Sludge-Based Superparamagnetic Nano-Sorbent Functionalized by Lanthanum Silicate Nanorods for Phosphorus Adsorption and Fertilization

Abstract

Phosphorus (P) recovery from wastewater is considered to be a positive human intervention towards sustainable P use in the global P cycle. This study investigated the feasibility of synthesizing a superparamagnetic nano-sorbent that was functionalized by lanthanum silicate nanorods (NR_La-Si) using drinking water treatment sludge (DWTS), evaluating both its P adsorption capacity and fertilization effect. The DWTS-based La-modified P nano-sorbent (P-sorbent _D) exhibited complicated but single-layer-dominant adsorption for phosphate, with a maximum adsorption capacity up to 26.8 mg/g, which was superior to that of most of the similar sludge-based P-sorbent. The NR_La-Si-modified P-sorbent _D was identified with several characterization techniques and the leaching metal elements from the nano-sorbent were tested, which were below the limits proposed by the Food and Agriculture Organization of the United Nations. In addition, the growth and vigorousness of Arabidopsis thaliana indicated that the exhausted P-sorbent _D could be used as a potential water-soluble moderate-release P fertilizer, which was also confirmed by the well-fitted P uptake model and the P desorption pattern from the sorbent–fertilizer. The doped lanthanum silicate nanorods could play the dual role of P complexation enhancement and health/growth promotion. In light of this, this study proposed a new way of reclaiming DWTS as a P-sorbent for fertilization, offering new insights into the path toward “closing the P loop”.

1. Introduction

The management of abundant drinking water treatment sludge (DWTS), the by-product generated during the production of drinking water in treatment plants, continues to pose a global challenge due to its potential risks to human health and the environment. The composition of DWTS primarily consists of ferric/aluminum oxy-hydroxides or oxides since it is the byproduct generated by the process of clarifying source water. This involves using aluminum or iron salts as coagulants to eliminate color, turbidity, and humic substances [1]. DWTS should be properly treated and disposed of. Recent studies have explored the potential of utilizing it for pollutant control [2,3]. For instance, it was reported that DWTS could be effectively transformed into value-added eco-concrete blocks through CO₂ curing. Additionally, it could also be repurposed as an adsorbent to capture heavy metals (Pb(II), Cd(II), and Ni(II)) from wastewater [4]. Sun et al. conducted experimental investigations to assess the feasibility of employing iron-rich DWTS to control sulfide levels in sewers [5]. More recently, Xie et al. studied the anerobic co-digestion of primary sludge with DWTS and observed significant levels of removal regarding phosphate and hydrogen sulfide after the addition of DWTS [6]. Extensive studies have also demonstrated the effectiveness of DWTS use in soils or filter media for P adsorption or immobilization from wastewater due to the presence of Al species in DWTS [7,8]. However, there have been limited efforts to re-synthesize and convert the DWTS into superparamagnetic nano-sorbents for P capture and no attempts have been made to evaluate the fertilization efficiency of P-loaded sorbents.

Actually, the components found in DWTS, such as trivalent ions (Al³⁺, Fe³⁺, etc.) and divalent ions (Ca²⁺, Mg²⁺), originating from the raw source water, bear a resemblance to the composition of LDHs (layered double hydroxides) and other metal hydroxides [9,10,11]. This provides us with an important insight into an innovative approach to reusing DWTS as an LDH-based superparamagnetic nano-sorbent for P fertilization. In recent decades, researchers have attempted to optimize the synthesis process and enhance adsorption capacity and selectivity by doping transitional elements onto LDHs. Up to now, a series of nanoparticles consisting of different LDHs and doped with oxides of different transitional elements, like zirconium, cerium, lanthanum (La), and hafnium, have been synthesized and assessed. Among these, the La-doped nano-sorbent seems superior due to its high P adsorption capacity, high selectivity for phosphate, excellent reusability, and minimal generation of secondary pollutants [12,13]. In addition, it is noteworthy that the application of La nanorods to improving plant yields has shown remarkable effectiveness in disease suppression, suggesting that it could serve as an effective and sustainable strategy with which to safeguard crops [14]. In addition, fertilization with La-doped Fe₃O₄@SiO₂ nanoparticles offered advantages over using fresh or semi-fresh dewatered DWTS as it effectively avoided adverse effects on the plant, which, due to the synthesis and preparation, could filter out the acrylamide and heavy metals [15]. Furthermore, some core–shell nanocarriers showed potential in enhancing the efficacy of various pesticides and enhancing disease resistance [16,17]. Until now, though a lot of work has been conducted for P adsorption with various sorbents [18], the re-utilization of exhausted DWTS-based nano-sorbents as P fertilizers and the fertilization efficiency of novel water-soluble P fertilizer has rarely been reported. To fill this gap, this study focused on whether DWTS-based paramagnetic nanoparticles doped with La can fill the gap of sludge disposal and P fertilizer reuse. Therefore, this study aimed to (1) synthesize La-doped superparamagnetic nano-sorbents using DWTS; (2) investigate the P adsorption performance and physiochemical properties of the new P-sorbent; (3) address the impact of the new sorbent–fertilizer on crop growth and nutrient uptake; and (4) model the P uptake by the crop and compare it with the P desorption pattern seen for the sorbent–fertilizer.

2. Results and Discussion

2.1. Phosphate Adsorption by DWTS-Based Sorbent

The P adsorption performance of both P-sorbent _D and P-sorbent _P are presented in Figure 1. The results showed that 1 g of P-sorbent _D was able to recover 97.2% of P from 100 mg of P/L phosphate solution. Adsorption isotherm data were utilized for modeling and the isotherm parameters were shown in

Table 1. Langmuir and Freundlich adsorption isotherm parameters of the P-sorbent _D and P-sorbent _P.

Items		P-Sorbent D	P-Sorbent P
Langmuir model	qm	26.8	165.5
	KL	0.54	0.67
	R2	0.989	0.979
Freundlich model	Kf	3.0	1.25
	n	4.18	1.77
	R2	0.955	0.945
Temkin model	KTlnA	7.73	44.36
	KT	0.039	0.25
	R2	0.867	0.771
Dubinin–Radushkevich model	KD	3905.39	3174.99
	qm	20.17	119.64
	R2	0.834	0.898

, suggesting the P adsorption fitted well to both Freundlich and Langmuir models for both types of adsorbents. Additionally, the Langmuir model exhibited a slightly higher correlation coefficient, suggesting better fitting compared to the Freundlich model. The Freundlich isotherm model allowed for multiple-layer adsorption of phosphate onto the nano-sorbent surface and properly represented the adsorption data at low and intermediate concentrations on heterogeneous surfaces. The Langmuir model also fitted the experimental data well, indicating that the binding energy on the whole surface of the P-sorbent p was fairly uniform. Namely, the whole surface exhibited identical adsorption behaviors, indicating that the adsorbed phosphates formed an almost-complete monolayer covering of the nano-sorbent particles and did not interact or compete with one another. Furthermore, the principal P adsorption mechanism was identified as chemosorption using well-fitting Langmuir model data. In contrast, as shown in Figure 1c,d, the adsorption isotherms of Temkin and Dubinin–Radushkevich were found to be less accurate in representing the equilibrium data for both P-sorbent _D and P-sorbent _P. Meanwhile, the isotherm parameters shown in Table 1 suggested that P adsorption was not fitted to Temkin model and Dubinin-Radushkevich model, which showed that the adsorption of phosphorus by P-sorbent _D did not belong to the adsorption reaction of molecules and micropore adsorption in the multi-substance system.

As indicated by q_max of 26.8 mg of P/g P-sorbent _D and 165.5 mg of P/g P-sorbent _P, respectively, the doped La silicate nanorods could efficiently enhance the complexation with H₂PO₄⁻ or HPO₄⁻, demonstrating stronger stability and higher binding energy, as revealed in the previous computational modeling [12]. Therefore, the DWTS-based sorbent was about 5 times less common in adsorption sites compared to the sorbent synthesized with pure chemicals.

In contrast to other synthetic DWTS-based sorbents (shown in

Table 2. Comparison of DWTS-based sorbent with other sorbents on P adsorption.

Absorbent	Adsorption Capacity (mg P/g)	Dosage (g/L)	C0 (mg/L)	pH	Reference
CSA	2.99	2.5	35	2~9	[19]
SBC	4.29	1	5	3~11	[20]
Fe-WAS	8.5	2	20	2~8	[21]
CSW-T	15.7	7.5	130	3~9	[22]
Fe-FSC	46.8	1	40	3~8	[23]
DSBC-700 °C	51.793	0.5	200	3~9	[24]
WAS-Ca900 °C	83.95	0.3	80	2~11	[3]
P-sorbent D	7	1	50	7	This study

Note: Abbreviations in Table 2: CSA (carbonized sludge adsorbent); SBC (modified sludge/biomass ash ceramsite); Fe-WAS (waste–activated sludge); CSW-T (the construction and demolition waste sludge); Fe-FSC (Fe-loaded sulfoaluminate cement).

), the P-sorbent _D in this study exhibited fairly high P adsorption content. Though it had a substantially greater P capture efficiency, its synthesis procedures were far simpler and more energy-saving than those of the sorbents DSBC-700 °C (dewatered dry-sludge biochar) and WAS-Ca900 °C (dewatered municipal sludge for calcium-based biochar adsorbent).

2.2. Fertilization Effect of Exhausted P-Sorbent _D in Hydroponic Culture

How Arabidopsis thaliana responded to external stimuli of P-loaded sorbents, such as the new P fertilizer, is shown in Figure 2. It is indicated that varied amounts of fertilizer triggered varying growth promotion rates for Arabidopsis thaliana. In the initial 10-day period, the blank group’s Arabidopsis thaliana leaves numbered just 2, whereas the leaf counts of the other groups treated with 1 g, 2 g, 3 g, and 4 g sorbent–fertilizer totaled 3 (Figure S2). In terms of leaf production, the Arabidopsis thaliana fertilized with 1 g, 2 g and 3 g P-loaded sorbents showed notable growth. Day 10–20 of the vegetative phase saw the fastest development of the vegetative (cauline) leaves when 4 g of sorbent–fertilizer was applied (Figure S3). As seen in Figure 2 and Table S1, 4 g of P-loaded sorbent generated the fewest rosette leaves and resulted in the lowest survival rate on days 20~30. This may be closely related with the phosphate leached from the sorbent–fertilizer, which served as the indispensable macro-nutrient for the plant during the rapid growth. In addition, the reduced growth might be attributed to the insufficient micro-nutrients in the liquid media, such as B, Mn, Zn, etc. for the development of flower and fruit [25,26]. P-starvation might bring about detrimental effects, such as the inhibition of N uptake and assimilation. This is because N is not only a required element for the synthesis of amino acids, but as some N-containing substances like nitrate may also act a signal molecule, modulating phosphate response, or play a significant role in flowering [27]. As opposed to the plants nourished with 3 g of P-loaded P-sorbent _D, Arabidopsis thaliana grew poorly following the dosing of 3 g of P-loaded La-free DWTS-based nano-sorbent, suggesting that NR_La-Si exhibited greater disease control ability, even in the hydroponic culture, than the commonly applied foliar spray mode. Therefore, the release of P from the P-loaded sorbent was substantially correlated with the growth of Arabidopsis thaliana, especially during the initial 3~4 weeks of growth. Since 3 g of P-loaded sorbent released enough phosphate to support one Arabidopsis thaliana plant, P-sorbent could serve as a special water-soluble fertilizer.

Figure 3 depicts the growth of seedling roots of Arabidopsis thaliana as impacted by the P-loaded sorbent. The root hair lengths were measured as 1.51 cm, 1.52 cm, 1.54 cm, 1.53 cm, and 1.49 cm on day 10 for the blank, while the root hair increased to 3.33 cm, 5.28 cm, 5.56 cm, 4.33 cm, and 4.18 cm on day 30 for groups dosed with 1 g, 2 g, 3 g, and 4 g, respectively. Therefore, it was evident that P-loaded sorbent had a stimulating impact on the root elongation of Arabidopsis thaliana. The elongation of seedlings was promoted most when using 2 g and 3 g of P-sorbent _D, indicating that the phosphate gradient in the liquid phase was likely responsible for root cell elongation rather than large swelling at the tip. This was consistent with previous studies that revealed longer and thicker roots to be beneficial for the total bacteria in the rhizosphere, playing a key role in enhancing plant growth and productivity [28].

2.3. Characterization of P-Sorbent _D and Property Variation during Adsorption and Fertilization

The SEM and EDS images of pure P-sorbent _D and exhausted P-sorbent _D loaded with P, that is, before and after the P adsorption test, are shown in Figure 4. As Figure 4a shows, the virgin P-sorbent _D had an irregularly rough and porous surface and therefore a high specific surface area, and resembling P-sorbent _P in appearance [11]. After phosphate adsorption, the exhausted P-sorbent _D exhibited a surface pattern resembling that depicted in Figure 4b. According to the EDS spectrum of the virgin and exhausted P-sorbent _D (Figure 5), the constituents of the synthesized adsorbents were mainly elements of La, O, Ca, Si, Al, Fe, Mg, and P, suggesting that the P was successfully captured and loaded onto the nano-sorbents and from aqueous solutions.

According to previous research, the DWTS primarily consisted of the components of Al, Fe, Ca, and Mg, but it also occasionally contained Cd, Cr, Pb, As, Ni, etc., which might pose adverse effects on the plants or crops [29]. However, according to the EDS spectrum of the P-sorbent _D, the heavy metal impurities, even if extant in DWTS, were all sequestered out in the synthesis and preparation procedures.

The magnetic P-sorbent _D was confirmed by the XRD analysis. As shown in Figure 6, it was found that peaks (111), (220), (311), (400), (422), (511), and (440) corresponded to MgFeAlO4 (PDF#11-0009), indicating that the P-sorbent _D mix was primarily dominated by metal oxide. Additionally, the presence of Fe₃O₄ was also confirmed by the peaks (311), (400), (511), and (440) (PDF#88-0315-1). It is worth noting that the XRD pattern in the P-sorbent _p, reported by Zhao et al. [30], presented ordinary metal oxides rather than obvious, typical, and distinct reflections of the LDH structure. Furthermore, the reflection peaks of (220), (400), (222), and (622) matched the existence of doped NR_La-Si and corresponded to La_9.33Si₆O₂₆ (PDF#49-0443). And more disordered peaks were observed in the P-sorbent _D than in P-sorbent _p. All these facts might be attributed to the impurities originating from the DWTS that could interfere the crystallization during preparation, even though these impurities were detected at trace level. After P adsorption, the sorbent showed a crystalline surface (301), which corresponded to MgAl₂FePO₄ (PDF#71-1233). Phosphate species were actually preferentially adsorbed due to inter-complexation with LDH components, which were predominantly protonated below pH values of 7~9 [2,31]. In contrast, the P-sorbent _D sequestered phosphates via more complicated complexation of the P-to-hydroxide (Fe/Al/Mg/La) bonds [32]. Electrostatic attraction, surface precipitation, and ionic ligand hydroxylation should also be taken into account during P capture [11,23,24].

Figure 7 presented the FTIR spectra of P-sorbent _p before and after P adsorption. The main absorptions bands and their attributions occurred at 3440 cm⁻¹ and 1623 cm⁻¹ in relation to the OH group; at 1448 cm⁻¹ in association with C-C, C-O, or C-N stretching vibrations, which might be attributed to the organic matter or carbonate ions within the anionic layer; at 580 cm⁻¹ due to Fe-O of Fe₃O₄; and at 1080 cm⁻¹ for the asymmetric vibrational absorption peak corresponding to Si-O-Si in Fe₃O₄@SiO₂ [33,34,35]. The spectra observed after P adsorption were characterized by an increase in the intensity of peaks at 1007–1035 cm⁻¹ (band of P-O). A new weak peak detected at 659 cm⁻¹ could be due to the presence of P-O from HPO₄²⁻. These pronounced peaks observed on the spectrum of phosphate adsorbed nano-sorbent could be characteristic of phosphate [35].

The XPS spectrum of virgin P-sorbent _D, P-sorbent _D loaded with P, and P-sorbent _D after P release is shown in Figure 8. The vibration of energy bonds at 133 eV for the exhausted P-sorbent _D sample demonstrated that P had been successfully loaded onto the nano-sorbent. This was consistent with the distinct P_2p peak for the samples after P adsorption, shown in Figure 8b. The P_2p peak of the exhausted sorbent was reasonably less significant than the P-loaded sorbents but more noticeable than the virgin ones; thus, the P desorption in deionized water might be a slow process. There were no significant differences in the O_1s spectra, while the peaks of Mg_1s, Ca_2p, Al_2p, and Fe_2p underwent a relatively small shift of about 0.4 eV after P adsorption, suggesting the occurrence of surface complexation. An intense satellite peak at 718.3 eV was also observed in Fe_2p, validating the supposed chemical precipitation.

According to the ICP-MS results for ion leaching from the sorbent during hydroponic culturing, phosphate was definitely the primary ion released from 4 g of sorbent with a P concentration of 43.5 μg/L after two weeks. Simultaneously, additional ions such as Ca, Fe, and Mg were also leached concurrently at concentrations of 21 μg/L, 12 μg/L, and 31.9 μg/L, respectively (

Table 3. ICP-MS results of leachate of the new P fertilizer on day 15.

Macro-Element	Concentration (μg·L−1)	Micro-Element	Concentration (μg·L−1)
Al	42.9810 ± 0.23	Cu	<3.17
Ca	21.2355 ± 0.13	Pb	<0.63
Fe	1.1752 ± 0.04	Cd	<0.16
Mg	31.8579 ± 0.25	Cr	<0.55
P	43.4612 ± 0.33	Zn	<7.62
La	10.6744 ± 0.17

P uptake kinetics by Arabidopsis thaliana with P-sorbent _D as fertilizer.

). The growth pattern of Arabidopsis thaliana (shown in Section 2.2) may be explained by these essential basic nutrients, released from the sorbent. For example, the germination failure rates would increase without sufficient Mg²⁺ and Ca²⁺ participation in shoot-specific processes. Additionally, biosynthesis and photosynthesis would also be hindered if Mg and Fe were not adequately supplied, particularly in the case of foliar fertilization [17]. The low concentrations of leached phosphate and Ca ions would prevent them from precipitating, given that the solubility product constant of Ca₃(PO₄)₂ was only 1.95 × 10⁻²⁹. Once the phosphate ions desorbed from the sorbent (fertilizer), they could soon be absorbed soon by the roots through active uptake and diffusion, driven by concentration gradients. We also tested 10.67 μg/L of La in the leachate, which was probably derived from the doped NR_La-Si. Therefore, the P-sorbent _D consistently provided a low dose of La ions that might activate antioxidative enzymes for the plant. It should also be mentioned that the leachate also contained Al ions, which might be toxic and inhibitive to the plants by suffocating or strangling roots, particularly in acidic environments [34].

This was consistent with the comparatively shorter root length in the group of Arabidopsis thaliana, which was overdosed with 4 g of sorbent–fertilizer in the hydroponic experiment. The concentrations of potential toxic elements, monitored on day 1, 10, and 15, were all below the limits recommended by Food and Agriculture Organization of the United Nations [36,37,38]. Without a leaching heavy metal able to persist in soil and act as a phytotoxic agent, the application of P-sorbent _D seemed safe for agricultural management.

The diffusion of phosphate ions toward the root tip from the hydroponic nutrient solution is driven by the concentration gradient. Figure 9a presents the plant uptake rate, U, which was determined and plotted against the phosphate concentration in hydroponic culture. The fitting curve was derived using the uptake kinetic equation in Equation (7). The results of the simulation indicated that the P uptake by Arabidopsis thaliana fitted the Michaelis–Menten kinetics model well. The kinetic parameter K_m was 7.74 μg/L (1.11 μmol/L) compared to the K_m of 3.4 μmol/L for the rice [39], suggesting a poorer ability for P uptake of Arabidopsis thaliana than rice at the same P nutrient level. Therefore, given that P plays a key role in various metabolic processes, it is possible that the intracellular phosphate transporter proteins in Arabidopsis thaliana are extremely active in P uptake.

Figure 9b–e indicate that the released P amount (in Table S2) from different masses of the exhausted P-sorbent _D contributed to varying levels of P uptake rates during the initial 10 days. The fitted results showed that 1~4 g P-sorbent _D led to 1.00~1.31, 1.56~2.55, 2.49~3.33, and 3.28~4.16 μg cm⁻¹ root h⁻¹ of P uptake rates, respectively. Taking into account the tip growth and root hair elongation, as revealed in Section 2.2, the dosed new fertilizer, i.e., P-sorbent _D, could maintain a phosphate concentration above 2.8 mg L⁻¹ in the nutrient solution during day 1~5 in order to achieve the maximum flourishing. In addition, it was also observed that applying 1~3 g of sludge-based fertilizer led to significant growth, while increasing the fertilizer amount to 4 g seemingly resulted in unexpectedly shorter root length as well as sparse or smaller ovate leaves. Therefore, overdosing on the sludge-based P fertilizer might pose adverse effects for the plant. The primary reason for this was presumably due to the metallic components present in the sorbent–fertilizer. For instance, excessive application of aluminum could cause Al toxicity, inhibit root growth, and reduce the uptake of an immobile nutrient such as P, which was one of the most ubiquitous soil-inherent issues [17,40]. It is noteworthy that the phosphate desorption in deionized water proceeded far more slowly than it did in the basic aqueous phase. This exhausted sorbent has the potential to be used as a special fertilizer, releasing soluble phosphate at a faster rate than the conventional slow-release fertilizers such as the struvite. Additionally, simulations could be employed to assess the doses of sludge-based fertilizer to be selected for certain type of plants and to evaluate the uncertainties associated with model predictions.

If applied in soil–water–plant systems, the fate and transport of the P and La elements released from this sorbent–fertilizer could be more complicated than that under hydroponic conditions. The dynamic distribution of the trace elements in crops would be strongly impacted by the characteristics of the soil. In contrast to insoluble P fertilizers like the citrate-insoluble P fertilizers, the sorbent–fertilizer can provide phosphate that is mobile and accessible for plant uptake and use, probably minimizing the loss due to adsorption or precipitation. The processes of P dissolution, immobilization, precipitation, and co-precipitation occurring at the soil–root interface might be characterized by a first-order kinetic equilibrium, which would be distinct from that seen after applying traditional P fertilizer [41]. The surface runoff due to irrigation and the weathering of bedrock, though negligible, might also contribute to the sources and sinks of P element. In any case, the P adsorption capacity and moderate-release property of the new P absorbent–fertilizer were demonstrated to be alternatives to DWTS reclamation and have the potential to lead to a paradigm shift in P-bearing water/wastewater treatment towards nutrient recovery and P sustainability. This novel sorbent–fertilizer for P reclamation from domestic wastewater, if applied to farmland, could serve as a human intervention in the P cycle to counterbalance the approximately 50% of mineral P fertilizer used in agriculture annually (Figure 10) [42]. The sludge-based P recovery process is a social-technical solution to sludge waste disposal and the preservation of limited or depleted P resources instead of complicated industrial processing for conventional P fertilizer.

3. Materials and Methods

3.1. Re-Synthesis of P-Sorbents with DWTS

The DWTS was collected from the Yuqing drinking water treatment plant in Jinan, Shandong province. The sludge was obtained from the plain sedimentation tank, where the coagulant of polymer aluminum chloride was dosed, with a slight brick-red color and a pH of about 8.23. The synthesis of Fe₃O₄ particles and Fe₃O₄@SiO₂ microspheres according to the co-precipitation method, as elaborated previously, was as follows [30,43]: The DWTS was dissolved with HCl (2 mol/L) with a ratio of DWTS to HCl solution of 0.1 g: 1 mL. It was then filtered with 0.45 μm membrane. The obtained filtrate was mixed and stirred agitatedly with 100 mL of basic solution containing 12.5 mmol of Na₂CO₃ (1.325 g) and 50 mmol of NaOH (2 g).

After stirring for 3 min, 8 mmol of LaCl₃·7H₂O and 0.572 g of SiO₂ were added to the precursor solution above, with the pH maintained at 14 by adding NaOH and stirring at 400 rpm for 50 min during the synthesis of NR_La-si. Afterwards, the mixture was crystalized in an 80 °C bath (LC-WB-4, Lichen Bath., Shanghai, China) for 24 h, centrifuged at 17,500× g for 15 min (Sorvall ST8R, Thermo Scientific., Changsha, China), and washed with boiling water/anhydrous ethanol several times until the pH of the filtrate became neutral. After freeze-drying for 24 h (Freeze dyer, LGJ-12., Yetuo, Shanghai, China), the obtained precipitates were the DWTS-based superparamagnetic P nano-sorbent (P-sorbent _D). P-sorbent _P was prepared with the same procedures, only with 3.26 g of chemical-grade MgCl₂·6H₂O and 1.94 g of AlCl₃·6H₂O instead of 0.1 of DWTS, as reported before [20]. All the reagents were of analytical grade and were purchased from Damao Co., Ltd., Tianjin, China.

3.2. Adsorption Experiments

A batch adsorption test was conducted by exposing 0.5 g–1.5 g of P-sorbent _D/P-sorbent _p, respectively, to a phosphate solution of 100 mg P/L in 250 mL Erlenmeyer flasks. These flasks were capped and shaken vigorously in a shaker at 120 rpm for 18 h to reach saturation; then, the phosphate concentration was determined according to the standard molybdenum blue method after the supernatant was decanted through a magnet [43]. Thus q_e (mg/g), the amount of phosphate loaded per unit mass of P-sorbent _D at the equilibrium, i.e., the maximum adsorption capacity, can be determined using Equation (1). The phosphate removal efficiency (%) can be calculated via Equation (2) [43]:

q_e = V(C₀ − C_e)/m

(1)

η = (C₀ − C_e)/C₀ × 100%

(2)

Adsorption isotherms models, i.e., the Langmuir, Freundlich, Temkin, and Dubinin–Radushkevich isotherm models shown in Equations (3)–(6), were, respectively, employed to describe the phosphate adsorption at different equilibrium concentrations:

q_e = q_m K_L C_e/(1 + K_LC_e)

(3)

q_e = K_f C_e^1/n

(4)

q_e = RT/b × lnA(C_e)

(5)

q_e = q_mexp(−Kε²)

(6)

where C_e (mg/L) is the phosphate equilibrium concentration. K_L (L/mg) and K_f ((mg/g)/(mg/L)1/n) are the Langmuir and Freundlich adsorption equilibrium constants, respectively. q_m in Langmuir isotherm model is the maximum adsorption capacity (mg/g). n is a constant, indicating the Freundlich isotherm curvature. R in the Temkin model is the gas constant (8.314 (J/(mol·K)), T is the temperature (K), b is the Temkin constant (J·mol⁻¹), and A is the Temkin isotherm constant (L/g), where K_T = RT/b. q_m in the Dubinin–Radushkevich Model is the adsorption amount at equilibrium, R is the gas constant (8.314 (J/(mol·K)), K is the adsorption isotherm constant (mol²/J²), T is the Kelvin temperature (K), ε is the adsorption potential (J/mol), and E is the energy (J/mol), where ε = RTln(1 + 1/C_e), E = 1/√2 K. Origin 2021 software was used to fit the models above with non-linear fitting. All the samples in the batch adsorption test were tested in triplicate.

3.3. Hydroponic Experiments

Arabidopsis thaliana seeds (purchased from Arabidopsis thaliana Home Co., Ltd., Kaifeng, China) were cultivated in a specialized hydroponic system consisting of (1) the seed holders, i.e., the cover of the 1.5 mL plastic centrifuge tube with central holes filled with 0.5% agar, and (2) the liquid medium container containing 750 mL of nutrient solution, in which 100 seed holders were dipped (shown in Figure S1). The seeds were sowed in the central holes of the soft agar so that the seed holders could ensure fully successful germination. Subsequently, the seedling roots grew through the agar into the solution and were measured at specific intervals.

The hydroponic nutrient solution was the P-free Hoagland’s nutrient-based solution, with the following constituents: K₂SO₄ 607 mg/L, MgSO₄ 493 mg/L, Fe-EDTA 20 mg/L, FeSO₄ 15 mg/L, boric acid 2.86 mg/L, borax 4.5 mg/L, MnSO₄ 2.13 mg/L, CuSO₄ 0.05 mg/L, ZnSO₄ 0.22 mg/L, and (NH₄)₂SO₄ 66.02 mg/L, as well as additional Ca(NO₃)₂ 1260 mg/L [40]. These salts were dissolved in distilled water and autoclaved at 120 °C for 30 min. After cooling, the solution was transferred into the liquid medium containers, with the liquid depth exceeding the seed holders’ bottom. Then, 1 g, 2 g, 3 g, and 4 g of exhausted P-sorbent _D, obtained in the adsorption test, were, respectively, dosed into the four containers above, with a fifth sample acting as the blank in absence of P fertilizer. They were grown at an 18 h: 6 h light–dark cycle at temperature of 22 °C for 30 days. After sowing, the survival rate, plant height, plant root length, and number of leaves of the Arabidopsis thaliana were observed and recorded every day during the growth period. The hydroponic cultures for the all the groups of P-sorbent _D were run in triplicate.

3.4. Characterization for the P-Sorbent _D

The surface morphology of the P-sorbent _D before and after adsorption was observed using a scanning electron microscope (SEM, Regulus8100, Hitachi Co., Ltd., Tokyo, Japan) and characterized using Fourier transform infrared techniques (FTIR, Nicolet submit, Thermoscientific Co., Ltd., Waltham, Massachusetts, MA, USA), aiming to analyze and determine the functional groups of P-sorbent _D before and after adsorption. The elemental compositions and distributions of the P-sorbent _D, before and after adsorption, were characterized with energy-dispersive X-ray spectroscopy (EDS, Octane Elect America EDAX Co., Ltd., Washington, DC, USA). In addition, X-ray diffraction (XRD, SmaitLab SE Co., Ltd., Rigaku, Japan) analysis was used to analyze the crystalized phases of powders. The element chemical states (Al, Fe, Mg, Ca, P, La, Cl, and O) of the virgin P-sorbent _D, the exhausted P-sorbent _D, and the sorbent after P release were analyzed via XPS (Thermos ESCALAB 250XI, Thermos Co., Ltd., Wortham, TX, USA).

3.5. Modeling for the P Uptake in P-Sorbent _D-Fertilized Hydroponic Culture

The P uptake simulation model was established to investigate the root P uptake after applying the new P fertilizers, i.e., the P-loaded nano-sorbents. The uptake of phosphate (released from the sorbent–fertilizer) under hydroponic conditions depended on the demand by Arabidopsis thaliana and the phosphate concentration in the liquid phases. Thus, the phosphate influx rate U at the root surface (μg cm⁻¹ root h⁻¹) of Arabidopsis thaliana could be described by the Michaelis–Menton reaction kinetics model, shown in Equation (5) [38]:

U = U_maxC_s/(K_m + C_s)

(7)

where C_s describes the phosphate concentration in solution and U_max is the maximum U when C_s is longer limiting. K_m is the root permeability coefficient (mM), corresponding to U at 50% of U_max.

The assumption conditions for the steady-state model above included the following: (1) The depletion of phosphate in the hydroponic culture could be regarded as being due to its complete uptake by the plant and the changes in the root system caused by root growth were also taken into account. (2) The root system had a uniform radius and the effects of root hairs and mycorrhizal fungi could be negligible. The plants were uniformly distributed throughout the growth system (hydroponics, in this case), and so they functioned as a uniform sink for water and ions at any given time. (3) The uptake of nutrients by the root system was not affected by root exudates and rhizosphere microorganisms.

In the modeling test, the Arabidopsis thaliana seeds that were washed and sterilized were sowed in the nutrient solution container with phosphate 0.5~10 mg/L. The other nutrients were set at the same concentrations as those in the hydroponic culture in the section “Hydroponic experiment”. The root length densities of Arabidopsis were measured every day. The phosphate influx rate U was calculated by measuring the change in concentration per unit of time (d) and root length via drawing out 100 μL with a pipette. The moisture loss in this process was negligible. “Cs” and “U” over time (d) were evaluated from the mass balance and then were simulated according to Equation (7). Meanwhile, 1 g, 2 g, 3 g and 4 g of P-loaded P-sorbent _D obtained in the adsorption test were, respectively, placed in deionized water in 4 microtubes in which no Arabidopsis thaliana existed. The phosphate concentration was assayed before draining and renewal with an initial amount of deionized water every day during day 10~20. The phosphate detected in the liquid phase was regarded as Ci, which should be equivalent to the released amount that allowed for active P uptake by Arabidopsis thaliana in the modeling test. Meanwhile, the ion composition and content of the desorption solution were also measured with inductively coupled plasma-mass spectrometry (ICP-MS, Agilent 7700, Agilent Co., Ltd., California, CA, USA) after 4 g of P-sorbent _D was dosed into 750 mL deionized water on days 1, 5, and 10. Sampling was conducted for two weeks after dosing sorbent fertilizer after drawing a 10 mL water sample.

4. Conclusions

The purpose of this study was to determine whether DWTS was a suitable raw material for preparing highly effective adsorbents for P capture and even for P fertilization. The fitting curve with Langmuir isotherms showed that the DWTS was converted into La-doped superparamagnetic nanoparticles using the effective co-precipitation approach. These superparamagnetic nanoparticles demonstrated effective P sequestration performance, with a q_max of 26.8 mg of P/g P-sorbent _D. It was superior to most of the other similar DWTS-based sorbents, with far simpler and less energy-intensive synthesis procedures. SEM-mapping, XRD pattern, EDS spectrum, and ICP-MS analysis confirmed that the preparation procedures filtered out hazardous metal ions and substances. The obtained nanosized P-sorbent _D exhibited a crystalline metal–hydroxide matrix that had a high affinity with phosphate and the NR_La-Si could enhance the ligand complexation of phosphate. The P-loaded sorbent–fertilizer employed in the hydroponic culture for Arabidopsis thaliana could stimulate plant growth and development due to both the released phosphate and the doped La. The phosphate leaching from 3 g of P-loaded sorbent seemed to be better suited for one Arabidopsis thaliana plant. Modeling for the P uptake by Arabidopsis thaliana indicated that appropriate dose of new fertilizer could maintain a phosphate concentration level in hydroponic cultures to support the maximum flourishment. This new type of DWTS-based sorbent could serve as a medium-release P fertilizer and provide a beneficial and sustainable solution for both DWTS disposal/recycling and P recovery.