidity of MgCO₃ column effluent was maintained below 1.0 NTU during the experimental period. It can be inferred that the high TSS in raw wastewater choked the pores of the MgCO₃-based pellets and blocked phosphorus adsorption during the short-time operation.

3.2. Effects of Feed Flow Rate on Phosphorus Recovery from Primary Effluent using MgCO₃-Based Pellets (Runs 2–4).

Raw wastewater was replaced with primary effluent (i.e., after primary settling, turbidity = 10.1 NTU) from the Mill Creek Wastewater Treatment Plant to avoid media pores’ clogging of the MgCO₃ column. The maximum TSS concentration that the MgCO₃ pellet can handle depends on both the feed flow rate and the operating period. It was found that over 300 mg/L TSS in raw wastewater easily clogged the MgCO₃ column at 10 L/day, but 203 mg/L TSS in primary effluent was acceptable for 10-day operation without clogging at the same feed flow rate.

As shown in Figure 2, the average sizes of particles in the primary effluent (Figure 2A) and the MgCO₃ effluent (Figure 2B) during the experimental period were in the range of 97–130 μm and 25–42 μm, respectively. It was found that after the primary settling, most of the large particles and microorganisms were removed (SI Figures S7 and S8). Table 4 shows the removal efficiencies of major constituents in primary effluent by the MgCO₃-based pellet column at various feed flow rates.

Figure 2.

Changes in particle size in MgCO₃ influent (A) and effluent (B) (Runs 2–4) (sample number = 10 per each run).

Table 4.

Removal Efficiencies of Major Pollutants in Primary Effluent Through the MgCO₃-Based Pellet Column at Various Feed Flow Rates (Runs 2–4, n = 10)

	run 2 (Q = 10 L/d)			run 3 (Q = 5 L/d)			run 4 (Q = 2.5 L/d)
parameter	in (mg/L)	out (mg/L)	Re. (%)	in (mg/L)	out (mg/L)	Re. (%)	in (mg/L)	out (mg/L)	Re. (%)
TSS	203.3 ± 106.6	56.6 ± 15.0	62.7	163.5 ± 73.3	21.0 ± 7.4	81.9	152.3 ± 46.7	37.6 ± 18.7	75.6
COD	378.6 ± 199.7	168.9 ± 23.6	44.6	335.6 ± 107.7	172.4 ± 42.2	44.7	327.2 ± 39.9	157.8 ± 31.0	50.6
NH3–N	11.0 ± 2.3	10.4 ± 1.8	11.0	18.9 ± 5.9	14.5 ± 5.2	24.7	18.6 ± 2.9	11.7 ± 1.8	36.4
TP	3.3 ± 1.0	2.4 ± 0.9	44.3	3.3 ± 1.2	0.9 ± 0.4	72.8	3.7 ± 0.5	0.9 ± 0.5	75.1
ortho-P	0.9 ± 0.3	0.4 ± 0.3	54.3	1.3 ± 0.4	0.4 ± 0.2	76.2	1.7 ± 0.4	0.1 ± 0.1	93.5

During the experimental periods, the TSS concentration of primary effluent was in the range of 152.3–203.3 mg/L, and 62.7–81.9% of TSS was removed by the MgCO₃ pellet column (sI Figure S9). The MgCO₃ pellets also showed significant removal of COD (45–50%) and NH₃–N (11–36%) from the primary effluent (SI Figures S10 and S11). In general, as the feed flow rate decreased from 10 L/d to 2.5 L/d, the overall removal efficiencies of TSS, COD, and NH₃–N increased. Turbidity of MgCO₃ column effluent was maintained below 1.0 NTU during the experimental period.

The effects of feed flow rate on phosphorus removal from primary effluent were also significant (Figure 3). At the flow rate of 10 L/d (Run 2), the average recovery efficiencies of TP (Figure 3A) and ortho-P (Figure 3B) were 44.3% and 54.3%, respectively. These results show that the fast flow rate might cause significant fluctuation in the removal efficiencies and affect phosphorus’s adsorption mechanism by reducing the contact time between MgCO₃-based pellets and phosphorus in wastewater. When the flow rate was reduced to 5 L/d in Run 3, the recovery efficiencies of TP and ortho-P increased to 72.8% and 76.2%, respectively. When the feed flow rate was further reduced to 2.5 L/d (Run 4), the recovery efficiencies of TP and ortho-P increased to 75.1% and 93.5%, respectively (Figures 3A,B).

Figure 3.

Distribution of TP (A) and ortho-P (B) in primary effluent and the MgCO₃ column effluent (Runs 2–4).

It could be assumed that the slower flow rate increased contact time between MgCO₃-based pellets and phosphorus, resulting in higher recovery efficiency (particularly for ortho-P, SI Figures S12–S14). Besides, the reduced flow rate had delayed the breakthrough of the MgCO₃-based pellets. However, the column was clogged by particles due to the high removal of TSS, and it was not operational after 10 days (Run 2), 13 days (Run 3), and 15 days (Run 4) (data not shown).

3.3. Enhanced Phosphorus Recovery from Primary Effluent using the MBR system and MgCO₃-based Pellets (Runs 5–6).

In Runs 5 and 6, the anoxic/aerobic MBR system was operated in combination with the MgCO₃-based pellet column. No significant changes were observed in the pH (~7.3) and temperature (21–23 °C) during the MBR operation. Table 5 shows the changes in TSS, NH₃–N, and ortho-P through the MBR system and the MgCO₃-based column reactor during Run 5 (operating time = 1–31 days) and Run 6 (operating time = 32–59 days).

Table 5.

Removal Efficiencies of Major Constituents in Primary Effluent Through the MBR System and the MgCO₃ Column Reactor (Runs 5 and 6). ND: Not Detected

	run 5 (1 – 31 days)			run 6 (32 – 59 days)
parameter	in (mg/L)	MBR out (mg/L)	Re. (%)	In (mg/L)	MBR out (mg/L)	Re. (%)	MgCO3 out (mg/L)	Re. (%)
TSS	840 ± 117	ND	> 99.9	622 ± 100	ND	> 99.9	ND	> 99.9
NH3–N	4.5 ± 1.3	2.0 ± 0.2	55.6	9.1 ± 0.6	4.1 ± 0.3	54.9	2.0 ± 0.3	78.0
ortho-P	0.59 ± 0.17	0.60 ± 0.18	NA	1.71 ± 0.31	1.72 ± 0.29	NA	0.46 ± 0.06	73.1

In Run 5, the average TSS in the influent, anoxic reactor, and the aerobic reactor were 840 mg/L, 12 188 mg/L, and 7068 mg/L, respectively (Figure 4). VSS concentration was approximately 60% of TSS concentration in influent, anoxic reactor, and the aerobic reactor. The MBR system provides solid-free wastewater (i.e., no TSS and VSS were observed, and turbidity was below the detection limit, 0.1 NTU) during the experimental period (SI Figure S15).

Figure 4.

Changes in TSS and VSS concentrations through the anoxic/aerobic MBR system (Runs 5–6).

It was also observed that NH₃–N was simultaneously removed through the MBR system. Over 55% removal efficiency of NH₃–N was achieved through the MBR system. In addition, approximately 23% of NH₃–N in influent was removed by the MgCO₃-based pellets. The overall NH₃–N removal efficiency of the MBR system and the MgCO₃-based pellets was found to be 78.0% (SI Figure S16).

The presence of phosphate and ammonia in an equal molar ratio and the MgCO₃ pellets could lead to the formation of struvite, resulting in the removal of ammonia. The high removal efficiency of soluble COD (>99%) was observed through the MBR system (>97%) and the column reactor (~2%) during the operation (data not shown).

Figure 5 shows that the MBR system achieved no net removal of ortho-P due to the long SRT (i.e., 60 days). In Run 6, the MBR system was combined with the MgCO₃ column. As a result, it was found that 73.1% of ortho-P in primary effluent could be recovered in Run 6, and the final ortho-P concentration was maintained at 0.5 mg/L during the operation.

Figure 5.

Changes in orthophosphate concentration through the anoxic/aerobic MBR system and the MgCO₃ column reactor (Runs 5–6).

The transmembrane pressure (TMP) of the PVDF membrane submerged in the aerobic reactor was maintained below 5 kPa at a low operating flux (3.2 L/m²/h) during the operation (data not shown), and the MgCO₃ column was operated for more than 20 days without the clogging issue (date not shown).

3.4. Characteristics of the MgCO₃-Based Pellets for Phosphorus Recovery.

The phosphorus recovery capacity of the MgCO₃-based pellets under the different operating conditions was calculated using eq 1 and data. The phosphorus recovery capacity of the MgCO₃-based pellets increased from 0.41 to 0.48 mg ortho-P/g MgCO₃-based pellet as the feed flow rate decreased from 10 to 2.5 L/d in Runs 2–4 (Table 6).

Table 6.

Phosphorus Adsorption Capacity of the MgCO₃-Based Pellets from Municipal Wastewater for 10 Days

run	feed flow rate (L/d)	EBCT (minute)	P recovery (mg ortho-P/g MgCO3-based pellet)
2	10	104	0.41
3	5	208	0.43
4	2.5	416	0.48
6	2.5	416	0.47

SEM-EDX images were taken to study the surface morphology of the MgCO₃-based pellets and confirm the phosphorus adsorption on the pellets after each run at various feed flow rates. As a result, it was found that magnesium and phosphorus were observed in abundance showing that there was phosphorus adsorption and still lots of magnesium available for adsorption (Figure 6A–C).

Figure 6.

SEM-EDX images of MgCO₃-based pellets after P adsorption at 10 L/d (Run 2) (A), 5 L/d (Run 3) (B), and 2.5 L/d (Run 4) (C).

These results indicate that phosphorus adsorption on MgCO₃-based pellets was not achieved at the maximum capacity and was limited by the hydraulic flow rate and TSS accumulation. Moreover, the MgCO₃ column was stably operated without significant clogging issues due to introducing particle-free membrane effluent into the MgCO₃ column.

The P-loaded MgCO₃ could be converted into struvite through precipitation within the pellets (Mg²⁺ + NH₄⁺ + PO₄³⁻ + 6H₂O → NH₄MgPO₄·6H₂O). However, in the present study, the concentration of struvite is shallow, and it was not detectable limits of XRD. The XRD patterns of the MgCO₃-based pellets after the experiments showed that hydromagnesite (Mg₅(CO₃)₄(OH)₂·4H₂O) was the dominant compound (Figure 7). It confirms that magnesium carbonate is present as hydromagnesite for phosphorus adsorption, and the pellets were not saturated during the experimental period.

Figure 7.

XRD patterns of MgCO₃-based pellets before and after adsorption at various flow rates.

The recovered phosphorus and ammonia as struvite can be used as a slow-release fertilizer. A recent study has indicated that struvite is slowly soluble in the presence of soil organic acids such as oxalic acid, malic acid, acetic acid, or citric acid and then uptake by roots.²⁶

The release of phosphorus is prolonged compared with conventional soluble fertilizers. Although struvite is only slightly soluble in water (1–5%), much previous laboratory-based work has suggested that struvite is as effective as highly soluble mineral P fertilizer as a source of P for plants in the presence of soil matrix.²⁶

4. CONCLUSIONS

In this study, the anoxic/aerobic MBR system and MgCO₃-based pellets effectively recovered phosphorus from municipal wastewater. The treatment of raw wastewater (i.e., before primary settling) that had high suspended solids blocked the column packed with MgCO₃-based pellets after only 3 days of operation due to solid materials’ fast accumulation.

Raw wastewater was replaced with primary effluent (i.e., after primary settling) to avoid media pores’ clogging. The MgCO₃-based pellets also showed effective removal of COD (45–50%) and NH₃–N (11–36%) from the primary effluent. In general, as the feed flow rate decreased from 10 L/d to 2.5 L/d, the overall removal efficiencies of TSS, COD, NH₃–N, and phosphorus increased. At the feed flow rate of 2.5 L/d (Run 4), approximately 94% of ortho-P recovered using the MgCO₃-based pellets. However, the faster flow rate might cause significant fluctuation in the removal efficiencies and affect phosphorus’s adsorption mechanism by reducing the contact time between MgCO₃-based pellets and phosphorus in wastewater.

In Runs 5 and 6, anoxic/aerobic MBRs were operated in combination with the MgCO₃-based pellet column. The MBR system achieved high removal efficiency of NH₃–N and COD from the primary effluent. However, the anoxic/aerobic MBR system is limited in phosphorus removal. The MBR system provided particle-free water (i.e., no TSS and VSS were observed in the membrane permeate) for the following MgCO₃ column. As a result, the MgCO₃ column was stably operated without the clogging issue.

The combination of the MBR system with the MgCO₃ column enhanced phosphorus recovery up to 73.1%. The phosphorus recovery capacity of the MgCO₃-based pellets was maintained at 0.47 mg ortho-P/g MgCO₃-based pellet during the continuous operation for 10 days. Overall, combining MBR with MgCO₃-based pellets could be an effective alternative method of phosphorus recovery from wastewater. However, it would be necessary to optimize EBCT by controlling the column dimension and flow rate for the maximum use of the P-adsorption capacity of MgCO₃-based pellets in further study.