1. Introduction
As an essential element for food production, the demand for phosphorus has been rising sustainedly in recent years [
1]. However, phosphorus resources are non-renewable, and global phosphorus resources with commercial value are expected to be depleted in the next 50–100 years [
2]. Sewage treatment plants are an important phosphorus sink. Approximately 2.7 million tons of phosphorus enters sewage plants globally each year, representing 15% of total phosphorus emissions [
3]. More than 90% of phosphorus in sewage ends up in sludge, where the mass fraction of phosphorus can be as high as 5–15% [
4]. Therefore, the recovery of phosphorus from sludge would effectively alleviate the phosphorus resource crisis.
Sewage sludge contains pathogenic bacteria, microplastics, and heavy metals, thus its direct use as a fertilizer has been banned in many countries [
5,
6]. Therefore, to recover phosphorus in sludge, the release or solubilization of phosphorus from the waste solid is usually though to be the first step, though which phosphorus is separated from the co-existing hazardous materials [
7]. Among various kinds of methods promoting the release of phosphorus from sludge, hydrothermal treatment at a relatively low temperature has been chosen as a priority technique due to its satisfactory phosphorus releasing effect and mild reaction condition [
7,
8,
9,
10,
11,
12,
13]. Kuroda et al. reported that nearly all poly-phosphorus in activated sludge could be released under hydrothermal treatment at 70 °C for 1 h [
8]. Xue et al. found that the release of orthophosphate in concentration is 90 mg/L when the treatment temperature and time were 50 °C and 1 h, respectively [
9]. Similarly, a 51.6 mg/L of orthophosphate in the sludge supernatant was observed by Zhang and Kuba after hydrothermal treatment at 70 °C [
10]. In addition, the release performance is usually intensified by adding acid-base reagents, since phosphorus bonded with metal ions and organic matter tend to be dissolved in acidic or alkaline environment [
14,
15,
16,
17]. Cheng et al. reported that compared to that under neutral conditions, the phosphorus release from sludge increased by 30% and 34% when the pH was adjusted to 2 and 12, respectively [
17].
Both hydrothermal and acid-base treatment would cause simultaneous leaching of Fe, Al, Ca, and organic matter from sludge, lowering the purity of recovered phosphorus. Li et al. reported that the purity of recovered struvite was less than 20% when coexisting Ca
2+ was more than 500 mg∙L
−1 in wastewater [
18]. In addition, Fe
2+ and Al
3+ impurities might result in the formation of non-bioavailable phosphorus in the recovered product [
19]. Therefore, purification of the sludge derived phosphorus solution is required before recovery [
7]. Na
2S is added to the hydrothermally treated sludge supernatant (HTSS) to precipitate coexisting metal cations (by the reaction with S
2−) to obtain a liquid containing only phosphorus. However, the addition of S
2− to the acidified HTSS may result in the formation of toxic H
2S [
20]. Adjusting the pH of the acid leachate of sludge to neutral allows phosphorus in the solution to be precipitated as Al–P and separated from the heavy metals that are still dissolved. However, subsequent conversion of Al–P to more readily bioavailable Ca–P is still required, thereby complicating the process [
7]. In addition, coexisting ions and organic matter may be removed using nanofiltration and cation exchange resins, albeit at a high cost [
21,
22].
Selective adsorption of phosphorus from the HTSS followed by desorption to obtain a phosphorus-only solution is a possible route of purification [
19,
23]. Biswas et al. prepared Zr(IV) modified orange waste gel to adsorb phosphorus from an acid leachate of sludge incineration ash, resulted in efficient elimination of Ca, Fe, and Al [
24]. Yu et al. used a Zirconium-containing material to selectively adsorb phosphorus from sludge leachate, followed by desorption and precipitation to obtain hydroxyapatite containing heavy metals (As, Cd, Cr, and Pb) at concentrations below the fertilizer safety standard values [
6]. Lundehøj et al. investigated the adsorption performance of Zn
2-Al-LDH on phosphorus in the acid leachate of sludge and found that Zn
2-Al-LDH selectively adsorbed phosphorus from the leachate, but could not effectively eliminate SO
42− [
25].
Purification by adsorption-desorption is currently performed primarily in adsorption columns, which might suffer clogging during long-term operation [
26]. Magnetic adsorbents can effectively resolve this issue. Magnetic adsorbents are dispersed in the liquid during adsorption process, and are magnetically separated after the reaction, thereby the clogging of adsorption column could be avoided [
27]. Previously, we prepared a novel adsorbent, magnetic zirconia (MZ). The saturated magnetization of MZ is 23.65 emu/g, which enabled the magnetic separation after use [
27]. The maximum P adsorption capacity is 42.19 mg/g, being higher or at least comparable with the recently reported Zr-containing adsorbent [
26]. MZ exhibited good selectivity in the extraction of phosphorus from the secondary effluent of a sewage plant with a recovery rate higher than 90%. MZ could be repeatedly used in several cycles and its performance was rather stable [
26,
28]. Therefore, MZ is a promising adsorbent for the purification of the HTSS.
In this study, we investigated the selective adsorption of phosphate in the HTSS by MZ under acidic conditions and evaluated the purification effect of the adsorption-desorption treatment by testing the purity of the sludge-derived struvite. The adsorption-desorption mechanisms were also investigated though spectrum analysis.
2. Materials and Methods
2.1. Sludge Source
The excess sludge was taken from the secondary sedimentation tank of Xi’an No. 4 Wastewater Treatment Plant. Solid content of the raw sludge is 9.54%, which was higher than the reported average level of excess sludge [
8,
9]. So, it was diluted with distilled water to achieve a 5% solid content. The properties of the sludge mixture are presented in
Table 1.
2.2. Hydrothermal Treatment
The hydrothermal treatment of sludge was performed in a glass flask placed in a water bath to control the reaction temperature. Mechanical stirring at 200 rpm ensured the uniformity. Based on the previous studies, the hydrothermal treatment was performed at 75 °C for 1 h [
29]. To find the optimized pH for phosphorus release during the hydrothermal treatment, the pH of the sludge mixture was adjusted to 3, 4, 5, 6, 7, 8, 9, 10, and 11 with 1 mol/L of HCl and NaOH. After treatment, the sludge mixture was centrifuged at 300 rpm for 10 min, after which the total phosphorus (TP) and orthophosphate (PO
43−) concentrations in the hydrothermally treated sludge supernatant (HTSS) were determined using a colorimetric method. The solid residue of hydrothermal treatment at pH 3 was cleaned, dried, ground, and passed through an 80 mesh sieve; its phosphorus fractions were determined using the SMT method [
15].
2.3. Preparation of MZ
FeSO
4∙7H
2O, FeCl
3∙6H
2O, and ZrOCl
2∙8H
2O were dissolved in 500 mL of ultrapure water to achieve a molar ratio of Fe
2+/Fe
3+/Zr = 1.33/2.66/1. The pH of the mixture was neutralized to 7.6 using 6 mol/L NaOH solution under mechanical stirring at 300 rpm. After aging at 60 °C for 16 h, the solid product was washed, dried, and sieved [
26]. The obtained MZ was characterized by X-ray fluorescence (PW2404, PHILIP, Amsterdam, The Netherlands) and X-ray diffraction (D8ADVANCE, BRUKER-AXS, Berlin, Germany) to investigate its chemical composition and crystal phase. The magnetism and surface charge of MZ were measured by vibrating sample magnetometer (Lakeshore, Columbus, OH, USA) and zeta potential analyzer (BECKMAN COULTER, Brea, CA, USA).
2.4. Adsorption Experiment
A certain mass of MZ was weighed and placed in a 50 mL centrifuge tube, and 40 mL of HTSS was added. The HTSS was obtained under the following reaction conditions: treatment time of 1 h, reaction temperature of 75 °C, and initial pH of 3. The mixture was reacted at 180 rpm at 25 °C for a certain time and then centrifuged to determine the phosphate concentration in the supernatant. The calculation equations for phosphate adsorption and adsorption rate are given in Equations (1) and (2):
where
Qe represents the phosphate adsorption amount (mg/g),
η denotes the phosphate adsorption rate (%),
C0 and
Ce represent the phosphate concentration in the supernatant before and after adsorption, (mg/L), and
m denotes the mass of the MZ added (g).
In the dosage experiments, MZ was added at 0.1–0.8 g to attain the dosage of 2.0–20 g/L, and the reaction time was 48 h. In the kinetic experiments, 15 g/L MZ was added, and the reaction time was 0.5–48 h. In the coexisting ion effect experiments, solutions with single and multiple coexisting ions were prepared according to the concentrations of phosphate and coexistence ions, as well as pH in the HTSS (
Table 2). The dosage and reaction time of MZ was 15 g/L and 48 h, respectively.
2.5. Desorption Experiment
MZ reacted with the HTSS at a dosage of 15 g/L, and the mixture was magnetically separated after 48 h. The phosphate concentration in the supernatant was measured, and the adsorption amount
Qe was calculated according to Equation (1). Subsequently, 40 mL of desorption agent was added and reacted for 48 h. After the reaction, the phosphate concentration in the solution was measured by solid–liquid separation, and the desorption amount was calculated. The desorption amount and desorption rate were calculated according to Equations (3) and (4):
where
Qd represents the amount of phosphate desorbed (mg/g),
Cd represents the concentration of phosphate in the desorption solution (mg/L), and
m represents the mass of MZ added (g).
where
η′ represents the phosphate desorption rate (%).
Unitary and binary desorption agents were used. The unitary desorption agents were NaOH and NaCl solutions with concentrations of 0.1, 0.5, and 1 mol/L. The binary desorption agents were mixtures of NaOH and NaCl with the following compositions: 0.1 mol/L NaOH + 0.1 mol/L NaCl, 0.1 mol/L NaOH + 0.5 mol/L NaCl and 0.1 mol/L NaOH + 1 mol/L NaCl. The concentrations of coexisting ions in the desorption solutions were also determined using ion Chromatography (Essentia IC-16, Shimadzu, Japan).
2.6. Recovery of Struvite from Desorption Solution
Phosphorus was recovered under the following conditions: Mg:P:N molar ratio of 1.2:1:1.6, pH = 9, reaction temperature of
T = 25 °C, stirring rate of 200 rpm, and stirring time of
t = 10 min [
22]. MgCl
2 and NH
4Cl were used as the source of Mg
2+ and NH
4+. The pH of the desorption solution was adjusted with HCl and NaOH solutions. After precipitation, the obtained product was separated, then its crystal phase and the surface morphology were characterized using X-ray diffraction patterns (D8ADVANCE, BRUKER-AXS, Berlin, Germany) and scanning electron microscopy (SEM; HATACHI TM3000, Kyoto, Japan), and the phosphorus recovery efficiency (PRE) was calculated according to Equation (5):
where
R represents the PRE (%);
Cb represents the phosphate concentration in the solution before recovery (mg/L), and
Ca represents the phosphate concentration remaining in the supernatant after recovery (mg/L).
To measure the struvite purity, the precipitate was dissolved by HCl, and the NH
4+-N content of the solution was determined using ion chromatography (Essentia IC-16, Shimadzu, Kyoto, Japan). The purity of struvite product was calculated according to Equation (6) [
18]:
where
P represents the purity of struvite in the precipitate (%),
CN denotes the concentration of NH
4+-N in the product digestion solution (mg/L),
V represents the volume of the precipitate digestion solution (L), 14 is the atomic weight of nitrogen (g/mol),
MMAP is the molecular mass of struvite (g/mol), and
mprecipitate is the mass of the precipitate (g).
2.7. Reusability of MZ
The reusability of MZ was investigated in five continuous cycles of adsorption, desorption and regeneration experiments. MZ (0.6 g) and HTSS (40 mL) were added to a centrifuge tube and reacted at 25 °C and 180 rpm. After 48 h, the adsorbent was magnetically separated, and the remaining phosphate concentration in the supernatant was measured. The adsorption amount was calculated according to Equation (1). Upon washing with deionized water, the desorption agent was added. After 48 h of desorption, the adsorbent was magnetically separated, and the phosphate concentration in the desorption solution was determined. The amount of phosphate desorption was then calculated according to Equation (3). Regeneration was performed by adding 40 mL of deionized water to the adsorbent and adjusting the pH of the mixture to 4.5 using 0.1 and 0.01 M HCl. The adsorbent was then magnetically separated, and the supernatant was discarded, completing the regeneration. After regeneration, a new cycle of adsorption, desorption, and regeneration was conducted.
2.8. X-ray Photoelectron Spectroscopy (XPS) and Electron Dispersive Spectroscopy (EDS)
The adsorption/desorption mechanism was characterized by XPS (Axis Ultra, Kratos Analytical, Manchester, UK). XPS spectra were corrected using C 1s spectra (standard peak level 284.8 eV), and data processing and peak search were performed using Casa XPS software. The surface elemental compositions of unreacted MZ and MZ subjected to five adsorption–desorption–regeneration cycles were characterized using EDS (TM3000, Hitachi, Tokyo, Japan).