Adsorption of Arsenic from Water Using Aluminum-Modified Food Waste Biochar: Optimization Using Response Surface Methodology

Abstract

Aluminum-impregnated food waste was selected as a filter medium for removing As(III) from aqueous solutions. The modification of food waste and its carbonization conditions were optimized using the Box–Behnken model in the response surface methodology. Pyrolysis temperature and Al content significantly influenced the As(III) adsorption capacity of aluminum-modified food waste biochar (Al-FWB), but the pyrolysis time was insignificant. Several factors affecting the adsorption capacity of the Al-FWB, including the pH, contact time, dosage, competitive anions, and reaction temperature, were studied. The low solution pH and the presence of HCO₃⁻, SO₄²⁻, and PO₄³⁻ reduced the As(III) adsorption onto Al-FWB. The pseudo-second order model showed a better fit for the experimental data, indicating the dominance of the chemisorption process for As(III) adsorption. Langmuir and Freundlich isotherm models fit the adsorption data, but the Langmuir model with a higher (R²) value showed a better fit. Hence, As(Ⅲ) was adsorbed onto Al-FWB as a monolayer, and the maximum As(Ⅲ) adsorption capacity of Al-FWB was 52.2 mg/g, which is a good value compared with the other porous adsorbents. Thus, Al-FWB is a promising low-cost adsorbent for removing As(III) from aqueous solutions and managing food waste.

1. Introduction

The contamination of natural water bodies and wastewater by arsenic is a serious problem that originates from natural and human activities. The composition of natural ores that undergo mining activities and erosion and the use of pesticides are known sources of As [1]. As(V) predominantly exists in surface water bodies, but As(III), with a lower redox potential, is a dominant species in groundwater [2,3]. As(III) is more toxic and mobile than As(V), and it is difficult to remove the As(III) present in wastewater [4]. As is a carcinogen and phytotoxic element with harmful effects on human health and ecosystems [5]. Considering the adverse health and environmental effects, as well as the vast natural exposure to As [6], it is important to investigate and develop methods and technologies to reduce As to the threshold limit (0.01 mg/L) recommended by the World Health Organization [7].

Recent investigations have led to the introduction of various technologies for removing As from water and wastewater, including oxidation-coagulation [8], electrocoagulation [9], ion exchange [10], precipitation [11,12], ultrafiltration [13], reverse osmosis [14], and adsorption [15]. However, owing to the high cost, the requirement for more equipment and different chemicals, and the complexity of these techniques and operating procedures, less attention has been paid to some of these technologies [16]. Consequently, the most utilized and promising technology is adsorption because of its low cost and energy consumption for removing contaminants from water and wastewater. Adsorption technology is widely applicable for environmental remediation and has simple application procedures with a higher efficiency than other conventional technologies [17,18,19]. Using adsorption technology for environmental remediation minimizes environmental impacts and waste production [20]. The commonly employed adsorbents for removing As from aqueous solutions are iron oxyhydroxides [21], iron oxide-modified adsorbents [22], activated alumina [23], and iron-modified activated carbon [24].

Biochar is a low-cost adsorbent produced by pyrolyzing organic material in the absence of oxygen; it is highly efficient at removing contaminants from water and is therefore widely used in environmental remediation [25,26,27]. The properties of various organic residues, including sewage sludge [28], wood [29], rice husk [30], corn cob and stalk [31], Rhodes grass [32], spent mushroom compost [33], pomelo peel [34], orange peels [35], palm kernel shell [36], and palm waste [37] have been studied to produce biochar and have demonstrated different levels of utilization capability [38]. Previous studies have revealed that organic waste materials can be a great source of raw materials for producing biochar through pyrolysis [39].

Many studies have investigated the efficiency of biochar in removing contaminants consisting of cations and polar organic molecules, such as phenolics [40], halogenated compounds [41], and solvents [42], from aqueous solutions. Scientific studies have stated that the ability of biochar to remove positively charged species is dependent on both its high surface area and the negative charge on its surface; however, the negative charge on the biochar surface decreases its tendency for the sorption of anions and oxyanions, such as As [43]. To enhance the negative ion adsorption capacity of biochar, several methods have been developed to modify biochar, such as steam activation [44], alkali activation [45], metal oxide-modified biochar [46], and clay mineral and biochar composites [47].

In this study, aluminum-modified food waste biochar (Al-FWB) was prepared and utilized to remove As(III) from aqueous solutions. No artificially selected single source of food waste was used, but instead the actual food waste collected from food waste treatment plants was used. The effects of a set of experimental conditions, such as solution pH, competitive anions, adsorbent dose, reaction time, initial concentration, and reaction temperature, were studied to evaluate the arsenic removal capability of Al-FWB. Equilibrium, kinetic, and thermodynamic adsorption model analyses were performed further to characterize As(III) adsorption to Al-FWB. Consequently, the objectives of the present study were to (1) assess the As(III) removal ability of Al-FWB and optimize the adsorbent preparation using response surface methodology (RSM), which enables the analysis of the various independent variables influencing the dependent variables in one system with multiple experimental runs [48]; (2) study the effects of solution pH, ionic competition, adsorbent dosage, contact time, initial As(III) concentration, and temperature on the As(III) removal capacity of Al-FWB; and (3) gain insight into the adsorption capability of the biochar and the mechanism of As(III) adsorption.

2. Materials and Methods

2.1. Preparation and Characterization of Adsorbent

The food waste (FW) used in this study is described in our previous study [49]. The FW was collected from households and transported to an FW treatment plant. The water in the FW was partially removed by squeezing it through a screw press and then drying it at 150 °C using a steam boiler. The impurities in the dried FW were separated using a magnet and trommel separator.

The modification process was performed by adding 100 g of raw FW to 300 mL of the solutions prepared by separately dissolving 0.25 M, 0.5 M, and 0.75 M of AlCl₃ in deionized (DI) water, which resulted in 2, 4, and 6 wt% Al contents in a mixture, respectively. The AlCl₃ used for preparing Al-FWB was purchased from Sigma Aldrich (Burlington, MA, USA). The mixture of solution and FW was constantly stirred at 100 rpm overnight. The mixture was then dried at 100 °C for 24 h. Subsequently, the modified FW was carbonized at different temperatures (300, 450, and 600 °C) for various durations (0.5, 2, and 3.5 h) under nitrogen gas. The FW-derived biochar was then crushed, sieved (40 mesh), and stored in a desiccator.

A field emission scanning electron microscope (FE-SEM, S-4700, Hitachi, Japan) was used to investigate the morphologies of the Al-modified biochar. The elemental composition of Al-FWB was analyzed using an X-ray fluorescence (XRF) spectrometer (XRF-1700, Shimadzu, Japan) and an energy dispersive spectrometer (EDS) attached to the FE-SEM. The crystalline structure of impregnated Al and other inorganic minerals were analyzed using an X-ray diffractometer (XRD) (SmartLab; Rigaku, Japan). The functional groups on the surface of Al-FWB were identified using Fourier-transform infrared spectroscopy (FTIR) (Nicolet iS10 spectrometer, Thermo Scientific, Waltham, MA, USA) in transmittance mode. The specific surface area and porosity of the Al-FWB were measured by analyzing nitrogen adsorption with a Brunauer–Emmett–Teller (BET) surface analyzer (Quandrasorb SI, Quantachrome Instrument, Boynton Beach, FL, USA).

2.2. Response Surface Methodology (RSM)

The Box–Behnken model in the RSM was used to optimize the conditions for preparing and modifying Al-FWB for removing As(III) from aqueous solutions. Three independent variables, temperature (300, 450, and 600 °C), modified Al dosage (2, 4, and 6 w/w %), and contact time (0.5, 2, and 3.5 h), were selected, and the arsenic removal percentage was considered as the response variable. Overall, 17 trial combinations were analyzed using the Design-Expert statistical software (version 7.0.0; STAT-EASE Inc., Minneapolis, MN, USA). A quadratic model generated from the experimental data was also used for the analysis of variance (ANOVA) to determine the significance of the achieved model. The recommended highest-order polynomial model of the statistical software was not aliased and had a significant additional term. A quadratic model was generated from the experimental data based on the design and according to the following polynomial equation:

$$y=a_0+a_{1}A+a_{2}B+a_{3}C+a_{12}AB+a_{13}AC+a_{23}BC+a_{11}{A}^2+a_{22}{B}^2+a_{33}{C}^2$$

(1)

where (y) is the predicted amount of As(III) adsorbed onto the unit mass of Al-FWB biochar (mg/g), and (a₀) is the constant coefficient. A, B, and C are the independent variables, namely, temperature (°C), reaction time (h), and impregnated Al dosage (w/w %), respectively. In addition, a_i is a model coefficient parameter. The response variables and design of the present experiment are listed in

Table 1. Box–Behnken design: list of independent variables and levels.

Independent Variables	Symbol	Values of the Coded Factor Levels
		−1	0	+1
Temperature (°C)	A	300	450	600
Pyrolysis time (h)	B	0.5	2	3.5
Al content (w/w %)	C	2	4	6

. The goodness of fit of the model, regression coefficient (R²), lack of fit, and adjusted and predicted R² were also evaluated.

2.3. Batch Adsorption Experiments

For batch adsorption, a set of conical tubes (50 mL), each containing 30 mL arsenic contaminated solution with an initial concentration of 300 mg/L As(III) and 0.1 g Al-FWB, were used. An As(III) stock solution of 1000 mg/L was prepared by dissolving 2.6379 g of As(III) oxide (Sigma Aldrich, Burlington, MA, USA) in DI water. As(III) solutions of various concentrations were prepared by diluting the As(III) stock solution with DI water. The pH of the experimental solution was adjusted by adding 0.1 M NaOH or HCl solution. Adsorption was performed using a digital shaking incubator to maintain a 100 rpm agitation rate and 25 °C temperature for 24 h. Thereafter, the solution was filtered through a 5–10 μm pore size qualitative filter paper (Advantec, Japan) and was filtered using a 0.45 μm syringe filter (25HP045AN, Advantec, Japan). Subsequently, the concentration of As(III) in the experimental solutions was determined using inductively coupled plasma-optical emission spectrometry (ICP-OES; Agilent 5100; Agilent Technologies, USA) after adsorption. The experiment was performed in triplicate.

The effect of the solution pH was examined at a pH range of 3–11, with 100 mg/L initial As(III) concentration, 3.33 g/L adsorbent dosage, 25 °C temperature, and 100 rpm agitation for 24 h. The influence of the co-existing anions on As(III) adsorption was tested in the presence of NO₃⁻, HCO₃⁻, SO₄²⁻, and PO₄³⁻ anions at initial concentrations of 1 mM and 10 mM, with 100 mg/L initial concentration As(III), 3.33 g/L adsorbent dosage, 25 °C temperature, and 100 rpm agitation for 24 h. The 1 mM and 10 mM stock solutions of NaNO₃, Na HCO₃, and Na₂SO₄ were prepared by mixing a 1:1 volume ratio of 2 mM and 20 mM of each chemical stock solution with 200 mg/L of As(III) stock solution. To test the effect of the adsorbent dosage on arsenic adsorption, a series of adsorbent dosages (1.67, 3.33, 6.67, 10.00, and 13.33 g/L) were used with initial As(III) concentrations of 100 and 300 mg/L. The experiment was performed at 25 °C with agitation at 100 rpm for 24 h.

Kinetic, equilibrium, and thermodynamic studies of As(III) adsorption were also conducted. The adsorption kinetics of As(III) in the batch experiment was similar to that in the equilibrium test. However, the batch experiment was conducted for a series of 21 samples, each containing 30 mL of As(III) solution and 0.1 g adsorbent in a 50 mL conical tube, with 0.5, 1, 2, 3, 6, 12, and 24 h contact time. In contrast, for the equilibrium experiment, the initial As(III) concentration varied between 10–1000 mg/L, and the contact time was 24 h. Thermodynamic experiments were performed using 0.1 g FW biochar with 30 mL As(III)-contaminated solution at different temperatures (15 °C, 25 °C, and 35 °C).

2.4. Analysis of the Experimental Data

The adsorption kinetics data were analyzed using pseudo-first order (Equation (2)) and pseudo-second order (Equation (3)) models:

$$q_t=q_e\left(1-e^{-k_{1}t}\right)$$

(2)

$$q_t=\frac{k_2{q}_e^{2}t}{1+k_2{q}_{e}t}$$

(3)

where q_t is the amount of As(III) adsorbed at time t (mg As(III)/g), q_e is the amount of As(III) removed at equilibrium (mg As(III)/g), k₁ is the pseudo-first order rate constant (1/h), and k₂ is the pseudo-second order rate constant (g/mg As(III)/h).

The adsorption equilibrium data were analyzed using the Langmuir model (Equation (4)) and the Freundlich model (Equation (5)):

$$q_e=\frac{Q_m{K}_L{C}_e}{1+K_L{C}_e}$$

(4)

$$q_e=K_F{C}_e^{\frac{1}{n}}$$

(5)

where C₀ and C_e are the concentration of As(III) in the aqueous solution at initial and equilibrium (mg As(III)/L), respectively; K_L is the Langmuir constant related to the binding energy (L/mg As(III)); Q_m is the maximum amount of As(III) removed per unit mass of Al-FWB (mg-As(III)/g); K_F is the distribution coefficient (L/g); and n is the Freundlich constant.

The adsorption thermodynamics data were analyzed using the following equations:

$$\Delta G^0=\Delta H^0-T\Delta S^0$$

(6)

$$\Delta G^0=-RTIn{K}_e$$

(7)

$$\ln K_e=\frac{\Delta S^0}{R}-\frac{\Delta H^0}{RT}$$

(8)

$$K_e=\frac{\alpha q_e}{C_e}$$

(9)

where $\Delta G^0$ is the change in Gibbs free energy (kJ/mol), $\Delta S^0$ is the change in entropy (J/mol K), $\Delta H^0$ is the change in enthalpy (kJ/mol), R is the universal gas constant (=8.314 J/K·mol), T is the absolute temperature (K), K_e is the equilibrium constant, and $\alpha$ is the amount of the adsorbent (g/L).

3. Results and Discussion

3.1. Response Surface Methodology (RSM) Study

In this study, a quadratic model comprising 17 trials (experimental results) was generated using the Box–Behnken design of RSM. The desired range and levels of all three variables, namely, pyrolysis temperature, time, and Al content, are shown in

Table 2. ANOVA results obtained from the response surface model for analyzing the amount of As(III) adsorbed per unit mass of Al-FWB.

Source	Sum of		Mean	F	p-Value
	Squares	df	Square	Value	Prob > F
Model	878.9	9	97.7	14.7	0.0009
A: Temp	44.4	1	44.4	6.7	0.0362
B: Time	12.4	1	12.4	1.9	0.2136
C: Al Content	183.9	1	183.9	27.7	0.0012
AB	0.34	1	0.34	0.05	0.8281
AC	2.64	1	2.64	0.4	0.5479
BC	3.2	1	3.2	0.5	0.5132
A2	591.7	1	591.7	89.2	<0.0001
B2	6.5	1	6.5	0.97	0.3565
C2	25.03	1	25.03	3.8	0.0932
Residual	46.5	7	6.64
Lack of Fit	43.5	3	14.5	19.3	0.0076
Pure Error	3.0	4	0.75
Cor Total	925.4	16

. ANOVA was used to perform an F-value test in order to calculate the significance of the employed RSM model of the Design Expert statistical software (version 7.0.0. STAT-EASE Inc., MN, USA). The highest-order polynomial quadratic model recommended by the results of the dependent variable (Y_A) indicated that the model was not aliased, the additional terms were significant, and the criteria were satisfied. The predicted responses were obtained using the following quadratic model equation of the Box–Behnken design:

$$Y_A=25.16+2.36A+1.25B+4.79C-0.29AB+0.81AC+0.88BC-11.855A^2+1.24B^2-2.44C^2$$

(10)

Terms with a positive sign in front of them show the synergistic effect of the term, while those with a negative sign indicate the antagonistic effect. The significance of the values showing the quality of the model developed and the amount of As(III) adsorbed were determined using $F$-value, $R^2$, adjusted $R^2$, lack-of-fit, and adequate precision examinations. As shown by the ANOVA analysis in Table 2, the F-value of the model, calculated by dividing the mean square of each variable’s effect by the mean square, was 14.72, implying that the model was significant. There is only a 0.09% chance that a “Model F-value” this large could occur because of noise. The model probability was 0.0009, showing a value of less than 0.05, indicating that the calculated values for all model terms were significant. The $R^2$ value showing goodness-of-fit was 0.9498. The $R^2$ indicated that 94.98% of the total variation in the adsorbent capability was ascribed to the independent variables in the adsorbent preparation process. Furthermore, the standard deviation in Equation (10) was 2.58. The closer R² is to one, along with a small standard deviation, the better the model fit will be in providing predicted values closer to the actual values for the response. The adjusted R² value of the model was 0.89. The adequate precision value was 13.34. Adequate precision measures the signal to noise ratio; a ratio greater than 4 is desirable and indicates an adequate signal. According to Table 2, the result of the lack-of-fit test was significant, with a high value of 14.48. Lack-of-fit examination defines model adequacy, and a significant lack-of-fit shows poor fit.

The interaction of the independent variables, namely pyrolyzing temperature (A), time (B), and aluminum content (C), affected the adsorption capacity of the biochar, as illustrated by the three-dimensional response surface curves (Figure 1a–c). Each curve was plotted for the combined effect of a pair of factors, whereas the remaining parameter was set at the center point. The negative and positive signs of the factor coefficient in the quadratic model equation (Equation (10)) indicated whether the effect of a factor on As(III) adsorption onto Al-FWB was synergistic or antagonistic. ANOVA for the response surface quadratic model indicated the significance of the first-order and second-order effects of the variables (p < 0.05). As seen in Table 2, the first-order effect of the temperature and aluminum content showed significant values, while only the Al content was significant in the case of the second-order effect of the variables. There was no significant interaction effect between each pair of variables. The insignificant effect (p > 0.05) of pyrolysis time (B) on the adsorption capacity of metal-impregnated biochar has also been observed in previous studies [50,51,52].

From the positive signs of A and C in the regression model equation, it was inferred that the amount of As(III) adsorbed by Al-FWB increased with the increasing temperature (from 300 °C to 600 °C) and impregnated Al content (from 2 to 6 w/w %). The positive sign of A and the negative sign of A² indicated that the maximum As(III) adsorption capacity of Al-FWB could be obtained between the minimum (300 °C) and maximum (600 °C) values of A. This is consistent with the curvature of the estimated response surface of the As(III) adsorption as a function of A (pyrolysis temperature), as shown in Figure 1a,c. The positive sign of the first and second order of the C factor indicated that the As(III) adsorption capacity of Al-FWB could be enhanced by increasing the Al content. Another study reported the effect of a modified aluminum fraction for improving the adsorption capacity of agricultural biochar [53]. They explained that the capacity of the biochar for As(V) adsorption increased from 0.051 mg/g to 0.48 mg/g as the Al content was raised from 0.00 to 0.5 M, and the enhanced efficiency was attributed to the presence of Al oxide/hydroxide on the surface of the biochar. Furthermore, as illustrated by the curvature of each respective three-dimensional response surface plot presented in Figure 1b,c, the As(III) adsorption onto Al-FWB was dependent on the Al content and was enhanced by the increase in Al content.

The optimum conditions for the synthesis of the with maximized As(III) adsorption were tracked using the Box–Behnken model of the RSM. The highest As(III) adsorption (31.1 mg/g) occurred under the conditions of 3.5 h pyrolysis time, 468.36 °C pyrolysis temperature, and 6 w/w % Al content. Further experiments were performed using Al-FWB synthesized under optimized conditions.

3.2. Physical and Chemical Characteristics of Al-FWB

The surface morphology of Al-FWB was analyzed using FE-SEM, as shown in Figure 2. Al-FWB showed a rough surface with the clumps of Al (hydr)oxide formed by Al impregnation on the Al-FWB. The presence of Al (hydr)oxide on the surface of the biochar is identified by the results of energy dispersive spectroscopy (EDS). As shown in

Table 3. Elemental composition (weight%) of Al-FWB analyzed using an energy dispersive spectrometer (EDS) and an X-ray fluorescence (XRF) spectrometer.

	C	O	Al	Cl	Ca	P	K	Na
EDS	46.8 ± 12.0	15.9 ± 6.7	9.2 ± 3.2	16.3 ± 10.2	8.6 ± 4.4	1.5 ± 0.9	1.8 ± 1.6	0.4 ± 0.1
XRF	41.2	21.5	11.0	12.4	8.6	1.6	1.6	1.3

, 9.2% of Al was present on the surface of Al-FWB. The Al amount of Al-FWB analyzed using XRF also showed a similar result. The Al amount analyzed using EDS and XRF was higher than the Al amount mixed with the food waste (6 w/w%). This result can be explained by the fact that the carbon and water were lost from Al-FWB during the pyrolysis while the Al amount was maintained. The Cl on the surface of the Al-FWB originated from the AlCl₃ and was used to modify the adsorbent. The XRD analysis of Al-FWB performed at 5°–90° did not show any distinct peaks related to Al (Figure 3), because Al is present as an amorphous form on the surface of Al-FWB. Previous studies [53,54] also reported the metals impregnated in the biochar were present in the amorphous form via metallic complexes on the biochar, and the XRD analysis showed no distinct peaks but broad humps. The BET-calculated surface area and pore volume of the Al-FWB were 3.74 m²/g and 0.0087 cm³/g, respectively. The average pore diameter of the Al-FWB was 9.32 nm, corresponding to a mesopore (2 < diameter < 50 nm) classified by the International Union of Pure and Applied Chemistry [55]. The functional groups on the surface of the Al-FWB were investigated by FTIR spectra (Figure 4). A single bond such as O–H (3350 cm⁻¹) and C–O–C (1000–1100 cm⁻¹) stretching vibrations is attributed to functional groups of hemicellulose and cellulose [56]. The peaks at 1918 and 670 cm⁻¹ were attributed to the C-H bending of the aromatic compound, and the peak at 1605 cm⁻¹ corresponded to C=C-C of the aromatic ring stretch [57].

3.3. Effect of Solution pH

The influence of the solution pH on As(III) removal by Al-FWB (Figure 5) was evident as the initial pH increased from 3 to 11. The effect of the solution pH on As(III) adsorption on Al-FWB was studied at an initial As(III) concentration of 100 mg/L and a pH ranging from 3–11. As the initial solution pH increased from 3 to 11, the final pH increased from 6.0 to 9.4, showing that Al-FWB had a buffering capacity during the adsorption process. This buffering effect can be due to the amphoteric nature of the Al-FWB [52]. As depicted in Figure 5, Al-FWB showed a higher efficiency for As(III) adsorption at higher pH values than at lower pH values. Adsorption increased significantly at pH 3 and reached a maximum at pH 11. The adsorption of As(III) onto the biochar surface increased from 13.5 to 16.2 mg/g through the increase of the initial pH from pH 3 to pH 5. A subsequent increase in pH revealed an even higher adsorption capacity, approaching the maximum of 18.5 mg/g at pH 11. As(III) species exist predominantly as H₃AsO₃⁰ in the pH range of 2.0–9.2, but H₂AsO₃⁻ is the dominant form at pH > 9. Therefore, the low adsorption observed in the acidic pH range was attributed to the As(III) species present at pH < 9; there were no electrostatic interactions between H₃AsO₃⁰ and Al-FWB, making the adsorption very weak [58]. The pH of the solution could significantly alter the speciation of As and the composition of the adsorbent surface. H₂AsO₃⁻, the dominant form above pH 9, was preferred for ionic exchange, favoring greater As(Ⅲ) adsorption [59,60].

Owing to the heterogeneous surface of the Al-modified adsorbent, the active sites of the adsorbent for As(III) adsorption had a partially positive charge. Furthermore, H₂AsO₃⁻ could be coordinated to the surface of adsorbents, and the adsorption energy was sufficiently large to overcome acid dissociation [61]. Again, as the pH of the solution increased, the adsorption capacity also increased, and the maximum As(III) uptake occurred at pH 11, proving the variation of As(III) from neutral to negatively charged species. This finding is consistent with those of previous studies investigating the adsorption of As(III) [62,63,64].

3.4. Effect of Competitive Anions

In real water systems, some anions might be present in the surface water and compete with As(III) for the active sites during adsorption [65,66]. In order to study the influence of competing anions on As(III) adsorption onto Al-FWB, the co-existence of anions such as NO₃⁻, HCO₃⁻, SO₄²⁻, and PO₄³⁻, which are suspected of having significant impacts on arsenic adsorption [58], were investigated by mixing 30 mL As(III) solution with the initial concentration of 100 mg/L and 1 mM and 10 mM of different anions in 50 mL conical tubes. As shown in Figure 6, the effect of 1 mM NO₃⁻ on the adsorption of As(III) onto Al-FWB was insignificant, irrespective of the initial concentration, as As(III) is known to form weak bonds with NO₃⁻. This finding is consistent with a previous study that reported an insignificant effect of NO₃⁻ on As(III) adsorption [67]. In the case of HCO₃⁻, PO₄³⁻, and SO₄²⁻, the amount of As(III) adsorbed decreased from 16.3 mg/g to 13.8 mg/g, 14.4 mg/g, and 15.3 mg/g at an anionic concentration of 1 mM and 7.6 mg/g, 13.7 mg/g, and 14.5 mg/g at an anionic concentration of 10 mM, respectively. Therefore, HCO₃⁻, PO₄³⁻, and SO₄²⁻ considerably affected As(III) adsorption in the following order: HCO₃⁻ > PO₄³⁻ > SO₄²⁻ > NO₃⁻ (Figure 6). Kong, et al. [68] reported similar results and attributed the effect to the formation of arseno-carbonate complexes, such as As(CO₃)²⁻, As(CO₃)(OH)₂²⁻, and AsCO₃^⁺, which prevents As(III) adsorption onto the adsorbent surface. The PO₄³⁻ and SO₄²⁻ inhibited the As(III) adsorption by forming a complex with Al present on the surface of Al-FWB [69]. In our previous study [52], the impact of anions on fluoride adsorption by Al-FWB followed the order of PO₄³⁻ > SO₄²⁻ > HCO₃⁻ > NO₃⁻. The higher impact of HCO₃⁻ on the As(III) adsorption by Al-FWB than fluoride adsorption also supports the formation of arseno-carbonate complexes.

3.5. Effect of Adsorbent Dosage

The influence of the adsorbent dosage on As(III) adsorption was examined by varying the amount of adsorbent (1.7–13.33 g/L) at initial arsenic concentrations of 100 and 300 mg/L. The maximum As(III) capacities of 1.7 g/L Al-FWB were 29.99 mg/g and 130.84 mg/g for 100 and 300 mg/L As(III), respectively (Figure 7a,b). The As(III) removal percentage increased with the increasing adsorbent dosage, and the maximum removal percentages were 92.5% and 88.1% for 100 and 300 mg/L of As(III) initial concentrations, respectively, at 13.3 g/L adsorbent dose (Figure 7a,b). The increase in the As(III) removal percentage with the addition of more adsorbents can be explained by the increase in the number of adsorption sites for a fixed As(Ⅲ) concentration [4,70]. The lower adsorption capacity of the adsorbent at higher adsorbent dosages is probably due to the presence of unsaturated adsorption active sites resulting from the increased adsorbent amount at a particular initial As(III) concentration and the interaction between adsorbent molecules (intermolecular attraction) [4,71].

3.6. Adsorption Kinetics

The effect of the contact time on As(III) adsorption was studied over a time range of 0.5–168 h and an initial As(III) concentration of 100 mg/L. Thereafter, the kinetic data were modeled using the pseudo-first order and pseudo-second order kinetic models (Equations (2) and (3)). Insights into the nature of the adsorption process and kinetic parameters are presented in

Table 4. Kinetic model parameters of As(III) adsorption onto Al-FWB at a 100 mg/L initial concentration.

Model	Parameters
Pseudo-first order	qe (mg/g)	k1 (/h)	R2
	19.1	0.298	0.815
Pseudo-second order	qe (mg/g)	k2 (g/mg/h)	R2
	20.5	0.019	0.925

and Figure 8. Figure 8 shows that As(III) adsorption reached 90% of the equilibrium concentration within 48 h at the 100 mg/L initial concentration. The determination coefficients (R²) for the pseudo-first order (R² = 0.815) and pseudo-second order (R² = 0.925) models revealed that the experimental data of As(III) adsorption fit better to the pseudo-second order model, suggesting that the chemisorption, which involves the sharing of electrons between the absorbent and As(III), determines the rate of As(III) adsorption [72,73,74]. The pseudo-second order model predicted the amount of As(III) adsorbed onto Al-FWB at equilibrium to be 20.5 mg/g (Table 4), which is similar to the experimental As(III) adsorption amount at equilibrium (19.9 mg/g).

3.7. Adsorption Equilibrium

Figure 9 shows the As(III) adsorption by Al-FWB under different initial As(III) concentrations (10–1000 mg/L). The adsorption increased considerably with increasing the initial As(III) concentration up to 500 mg/L and almost reached a constant value above 500 mg/L. A decrease in the As(III) adsorption amount above 500 mg/L indicates the saturation of the As(III) ion adsorption sites on the surface of Al-FWB [4,75]. The commonly used equilibrium models, namely, the Langmuir and Freundlich models (Equations (4) and (5)), were used to model the experimental data.

Table 5. Equilibrium models fitted As(III) adsorption onto Al-FWB.

Model	Parameter
Langmuir	Qm (mg/g)	KL (L/mg)	R2
	52.2	0.0186	0.990
Freundlich	KF (L/g)	1/n	R2
	6.1475	0.325	0.964

shows that both equilibrium models fit the experimental data fairly well, as their corresponding determination coefficients (R²) were 0.990 and 0.964 for the Langmuir and Freundlich models, respectively. However, the Langmuir model with a higher R² value (0.990) showed a better fit than the Freundlich model, indicating that As(Ⅲ) was adsorbed onto the surface of the Al-FWB as a homogeneous monolayer [4,76].

As a waste-derived adsorbent, Al-FWB showed a high adsorption capacity compared with other adsorbents reported in the literature (

Table 6. Comparison of the adsorption capacity of different adsorbents for As(III) removal.

Adsorbent	Adsorption Capacity (mg/g)	pH	Temperature	Reference
Fe-Mn binary oxide zeolite	296.23	7	25 °C	[68]
Fe impregnated food waste biocahr	119.5	7.0	25.0	[51]
α—Fe2O3/MCM-41	102.1	8	25 °C	[77]
Al-FWB	52.2	7	25 °C	This study
Fe coated empty fruit bunch and rice husk biochar	30.7–31.4	8	25 ± 1 °C	[64]
ZnCl2-biogas residue biochar	27.67	7 ± 1	25 ± 1 °C	[78]
Magnetic pine cone biomass	18.02	7	25.85 °C	[79]
Fe-Mn-La-biochar composite	14.9	3–7	25 °C	[76]
Almond shell biochar	4.86	~7.2	20 ± 2 °C	[4]
Cerium-loaded cation exchange resin	2.5297	6	25 °C	[65]
Synthetic Kan grass biochar	2.004	7	25 °C	[16]
Ca-Jatrofa charcoal	0.002	10	28 ± 5 °C	[80]
GAC-Cu	0.0144	9–11	29 ± 1 °C	[62]

). The maximum As(III) adsorption capacity of Al-FWB was 52.2 mg/g. The granular nature of Al-FWB is also an important property for As (Ⅲ) removal because it allows for easy separation after adsorption and for use as a filter medium in a fixed bed reactor. Furthermore, Al-FWB is a good substitute for removing As(III) in real water systems by utilizing waste as a resource for producing new products to maximize environmental sustainability.

3.8. Adsorption Thermodynamics

Changes in free energy (∆G°), enthalpy (∆H°), and entropy (∆S°) during the adsorption of As(III) on Al-FWB are presented in

Table 7. The estimated thermodynamic parameters for the adsorption of As(III) on Al-FWB.

Temperature	∆H°	∆S°	∆G°
°C	(kJ/mol)	(J/K mol)	(kJ/mol)
15	5.1	31.6	−3.97
25			−4.29
35			−4.61

. The negative value of (∆G°) for all three temperatures indicated that the adsorption of As(III) on Al-FWB was spontaneous and feasible [81]. The positive value of the enthalpy (∆H°) indicated that the process was endothermic [82], while the positive value of entropy (∆S°) indicated that the randomness at the solid–solution interface increased during the adsorption of As(III) on Al-FWB.

4. Conclusions

RSM was employed to optimize the synthesis conditions, including the pyrolysis time, temperature, and Al content, for preparing Al-FWB. The pyrolysis temperature and Al content remarkably influenced the As(III) adsorption capacity of Al-FWB, but the effect of the pyrolysis time was not statistically significant. Al-FWB showed the highest adsorption capacity under the conditions of 3.5 h pyrolysis time, 468.36 °C pyrolysis temperature, and 6 w/w % of Al content. The characteristics of As(III) adsorption on the optimized Al-FWB were investigated by varying the solution pH, presence of competitive anions, reaction time, initial As(III) concentration, and reaction temperature. The increase in solution pH enhanced As (Ⅲ) adsorption to Al-FWB because of the change in As(III) species to H₂AsO₃⁻ at pH > 9.2. As(Ⅲ) adsorption was reduced in the presence of anions; their impacts followed the order: HCO₃⁻ > HPO₄²⁻ > SO₄²⁻ > NO₃⁻. Kinetic adsorption experiments using the model analysis showed that As(III) adsorption was mainly controlled by chemisorption. The better fit of the Langmuir model for As(III) adsorption than the Freundlich model revealed monolayer As(Ⅲ) adsorption to Al-FWB, and its maximum adsorption capacity was 52.2 mg/g, which is superior to that of other adsorbents reported in the literature. The adsorption of As(III) to Al-FWB was endothermic and spontaneous under experimental conditions. Al-FWB is a promising adsorbent for removing As(III) from aqueous solutions and managing FW.