Introduction

Discharge of wastewater enriched with high levels of nitrogen, phosphorus, and other nutrient elements into rivers, lakes, and additional surface water bodies1 can trigger extensive algal blooms, notably of cyanobacteria, subsequently altering the physico-chemical characteristics of these aquatic environments2. Such processes lead to a substantial depletion of dissolved oxygen, an intensified odor, and overall degradation in water quality, thereby disturbing the equilibrium of aquatic ecosystems and contributing to eutrophication3. Phosphorus enrichment is a recognized cause of eutrophication in lakes, underscoring that controlling phosphorus levels is critical to mitigating eutrophication; hence, reducing the phosphorus concentration in wastewater remains a pivotal aspect of wastewater treatment.

Currently, wastewater phosphorus removal methods primarily comprise biological phosphorus removal4 and physicochemical phosphorus removal, which includes chemical precipitation5crystallographic method, ion exchange phosphorus removal, and adsorption phosphorus removal6. Relative to conventional biological phosphorus removal, physicochemical approaches demonstrate more consistent phosphorus removal in sewage7. Moreover, the physicochemical process system is comparatively simpler and incurs lower short-term costs than biological methods, thereby offering significant economic benefits. In practical applications, the adsorption method exhibits broad applicability, effective phosphorus removal, high removal efficiency, ease of control, low cost, and enables phosphorus recovery8.

Significant advancements have been made in recent years toward developing novel adsorbents for phosphorus removal. Particularly, composite materials synthesized by integrating metal oxides/hydroxides onto carrier surfaces or within porous frameworks have drawn considerable research interest. Clay-derived substrates are increasingly recognized for in situ phosphate immobilization and remediation, owing to their economic viability, natural abundance, non-toxic properties, robust chemical-mechanical stability, and environmental compatibility with soil/sediment systems9,10. Montmorillonite has been demonstrated to be an efficient and economical phosphorus adsorption material due to its layered structure, high specific surface area, strong ion exchange capacity, and ease of modification. Through targeted modification, its adsorption capacity and selectivity can be further enhanced to address diverse environmental treatment requirements. Kaolin possesses a low specific surface area and cation exchange capacity (CEC), resulting in its capacity for adsorption being limited. Illite exhibits a fixed interlayer structure, poor swelling, and a paucity of adsorption sites. Aconite and seafoam, on the other hand, possess a high specific surface area; however, they are not as widely distributed as montmorillonite. Nevertheless, native clays predominantly exhibit negative surface charges under neutral pH conditions, resulting in limited phosphate adsorption capacities2. To enhance phosphorus sequestration efficiency, numerous modified clay formulations have been systematically investigated over the past twenty years11,12. Concurrently, magnetic adsorbents have emerged as a promising solution for aqueous phosphate extraction, particularly valued in water treatment applications for their environmental retrievability and reusability advantages13,14. Research has demonstrated that magnetite/lanthanum hydroxide hybrid materials possess exceptional magnetic responsiveness and phosphorus-binding capabilities, enabling efficient recovery through external magnetic fields15,16. Meanwhile, γ-alumina has been identified as a superior adsorbent due to its chemical inertness, non-toxicity, stable physical characteristics, and high removal efficiency17,18. Extensive studies have focused on γ-alumina’s application in wastewater remediation, with experimental evidence confirming activated alumina’s effectiveness in phosphorus removal. For instance, Jiang Diao et al. implemented activated alumina for phosphorus extraction from steel slag leachate, achieving 88% elimination efficiency at an initial concentration of 25.95 mg L− 1 under optimized static adsorption conditions19. Similarly, Wang Ting et al. investigated activated alumina’s performance in domestic wastewater treatment, reporting 93.85% phosphorus removal from wastewater containing 12.36 mg L− 1 initial phosphorus concentration under optimal experimental parameters. The significant enhancement of the adsorption capacity of Fe-Al-modified montmorillonite (Fe/Al-Mt) is attributable to the synergistic effect of its structural remodeling and the optimization of its surface chemical properties. Structurally, the hydroxy-Fe/Al columns effectively enlarged the interlayer spacing of montmorillonite (from 1.5 nm to 2.0–2.3 nm), and constructed a “card house” framework with a high specific surface area (50–800 m2/g) and mesopore structure (pore size of 2.0–2.3 nm)20,21greatly increasing the adsorption capacity. The “card house” framework has been demonstrated to significantly enhance the quantity and ease of access to adsorption sites. Concurrently, the composites exhibit remarkable thermal stability (structural stability at 350 °C) at reduced Fe/Al ratios (Fe/(Fe + Al) = 0.05–0.3), thereby expanding their application scope at elevated temperatures (e.g., cigarette filters)20. From a chemical perspective, the introduction of Fe-Al doping results in the generation of a substantial number of Lewis acid sites, namely Fe3+ and Al3+ null orbitals. This phenomenon has been observed to lead to a notable enhancement in the ligand adsorption capacity for oxygen-containing anions. For instance, The presence of CrO42− and PO43− in the material has been identified as a key factor in its coordination adsorption capacity20,21. A notable finding is the inversion of the surface charge of the material, as evidenced by the increase in zeta potential from − 30 to + 15 mV for natural montmorillonite at a pH of 722. This modification has been shown to enhance the electrostatic attraction to anionic pollutants, suggesting a potential for enhanced removal or remediation processes. In terms of the performance of the adsorption process, these advantages manifest as rapid adsorption kinetics (90% of the contaminants were captured within 30 min21 and exceptional regeneration capacity (> 85% adsorption retention of Cr(VI) after acid-base regeneration22. These benefits are primarily attributable to the high reactivity of the Fe-Al active sites and their stable bonding with the montmorillonite backbone.

Nevertheless, activated alumina demonstrates constrained adsorption performance when applied to low-concentration phosphorus-containing wastewater remediation. This limitation primarily originates from inherent material characteristics, specifically the suboptimal particle dimensions of iron oxide constituents and insufficient structural integrity of the composite system. Direct application in an aqueous phase adsorption system results in significant loss and renders recovery challenging. Moreover, natural mineral soil with large particle sizes exhibits low adsorption performance. In the present study, a composite adsorbent, Fe/AlPMt, was prepared by combining iron-modified montmorillonite with Al2O3. The resulting adsorbent demonstrated superior phosphorus removal efficiency, facile separation and recovery, and cost-effectiveness. Table S1 summarizes the adsorption-desorption properties of the adsorbents of this work and other similar materials. This approach was selected based on a comparative evaluation of phosphorus removal performance. Comprehensive analysis of the composite adsorbent’s architectural and surface features was performed, accompanied by systematic investigation into its regenerative potential. Systematic batch experimental studies were designed to assess phosphate sorption properties, including kinetic behavior, equilibrium isotherm patterns, thermodynamic parameters, and pH-dependent adsorption phenomena.

Materials and methods

Materials

Montmorillonite employed for experimental purposes was sourced from De Hang Mineral Products Co., Ltd. (Shijiazhuang, Hebei, China). Chemical reagents including aluminum chloride hexahydrate (AlCl3·6H2O), ferric chloride hexahydrate (FeCl3·6H2O), sodium hydroxide (NaOH), anhydrous ethanol (C2H5OH), concentrated HCl (37% w/w), concentrated H2SO4 (98% w/w), ascorbic acid (C6H8O6), ammonium molybdate ((NH4)6Mo7O24·4H2O), potassium antimonyl tartrate (C8H4K2O12Sb2), potassium dihydrogen phosphate (KH2PO4), potassium persulfate (K2S2O8), sodium carbonate (Na2CO3), and sodium bicarbonate (NaHCO3) were procured commercially. All chemical substances, unless specified, met analytical-grade standards and were utilized without additional purification. Aqueous solutions throughout the experimental procedures were formulated using 18.2 MΩ·cm deionized water.

Preparation adsorbent

The synthesis of iron-modified montmorillonite (Fe-MMT) commenced with dispersing 1 g MMT into 100 mL of 0.5 wt% FeCl3·6H2O solution under mechanical agitation. This mixture underwent continuous stirring at 25 ± 1 ℃ for 24 h before being subjected to centrifugal rinsing and vacuum desiccation. Subsequently, 4.83 g AlCl3·6H2O was dissolved in 200 mL deionized water, followed by ultrasonic dispersion of 0.5 g Fe-MMT to create homogeneous suspension. A separate alkaline solution containing 2.48 g NaOH dissolved in 200 mL DI water was incrementally introduced into the suspension under thermostatic control at 50 ℃ over 3 h. Following sedimentation, the resultant slurry underwent hydrothermal treatment in a 100 mL autoclave at 180 ℃ for 10 h. Post-reaction processing involved sequential centrifugal purification and drying cycles. The intermediate product was then transferred to crucibles for thermal activation in a muffle furnace (500 ℃/10 h). Final functionalization was achieved by immersing 1 g of the calcined powder into 100 mL 0.5 wt% FeCl3·6H2O solution under ambient temperature agitation (25 ℃/24 h), concluding with centrifugal separation at 60 ℃ and 12-hour desiccation to yield the iron-alumina modified montmorillonite composite (Fe/AlPMt).

Characterizations

Characterization of Fe/AlPMt morphology by scanning electron microscopy (SEM; Sigma 500) and elemental analysis by energy dispersive X-ray spectroscopy (EDS). The sample was scattered on a conductive carbon strip and coated with a thin layer of gold to make it conductive. The surface area, pore size, and volume of Fe/AlPMt were determined from N2 adsorption–desorption experiments by using a surface-area and pore-size analyser (ASAP 2460, Atlanta, America). The chemical composition and state of the adsorbent were studied by X-ray photoelectron spectroscopy (XPS, FEI ESCALAB). The Zeta potential before and after adsorption was analyzed by a Zeta potential analyzer (Zetasizer Nano Z, Malvern, UK). The concentration of phosphate was determined by ammonium molybdate spectrophotometry using a spectrophotometer at 700 nm (UV-2550, Shimadzu, Japan).

Batch adsorption experiment

Batch adsorption experiments were performed under controlled thermal conditions (25.0 ± 0.5 °C) using a thermostatic orbital shaker, with Fe/AlPMt loading maintained at 1 g L− 1 under continuous agitation (200 rpm). Both initial and equilibrium phosphorus concentrations were measured after solution filtration through 0.45 μm GE cellulose-nylon composite membranes. The adsorbent’s phosphorus uptake capacity (q, mg g⁻¹), defined as mass of phosphorus retained per unit mass of adsorbent, was quantified through Eq. (1):

$${{\text{q}}_t}=\frac{{V\left( {{C_0} - {C_t}} \right)}}{m}$$
(1)

where C0 (mg/L) and Ct (mg/L) are the initial P concentration and equilibrium P concentration, respectively. V (L) is the solution volume, and m (g) is the dosage of the adsorbent.

The pH influence on adsorption was systematically examined through adjustment of aqueous solution acidity from 2.0 to 12.0 using HCl/NaOH. To evaluate competitive anion effects during phosphate sequestration, controlled additions of Cl−, NO3−, and SO42− were introduced at predetermined concentrations. All experimental trials maintained an initial phosphorus level of 20 mg L− 1. Adsorption kinetic studies employed phosphate solutions with initial concentrations spanning 20–30 mg L− 1 (20, 25, 30 mg L− 1) under controlled acidic conditions (pH 3.0). Thermodynamic investigations were conducted across three temperature regimes (25 °C, 35 °C, 45 °C) with constant adsorbent loading and pH 3.0, exploring phosphorus concentrations from 10 to 100 mg L− 1. All experiments were repeated independently three times (n = 3) under identical conditions.

Regeneration experiments

In order to study the effect of desorbing agent, a 0.1 mol/L NaOH solution, a Na2CO3 solution, and a NaHCO3 solution were configured as desorbing agents. The solid-liquid ratio of Fe/AlPMt to desorbing agent was taken to be 1 g/L after adsorption of phosphorus. The desorption time was chosen to be 24 h at 25 °C. An investigation was conducted into the effects of varying concentrations of desorbents, with three concentrations (0.1 mol/L, 0.5 mol/L, and 1 mol/L) utilized. The sampling times encompassed 1 h, 3 h, 6 h, 12 h, and 24 h. The study further examined the impact of the number of cycles on desorption, five consecutive adsorption-desorption cycles were implemented to assess the material’s regenerative capacity and structural integrity. Experimental conditions maintained an initial phosphorus concentration of 20 mg L− 1 throughout the investigation. Post-adsorption treatment involved triple rinsing of the saturated material with deionized water. Subsequent regeneration required immersion in 0.5 M Na2CO3 solution at 1 g L− 1 solid-liquid ratio. This restorative phase proceeded under ambient thermal conditions (25 ± 1 °C) for 12 h. Following regeneration, exhaustive washing with deionized water continued until effluent pH stabilized at neutrality, succeeded by thermal dehydration at 60 °C preparatory to subsequent utilization cycles.

Results and discussions

Optimization of the Preparation conditions

Montmorillonite was initially employed as the substrate, and ferric chloride was used as the iron salt modifier. A predetermined mass of montmorillonite was measured, and four types of Fe-modified montmorillonite adsorbents, each with a distinct iron mass percentage, were synthesized. The iron concentration (wt%) was determined based on the casting mass of the ferric chloride solution, with values of 0.5%, 1%, 2%, and 3%, respectively, by employing a water bath heating method. As shown in Fig. 1a, the Fe-MMT prepared using a 0.5 wt% Fe salt solution exhibited superior phosphorus removal performance compared with the other three concentrations. To further investigate the effects of different compounding methods for Fe-modified montmorillonite and to identify the optimal preparation procedure, three adsorbents with distinct addition methods were fabricated. Moreover, the impact of calcination on the phosphorus removal efficacy of the adsorbents was examined, resulting in the preparation of six adsorbents, as depicted in Fig. 1b. The first three adsorbents were produced both before and after calcination, whereas the third adsorbent was generated exclusively after calcination. Experimental results indicated that the incorporation of Fe-MMT was the preferred route to form the proposed thin hydrotalcite. The subsequent experimental procedure involved adding Fe-MMT to an aluminum salt solution followed by the addition of an alkali. To assess the variation in adsorption performance for different theoretical mass percentages (wt%) of Fe-MMT, the optimal theoretical loading was determined, as illustrated in Fig. 1c. Consequently, the adsorbent prepared with a theoretical mass percentage of 30 wt% Fe-MMT was selected for phosphorus removal in subsequent experiments. Finally, the calcined adsorbent underwent a secondary iron modification (Fig. 1d) to yield the final adsorbent.

Fig. 1
figure 1

(a) Comparison of phosphorus removal from iron-modified montmorillonite with different iron salt concentrations.The horizontal coordinates of the graph correspond to: Fe-1: 0.3 wt% iron salt solution, Fe-2: 0.5 wt% iron salt solution, Fe-3: 1 wt% iron salt solution, Fe-4: 2 wt% iron salt solution; (b) Comparison of phosphorus removal by adsorbents with different Fe-MMT addition methods.The horizontal coordinates of the graphs correspond to: 1: Fe-MMT added directly after mixing aluminum salt solution and lye, 2: Fe-MMT added to aluminum salt solution and then lye, 3: Fe-MMT added to hydrothermally heated böhmite solution, 4: calcined vs. 1, 5: calcined vs. 2, 6: calcined vs. 3; (c) Comparison of phosphorus removal by adsorbents with different theoretical Fe-MMT contents; (d) Comparison of phosphorus removal by secondary iron-modified adsorbents.The horizontal coordinates of the graph correspond to: 1: without secondary iron salt modification, 2: with secondary iron salt modification.

Full size image

Characterizations of adsorbents

Morphological characteristics and elemental composition of Fe/AlPMt were analyzed using SEM-EDS coupled with elemental mapping, revealing homogeneous spatial dispersion of Fe, Al, O, and Si throughout the specimen’s exterior (Fig. 2a–e). Fe/AlPMt exhibits a rough surface with a loose, porous structure, and a layer of lamellar loadings is visible on the surface. Voids can be observed between the lamellae, exposing the loose structure of the activated alumina underneath (Fig. 2f). Following the adsorption of phosphate by Fe/AlPMt, some of the lamellae form a stacking structure, which reduces the size of the intra-particle voids (Fig. 2g).

Fig. 2
figure 2

(a–e) The Mapping and EDS of Fe/AlPMt; (f) SEM morphology before adsorption; (g) SEM morphology after adsorption.

Full size image

Adsorption experiments

Influence of pH value on phosphate adsorption

Solution pH exerts significant influence on physicochemical interactions at liquid-solid interfaces. Figure 3a illustrates phosphate adsorption patterns by Fe/AlPMt across the pH spectrum (2–12), demonstrating pronounced pH-dependent adsorption behavior. The gradual decline in phosphorus adsorption capacity with pH elevation primarily originated from composite surface deprotonation-induced electrostatic repulsion between Fe/AlPMt and phosphate species. Progressive alkalization enhanced hydroxyl ion (OH⁻) concentrations, which competitively inhibited phosphate anion adsorption through dual mechanisms: suppressing hydrolysis of surface-bound metal oxides and impeding phosphate-metal ion complexation processes. This competitive adsorption phenomenon significantly compromised phosphorus removal efficiency under alkaline conditions23.

Fig. 3
figure 3

(a) Effects of pH on the P adsorption capacity of Fe/AlPMt; (b) Zeta potential of Fe/AlPMt; (c) Effects of coexisting electrolytes on P adsorption. Error bars represent the standard deviation (n = 3 independent experiments).

Full size image

As depicted in Fig. 3b, Fe/AlPMt demonstrated an isoelectric point(IEP)24 at pH 6.0. Under acidic conditions (pH < IEP), the material’s surface acquired positive charges, facilitating electrostatic attraction toward anionic phosphate species. Conversely, alkaline environments (pH > IEP) induced negative surface charges, creating electrostatic repulsion between phosphate anions and the adsorbent that substantially reduced adsorption efficiency. This pH-dependent charge reversal mechanism explains why materials with elevated IEP values exhibit enhanced affinity for phosphate anions. The observed pH-adsorption correlation aligns precisely with zeta potential measurements.

Effect of coexisting anions on phosphate adsorption

Competitive adsorption between phosphate and coexisting anions on Fe/AlPMt surfaces may reduce phosphate sequestration efficiency. To assess these interference effects, environmentally prevalent anions (Cl−, NO3−, SO42−) were systematically examined for their impacts on phosphate adsorption (Fig. 3c). Chloride ion concentrations exhibited concentration-dependent interference, with phosphorus removal efficiency displaying gradual yet limited reduction (3.2–8.7% across 0–100 mg/L Cl−). Comparatively, nitrate demonstrated weaker competitive effects than chloride, inducing merely 2.1–5.9% efficiency loss under equivalent concentration gradients. Nitrate exposure induced minor variations in Fe/AlPMt’s phosphate sequestration efficiency, with concentration-dependent NO3− levels (0–100 mg/L) causing merely 2.1–5.9% efficiency decline. This limited response confirmed nitrate’s insignificant influence on the composite’s binding capability. Sulfate-containing systems exhibited exceptional performance stability, maintaining > 98% phosphorus retention relative to sulfate-free controls across 0–100 mg/L SO42− concentrations, with adsorption fluctuations constrained below 1.5%. These findings demonstrate sulfate’s negligible impact (< 1.5% variation) on Fe/AlPMt’s phosphate uptake capability. Ionic strength elevation revealed minimal interference from sulfate and nitrate anions, with adsorption capacity fluctuations remaining below 2.8% across tested concentrations. This ionic strength independence strongly suggests ligand exchange mechanisms dominate phosphate binding over electrostatic interactions. In contrast, chloride exhibited measurable competition, reducing adsorption efficiency by 6.3–12.1% at equivalent ionic strengths. The composite’s exceptional anion resilience (maintaining > 92% efficiency at 100 mg/L competing ions) confirms its selective phosphate capture through surface complexation, highlighting promising applications for phosphate recovery in complex aqueous matrices containing multiple interfering anions.

Adsorption kinetics

The kinetic behavior of phosphate adsorption on Fe/AlPMt under varying initial concentrations was analyzed through adjusting contact duration, with experimental outcomes presented in Fig. 4a–c. The adsorption process exhibited distinct dual-phase characteristics: an initial rapid stage succeeded by a gradual deceleration until equilibrium attainment over temporal progression. This accelerated adsorption kinetics may be partially ascribed to the nanoscale particulate dimensions and enhanced dispersion characteristics of Fe/AlPMt, thereby increasing interfacial interactions with phosphate species and promoting solute transport from bulk solution to reactive surfaces25. Quantitative analysis revealed that adsorption performance demonstrated significant concentration dependence, with enhanced uptake performance at elevated solute levels.

Fig. 4
figure 4

Adsorption kinetic model fittings: (a) the pseudo-first-order; (b) pseudo-second-order model; (c) the intra-particle diffusion model. The adsorption isotherm of (e) Langmuir, (d) Freundlich and (f) Tempkin for phosphate onto Fe/AlPMt at 25, 35 and 45 °C.

Full size image

To comprehensively assess the adsorption characteristics of Fe/AlPMt toward phosphate species, graphical representations of pseudo-first-order and pseudo-second-order kinetic fits were displayed in Fig. 4a-b. Quantitative descriptors including parametric variations and associated correlation metrics for both models were systematically tabulated in Table S2. The respective linearized rate equations corresponding to each kinetic framework could be formulated mathematically as:

$$\text{Pseudo-first-order: } \ln \left( {{q_e} - {q_t}} \right)=\ln {q_e} - {k_1}t$$
(2)
$$\text{Pseudo-second-order: }\frac{t}{{{q_t}}}=\frac{1}{{{k_2}q_{e}^{2}}}+\frac{1}{{{q_e}}}t$$
(3)

where qe and qt are the adsorption capacities (mg g− 1) at equilibrium and any time t (min), respectively. k1 (min− 1) and k2 (g mg− 1 min− 1) are the rate constants.

The experimental data revealed superior fitting accuracy of the pseudo-second-order kinetic model, with R² exceeding 0.99 across initial phosphate concentrations of 20 mg L− 1, 25 mg L− 1, and 30 mg L− 1. In contrast, the pseudo-first-order model exhibited poor correlation with experimental values. This mechanistic divergence indicates chemisorption dominance between phosphate ions and Fe/AlPMt, primarily mediated through electron-sharing interactions or ligand exchange processes where phosphate substitutes hydroxyl groups (-OH), governed by valence force mechanisms26.

To systematically investigate the dynamic mechanisms governing solute transport and adsorption kinetics, the experimental adsorption datasets underwent intra-particle diffusion modeling analysis based on Eq. (4).

$$\text{Intra-particle model: }{q_t}={k_p}{t^{0.5}}+a$$
(4)

where kp is the ground intra-particle diffusion coefficient (mg g− 1 min− 1/2) and a denotes the thickness of the boundary layer (mg g− 1).

The adsorption dynamics of phosphate on Fe/AlPMt exhibited a dual-linear pattern, graphically illustrated in Fig. 4c, suggesting the coexistence of two distinct mechanisms (film diffusion and intra-particle diffusion) during the sorption process. Parametric descriptors for these mechanisms and their associated statistical metrics were quantitatively documented in Table S2. Within the intraparticle diffusion framework, the mass transfer coefficient (a) demonstrated proportional dependence on surface liquid film thickness. Elevated coefficient magnitudes corresponded to enhanced resistance against solute diffusion processes27. Conversely, elevated magnitudes of the k coefficient signify predominant control by intra-particle diffusion mechanisms. Initial adsorption predominantly occurs through rapid occupation of surface-active sites via external adsorption phenomena. Following saturation of superficial adsorption regions, phosphate anions migrate into Fe/AlPMt’s structural matrix, subsequently binding to internal adsorption sites. Progressive accumulation of phosphate within internal matrices elevates diffusive impedance, consequently diminishing intra-particle transport kinetics. Stage II adsorption evolves into a hybrid mechanism integrating both intra-particle migration and interfacial film diffusion processes28. Quantitative analysis confirmed primary adsorption completion during the initial phase, with intra-particle mechanisms demonstrating predominant influence over sorption efficiency.

Adsorption isotherms

The phosphate adsorption performance of Fe/AlPMt was systematically evaluated through equilibrium isothermal adsorption analysis, with experimental outcomes graphically and numerically summarized in Fig. 4d–f and Supplementary Table S3. Three thermodynamic models—Freundlich, Langmuir, and Tempkin—were applied for modeling the experimental dataset, with their respective mathematical formulations corresponding to Eqs. (5), (6), and (7).

$$\text{Langmuir: } {q_e}={q_m}\frac{{{K_L}{C_e}}}{{1+{K_L}{C_e}}}$$
(5)
$$\text{Freundlich: }{q_e}={K_F}{C_e}^{{\frac{1}{n}}}$$
(6)
$$\text{Tempkin: }{q_e}=\frac{{RT}}{b}\ln \left( {A{C_e}} \right)$$
(7)

where qe denotes the equilibrium adsorption quantity (mg g⁻¹) of phosphate species; Ce corresponds to the residual phosphate concentration (mg L⁻¹) in equilibrated solutions; KF and n serve as characteristic parameters within the Freundlich model; A and b function as thermodynamic coefficients specific to the Tempkin isotherm; KL represents the adsorption affinity coefficient defined by the Langmuir equation; qe (mg g⁻¹) quantifies the maximum monolayer coverage capacity according to Langmuir theory.

Supplementary Table S3 comprehensively compiles the model-specific constants and regression coefficients derived from tri-isothermal analyses. Phosphorus sequestration by Fe/AlPMt exhibited poor alignment with the Langmuir adsorption framework. Thermodynamic evaluations revealed an inverse correlation between operational temperature and maximum adsorption performance, with enhanced phosphorus retention observed under lower thermal conditions. Comparative modeling demonstrated superior data congruence for the Freundlich isotherm, achieving determination coefficients (R² ≈0.90) significantly exceeding those of Langmuir predictions. This empirical evidence substantiates Freundlich-type multilayer adsorption as the governing mechanism for phosphorus immobilization29. The dimensionless parameter 1/n maintained values within 0.1–0.5 across tested thermal conditions, demonstrating favorable thermodynamic feasibility for phosphorus sequestration using Fe/AlPMt. Comparative analysis confirmed enhanced congruence between experimental data and Tempkin isothermal modeling relative to alternative frameworks. These observations collectively suggest that intermolecular interactions among adsorbed phosphorus species significantly modulate interfacial adsorption mechanisms during contaminant removal processes.

Application of Fe/AlPMt to remove phosphate from actual water use

Ecological restoration of river systems requires wastewater treatment infrastructures to achieve stringent phosphate thresholds (≤ 0.05 mg L− 1) while sustaining minimal effluent concentrations. To validate Fe/AlPMt’s operational efficacy under realistic conditions, a field-simulated batch experiment was implemented using Guangxi rural lacustrine water samples containing 0.507 mg L− 1 initial phosphate (Fig. 5). Application of Fe/AlPMt at 0.3 g L− 1 enabled dramatic phosphate reduction from 0.507 mg L− 1 to 0.005 mg L− 1, attaining 99% contaminant elimination efficiency. Even at reduced dosage (0.01 g L− 1), residual phosphate concentrations decreased to 0.208 mg L− 1-substantially below China’s regulatory limit (0.5 mg L− 1) for recycled municipal wastewater in landscape applications. These findings demonstrate Fe/AlPMt’s robust capability as a rapid-response adsorbent for trace-level phosphate remediation.

Fig. 5
figure 5

Effectiveness of Fe/AlPMt on lake water treatment.

Full size image

Mechanical study

Chemisorption mechanism

Collective analysis of sorption behaviors, equilibrium isothermal characteristics, and kinetic profiles collectively revealed ligand exchange as the primary sequestration mechanism for phosphate removal by Fe/AlPMt30,31. The mechanism initiates with phosphate solution saturation of surface hydroxyl moieties on Fe/AlPMt, progressing into selective phosphate capture through ligand substitution dynamics. During this process, structural hydroxyl groups undergo displacement by phosphate anions, facilitating covalent bond formation between metal centers and ligand species32,33. BET surface characterization (Table 1) revealed that Fe/AlPMt’s structural architecture demonstrated enhanced microtopographic complexity, optimized pore geometry, and markedly diminished specific surface area. Consequently, Fe/AlPMt demonstrated selective ligand exchange functionality, enabling efficient phosphate adsorption even under low ligand availability conditions. The adsorption mechanism predominantly involved dynamic ligand substitution between surface hydroxyl groups (-OH) on montmorillonite (MMT) and aqueous PO43− species. Specifically, hydroxyl moieties at metal-coordinated surface sites underwent ligand displacement with negatively charged phosphate ions, where the exchange process between these surface -OH groups and phosphate anions significantly enhanced adsorption capacity. Furthermore, phosphate sequestration by modified MMT extended beyond singular chemical processes, encompassing synergistic coupling of multiple interfacial interactions that collectively governed adsorption dynamics under specific environmental parameters. The experimental system revealed two critical auxiliary mechanisms: (1) electrostatic attraction between protonated surfaces and anionic phosphate species, and (2) surface precipitation phenomena. Solution pH exerted significant control over surface charge characteristics, with acidic conditions enhancing material protonation to strengthen electrostatic adsorption capacity34. Nevertheless, the magnitude of electrostatic interactions could be diminished due to solvent-induced screening effects, competitive ion interference, and other solution-phase phenomena. Surface precipitation phenomena significantly contributed to phosphate sequestration efficiency. Metallic species (Al3+, Fe2+) present on Fe/AlPMt surfaces underwent complexation with phosphate anions, generating insoluble precipitates or thermodynamically stable compounds in aqueous environments. These mechanistic investigations revealed ligand exchange as the predominant pathway for phosphate immobilization on Fe/AlPMt, operating concurrently with auxiliary electrostatic attraction and surface precipitation mechanisms throughout the adsorption sequence.

Table 1 Comparison of the BET surface areas, total pore volumes, and average pore sizes of fe/alpmt and fe/alpmt-P.
Full size table

XPS study

XPS characterization was conducted to elucidate the molecular-level mechanisms underlying phosphate adsorption on Fe/AlPMt. Figure S1 presents comparative full-range spectral profiles of Fe/AlPMt in pre- and post-adsorption states. Elemental composition analysis via wide-scan spectra revealed consistent presence of principal constituents throughout the adsorption process. Notably, a distinct spectral feature emerged at 133.0 eV in post-adsorption specimens (Fig. 6a), corresponding precisely to the characteristic P 2p binding energy signature, thereby providing direct evidence of successful phosphorus immobilization on Fe/AlPMt surfaces35. Following phosphate sequestration using Fe/AlPMt, the P 2p spectral signatures (134.26 eV, 133.36 eV) manifested within the 132.3–135.2 eV range, corresponding to electrostatic interactions between surface metallic species and HPO₄²⁻/H₂PO₄⁻ anions (Fig. 6b). Analytical data confirmed the coexistence of both phosphate species on Fe/AlPMt surfaces, facilitating Fe-P and Al-P precipitate formation. Adsorption processes induced a 0.5–1.2 eV upshift in Fe 2p binding energy, indicative of valence-band electron redistribution and subsequent Fe-O-P inner-sphere complexation (Fig. 6c and d). Deconvolution of O 1 s spectra resolved three oxygen states: MO (metal-oxygen bonds), MOH (metal-hydroxyl groups), and adsorbed H₂O. High-resolution XPS characterization was employed to investigate interfacial oxygen speciation of Fe/AlPMt pre- and post-phosphate adsorption, with spectral profiles presented in Fig. 6e and f. Post-adsorption quantification revealed an elevation in M-O component proportion from 22.17 to 26.01%, contrasted by a concurrent reduction in M-OH contribution from 48.30 to 47.39%. These inverse trends substantiated the critical function of surface M-OH groups in phosphate sequestration through hydroxyl-phosphate ligand substitution dynamics. Solution alkalinity elevation during adsorption exhibited mechanistic consistency with ligand exchange processes involving surface hydroxyls, thereby confirming inner-sphere complex formation36. This observation aligned with characteristic O-P-O vibrational signatures identified in FTIR spectra. Quantitative variations in surface hydroxyl stoichiometry between pre- and post-adsorption states enabled determination of inner-sphere complex configurations. Comparative analysis revealed stoichiometric ratios of 0.5, 1, and 2 for pristine versus phosphate-adsorbed Fe/AlPMt, corresponding respectively to monodentate mononuclear, bidentate mononuclear, and bidentate binuclear coordination modes37,38. The hydroxyl content ratio between original (48.3%) and phosphate-adsorbed (47.39%) Fe/AlPMt approximated unity (≈ 1), confirming predominant phosphate adsorption through bidentate mononuclear coordination with surface metal-hydroxyl sites. Collective XPS evidence established phosphate sequestration by Fe/AlPMt through three synergistic mechanisms: ligand exchange dynamics, electrostatic attraction, and surface precipitation phenomena.

Fig. 6
figure 6

Comparative XPS characterization of Fe/AlPMt in pristine and phosphate-adsorbed states. (a) Localized XPS survey scans of Fe/AlPMt pre- and post-phosphate exposure; (b) Post-adsorption P 2p spectral signature; (c, d) Fe 2p spectral comparisons before and after phosphate interaction; (e, f) O 1s spectral resolution under phosphate-free and phosphate-loaded conditions.

Full size image

Regeneration study

In order to regenerate and reuse the Fe/AlPMt adsorbent following adsorption saturation, it is first necessary to select the appropriate adsorbent for desorption of Fe/AlPMt. As demonstrated in Fig. S2a, for a given desorption time, the desorption effect of Na2CO3 on phosphorus was superior to that of NaOH and NaHCO3. As demonstrated in Fig. S2b, the desorption rate of phosphorus exhibited an increasing trend with the passage of time, depending on the varying concentrations of desorbents utilized. The rate of desorption exhibited a marked increase at the onset, subsequently decelerating. This observation was indicative of a comprehensive evaluation of the selection of 0.5 mol/L Na2CO3 desorption agent and the desorption time selection of 12 h, which was identified as optimal for achieving the desired desorption effect.

Progressive depletion of active sites and residual phosphorus retention during regeneration cycles led to gradual attenuation in Fe/AlPMt’s phosphorus sequestration efficiency39,40. Remarkably, the material exhibited sustained performance through five consecutive regeneration sequences (Fig. 7a), retaining 95.8% phosphorus adsorption efficiency under low-concentration conditions. The desorption rates from the first to the fifth cycle were 85.0%, 84.5%, 84.0%, 83.0%, and 80.0%. The desorption rate was found to be 80% after five cycles. In addition, phosphorus recovery can be achieved by regeneration operation. The results of phosphorus removal after regeneration of different adsorbents reported in the literature are shown in Fig. 7b41,42,43,44,45,46,47,48,49. The ability of Fe/AlPMt to remove phosphorus after regeneration is better than most of the previous adsorbents.

Fig. 7
figure 7

(a) Regeneration performance of Fe/AlPMt in a five-cycle operation; (b) Comparison of phosphorus removal capacity after regeneration of various adsorbents. Error bars represent the standard deviation (n = 3 independent experiments).

Full size image

Conclusions

This investigation advances the mechanistic understanding of phosphate sequestration using metal oxide-modified mineral clay composites in aquatic remediation. The Langmuir model demonstrated an adsorption capacity of 51.407 mg P g− 1. Zeta potential analyses and XPS characterization confirmed hydroxyl-phosphate ligand substitution as the predominant adsorption mechanism on Fe/AlPMt, with spectroscopic evidence of Fe/Al-phosphate complexation. Operational parameters optimization revealed dose-dependent performance: At 0.3 g L− 1 Fe/AlPMt dosage, phosphate concentration in lake water decreased from 0.507 to 0.005 mg L− 1 (99% elimination efficiency). Reduced dosage (0.01 g L− 1) achieved phosphate reduction from 0.507 to 0.208 mg L− 1. These findings establish a framework for engineering metal oxide composites with tailored surface properties, enabling selective pollutant adsorption in complex aqueous matrices through rational material design.