1 Introduction

Phosphorus (P) is a nonrenewable resource, and it is predicted that global P reserves will be depleted within 100 years if no solution is taken (Smit et al. 2009). Since P is essential for plant growth, large amounts of P continuously enter the environment with the development of agriculture, causing the wastage of P resources and water eutrophication (Conley et al. 2009). Recently, several countries have implemented strict thresholds for P concentrations in the environment. Total phosphorus (TP) emissions should not exceed 1 mg L−1 in class 1B and 0.5 mg L−1 in class 1A (MEP 2002). The US EPA advised that TP levels in streams entering lakes should be below 50 μg of P L−1 (Loganathan et al. 2014). In response to the upgrading of emission standards, the development of more effective measures to mitigate P pollution and promote P recycling is urgently required.

Ecological ditches are one of the most effective methods for mitigating agricultural P pollution by virtue of their ability to remove P (Kröger and Moore 2011; Kumwimba et al. 2017). Emergent plants, in particular, are frequently utilized in environmental remediation because of their high P uptake, low cost, and visual attractiveness (Singh et al. 2023). Emergent plants can adsorb P directly and store it in their biomass. The P uptake rates of emergent plants have been reported to be 1–4 kg of dry matter m−2 yr−1 or more (Iamchaturapatr et al. 2007; Christopher 2013; Wu et al. 2023). However, many emergent plants wilt and rot during winter, which poses a risk of secondary water contamination; thus, they must be harvested regularly (Yu et al. 2019). Current methods for treating emergent plants mainly include composting and incineration; however, these methods suffer from inefficiencies and long processing times (Gusain et al. 2018). Therefore, the next crucial step is to create efficient and eco-friendly alternative utilization strategies for emerging plant residues.

Biochar is expected to be a reliable P-adsorption material because of its cost-effectiveness and unique properties that can be easily obtained by biomass pyrolysis in an anaerobic or oxygen-limited environment (Klinar 2016). With in-depth research into engineered biochar, the feedstock options for raw biochar are continuously increasing (Cheng et al. 2023), however, these were found to have a limited P adsorption ability because of the negative charge on their surface (Shepherd et al. 2017; Liu et al. 2022). To overcome this problem, metal-modified biochars have received increasing attention owing to their strong affinity for P (Palansooriya et al. 2021; Padilla et al. 2023). Currently, metals including iron (Fe), calcium (Ca), and magnesium (Mg) are frequently utilized to modify biochar because of their low cost and lack of secondary pollution (Jiang et al. 2018; Mitrogiannis et al. 2018; Chen et al. 2022). In addition, Fe-based biochars have the advantage of magnetic recovery, offering the ability to separate these from solution (Huang et al. 2019; Palansooriya et al. 2021). According to previous studies, Fe-modified biochar can increase the surface positive charge, thereby improving the anion exchange capacity and redox properties of the biochar (Ou et al. 2023). Wang et al. (2020a, b)’s study validated that Fe-modified biochar had a greater adsorption capacity (1.84 mg g−1) for P than raw biochar (no effect) and Palansooriya et al. (2021) also found comparable outcomes in their study of Fe(III)-loaded chitosan biochar. Mixed metal oxides/hydroxides containing multiple metal components possess the benefits of parent metals and produce synergies that improve their adsorption properties (Wu et al. 2020). Although reports are available on the effect of Fe-modified materials since over two years ago, little attention has been paid to Fe(II) and Fe(III) modification of biochars for P adsorption.

Additionally, after P adsorption, biochar has the potential to be functionalized as a P fertilizer. Ai et al. (2023) used corn stalk biochar after P adsorption as a fertilizer and observed that it significantly promoted plant growth. Nevertheless, the variation in the bioavailability of the P species adsorbed by biochar in actual water is not clear because most studies tested the performance of biochar as a P adsorbent in phosphate solutions. In this study, to reduce P discharge into external water bodies and to reuse plant residues, we prepared Fe(II)/Fe(III) co-modified biochars using three types of emergent plant residues for application in P adsorption from wastewater. The aims of this study were: (i) to investigate the P adsorption capacity of biochars derived from emergent plant residues, (ii) to explore the P adsorption processes and mechanisms of magnetic biochars (MBs), and (iii) to identify the P species in P-laden biochars and provide data for the effective support of using these as P fertilizers. The results of this study offer an efficient adsorbent for P adsorption and a greater understanding of the P species adsorbed by MBs, which will be helpful for the further utilization of MBs in P adsorption, fertilizer, and resource cycling engineering.

2 Materials and methods

2.1 Preparation of RBs

Three emergent plants (Acorus calamus L., Canna indica L., and Thalia dealbata Fraser) were collected from ecological ditches in Pinghu County, Zhejiang Province, China. Emergent plant residues were cleaned with deionized (DI) water before use, dried in an oven for 2 h at 105℃, and further dried at 60℃. The dried residue was cut into fragments smaller than 5 cm, crushed, and filtered through a 0.15 mm pore size filter. The treated residue was pyrolyzed in a programmable electric tube furnace (OTF-1200X, Hefei kejing materials technology Co., Ltd, China). The system temperature was raised to 650 °C at a pace of 10 °C min−1 and the high temperature was maintained for 0.5 h under N2 conditions (Zhao et al. 2022). The yields of the three biochars were approximately 29%. These raw biochars (RBs) were denoted as RB-A, RB-C, and RB-T and were produced from the residues of Acorus calamus L., Canna indica L., and Thalia dealbata Fraser, respectively.

2.2 Preparation of MBs

The MBs were prepared as follows: 2.5 g of each RB was placed into 50 mL of mixed iron (Fe) solution (0.01 mol FeSO4·7H2O and 0.01 mol FeCl3), heated at 60℃ for 30 min under slow stirring, cooled to 40℃, and then 5 M NaOH was added to maintain the pH at 10–11 (Dai et al 2020). The mixture was stirred for 10 min and aged for 24 h. The black precipitate was collected and cleaned with DI water and ethanol. The MBs were finally obtained after filtration and drying at 60℃ for 24 h. Modified RB-A, RB-C, and RB-T were named MB-A, MB-C, and MB-T, respectively. By dispersing the MBs evenly in pure water, aggregation of the suspended powder was observed under the attraction of magnets. Thus, the MBs were confirmed to be highly magnetic and could be successfully separated from water via magnetic separation (Additional file 1: Fig S1).

2.3 Adsorption experiments

The P adsorption ability of the MBs was assessed using batch adsorption tests. All the solutions used for the adsorption experiments were obtained from a livestock breeding farm in Kaihua, Zhejiang Province, China. The experiments were conducted at 120 rpm and 25 °C for 24 h. To evaluate the adsorption capacity of the  RBs, 100 mg of RBs were added to 40 mL of the diluted solution. Similarly, 100 mg of RBs were added to 40 mL of solution with the initial P concentration in the range from 0.8 mg L−1 to 43 mg L−1 to examine the adsorption abilities of the RBs. The adsorption kinetic tests were completed at 25 °C with the TP concentration of the solution at 20 mg L−1. Samples were periodically collected at a mixture time after 2, 4, 6, 8, 10, 20, 30, 60, 90, 180, 360, 720, and 1440 min. The P content of the solution was analyzed using the molybdenum blue colorimetric method (Lei et al. 2021). The experimental data were fitted with pseudo-first-order, pseudo-second-order, and intra-particle diffusion models to understand the adsorption kinetics.

The influence of the solution pH on adsorption was investigated by conducting experiments at pH levels between 3 and 11. The pH of the solution was adjusted using 0.2 M HCl or NaOH solutions. The method of pH drift was used to quantify the point of zero charge (pHPZC) of the adsorbent (Villanueva et al. 2014). In addition, possible interferences between co-existing anions (sulfate (SO42−), chlorine (Cl−), nitrate (NO3−), and bicarbonate (HCO3−)) and P adsorption by MBs were investigated. Different co-existing anions of the same concentration (1 M) were added to the solution under the same conditions, including pH. The P adsorption experiment followed the same procedure, with 100 mg of MBs added to 40 mL of solution and oscillated for 24 h at 120 rpm and 25 °C.

In the above experiments, the amount of P in the solution was determined using a UV–vis spectrometer (UV-9000, Shanghai Metash Instruments Co., Ltd., China) following the molybdate blue spectrophotometric method. The pH of each solution was measured using an HQ 11D pH meter (HACH, USA). The average values for each experiment were obtained from three replicates.

2.4 Characterization of biochars

Microscopic features and morphology of the biochars were characterized using scanning electron microscopy-energy dispersive spectroscopy (SEM–EDS; Gemini SEM–EDS 300; ZEISS, Germany). The crystal structures of the MBs were analyzed by X-ray diffraction (XRD; Bruker D8 Advance, Germany). The surface functional groups of the RMs and MBs were examined by Fourier transform infrared spectroscopy (FTIR, Nicolet is50 FTIR, Thermo Scientific, USA) with a wavenumber range of 400–4000 cm−1. X-ray photoelectron spectroscopy (XPS, Thermo Scientific Escalab 250xi, USA) was used to analyze the elemental states on RMs before and after adsorption.

2.5 Phosphorus species of the MBs before and after adsorption

The P species were determined using a slightly altered version of the Cade-Menun and Preston (1996) approaches. Specifically, P on the MBs before and after adsorption was extracted and examined using solution P-31 nuclear magnetic resonance (31P-NMR) spectroscopy. First, 2.0 g of samples and 40 mL of NaOH-Na2EDTA mixture solution (0.25 M NaOH and 0.05 M Na2EDTA) were put in centrifuge tubes, shaken gently at 20 °C for 16 h, then centrifuged at 5000 rpm for 20 min. One milliliter of the supernatant was used for inductively coupled plasma optical emission spectroscopy (ICP-OES) analysis. Freeze-dried samples were dissolved in 0.65 mL of deuterium oxide (including 0.84 mg mL−1 methylenediphosphonic acid (MDPA)), 0.65 mL of distilled water, 0.4 mL of 10 M NaOH, and 0.65 mL of NaOH-EDTA solution. The dissolved samples were vortexed for 10 min and centrifuged at 5000 rpm for 20 min, then 0.7 mL was transferred into 5-mm NMR tubes. The NMR parameters included 90° pulse, 0.68-s acquisition time, 4.0-s pulse delay, 10 Hz spinning, 20 °C, and 2400–2800 scans for different samples depending on the P concentration. The quantities of various P species were analyzed by calculating the relative integral of the peak area versus MDPA using MestReNova (version 7.0, Mestrelab Research, Spain). The position of orthophosphate was adjusted at 6.0 ppm, the peaks between 5.0 and 4.3 ppm were sequentially determined as orthophosphate monoesters, the peak around -0.7 ppm was determined as DNA, and the peaks from −4.0 to −4.2 ppm were determined as pyrophosphate (Liu et al. 2018).

2.6 Phosphorus release tests

The P release capability of the MBs after P adsorption (P-laden MBs) was measured using DI water as the release medium. The specific steps were as follows: 100 mg of P-laden MBs was added to 40 mL of DI water, and the mixture was shaken at 25 °C and 120 rpm for sampling at 1, 2, 3, 6, 10, 24, 48, 72, 120, 144, 168, and 192 h. The results were fitted to second-order release kinetics. The P-laden MBs were reintroduced into distilled water for cyclic P release at 72 h. The P content was determined using molybdate blue spectrophotometry. The P release experiments were conducted three times, and the average and standard deviation were calculated.

2.7 Statistical analysis

Data were analyzed using SPSS (version 26.0, IBM, USA) and the results were visualized using Origin 2021 (Origin Lab, USA). Significant differences in the P adsorption content were tested using a one-way analysis of variance (ANOVA) with the least significant difference (LSD) test at the 5% level (p < 0.05).

3 Results and discussion

3.1 Characteristics of biochars

SEM was used to observe the microstructures of RBs and MBs (Fig. 1). As shown in Fig. 1a–c, the surface morphology of the RBs appeared smooth and irregular, whereas the MBs exhibited disordered particles on the surface (Fig. 1d–f). The EDS results (Fig. 1j–l) confirmed that the Fe content in the MBs exceeded 20%, indicating that Fe-related complexes were successfully loaded onto the MBs surface. Simultaneously, the disordered particles on the MBs were iron oxides, which provide binding sites for P adsorption. Notably, the Fe content in MB-C reached 27.3%, indicating its potential for a higher adsorption capacity than that of the other MBs (Huang et al. 2021).

Fig. 1
figure 1

SEM images (a–f) and EDS spectra (g–l) of RBs and MBs

Full size image

The XRD patterns of the RBs and MBs revealed six peaks corresponding to crystalline substances (synthetic minerals) (Fig. 2a). The RBs were mainly composed of potassium chloride (KCl) and calcite (CaCO3), whereas the MBs contained magnetite (Fe3O4) and goethite (FeO(OH)). These results show that Fe modification changed the predominant mineral composition of biochars, with the major substance changing from KCl or CaCO3 to Fe3O4 or FeO(OH). The transformation of Fe(III) into FeO(OH) and Fe3O4 was consistent with the previous report of: FeCl3 → Fe(OH)3 → FeO(OH) → Fe3O4 (Li et al. 2016). Fe(III) loaded onto biochar reacted directly with P to form precipitates, whereas FeO(OH) interacts with P to produce complexes (Yang et al. 2018; Min et al. 2020). Furthermore, Fe(II) also contributed to the doping of Fe3O4 on biochars, enhancing the magnetic separation and adsorption capabilities of the biochar (Reguyal et al. 2017; Zhou et al. 2017; Dong et al. 2018). The Fe oxide layer on the MBs surface was the primary reason for the enhanced P adsorption.

Fig. 2
figure 2

The XRD spectra of RBs and MBs. Notes: ♣ Sylvite-KCl; ◆ Calcite-CaCO3 ♥ Magnetite-Fe3O4; ♠ Goethite-FeO(OH); • Chloromagmesite-MgCl2; ⁎ Magnesium Chlorate-Mg(H2O)6(ClO2)2

Full size image

3.2 Phosphorus adsorption capacities and mechanism of biochars

The capability of RBs for P adsorption was investigated using a 2.5 mg L−1 P solution, however, it demonstrated a desorption phenomenon of P from raw materials (Fig. 3a), possibly owning to the negative surface charge on RBs, which was not conducive to P adsorption (Ahmad et al. 2014). Therefore, RBs were excluded from the subsequent P adsorption studies. The P-adsorption performance of the MBs is shown in Fig. 3b. At 0.8 mg L−1 (low concentration of P), the adsorption efficiency of all MBs was above 90%. As the solution's P content increased from 0.8 to 43.0 mg L−1, the P adsorption efficiency decreased owing to the limited availability of adsorption sites. However, at the initial concentration of 22.0 mg L−1, the adsorption efficiency of MB-A, MB-C, and MB-T was still 61.7%, 76.0%, and 57.2%, respectively. The results showed that the MBs had a stronger P-adsorption capacity than the RBs, and that MB-C was the most efficient among the three MBs.

Fig. 3
figure 3

P adsorption performance of RBs (a) and MBs (b); adsorption kinetics of MBs fitting by Pseudo-first-order and Pseudo-second-order model (c); adsorption kinetics of MBs fitting by intraparticle diffusion model (d)

Full size image

To enhance our comprehension about the P adsorption process onto MBs, adsorption studies were conducted with varying contact times. Kinetic models play a significant role in elucidating the pathways and mechanisms associated with adsorption processes (Gong et al. 2019). As depicted in Fig. 3c, the P adsorption of all MBs was rapid during the first two hours before slowing down and reaching equilibrium at approximately 6 h. The initial rapid adsorption can be attributed to the large number of binding sites and the electrostatic attraction between the phosphates and positively charged Fe or Fe oxides (Shan et al. 2021). Notably, this behavior was consistent across different types MBs. The adsorption parameters for the pseudo-first-order and pseudo-second-order models are listed in Table 1. Among all MBs, the pseudo-second-order model exhibited a better fit (R2 ≥ 0.87) than the pseudo-first-order model (R2 ≥ 0.63), indicating that chemical adsorption was the rate-determining step (Yi et al. 2019).

Table 1 The parameters of kinetic and intra-particle diffusion models about P adsorption onto MBs
Full size table

The experimental data were fitted using an intra-particle diffusion model to enhance our understanding of the rate-limiting steps. This model was used to explain the diffusion of P onto the adsorbent particle pores (Tran et al. 2017). The P adsorption process was divided into three sections (Fig. 3d), and the related parameters  are listed in Table 1. The results revealed that the slope of the first stage was the highest of the three stages, implying a gradual decrease in the reaction rate with increasing contact time (Malash and El-Khaiary 2010). The rapid adsorption stage was caused by an external mass transfer phase, which may have been influenced by the high density of P-binding sites on the surface of the MBs (Fang et al 2022). Subsequently, the rate of P adsorption decreased, and P began to diffuse within the solid biochar particles, which is known as an intra-particle diffusion process (Yu et al. 2021). The third stage of adsorption reached equilibrium when no accessible adsorption sites were present on the MBs (Yao et al. 2013). Additionally, the first lines did not pass through the origin, indicating that intra-particle diffusion was not the sole rate-limiting process (Li et al. 2019).

The functional groups of RBs, MBs and P-laden MBs were examined by FTIR (Fig. 4b–d). Wide absorption bands approximately 3390 cm−1 were observed in MBs samples and were linked to the stretching vibrations of −OH, which were vibrationally enhanced with the addition of Fe (Yin et al. 2018). The 1630 cm−1 peak corresponded to the stretching vibration of  −OH, arising from the distortion and movement of adsorbed water molecules (Ajmal et al. 2018). The peak at approximately 1085 cm−1 was associated with the deformation vibration of  P−O in iron phosphates, providing evidence for the presence of phosphate in the MBs (Zhang et al. 2019). After adsorption, these peaks became more pronounced, indicating successful adsorption of P onto the MBs surface. The characteristic vibration of  C−O bonds was attributed to the peaks observed at 980 cm−1 and 797 cm−1 (Zhao et al. 2022). These two peaks weakened after P adsorption, suggesting that phosphate entered the exchange layer and was exchanged (Jiang et al. 2020). The peak at 580 cm−1 was attributed to the Fe–O stretch vibration of magnetite and it disappeared after P adsorption, which might be attributed to the strong adsorption of phosphate on the metal oxide surface (Wang et al. 2023). A novel peak at 1390 cm−1 was related   to the stretching vibration of Fe–O–P, which belonged to inner-sphere surface complex (Wang et al. 2021a, b). P adsorption onto MBs was controlled by the inner-sphere surface complex formation process while maintaining structural and chemical stability.

Fig. 4
figure 4

The FTIR spectra of RBs (a), MBs (b), and P-laden MBs (c)

Full size image

Figure 5 shows that the P 2p peak intensity of MBs was stronger after P adsorption, demonstrating that P was adsorbed on the MBs surface. The high-resolution P 2p spectrum (Fig. 5b) showed a clear separation, revealing two individual peaks (2p1/2 and 2p3/2). Novel bonds were created by P adsorption. The primary component at binding energy (BE) = 133.0–133.4 eV was pentavalent tetra coordinated P (PO43−) and the peak around 134.0 eV was attributed to HPO42− (Xie et al. 2014). Moreover, the BE of the P 2p peaks of the three MBs differed, with MB-CP (P-laden MB-C) having the highest BE, indicating that it exhibited the strongest P binding (Gao et al. 2019). Additionally, Fe peaks were clearly visible in the XPS survey of the MBs, indicating that Fe was successfully loaded onto the MBs (Fig. 5c, d). The two peaks visible in the Fe 2p spectra were Fe 2p3/2 and Fe 2p1/2, with the former being narrower and stronger than the latter (Yamashita and Hayes 2008). The 711.4 eV and 713.8 eV of the Fe 2p3/2 spectra were associated with Fe(II) and Fe(III), respectively (Tang et al. 2023). The satellite peaks observed at approximately 719.0 eV and 732.0 eV were attributed to Fe 2p3/2 and Fe 2p1/2 respectively, confirming the Fe loaded on the MBs existed in the form of γ-Fe2O3 (Hao et al. 2018; Tu et al. 2020). After P adsorption, the peak position of Fe 2p remained similar, indicating that Fe maintained its oxidation state during the adsorption process (Yang et al. 2022). The binding energies of the main peaks were consistent with the slightly shifted satellite peaks. These phenomena revealed that the MBs adsorbed P by relying on active Fe sites.

Fig. 5
figure 5

XPS spectra of  MBs (a, c) and P-laden MBs (b, d)

Full size image

pH is a key influencing factor that affects the charge on the adsorbent’s surface and the dissociation equilibrium of phosphate; therefore, the effect of starting pH levels on P adsorption by MBs and the equilibrium pH of the solution after P adsorption were explored (Fig. 6a–c). Moreover, we found that the pHPZC values of MB-A, MB-C and MB-T were 3.85, 3.33, and 3.28, respectively (Fig. 6d). As the initial pH rose from 3 to 11, the P adsorption capacity of MB-A, MB-C, and MB-T decreased from 7.8 to 5.5 mg g−1, 7.9 to 5.2 mg kg−1, and 6.4 to 5.4 mg g−1, respectively. When the solution pH fell below the pHPZC, a strong electrostatic attraction developed between the binding sites on the MBs and H2PO4− ions, increasing P adsorption (Hartley et al. 1997). As the pH increased, the charge on the surface of the adsorbent shifted in the negative direction and formed a repulsive charge with the phosphates, causing the P adsorption capacity to decline (Liu et al. 2021). Additionally, the solution pH increased after P adsorption, which could be attributed to the complexation processes between P and the hydroxyl groups on the MBs, which consisted of the formation of inner-sphere surface complexes (Xu et al. 2019). In the preliminary stage, electrostatic attraction occurred between the negatively charged phosphates and the electropositive MBs surfaces loaded with amorphous hematite with a high zeta potential. During the subsequent adsorption process, inner-sphere complexation through ligand exchange played a significant role between the surface-bound Fe oxides/hydroxyl groups and phosphates (representative equilibrium; Eq. (1) is given below, taking H2PO4− as an example).

$${\text{B}}_{{\text{s}}} - {\text{Fe}} - {\text{OH}}\, + \,{\text{H}}_{{2}} {\text{PO}}_{{4}}^{ - } \, = \,{\text{B}}_{{\text{s}}} - {\text{Fe}}\left( {{\text{H}}_{{2}} {\text{PO}}_{{4}} } \right)\, + \,{\text{H}}_{{2}} {\text{O}}$$
(1)
Fig. 6
figure 6

Effects of pH on P adsorption capacity and equilibrium pH of MB-A (a), MB-C (b), MB-T (c); Zeta potential of MBs (d)

Full size image

As shown in Table 2, the P adsorption capacity was negatively affected by the ions SO42−, Cl−, NO3−, and HCO3−. Of these, HCO3− imposed the most significant effect on the P adsorption capacity, decreasing it by 30.3%, 11.5%, and 27.0% for MB-A, MB-C, and MB-T, respectively. Because HCO3− had a higher distribution coefficient, causing significant interference to phosphate adsorption (Pitakteeratham et al. 2013). Another non-negligible negative effect was from SO42− on P adsorption with obvious reduction rates of 19.1%, 9.7%, and 19.7% for MB-A, MB-C, and MB-T, respectively. This may be owing to SO42− competing for complexation reaction sites with phosphate (Drenkova-Tuhtan et al. 2013). Conversely, Cl− and NO3− were related to outer sphere compounds forming (Wang et al. 2020a, b). They showed a slight adverse impact on P adsorption, indicating that the affinity for Cl− and NO3− was relatively low and inner-sphere adsorption plays a primary role in P adsorption (Fang et al. 2022). MB-C adsorbed more P than MB-A and MB-T for the different co-existing anions, validating that MB-C had the strongest P adsorption ability.

Table 2 The effects of co-anions about P adsorption onto MBs
Full size table

3.3 Phosphorus species of MBs before and after adsorption

The NaOH-EDTA extracted 59.5–66.0% of the TP from the MBs before and after P adsorption, with similar P species extracted. The results of the 31P-NMR spectroscopy analysis showed that MBs contained only orthophosphate and pyrophosphate before adsorption, while novel peaks of orthophosphate monoesters (5.0−4.3 ppm) and DNA (~ 0.7 ppm) were observed in the MBs after adsorption (Table 3, Fig. 7) (Liu et al. 2018), indicating that solutions from livestock and poultry farms contain orthophosphate monoesters and DNA. Pyrophosphate has been reported to be directly involved in microbial activities and have a significant impact on P cycling (Xie et al. 2019). Furthermore, orthophosphate was the dominant P species in MBs  both before and after adsorption, ranging from 0.30–1.37 mg g−1 in MBs and 7.99–9.44 mg g−1 in P-laden MBs. Orthophosphate is the major P species directly available for plant uptake(Nash et al. 2014), which suggests that MBs possess great potential as soil fertilizer to facilitate P resource recycling.

Table 3 P species of MBs before and after P adsorption
Full size table
Fig. 7.
figure 7

31P-NMR spectra of MBs (a) and P-laden MBs (b)

Full size image

3.4 Phosphorus release from P-laden MBs

The desorption kinetics of the P-laden MBs revealed a slow release of P and the release rate gradually decreased within 192 h (Additional file 1: Fig S2). This might be attributed to the high concentration of P in the solution, where P release and adsorption approached a dynamic equilibrium. The data showed a good fit with the second-order kinetic model, with R2 values of 0.98, 0.96, and 0.97 for MB-AP, MB-CP, and MB-TP, respectively. The P release capacities of MB-AP, MB-CP, and MB-TP were 8.24, 9.61, and 7.27 mg L−1, respectively. In three consecutive cyclic release experiments, the amount of P released in each cycle was approximately consistent, indicating the potential slow-release capability of the P-laden MBs (An et al. 2021). Given that the P released into the soil by P-laden MBs is promptly absorbed by plants, possibly intensifying the further release of P at soil–water interfaces (Yao et al. 2013), the slow release of P will contribute to reduce P loss and enhance P bioavailability for plants (Wang et al. 2021a, 2021b). The release patterns of the three treatments were consistent, with MB-CP exhibiting the best P release capability.

3.5 Environmental implication

In this study, MBs with stronger P adsorption capacities and magnetism were prepared from three types of emergent plant residues, enhancing P adsorb and enabling efficient recovery using magnets. The 31P-NMR results showed that orthophosphate accounted for the largest proportion of the P constituents in the P-laden MBs and is considered an ideal P fertilizer. Developing sustainable fertilizer as an alternative to conventional P fertilizer will promote environmentally friendly P supply and plant uptake efficiency. We anticipate that MBs-based absorbent will facilitate the reuse of P in additional scenarios, which are also under investigation by our research group.

4 Conclusions

In this study, we provide a novel approach for utilizing emergent plant residues from ecological ditches as P-adsorption materials. By successfully loading Fe compounds onto the surface of the three biochars, as validated by adsorption kinetics, MB-C was observed to have the best P adsorption capacity. Combined with the SEM-EDS, XRD, FTIR, and XPS characterization data analysis, inner-sphere complexation and electrostatic attraction were confirmed to play major roles in controlling P adsorption by the MBs. The P species adsorbed by the MBs was mainly orthophosphate, proving the potential of the MBs as carbon-based P fertilizer. Overall, this study provides an innovative idea for utilizing emergent plant residues as sustainable P resources, as well as proposes a feasible and eco-friendly scheme to reduce P pollution and promote P recycling in agricultural.