Kinetics, Isotherms and Adsorption–Desorption Behavior of Phosphorus from Aqueous Solution Using Zirconium–Iron and Iron Modified Biosolid Biochars

Abstract

Excessive discharge of phosphorus (P) to aquatic ecosystems can lead to unpleasant eutrophication phenomenon. Removal and recovery of P is challenging due to low C/N ratios in wastewater, hence the development of efficient removal and recovery of P strategies is essential. In this study, zirconium–iron (Zr–FeBC) and iron modified (Fe–BC) biosolid biochars were examined to investigate their capacity for the removal of P by batch experiments. The influence of solution pH, biochar dose, initial P concentration, ionic strength, interfering ions and temperature were also studied to evaluate the P adsorption performance of biochars. The P experimental data were best described with pseudo-second order kinetics and the Freundlich isotherm model. The maximum P adsorption capacities were reached to 33.33 and 25.71 mg g⁻¹ for 24 h by Zr–FeBC and Fe-BC at pH 5 and 4, respectively. Desorption studies were performed to investigate the reusability, cost-effectiveness and stability of the adsorbents Zr–FeBC and Fe-BC. The adsorption–desorption study suggests that both examined biochars have considerable potentiality as adsorbent candidates in removing as well as recovery of P from wastewaters. Results also reveal that the regenerated Zr–FeBC and Fe–BC could be utilized repetitively in seven adsorption–desorption cycles using NaOH as a desorbing agent, which greatly reduces the P-removal cost from wastewaters. Thus, P enriched biochar could potentially be used as fertilizer in the agriculture sector.

1. Introduction

Phosphorus (P) is one of the essential primary nutrients controlling soil fertility, plant growth and agricultural productivity [1]. However, the quality of water may deteriorate and cause eutrophication when excess (due to the emission of wastewater discharge or runoff system) P is present in aqueous systems [2,3]. Excess P in aquatic systems may occur from excessive fertilization in agricultural systems or release from point sources, such as municipal waste treatment plants, industrial sources (e.g., cheese factories, abattoirs) and animal husbandry (e.g., dairy and piggery operations). The excess inputs may indirectly impact humans and animal health via stimulation of cyanobacteria populations. Moreover, the effective recovery and utilization of P from wastewater sources is critical for the continued productivity of agricultural systems. Therefore, P removal and recovery for beneficial reuse is of critical importance globally, given the limitation of P sources into the future [4,5].

The coagulation–flocculation, co-precipitation and crystallization, adsorption, ion exchange, complexation–sequestration, membrane technologies, biological and electrochemical treatment are commonly researched methods used in removing P, including heavy metals (arsenic, copper, chromium, nickel, zinc, mercury etc.) from aqueous solutions [6,7,8,9,10,11,12]. Of these, adsorption is a promising approach due to it being quick, simple and economical for P removal [13].

Single and bimetal adsorbents have been reported for P removal and recovery from waters. Metal-based adsorbents have included metals such as Fe, zero-valent iron (ZVI) [14], magnesium oxide (MgO) [15], aluminum (Al), zirconium (Zr) [16,17], cerium (Ce), and lanthanum (La) [18]. Zr-modified merlinoite [19], ZrO nanoparticles [16], calcium-alginate beads [20], Zr-modified chitosan beads [21], Zr-loaded carbon nanotubes [22], Zr-bentonite alginate hydrogel beads [23], Fe-La doped porous silica [24], Fe-Al-hydroxides [25], Zr/Al-pillared montmorillonite [26], Ni-Al-Zr hydroxide [27], Mg(OH)₂/ZrO₂ [13], zinc ferrite/activated carbon hybrid complex [28], Zr-Ca montmorillonite [29], silver nanoparticles-loaded activated carbon [30], Fe and Zr treated activated carbon nanofibers [31] have been examined to recover P from water. Several studies have investigated the metal/metal oxide modified materials, including activated carbon, modified clay, modified biochars, and nanomaterials used for remediation of pollutants such as As, Cr, Sb, P, V, Cd, Ni, Pb, etc., in wastewater and found that the leaching of metals from modified materials is minimal and biocompatibility tests showed low toxicity [32,33,34].

There is an important difference between the conventional P removal techniques and P recovery processes. While P removal aims to obtain a P free effluent by transferring P to sludge using chemical and biological processes, P recovery focusses on obtaining a P-containing product that can be used as a fertilizer or as a soil conditioner. A substantial proportion of the available literature on P adsorption technologies focuses on removal rather than recovery and reutilization. One of the commercially utilized processes for P removal from wastewaters is chemical precipitation, which involves adding cationic salts such as calcium (Ca), magnesium (Mg), iron (Fe) or aluminium (Al), to water containing phosphate, precipitating insoluble phosphate complexes, and separating precipitates from the aqueous phase [35,36]. Adsorption based products show considerable potential in the recovery of P for reutilization for agriculture.

Biochar has attracted a wide range of treatment technologies in removing inorganic and organic contaminants due to its low cost, available sources, environmental friendliness, multiple surface functional groups, large specific surface area (SSA), and facile fabrication [37,38]. However, the performance of pristine biochar is still low in removing P. The P removal capacity has been reported to be of 0.132, 0.036 and 0.296 mg g⁻¹ for biochars produced from garden wood waste, corncobs and wood chips, respectively [39]. Biochar with metal oxy(hydrox)ides in single or binary-surface modifications could significantly improve the P adsorption capacity. For example, the P sorption capacity was reached to 57.39 mg g⁻¹ by goethite-corn straw biochar composite [40]. Another study showed a maximum P sorption capacity of 5.07 mg g⁻¹ using FeO-water hyacinth biochar [41]. Following this, Jack, et al. [42] synthesized a magnetic biochar using fungal encapsulated-Fe waste biomass (Neurospora crassa a Fe containing wastewater effluent), which reached up to 23.9 mg g⁻¹ P sorption capacity. Biochar/AlOOH [43] and Mg-O [44] nanocomposites were also reported increased P removal efficiency. La-loaded magnetic (using 1:1 Fe²⁺ and Fe³⁺) pineapple biochar showed improved P adsorption capacity up to 101.16 mg L⁻¹, which was 27 times higher than its pristine biochar [45]. Hydrous ZrO₂ was synthesized by Rodrigues et al. [17] and Su et al. [16], where the P sorption capacity was reported to 67 and 99.01 mg g⁻¹, respectively.

In addition, the oxides of Fe–Zr have been effectively utilized for PO₄³⁻ [31,46], AsO₄³⁻ [32,47], SbO₄³⁻ [48,49] and Cr⁶⁺ [31] removal. As(V) and P are chemical analog elements with many similarities [50]. Hence, it is potentially feasible that Fe–Zr-treated biochar composites could significantly contribute to P removal from wastewater.

In this study, Zr, Zr–Fe and Fe-reacted biosolid biochar (BC) were synthesized and characterized. The removal of P from aqueous solutions was explored using batch adsorption methods. The influence of solution pH, initial P concentration, biochar density, ionic strength, interfering anions and cations were examined on P removal. Finally, the desorption behavior was investigated for the stability and reusability of Zr, Zr–Fe and Fe- reacted BC’s.

2. Materials and Methods

2.1. Reagents and Materials

All analytical grade reagents were purchased from Sigma–Aldrich, USA. The glassware used during the experimental and analytical work were washed in 3% HNO₃ solution followed by rinsing three times with ultrapure water. A stock solution of 1000 mg P L⁻¹ was prepared from potassium phosphate monobasic (H₂PO₄K). The biosolid biomass (BM) was obtained from Winmalee sewage treatment plant in Winmalee, New South Wales, Australia. The working BM was stored at ambient temperature (approx. 23 °C) after being oven dried at 80 °C for 12 h and ground to a powder (<1 mm, 50 mesh).

2.2. Preparation of BC

50 g air-dried and ground (<1 mm, 50 mesh) BM was employed in a ceramic crucible covered with a lid and heated in a muffle furnace under a N₂ atmosphere. The BC was prepared as per the procedure of Rahman, et al. 2021 [51], where slow pyrolysis was performed at a rate of 7 °C min⁻¹ holding at a peak temperature 300 °C for 30 min. The resulting BC samples were allowed to cool at room temperature inside the furnace. Afterwards, the BC was removed from the furnace, stored in airtight plastic containers and preserved in a desiccator for further experiments.

2.3. Synthesis of Zr–BC, Zr–FeBC and Fe–BC

BC was modified by using Zr-salt (ZrOCl₂.8H₂O), Fe-salt (FeCl₃.6H₂O), and combination of Zr and Fe-salt, where subsequently denotes as Zr–BC, Fe–BC and Zr–FeBC. All BC composites were synthesized by the in-situ co-precipitation method [32]. Concisely, 10 g of ground BC was emerged separately into 100 mL ZrOCl₂·8H₂O solution (0.1 M), and a combination of 100 mL 0.1 M ZrOCl₂·8H₂O and 0.1 M FeCl₃·6H₂O (biochar: solution ratio was 1:10) adjusting solution pH to 6.5 through dropwise addition of NaOH (1.0 M). The resulting suspension was aged for 12 h at approx. 24 °C. The synthesized Zr–BC and Zr–FeBC composites were rinsed several times by purified water to remove impurities after centrifuging at 5500 rpm for 10 m followed by decanting the solution. Finally, BC composites were dried in an oven at 80 °C and preserved in a desiccator for further batch adsorption studies.

Seven types of Zr–FeBC composite biochar’s were synthesized in varied molar ratio of Zr to Fe, namely 1:1, 1:2, 1:5, 1:10, 1:20, 1:50 and 1:100. The composite biochar products were labelled as Zr–FeBC1, Zr–FeBC2, Zr–FeBC5, Zr–FeBC10, Zr–FeBC20, Zr–FeBC50 and Zr–FeBC100, respectively. However, based on our preliminary adsorption results (data not shown), Zr–FeB20 showed the best performance, hence we continued all subsequent experiments with Zr–FeB20 referred to as Zr–FeBC. Additionally, an Fe-coated BC was produced using FeCl₃.6H₂O solution as described above and referred to as Fe–BC. Surface charge characteristics, zeta potential and elemental composition (C, N, S) of all synthesized materials were performed by Surface area analyzer (Tristar^TM II 3020, Micromeritics, USA), NanoPlus HD analyzer (Micromeritics, Norcross, GA 30093, USA), and LECO TruMac C/N/S analyzer. Specific surface area (SSA), pore volume and pore size distribution were estimated using N₂ sorption (Tristar^TM II 3020, Micromeritics, USA). Other characterizations such as SEM, EDS, TEM, TEM-EDS, XRD and FTIR were reported in our previous publications [49].

2.4. Batch Adsorption Setup: pH, Kinetics and Isotherms

Sorption edge (i.e., effect of pH) experiments were conducted in a pH range of 2–11 at an initial P concentration of 10 mg L⁻¹ and biochar suspension composite density of 2 g L⁻¹ at room temperature (22 ± 0.5 °C). The pH of biochar suspensions was controlled by adding 0.1 M HNO₃ and/or 0.1 M NaOH. Kinetics studies were conducted using 0.1 g composite biochar, which was added to 50 mL Falcon tubes (biochar to solution ratio = 1:500) containing 10 mg L⁻¹ P for 3 days at room temperature (22 ± 0.5 °C). To maintain a constant ionic strength, the background solution contained 0.01 M NaNO₃. Suspensions were centrifuged at 5000 rpm for 20 min and the supernatants were filtered through 0.22 µm PES membrane syringe filters. Adsorption isotherms were conducted with the similar procedure as the kinetic study, except for the variable initial P concentrations (5 to 300 mg L⁻¹). The reaction was stopped after 24 h.

In this study, kinetic data were fitted with four classical kinetic models, namely pseudo-first-order, pseudo-second-order, Elovich and intraparticle diffusion models (detailed information of all models are described in Supporting Information). In addition, three sorption isotherm models were used to fit the equilibrium experimental data of P sorption to modified biochars, namely the Langmuir, Freundlich and Temkin models (see detailed information of all models described in Supporting Information). The amount of adsorbed P onto the biochars was calculated using the following Equation (1).

$$q=\frac{ \left(C_i-C_f\right)\times \mathrm{V}}{\mathrm{M}}$$

(1)

where C_i and C_f are the initial and final P concentration (mg L⁻¹) in the solution; V (L) is the volume of the P solution; M (g) is the mass of the biochars and q is the adsorbed amount of P (mg g⁻¹). The concentration of P was estimated by the inductively coupled plasma optical emission spectrometry (Avio 200, PerkinElmer, Waltham, MA, USA).

2.5. Effect of Major Anions and Cations on P Adsorption

The coexisting anions (Cl^–, NO₃^–, SO₄^2–, and CO₃^2– at concentrations of 0.01 M) and cations (Na, K, Mg, and Ca at concentrations of 0.01M) were also studied. Biochar suspensions were adjusted to pH 5 (Zr–FeBC) and pH 4 (Fe-BC) using 0.1 M HNO₃ and/or 0.1 M NaOH. The biochar density was 2 g L⁻¹ and initial P concentration was 20 mg L⁻¹ at 22 ± 0.5 °C.

2.6. Effect of Ionic Strength and Thermodynamics Studies

The influence of ionic strength (0.01, 0.05, 0.1, 0.5 and 1.0 M of NaNO₃) has been studied. The change of entropy (∆S) and enthalpy (∆H) are important thermodynamic parameters for the identification of spontaneous reactions. The thermodynamic parameters ∆S, ∆H and the Gibbs free energy (∆G) can be calculated using the linearized van’t Hoff equations as follows (Equations (2)–(4)) [52]:

∆G = −RTlnK

(2)

Kc = q_e/C_e

(3)

∆G = ∆H − T∆S

(4)

where R is the ideal gas law constant 8.314 × 10⁻³ kJ mol⁻¹ K⁻¹, T is the absolute temperature (K) and Kc is the distribution coefficient, which is the ratio of the equilibrium sorption quantity to the equilibrium concentration of P. The final equation can be written as:

lnK = ∆S/R − ∆H/(RT)

(5)

Based on equation Equation (2), ∆H and ∆S parameters can be calculated from the slope and intercept of the plot of lnKc versus 1/T, such as using Equation (6):

∆H = −Slope × R and ∆S = Intercept × R

(6)

2.7. Desorption Study

To investigate the reusability and release of P from biochars, seven adsorption–desorption cycles were performed utilizing desorbing agents such as 0.05 M ammonium sulfate [(NH₄)₂SO₄], and 0.1 M NaOH (biochar density was 2 g L⁻¹ and initial P concentration was 20 mg L⁻¹) after washing and oven drying followed by continuous shaking for 24 h at 22 ± 0.5 °C. To check the reproducibility of biochars, each experiment was carried out under identical conditions. The desorbed P was tested by ICP-OES and the desorption efficiency of adsorbent was calculated employing the following Equation (7) [32,53]:

% desorption, P = C_des/C_ads × 100

(7)

where, C_des and C_ads are the desorbed amount (mg L⁻¹) of P in the solution and adsorbed amount of P by biochars.

All experiments (pH edge, kinetics, adsorption isotherm, anions and cations including ionic strength and thermodynamics) were conducted in triplicates and average results were reported.

3. Results and Discussion

Biochars physico-chemical characteristics are listed in

Table 1. Physico-chemical characteristics of biochars.

Biochar	Specific Surface Area (BET), (m2/g)	Pore Volume (cm3/g)	Pore Size/ Diameter (nm)	Average Particle Size (nm)	pHPZC	EC (mS/cm)	pH		CEC (cmol(+)/kg)
							In H2O	In CaCl2
BC	4.64 ± 0.1	0.006 ± 0.002	6.51 ± 1.41	1292 ± 76	3.6 ± 0.15	0.32 ± 0.09	7.12 ± 0.12	5.98 ± 0.13	8.1 ± 0.61
Zr–BC	75.85 ± 2.73	0.06 ± 0.005	3.37 ± 1.01	79 ± 10	3.8 ± 0.1	10.34 ± 1.15	8.97 ± 0.14	8.37 ± 0.15	6.24 ± 0.68
Zr–FeBC	25.51 ± 2.81	0.027 ± 0.003	3.92 ± 0.21	235 ± 32	5.7 ± 0.08	18.57 ± 1.6	5.64 ± 0.12	5.17 ± 0.11	5.63 ± 0.57
Fe–BC	24.02 ± 2.17	0.019 ± 0.002	4.80 ± 0.46	249 ± 12	4.5 ± 0.09	20.54 ± 2.27	5.88 ± 0.13	5.45 ± 0.16	5.06 ± 0.46

. The pH of biochars indicated that the modified Zr–FeBC and Fe–BC are acidic characteristics whereas BC and Zr–BC showed basic properties (Table 1). BET analysis demonstrated that both the SSA and pore volume of metal incorporated biochar are much higher than pristine biochar (Table 1). The reason is due to the loading with the Zr or Fe particles; and combination with Zr and Fe particles on the biochar surface. Thus, loading with metal (single or dual) on biochar surface greatly affected the structure of the pore size to be opened on the surface of the Zr, Fe or Zr–Fe modified biochar [54]. Therefore, the SSA of BC, Zr–BC, Zr–FeBC and Fe–BC was 4.64, 75.85, 25.51 and 24.02 m² g⁻¹, which increases with smaller particle size [54] and the average particle size was determined to 1292, 79, 235 and 249 nm, for BC, Zr–BC, Zr–FeBC and Fe–BC, respectively (Table 1). Additionally, SSA positively correlated with an increased pore volume of biochar [55].

3.1. Batch Experiments: P adsorption on Biochars

3.1.1. Influence of pH

The pH is a significant and primary parameter during adsorption. In this study, results indicated that P sorption efficiency increases with increasing pH up to 5. Maximum adsorption for Zr–FeBC and Fe–BC was observed at pH 5 and 4, respectively. The P removal efficiency decreased almost 3-fold (from 3.94 to 1.11 mg g⁻¹ (93% to 30.13%) and 1.97 to 0.82 mg g⁻¹ (78.47% to 28.52%)) when pH values increased from 5 to 11 and 4 to 11 for Zr–FeBC and Fe–BC, respectively (Figure 1A). This can be explained by the pKa value of phosphate. The phosphate species mainly exists as PO₄^3–, HPO₄^2– and H₂PO₄^– (pKa values of 12.3, 7.2 and 2.1) in aqueous solution [56]. The common phosphate ions mostly exist as H₂PO₄^– at pH 4–6, that are comparatively more adsorbed than HPO₄^2– or PO₄^3– at pH > 6 [56,57]. High pH can lead to electrostatic repulsion between negatively charged phosphate ions and Zr–OH/Fe–O groups, which results in a declined trend of P adsorption since deprotonation occurs at the biochars surface at higher pH [57,58]. The higher P adsorption capacity of thermally treated alum sludge has been reported at pH 4–6 [57], which is supported in this study.

Zeta potential (ZP) results showed that surface of Zr–FeBC biochar possessed a positive charge between pH 2 to 5 (ZP value varied from +25.02 to +9.98 mV) and then the negative surface charge increased gradually with increasing pH from 6 to 11 (ZP value varied from −4.78 to −25.68 mV) (Table S1 in Supporting Information section). Similarly, positive surface charge increased up to pH 4 (ZP value varied from +12.23 to +5.66 mV) and then decreased onwards for Fe-BC (negative ZP value increased from −5.46 to −15.36 mV) (Figure 1B). Thus, the isoelectric point (pH_PZC) was found to be 5.7 and 4.5 for Zr–FeBC and Fe–BC, respectively.

The biochar surfaces are protonated at pH < pH_PZC and formed positive potentiality, which is partly responsible for enhancing P adsorption. At pH > pH_PZC, the P adsorption was decreased due to formation of the negative charge of biochars surface by deprotonation reaction, which repulsed between (–)ve biochar surface and phosphate anions [59]. In addition, another reason for reduced P sorption could be the competition between PO₄^3– and OH^– through ligand exchange on the biochar surface at higher pH [60,61]. Therefore, acidic pH (protonation of biochar surface) favored the P adsorption of Zr–FeBC and Fe–BC by the following Equations (8)–(10) [58]:

Zr–OH + H⁺ = Zr–OH₂⁺

(8)

H₂PO₄⁻ + Zr–OH₂⁺ = (H₂PO₄)⁻(Zr–OH₂)⁺

(9)

HPO₄²⁻ + 2Zr–OH₂⁺ = (HPO₄)²⁻(Zr–OH₂)₂⁺

(10)

The pH results also demonstrated that the pristine BC and modified Zr–BC could not sorb P from aqueous solution. This may be due to these two biochars being saturated with P content and a substantial negative P sorption was thus observed. Therefore, repulsive force between P and BC and Zr–BC occurred. The amount of P content was measured to 54.86, 26.83, 6.12 and 7.86 mg g⁻¹ in BC, Zr–BC, Zr–FeBC and Fe–BC, respectively (Table S2 in Supporting Information section). Next, all experiments were carried out with Zr–FeBC and Fe–BC.

3.1.2. Adsorption Kinetics and Fitted Model

To understand the P removal mechanisms and to evaluate the practical applicability of biochars, the investigation of kinetics is fundamental, as kinetic models offer valuable insights into the mechanisms of adsorption procedure. In this study, the amount of P removal increased with time and reached close to the equilibrium state within 24 h (Figure 2A). To probe the rate-controlling step of kinetic reaction by Zr–FeBC and Fe–BC, the pseudo-first order, pseudo-second order, intraparticle diffusion and Elovich model (Figure 2B–E) were utilized; kinetic model parameters are listed in

Table 2. The parameters of the kinetics model for P adsorption.

Biochar	qe-exp (mg g−1)	Pseudo First-Order			Pseudo Second-Order				Elovich			Intraparticle Diffusion			pH
		k1 (h−1)	qe-cal	R2	k2 (g mg−1.h−1)	qe-cal (mg g−1)	h (mg g−1.h−1)	R2	β (mg g−1)	α (mg g−1.h−1)	R2	kid (g mg−1 h−1/2)	C (mg g−1)	R2
Zr–FeBC	2.63	0.07	0.93	0.83	0.24	2.67	1.68	0.99	1.48	0.28	0.98	0.17	1.38	0.85	5
Fe–BC	1.72	0.08	0.92	0.84	0.07	1.88	0.26	0.99	0.21	0.36	0.96	0.21	0.19	0.94	4

. Results revealed that the kinetics data best fitted with the pseudo-second order model (R² = 0.99) (Figure 2C) followed by the Elovich model (R² = 0.96–0.98) (Table 2).

The Elovich model suggests heterogeneous adsorption on the biochar surface as well as describes a range of activation energy during sorption. The plot of q_t vs t_1/2 from the intraparticle diffusion model (Figure 2D) showed non-linearity with no intercept, which represented that the P adsorption process could be controlled by more than a single step. Based on the pseudo-second order model, the experimental sorption data in equilibrium (q_e-exp) were determined to be 2.63 and 1.72 mg g⁻¹, which is more consistent with the calculated value (q_e-cal) of 2.67 and 1.88 mg g⁻¹ by Zr–FeBC and Fe–BC, respectively (Table 2).

The closer the value between q_e-exp and q_e-cal supported the pseudo-second order kinetics with the chemisorption process [32,53,57]. The parameters of the pseudo-first order model suggest that the pseudo-first order model was not applicable (Table 2). In this study, kinetics results suggested that P adsorption was followed through surface complexation between PO₄^3– and Zr–FeBC/Fe-BC surfaces through sharing or exchange of electrons between the biochar surface and phosphate, which indicated that the chemisorption was the rate-controlling step during P removal [16,45]. Similar kinetics results were reported by Rodrigues and da Silva [62], Tang et al. [63] and Vu et al. [64] on various adsorbents, such as hydrous niobium oxide, zirconia functionalized SBA-15 and steel-making slag.

3.1.3. Influence of Initial P Concentration and Isotherms Study

The P adsorption increased rapidly with increasing initial P concentration up to 20 mg L⁻¹ at the equilibrium and then increased gradually (Figure 3A). Hence, the Langmuir, Freundlich and Temkin isotherm models were applied to fit for calculating the maximum P adsorption capacity by biochars as well as adsorbents sorption process (surface phenomenon whether homogeneous or heterogeneous). The isotherm model parameters were listed in the

Table 3. Model parameters of adsorption isotherms for P adsorption.

Biochar	qexp (mg g−1)	Langmuir Model Parameters					Freundlich Model Parameters				Temkin Model Parameters			pH
		qcal (mg g−1)	qm (mg g−1)	KL (L mg−1)	RL	R2	qcal (mg g−1)	KF (g mg−1.h−1)	1/n	R2	b (J mol−1)	A (L g−1)	R2
Zr–FeBC	30.89	28.89	33.33	0.04	0.1–0.75	0.95	30.16	3.45	0.42	0.99	392	0.51	0.92	5
Fe–BC	22.62	22.31	25.71	0.03	0.11–0.76	0.99	24.74	2.17	0.46	0.99	519	0.55	0.96	4

and Figure 3. In general, the Langmuir model described a monolayer adsorption process applied onto a uniform surface, while the Freundlich model is applied for heterogeneous distribution with multiple layers of adsorption [15,42]. In the current study, the calculated coefficient of determination (R²) by Langmuir, Freundlich and Temkin models were 0.95 and 0.99, 0.99 and 0.99, and 0.92 and 0.96 for Zr–FeBC and Fe–BC, respectively. Therefore, the experimental data were best described with the Freundlich model (R² = 0.99 for both biochars), which demonstrated the adsorption of P occurred via multilayer and heterogeneous surfaces of the biochars. The maximum P adsorption capacities were achieved to 33.33 and 25.71 mg g⁻¹ by Zr–FeBC and Fe–BC, respectively by the Langmuir isotherm model. The Freundlich constant K_F value of Zr–FeBC (3.45) was higher than Fe–BC (2.17). The biochar surface heterogeneity or adsorption intensity (1/n) for both Zr–FeBC and Fe–BC were less than unity, which indicated P adsorption is controlled by a chemisorption process [65].

The maximum P absorption capacity was reported as 4.18 mg g⁻¹ using zeolite/steel slag/fly ash/basalt composite fiber [66] at pH 4, which is 7.97 and 5.33 times lower than this study. However, the maximum adsorption capacity was reported to 46.9 mg g⁻¹ at pH 2 when biomagnetic nanoparticles were used (CoFe₂O₄@γ-Fe₂O₃) [67], which is 1.4 and 1.8 times higher than this study.

The dimensionless separation factor (R_L) was determined from the Langmuir model to predict the affinities between the P and biochar surfaces. The nature of the adsorption isotherms are favorable, linear and unfavorable if the R_L value is calculated to be of 1 > R_L > 0, R_L = 1 and R_L > 1 [45].

The R_L values ranged between 0.1–0.75 and 0.11–0.75 (Table 3) for Zr–FeBC and Fe–BC, respectively, which is less than unity and thus indicative the P sorption process was favorable in this study. During P sorption, the Temkin model suggested the structure of biochar surface was heterogeneous (R² = 0.92 and 0.96 for Zr–FeBC and Fe–BC) with a range of binding energy. The value of bonding energy (b) from the Temkin model were 392 and 519 J mol⁻¹ for Zr–FeBC and Fe–BC, which indicated that the involved adsorption process was both chemisorption and physisorption [21].

3.1.4. Effect of Biochar Concentration and Ionic Strength

The effect of biochar concentration on P removal was examined with a range of doses from 1–10 g L⁻¹ at pH 4 and 5 for Fe–BC and Zr–FeBC, respectively. Results showed that P adsorption capacity increased from 3.11 to 3.46 mg g⁻¹ (11.2%) and 2.23 to 2.55 mg g⁻¹ (14.3%) for Fe–BC and Zr–FeBC, respectively, when biochar concentration increases from 1 to 2 g L⁻¹ (Figure 4A). The maximum adsorption capacity of both Fe–BC and Zr–FeBC reached at biochar concentration of 2 g L⁻¹, which described that the biochar surface is fully saturated with the P molecules at this biochar concentration [68]. The enhanced P removal capacity could be attributed to the increased biochar surface area available, which provides more active adsorption sites for adsorption of the P [14,40]. However, the P adsorption capacity decreased to 61.6% (3.46–1.33 mg g⁻¹) and 56.1% (2.55–1.12 mg g⁻¹) with an increase in the biochar concentration from 2 to 10 g L⁻¹ for Fe–BC and Zr–FeBC, respectively using the same P concentration (20 mg L⁻¹) (Figure 4A). A decreasing trend of P removal efficiency may be associated with the unsaturated active sites on the biochar surface [69]. Therefore, the excess amount of biochar was not significant for the contribution of P adsorption from solution when all P was adsorbed by biochar surface. Thus the P adsorption was correlated negatively with biochar concentration, which is also wasteful [32].

To identify the adsorption mechanisms, the effect of ionic concentration on P sorption behavior can be applied as a significant tool [70]. Results demonstrated that with increasing NaNO₃ concentration from 0.01 M to 1.0 M, the P adsorption efficiency decreased by up to 51.1% and 58.8% by Zr–FeBC and Fe–BC, respectively (Figure 4B). This could be the formation of outer-sphere complex of P with Zr–FeBC and Fe–BC during adsorption process [16].

3.1.5. Effect of Co-Existing Ions and Thermodynamic Study

Sorption selectivity is a significant factor for the investigation of the influence of P removal effectiveness. Surface or wastewater generally contains common naturally occurring anions (Cl^–, NO₃^–, SO₄^2– and CO₃^2–) and cations (Na⁺, K⁺, Ca²⁺ and Mg²⁺), which may compete with P-sorption sites and significantly influence the P-sorption efficiencies. In this study, there was a limited decrease in P-sorption capacity of biochars induced by Cl^–, NO₃^– and SO₄^2–. However, P-removal declined from 80 to 66.77% and 98.43 to 90.23% when reacted to CO₃^2– by Zr–FeBC and Fe–BC, respectively (Figure 5A). This inhibition could be due to the strong competition between PO₄^3– and CO₃^2– for adsorption sites on biochars surfaces. In addition, the competition between PO₄^3– and CO₃^2– was found to be more prominent using Zr–FeBC compared to Fe–BC biochar, which could be influenced by the Fe or Fe/Zr ratios. The literature has shown that there was negligible interference from common ions such as Cl⁻, NO₃⁻, SO₄²⁻, HCO₃⁻, Ca²⁺ and Mg²⁺ with concentrations up to 100 mg/L, indicating the strong selectivity of the La(OH)₃/Fe₃O₄ material [71]. Another study showed that no significant decline in phosphate removal (initial phosphate concentration was 10 mg L⁻¹) to Cl⁻, NO₃⁻, SO₄²⁻ and HCO₃⁻ anions using La-modified bio-ceramisite [72]. Other adsorbents, such as La(OH)₃ nano composites, La-loaded mesoporous nanospheres and La(OH)₃ modified BC, showed excellent selectivity facilities on the recovery of phosphate results [45,73,74]. Among the analyzed common coexisting ions (NO₃⁻, CO₃²⁻, Cl⁻, SO₄²⁻), only SO₄²⁻ considerably affected the P adsorption, reducing the removal capacity in 7.3% for smaller nanoparticles and 12.4% for larger nanoparticles of CoFe₂O₄γ-Fe₂O₃ [67]. The La-based Fe-oxide nanosheets showed high selectivity for phosphate removal with less than 10% reduction in phosphate in the presence of NO₃⁻, HCO₃²⁻, Cl⁻, SO₄²⁻ and humic acids [75]. However, similar anionic effects on P removal were reported using zeolite/steel slag/fly ash/basalt composite fiber [66]. No inhibitory impact of NO₃⁻, Cl⁻ and SO₄²⁻ was observed on P removal using graphite modified magnesium oxide nanoparticles [76]. No significant inhibition or enhancement of P-removal was noticed when reaction with common cations, including Na⁺, K⁺, Ca²⁺ and Mg²⁺, in the current study (Figure 5A).

The thermodynamic parameters suggested that the P adsorption process is an endothermic reaction (∆H values > 0, indicated endothermic process) where the randomness of the P-molecules were increased at the biochar-solution interface due to having the positive values ∆S [15] (

Table 4. Thermodynamic parameters for the P adsorption on biochars.

Biochar	∆G (kJ mol−1)						∆H (kJ mol−1)	∆S (kJ mol−1 K−1)	pH
	277 K	288 K	293 K	298 K	303 K	313 K
Zr–FeBC	−0.67	−1.11	−1.79	−2.31	−2.77	−3.38	26.53	0.25	5
Fe–BC	−0.53	−0.97	−1.22	−1.45	−1.95	−2.47	18.27	0.14	4

). The energy balance of the whole adsorption process could be explained by ∆H values. The P species were solvated and lose part of their hydration shell to bind on the biochar surface in aqueous solution [67]. This process consumes energy, which is larger compared to the energy when the P-species are adsorbed [67,77]. Positive ∆S values for both biochars may be associated with extra translational entropy acquired by the released solvent particles to the biochar surface [67,78].

The negative ∆G values were increased with raised in temperature which indicating a spontaneous and feasible reaction during P adsorption. In addition, results revealed that the higher the temperature, the more ∆G, which demonstrates that the P adsorption is more favorable at greater temperature, in the investigated temperature range from 4 to 40 °C (Figure 5B). Thus, P adsorption efficiency is likely more favorable at elevated temperatures where chemisorption reaction to be the dominant reaction mechanism [69] interacting electrostatically with surface positive sites surrounded by solvated P ions. Guerra, et al. [67] synthesized biomagnetic nanoparticles (CoFe₂O₄@γ-Fe₂O₃) for P adsorption that showed similar thermodynamic characteristics.

3.2. Desorption and Reusability Study

Desorption tests were conducted to seven adsorption–desorption cycles for evaluating the regeneration possibility of biochars. Results demonstrated that for Zr–FeBC, the P removal efficiency drastically reduced from 70.43% to 52.33% from the first to seventh cycles using desorbing agent (NH₄)₂SO₄. However, NaOH is effective up to 74% removal efficiency of P, even up to seven cycles (Figure 6A). Initially the removal efficiency of P was 95.1% after the first cycle, 81.34% after the fourth cycle and 74% after the seventh cycle when using NaOH. Therefore, the reduction of P removal efficiencies was 13.74% after the fourth cycle and 21.13% even after seven cycles for Zr–FeBC. Similarly, the desorption efficiencies of Fe–BC was decreased from 83.43% to 67.33% using (NH₄)₂SO₄ and 98.1% to 78.97% using the NaOH solution from the first to the seventh trials, respectively (Figure 6B). The leached Fe ions from Zr–FeBC and Fe–BC were determined to be 0.36 and 0.75 mg/L after the first regeneration cycle, while the concentration of Fe was 0.45 and 0.87 mg/L after the seventh cycles at pH 5. Zr leached from Zr–FeBC was 0.1 mg/L and 0.15 mg/L after the first and seventh adsorption–desorption cycle, respectively at pH 5. The amount of recovered Zr–FeBC was 97.23, 95.17, 89.75, 85.41, 81.37, 78.71 and 75.11% from first to the seventh cycles, whereas the recovered Fe–BC was 99.12, 97.87, 95.55, 92.18, 88.73, 84.38 and 80.13%, respectively. Therefore, the recovered Zr–FeBC and Fe–BC could be reused for P-adsorption from wastewater, which significantly reduces the operational cost. It was reported that the phosphate desorption performance was higher than 70% for all five cycles by La(OH)₃/Fe₃O₄ nanocomposite [71] whereas the adsorbents used in this study were effective up to 78.97%, even after seven cycles.

Technology should be simple and economically viable for applicability of materials being proposed for the removal of P from aqueous solution. When taking cost effectiveness into account, the costs of Zr–FeBC and Fe–BC were calculated based on the prices of raw materials and processing expense. The costs of Zr–FeBC and Fe–BC were estimated to be AUD 0.21 and 0.17/kg, whereas the cost for treatment of P removal from water containing 20 mg/L of P was calculated to be AUD 0.32 and AUD 0.26 per 100 L, respectively.

4. Conclusions

This study reported the applicability of modified Zr–FeBC and Fe–BC for P adsorption using batch experiments. Kinetic data suggested more than a single step could control the P adsorption process. The study revealed that both modified BCs are effective in removing P from aqueous solution based on maximum P adsorption capacity. Carbonates significantly influenced the adsorption capacity for Zr–Fe and Fe modified biochar, with a slight reduction for Fe modified BC. As both modified BCs are effective in removing P up to seven recycle times, these can be utilized practically at field application (large scale set up). The recovered P can be utilized in the agricultural sector as P fertilizer, which may enhance the soil health and productivity as P sources are limited. Future studies should focus on the mechanism of adsorption and how microbes influence the adsorption process.