Use of Chitosan and Chitosan–Magnetite Spheres for Arsenic Groundwater Removal: Factorial Designs as Tools to Optimize the Efficiency of Removal

Abstract

The lack of access to drinking water is a problem affecting many regions worldwide. In Santa Fe, Argentina, the population uses groundwater as a source of drinking water. Unfortunately, it has high concentrations of As(V), which makes it unsuitable for consumption. Despite several methods for As(V) quantification and elimination, the high cost and technical difficulties in their implementation make many of them cost-ineffective, especially in small communities. In this work, a hybrid sorbent of magnetite–chitosan spheres (M-Q) is synthesized, and its sorption capacity is evaluated by employing groundwater with high conductivity (12.1 mS/cm) and hardness (1125 mg/L CaCO₃) and an As(V) concentration of 0.265 mg/L. A colorimetric analytical technique, which is sufficiently sensitive, simple, and economical to apply, is used for As(V) quantification. The experimental results indicate that the sorption capacity for As(V) removal is 80.1% (sorbent mass 0.466 g, time 85.3 min, and pH 5.9), with the advantage of the capability of being independent of its magnetic properties. The optimal experimental conditions for As(V) sorption (pH, time, and mass of the hybrid sorbent M-Q) are obtained by response surface factorial designs, which significantly reduce the total number of experiments and, at the same time, demonstrate that all the selected variables significantly affect the As(V) sorption percentage (response studied).

1. Introduction

Water is a natural and indispensable resource; it is of fundamental importance since life cannot exist without it [1]. Human exposure to high concentrations of arsenic (As) present in groundwater/aquifers is one of the most widespread environmental problems in many countries [2]. The International Agency for Research on Cancer (IARC) classifies As as group I (sufficient evidence of carcinogenicity in humans) [3], while the United States Environmental Protection Agency (USEPA) describes it as a group A human carcinogen [4]. As is a metalloid that is widely distributed in the environment in water, rocks, soil, and air. It contaminates plants and animals and, consequently, the food consumed by humans. As in water is present in its organic and inorganic forms, with the inorganic form being the most common. As has different oxidation states: arsenites (AsO₃³⁻) and arsenates (AsO₄³⁻) are widespread in water bodies. The toxicity of As increases considerably with the reduction of its oxidation state from As(V) to As(III); additionally, the organic forms are less toxic than the inorganic ones.

As represents a global and regional problem: in Argentina, the presence of As in the environment, specifically in water for human consumption, is essentially due to natural factors, with some anthropogenic contributions [5,6]. The detection and quantification of As in the aquifers of Argentina is the subject of intense scientific activity. The purpose of these studies is to define regions based on the degree of contamination so as to obtain an in-depth understanding of the dimensions of this problem and promote initiatives that provide improvement solutions [7]. The World Health Organization (WHO) recommends 0.01 mg/L as the maximum permitted level of As for human and agricultural use to reduce the negative impacts of this metalloid on human health [8,9]. Published studies indicate that, in groundwater from Buenos Aires, Santa Fe, Córdoba, and Tucumán, the predominant species is As(V) [10]. Concerning As(III), its prevalence was only recorded in some surface water areas of the provinces of San Juan, La Pampa, and Neuquén [5,10,11,12]. In Argentina, the permitted value for drinking water varies according to the province. The allowed limit of As in Santa Fe is 0.05 mg/L [13,14]. The concentrations in the groundwater of Western Santa Fe and the city of Córdoba exceed 0.1 and 0.5 mg/L [15], respectively. Different treatment technologies have been used to reduce its concentration and avoid its harmful effects on the environment and people’s health, which are being constantly studied and tested internationally [16] and are based on physical–chemical methods (oxidation/reduction, coagulation–filtration, precipitation, adsorption and ion exchange, solid/liquid separation, physical exclusion, membrane technologies [17,18]), as well as biological methods (phytoremediation, electrokinetic treatment [19,20]). All these methods have advantages and disadvantages [21]. For example, reduction–precipitation and coagulation–flocculation are the most common methods. However, their cost is high because they require the participation of chemical agents (which cannot be reused), and they generate toxic sludge that is difficult to handle and treat. In addition, the high cost of many of these conventional water methods is prohibitive, especially in small communities.

The reverse osmosis technique deserves special mention. Most of the towns near the central cities (Santa Fe, the capital, and Rosario) do not have drinking water networks due to the nonexistence of aqueducts. This long-standing lack of infrastructure is not related to the lack of water, since the Paraná River is large and long enough to supply the central drinking water generation plants; the problem lies in the fact that drinking water does not reach the towns, so they must source themselves from groundwaters that have high hardness, significant contaminants, and As in variable quantities (often exceeding the permitted limits). For this reason, reverse osmosis plays a prominent role. Most towns use it (El Trebol, San Jorge, Piamonte, Roldán, Andino, etc.). This technique provides the population with immediate drinking water, but some drawbacks exist. The main one is that the wastewater, which contains 90–95% of the ions in groundwater, is stored in large tanks or basins and, in the worst case, returned to the groundwater. In this way, the problem of providing drinking water is solved immediately, but the concentration of contaminants in the groundwater (including As(V)) remains unresolved.

The need for economical and effective methods for the elimination of different pollutants has resulted in the development of new separation technologies, such as remediation. This technique uses sorbents such as algae and agro-industry waste [22,23,24,25], biopolymers, or mixtures of biopolymers with metal oxides (hybrid sorbents) [26,27]. Several studies have been carried out during the last few years using hybrid materials to eliminate As from water [28,29,30,31]. These hybrid sorbent materials use chitosan (Q) [32] as a support or matrix for metal oxides (such as magnetite (M), maghemite, and others [33,34,35]). Our research group has been working on the elimination of As(V) in water for several years, first using chitosan [36,37] in discontinuous (batch) and continuous (column) systems. Then, we incorporated zerovalent nano-iron into our sorbent for As(V) removal, with continuous systems in all cases, using the sorbent in a powder or nano-aggregate form [38]. Recently, we have addressed the possibility of eliminating As(V) present in water using a hybrid sorbent synthesized based on Q and M. The first experimental results, obtained in distilled water, were promising [38], so it was decided to evaluate the removal capacity of As(V) in groundwater. In this way, this work is a continuation of a previous one [38,39], and it is demonstrated that incorporating M into the Q gel, generating a hybrid sorbent, M-Q, increases the As(V) sorption percentage and gives more mechanical stability to the M-Q sorbent during the stirring/contact time. Additionally, central composite experimental designs (CCD) were carried out to evaluate the sorption capacity of the sorbent, identify the factors that significantly influenced the response under study, detect possible interactions of factors with statistical significance, and optimize the response and percentage of As(V) sorbed.

2. Results and Discussion

2.1. Groundwater Analysis

High-hardness groundwater from the city of Andino was analyzed. The results of these analyses are shown in

Table 1. Groundwater characterization.

Analyte	Sample (mg/L)	Limit CCA (mg/L)	Method
pH	7.8 1	6.5–8.5 1	Potentiometric
Total dissolved solids	11,392.0	300–6000	Gravimetric
Nitrates	16.0	45.0	Selective electrode
Nitrites	0.003	0.10	Colorimetric
Fluoride	2.0	1.5	Colorimetric
Ammonia	ND	0.2	Colorimetric
Sulfates	240.7	400.0	Turbidimetric
Phosphates	2.2 × 10−3	1.5	Colorimetric
Silicates	41.6	10.0–30.0	Colorimetric
Conductivity	12.1 2	***	Conductimetric
Hardness	1125 3	51 3	Titrimeric
As (total)	0.015	0.01	Colorimetric

¹ pH units; ² mS/cm; *** until 2012, the current regulations of Santa Fe did not establish limits; ³ expressed as mg/L of CaCO₃; ND: not detected; CAA (Argentine Food Code, using the Spanish acronym).

.

Comparing the values in Table 1 with those established by the CAA (Argentine Food Code, using the Spanish acronym) shows that the groundwater at Andino is unsuitable for human consumption and must be purified [40]. The concentration of As(V) is below the limits permitted by the province of Santa Fe. However, in periods of scarce rainfall, the concentrations tend to rise. Additionally, the northwest of the province naturally presents higher values (>0.10 mg/L) [41]. For this reason, we decided to increase the concentration of As(V) to 0.265 mg/L. Significant amounts of phosphate, silicate, and sulfate species were also detected, high enough to interfere with some methods of determining As(V), such as the Molybdenum Blue method. For this reason, it was decided not to use it during the development of the experimental part, and it was replaced with the DDCAg method [42].

2.2. Sorbent Characterization

2.2.1. pH Value at pH_zc

The experimentally determined pH_zc value of the Q spheres was 7.25 and that for the Q-M spheres was 7.8, as shown in Figure 1. Therefore, at a working pH lower than these values, the surface of the sorbent is positively charged. Considering the diagram of As(V) species as a function of the pH [21], at this pH value, the negative H₂AsO₄⁻ species of As(V) predominates, being the one that is electrostatically attracted to the surface of the sorbent.

2.2.2. Iron Percentage Determination for M-Q Spheres

The average iron content in the sorbent spheres of M-Q, determined experimentally, was 51.0% w/w.

2.2.3. XRD Spectroscopy

Figure 2 shows the characteristic signals of M, which are in agreement with those indicated in the diffraction pattern library (COD—Crystallography Open Data Base). The diffractogram exhibits peaks at the corresponding angles and intensities at 2θ detailed below: 18.358°, 30.154°, 35.503°, 43.073°, 53.441°, 57.061°, 62.661°, 74.382°, and 89.676° for M and 20.050°, 22.412°, 20,050°, and 26.791° for Q. The presence of well-defined peaks in the diffractogram indicates that M is a material with high crystallinity. Figure 2 compares the diffractograms of Q, M, M-Q, M-Q-As (M-Q spheres treated with As(V)), and M-Q-As-T (M-Q spheres treated with As(V) and kept in a normal atmosphere for six months). In the diffractogram of Q, the presence of the characteristic signal of the polymer is observed at approximately 20° with the maximum intensity, which is related to the semi-crystalline order in the structure; this signal overlaps with the crystalline peaks of M and represents the disordered parts of Q [43,44,45]. The slight broadening of the peaks in the diffractogram of M-Q-As indicates a decrease in the degree of semi-crystallinity of the polymer. The XRD of a sample of a hybrid M-Q sorbent after being treated with an As(V) solution ([As(V)] = 0.265 mg/L) for 1 h and exposed to air for six months at a temperature of 25 °C (average) is shown. As can be seen, the peak pattern is not altered, so the oxidation of Fe(II) to Fe(III) does not occur, at least under these conditions. In other words, there is no evident conversion to hematite.

2.3. As Quantification

Numerous techniques can be applied for As quantification, such as inductively coupled plasma mass spectrometry (ICP-mass), electrochemistry, and atomic absorption, among others. They are all sensitive and precise. Their greatest drawbacks are the difficulties in access and high costs. This leads to the use of other spectrophotometric techniques, which are somewhat less sensitive but compatible with the current regulations: spectrophotometric techniques [46]. The spectrophotometric method using silver diethyldithiocarbamate (in pyridine) was used in this work to quantify As(V). This is a standard method, widely used in many laboratories [47]. The methodology and preparation of the calibration curve are explained in Section 3.4. The calibration curve, the deduction of the mathematical equations, and the corresponding parameters are summarized in the Supplementary Materials.

2.4. Optimization of the As(V) Sorption Process—Experimental Designs

As(V) removal studies were carried out in groundwater to determine the effects of inorganic salts on the sorption process. To increase the Q sorption capacity, M-Q spheres were prepared. Two central composite experimental designs (CCD) were carried out: one was used to evaluate the sorption capacity of the Q spheres, and the second was used for the M-Q spheres. The objective was to identify the factors that significantly influenced the response under study, detect possible interactions of the factors with statistical significance, and optimize the response and the As(V) removal percentage sorbed. Response surface modeling, applying design of experiments–surface response optimization (DOE-SRO), is more appropriate than the univariate method or one variable at a time (OVAT) when the objective is optimization because (1) SRO provides global information, while OVAT only provides local information; (2) SRO considers interactions between the factors and OVAT does not; and (3) the number of experiments required for SRO is lower than for OVAT [48,49].

Table 2. Values for each factor generated using CCD for the M-Q and Q sorbents.

M-Q Spheres				Q Spheres
Sorbent Mass (A)	Contact Time (B)	pH (C)	As(V) Removal Percentage	Sorbent Mass (A)	Contact Time (B)	pH (C)	As(V) Removal Percentage
g	min	pH	%	g	min	pH	%
0.625	80.0	6.5	63.4	0.426	80.0	4.0	5.4
0.340	80.0	9.0	45.7	0.426	80.0	6.5	15.5
0.340	80.0	6.5	69.1	0.213	120.0	8.0	17.2
0.510	120.0	8.0	52.8	0.426	80.0	6.5	15.8
0.340	12.7	6.5	35.5	0.426	13.0	6.5	16.2
0.170	40.0	8.0	37.4	0.784	80.0	6.5	33.0
0.170	40.0	5.0	30.9	0.639	40.0	8.0	22.2
0.340	80.0	4.0	65.2	0.213	120.0	5.0	29.0
0.510	40.0	5.0	74.0	0.068	80.0	6.5	13.2
0.510	120.0	5.0	78.8	0.639	40.0	5.0	16.3
0.340	80.0	6.5	75.5	0.426	80.0	9.0	2.3
0.340	80.0	6.5	71.0	0.213	40.0	5.0	3.0
0.340	80.0	6.5	73.6	0.426	80.0	6.5	16.2
0.170	120.0	5.0	28.0	0.639	120.0	8.0	29.5
0.510	40.0	8.0	34.3	0.639	120.0	5.0	30.6
0.054	80.0	6.5	23.6	0.426	80.0	6.5	17.0
0.170	120.0	8.0	45.4	0.426	147.0	6.5	40.8
0.340	147.3	6.5	50.1	0.213	40.0	8.0	2.5

shows the different experiments carried out, where the factors (variables) are combined as determined by the CCD and the experimental response (As(V) removal percentage).

The analysis of variance (ANOVA) is a fundamental statistical tool in the analysis of factorial designs. Response surface factorial designs are used to explore and model the relationships between independent variables (factors) and a dependent variable (response) to optimize the response.

Table 3. ANOVA for quadratic model. Response: As(V) removal percentage.

Source	Sum of Squares	df	Mean Square	F Value	p-Value Prob > F	Q-M	Sum of Squares	df	Mean Square	F Value	p-Value Prob > F	Q
Model	5831.31	9	647.92	121.36	<0.0001	S	2085.61	9	231.73	425.16	<0.0001	S
A—sorbent mass	1989.16	1	1989.16	372.57	<0.0001		146.71	1	146.71	269.17	<0.0001
B—time	205.34	1	205.34	38.46	0.0003		800.02	1	800.02	1467.79	<0.0001
C—pH	407.25	1	407.25	76.28	<0.0001		32.11	1	32.11	58.92	<0.0001
AB	41.41	1	41.41	7.76	0.0237		45.60	1	45.60	83.66	<0.0001
AC	1003.52	1	1003.52	187.96	<0.0001		36.55	1	36.55	67.06	<0.0001
BC	75.65	1	75.65	14.17	0.0055		41.86	1	41.86	76.80	<0.0001
A2	1197.45	1	1197.45	224.28	<0.0001		79.91	1	79.91	146.61	<0.0001
B2	1261.34	1	1261.34	236.25	<0.0001		248.34	1	248.34	455.63	<0.0001
C2	385.47	1	385.47	72.20	<0.0001		235.46	1	235.46	431.99	<0.0001
Residual	42.71	8	5.34				4.36	8	0.5451
Lack of Fit	18.85	5	3.77	0.4741	0.7823	NS	3.09	5	0.6186	1.46	0.4004	NS
Pure Error	23.86	3	7.95				1.27	3	0.4225
Cor Total	5874.03	17					2089.97	17

S: significant; NS: not significant.

shows the ANOVA of the quadratic model, which best fits the selected response’s experimental data (As(V) removal percentage), for both types of sorbent spheres: Q and M-Q.

An analysis of variance (ANOVA) is the most reliable way to evaluate the quality of an adjusted model, and it was used to verify the statistical significance of the ratio of the mean square due to regression and the mean square due to residual errors (F-value). The regression model equation can explain the variation in the response when a large F is acquired. An associated p is used to decide whether F is large enough to indicate statistical significance. If the p for a large F is less than 0.05, it means that there is 95% confidence, and the model is statistically significant. The probability p-value, standard error of coefficient (SE coefficient), and T-value were applied to determine the significance of the regression coefficients for each parameter. p is the minimum significance level in rejecting the null hypothesis and determining which variables are statistically significant, SE measures the variation in estimating the coefficient, and T is the ratio of the coefficients to the standard error. These effects are significant when their probability level is 5% (p < 0.05). Generally, the larger and smaller the values of T and p, respectively, the more significant the corresponding coefficient terms will be. A positive value for the regression coefficient means an ameliorating effect, while a negative sign indicates the opposite effect of the factor on the response.

The ANOVA results in Table 3 show that the quadratic model used is significant when applied to As(V) sorption with Q or Q-M. This is corroborated by the fact that the lack of fit is not substantial; in other words, the experimental data correlate well with the applied model. The variables studied (sorbent mass, pH, and reaction time), as well as their interactions, significantly affect the response, namely the As(V) removal percentage, as indicated by the large F values and the small p values (≤0.01). On the other hand, the statistical prediction parameters shown in

Table 4. Fit statistics.

M-Q				Q
Std. Dev.	2.31	R2	0.9927	Std. Dev.	0.7383	R2	0.9979
Mean	53.02	Adjusted R2	0.9845	Mean	18.09	Adjusted R2	0.9956
C.V. %	4.36	Predicted R2	0.9683	C.V. %	4.08	Predicted R2	0.9859
		Adeq. Precision	31.5957			Adeq. Precision	71.4052

indicate that (a) the R²-predicted vs. R²-adjusted are close enough (difference (≤0.01) to indicate the absence of outliers and (b) the adequate precision values of 30.1 (M-Q) and 74.5 indicate that the signal/noise ratio is satisfactory since they are greater than 4.

The second-order model expression for the parameters studied for the optimization of As(V) sorption can be expressed in coded units using Equation (1) for the hybrid M-Q sorbent’s sorption. See the Supplementary Materials for the As(V) sorption using Q.

As(V) removal percentage = 72.19 + 12.08 A + 3.88 B − 5.48C + 2.28 AB − 11.20 AC + 3.08 BC − 9.74 A² − 9.97 B² − 5.60 C²

(1)

Figure 3 and Figure 4 show the relationship obtained between the experimental data and those predicted by the model. The data shown for As(V) sorption with M-Q and Q fit well. No outliers were observed, with a significance level of 0.05, and the residuals were scattered randomly around ± 4.56.

From Equation (1), we can infer that the sorbent mass (including its interactions with the pH and time) is the factor that most affects the response. The sorbent mass positively affects the response, while the pH does so negatively. This means that if the pH increases, the response decreases, and if the pH decreases, the response increases. This is reflected in Figure 5.

The model optimizes a response of 80.1% sorption. A possible explanation would be the following: As(V) exists in aqueous media as an oxoanion (arsenate) due to its high charge/radius ratio. Speciation depends on the experimental pH. The H₃AsO₄ species has three pKa [50].

pK_a₁ = 2.19: this corresponds to the first deprotonation,

H₃AsO₄ → H₂AsO₄⁻ + H⁺

(2)

pK_a₂ = 6.94: second deprotonation,

H₂AsO₄⁻ → HAsO₄²⁻ + H⁺

(3)

pK_a₃ = 11.5: third deprotonation,

HAsO₄²⁻ → AsO₄³⁻ + H⁺

(4)

The proportion of species present in the reaction medium can be easily estimated from speciation diagrams obtained with the software “Medusa” https://www.kth.se/che/medusa/downloads-1.386254 accessed on 30 September 2024 (Make Equilibrium Diagrams Using Sophisticated Algorithms, Ignasi Puigdomenech, Royal Institute of Technology 100 44 Stockholm, Sweden). Figure 6 shows the speciation diagram and species fraction vs. pH (1–12) for As(V) with an initial concentration of 3.54 µM (0.265 mg/L). Medusa uses the pK_a 1–3 values to calculate the species fraction.

As shown in Figure 1, the surface of the M-Q sorbent is positively charged at pH values lower than 7.8; on the other hand, at pH 5, almost 100% of the As(V) is negatively charged (H₂AsO₄⁻). As the pH increases, the species composition of the mixture changes (the H₂AsO₄⁻ species decreases and the HAsO₄²⁻ species slightly increases), but the sorbent surface begins to lose its positive charge (Figure 1). To conclude, there is a range of favorable pH values for the sorption of As(V) within the acidic zone, which can be explained by the electrostatic attractions between the species of the As(V) oxoanions and the positive surface charge of the M-Q sorbent’s surface. Figure 7A shows this behavior: there is no significant variation in the response at pH 7.4–8. The As(V) sorption increases significantly as the pH decreases at first and then at a lower percentage (pH 5.6–5).

The sorbent mass and time also play important roles in the sorption of As(V). According to Equation (1) and Figure 7, a zone with the minimum sorption of As(V) is observed (pH = 5 and sorbent mass = 0.17 g). As the sorbent mass increases, the number of sorption sites also increases. This effect is quite evident in the pH range of 5.6–5 for a high sorbent mass. However, the sorption is almost independent of the pH at extremely low sorbent masses; see Figure 7.

The independence of As(V) sorption is also observed when the sorbent amount is very low. As the sorbent mass increases, the response increases; see Figure 8.

These results were validated in the laboratory in duplicate. The sorption percentage found experimentally for an M-Q sorbent mass of 0.466 g, time of 85.3 min, and pH 5.9 was 75%.

3. Materials and Methods

3.1. Reagents

3.1.1. Inorganic Reagents

Hydrochloric acid 36% w/w, δ = 1.18 g/mL (Cicarelli, Rosario, Argentine), sulfuric acid 98% w/w, δ = 1.84 g/mL (Cicarelli), nitric acid (Cicarelli) 65% w/w, δ = 1.53 g/mL, sodium hydroxide (Cicarelli), sodium nitrate (Anedra, Buenos Aires, Argentine), ammonium hydroxide 25–30% w/v, δ = 8.9 g/mL (Cicarelli), hydroxylamine hydrochloride (Anedra).

Reagents provided by GT Laboratory (GMP, ISO 9001:2008 and ISO 13485:2003 standards, Rosario, Argentine): KI 0.1 M, SnCl₂ 0.15 M, Zn (As-free) shots, lead acetate filters, silver diethyldithiocarbamate (DDCAg), pyridine solution p.a. (BioPak, Buenos Aires, Argentine) 0.982 g/mL, As(V) standard 100 mg/L.

Fe(III) standard 1000 ug/L (ChemLab, Zedelgem, Belgium). Ammonium ferrous sulfate heptahydrate Fe(NH4)₂(SO4)₂.6H₂O salts were used as a source of Fe(II) ions, and ferric chloride hexahydrate FeCl₃.6H₂O as a source of Fe(III) ions, which were provided, respectively, by Biopack and Cicarelli and were of analytical grade.

3.1.2. Organic Reagents

Acetic acid 99.5% w/w, δ = 1.67 g/mL (Cicarelli); chitosan (Parafarm, Buenos Aires, Argentine), with degree of acetylation of 98% and Mv = 1.61 × 10⁵ g/mol; pH buffer solutions (4.0, 7.0, and 10.0) (Cicarelli); antimony and potassium tartrate (Mallinckrodt, Dublin, Ireland); ortho-phenanthroline (Cicarelli); ammonium acetate (BioPak).

3.1.3. Control of Physical–Chemical Parameters

A Hanna Instruments 8519 pH meter and a Hanna Instruments Dist 4 conductivity meter were used to measure the pH, temperature, and conductivity of the solutions and samples. A Cole-Parmer^® 8891E-DTH sonicator bath (Vernon Hills, IL, USA) was used. The groundwater was characterized using the following commercial kits (using the methodology of the kit): nitrate measurements were performed with a HORIBA LACUA TWIN reader (selective electrode, Kyoto, Japan), and sulfate measurements were performed with an AQAssay kit (turbidimetric, by precipitation of barium sulfate [51], Rosario, Argentine). Phosphate measurements were performed with a kit from Wiener Lab (colorimetric, based on the ascorbic method [52], Rosario, Argentine). Silicate measurements were performed with an AQAssay kit (colorimetric, based on the molybdosilicate method [53], Rosario, Argentine). Ammonia measurements were performed with an AQAssay kit (colorimetric, based on the salicylate method [54], Rosario, Argentine). Nitrite measurements were performed using a colorimetric method, with an AQAssay kit (colorimetric, based on the sulfanilamide method [55], Rosario, Argentine). Fluoride measurements were performed with an AQAssay kit (colorimetric, based on the zirconium oxychloride and eriochrome cyanine method [56], Rosario, Argentine). Hardness measurements were performed with an AQAssay kit (titrimetric [57], Rosario, Argentine). TDS measurements were performed via gravimetry [58].

3.2. Sorbent Synthesis

3.2.1. Synthesis of Chitosan (Q) Spheres

First, 4.0 g of Q was dissolved in 100.0 mL of acetic acid (4.0% w/v at 40 °C). The solution was homogenized with a magnetic stirrer for 24 h and slowly dripped onto NaOH 4 M, with gentle and constant stirring. The obtained spheres were incubated for 24 h in the sodium hydroxide solution to complete the gelling process and washed with enough distilled water to a neutral pH. Finally, the Q spheres were dried at room temperature (25 °C) until they reached a constant weight and stored in caramel-colored glass containers [39,59].

3.2.2. Magnetite (M) Synthesis

All synthesis was carried out in an inert nitrogen atmosphere: 7.84 g of Fe(NH₄)₂(SO₄)₂.6H₂O and 10.80 g of FeCl₃.6H₂O were weighed and dissolved in 200 mL of distilled water (deoxygenated) to prepare 0.1 M and 0.2 M solutions in Fe(II) and Fe(III), respectively. To this solution, 70 mL ammonium hydroxide (20% v/v) was added, with stirring and under a fume extraction hood, up to pH 9.0, observing the formation of a black suspension; the mixture was then incubated at 80 °C for 30 min [60,61]. The solid obtained was separated by decantation and washed with distilled water (deoxygenated) until it reached a neutral pH. The solid obtained was dried at room temperature until it reached a constant weight, ground to a powder, and stored in a desiccator in a nitrogen atmosphere. The magnetite synthesis reaction is shown in Equation (5):

Fe(NH₄)₂(SO₄)₂ + 2FeCl₃ + 8NH₄OH → Fe₃O₄ + 4H₂O +2(NH₄)₂SO₄ + 6NH₄Cl

(5)

Theoretical mass = 4.64 g; mass obtained = 4.60 g and yield = 98.9%.

3.2.3. Synthesis of Magnetite–Chitosan (M-Q) Spheres

For this, 4.0 g of Q was dissolved in 100.0 mL of acetic acid (4.0% w/v at 40 °C). The solution was mixed with 0.702 g of M, stirred, and homogenized in a sonicator (40 min). The mixture was slowly dripped onto a 4 M sodium hydroxide solution with gentle stirring, observing the gelling of the drops as they fell. The mixed spheres obtained were left to rest in the sodium hydroxide solution for 24 h to complete the gelling process. They were left at room temperature (25 °C) until they reached a constant weight and stored in caramel-colored glass containers.

3.3. Characterization of Sorbent Material

Three different assays, described below, were performed to characterize the sorbent spheres.

3.3.1. Determination of Zero-Charge pH (pH_zc)

The Muller method was followed to determine the pH_zc of both synthesized sorbents [62]. Different solutions with pH values in the range of 1–9 were prepared, using H₂SO₄ and NaOH. Then, 50 mL of each solution was placed in a beaker and the initial pH (pH₀) was measured. The same mass for both sorbents (0.1 g) was added to every solution. After 10 min, 3.45 g of NaNO₃ was added to regulate the ionic strength of the system to a common value of 0.1 M and it was left to stir for 24 h, after which the final pH (pH_f) was recorded, calculating the ∆pH value as the difference between pH₀ and pH_f. The pH_zc value was determined from the ∆pH vs. pH₀ graph as the point at which the graph cut the x-axis (pH₀).

3.3.2. Iron Percentage Determination for M-Q Spheres

Ten spheres of Q-M, with a mass of 0.0262 g, were placed in a beaker with 2 mL concentrated HCl and 2 mL concentrated HNO₃ and heated in a fume extraction hood until total dissolution. The resulting solution was cooled to room temperature, and the volume was adjusted to 4 mL with distillate water and diluted 25 times with distilled water. The iron percentage was determined by the o-phenanthroline method [63]. This method allows the determination of the total iron, after reducing the Fe(III) present in the sample to Fe(II), with hydroxylamine, followed by a posterior reaction with the color reagent ortho-phenanthroline, giving a red–orange complex, whose intensity is measured at 510 nm. The absorbance was obtained with a JASCO V 530 UV–VIS spectrophotometer (JASCO INTERNATIONAL CO., LTD, Tokio, Japan) and was directly proportional to the concentration of iron. The determination was performed in duplicate and contrasted with a calibration curve.

3.3.3. Powder X-Ray Diffraction (XRD)

The diffractograms were obtained with a D2 Phaser X-Ray Diffractometer with a second-generation detector (Bruker, Billerica, MA, USA), applying the parameters in

Table 5. Instrumental setup for XRD.

Instrumental Setup/Methodology Used	Bragg Value/Concept
Geometry—Setup	Bragg Brentano θ-θ
2θ angular range	5–100°
Time per step	1 s
Step size	0.04°
Voltage	30 kV
Current	10
Divergence slot (primary)	mA
2θ angular range	1 mm

.

3.4. As(V) Quantification

For As(V) quantification, an economical, sensitive, and accepted spectrophotometric method, aligned with the current standards, was selected [42,47]. A detailed explanation is included in the Supplementary Materials.

A calibration curve was realized to determine the molar absorptivity of the complex in groundwater and verify the linearity of the experimental system. The detection limit (DL) and quantification limit (QL) were 0.011 mg/L (1.42 × 10⁻⁷ M) and 0.034 mg/L (4.32 × 10⁻⁷ M), respectively. Their calculation is described in the Supplementary Materials.

3.5. Experimental Design Applied to As(V) Sorption

A central composite rotating surface design (CCD) model was chosen. Three factors were considered as independent variables, the time, pH, and sorbent mass, at two levels and with four repetitions of the central point.

Table 6. (A) Coded levels used in the CCD for Q. (B) Coded levels used in the CCD for Q-M.

(A)
Factor	Symbol	Minimum	Maximum	Code Low	Code High	Mean
Sorbent Mass (g)	A	0.068	0.784	−1 ↔ 0.20	+1 ↔ 0.50	0.426
Time (min)	B	13.0	147.0	−1 ↔ 40.0	+1 ↔ 120.0	80.0
pH	C	4.0	9.0	−1 ↔ 5.0	+1 ↔ 8.0	6.5
(B)
Factor	Symbol	Minimum	Maximum	Code Low	Code High	Mean
Sorbent Mass (g)	A	0.060	0.630	−1 ↔ 0.17	+1 ↔ 0.51	0.341
Time (min)	B	13.0	147.0	−1 ↔ 40.0	+1 ↔ 120.0	80.0
pH	C	4.0	9.0	−1 ↔ 5.0	+1 ↔ 8.0	6.5

summarize the ranges and levels of the factors studied.

The total number of experiments in this model, 18, was calculated as 2^k + 2k + n (k is the number of factors and n is the repetitions of the central point). The Design-Expert 11 program was used to prepare the design and for the processing of the experimental information. The response optimization was resolved by setting a range for the three factors, the sorbent mass, time, and pH (within the range of the experimental values used). In addition, the alternative of maximizing the response was chosen (As(V) removal percentage). For this work, the desirability function (which is a mathematical structure) considered (a) restrictions concerning the factors and (b) the objective of maximizing the response [64].

The As(V) removal percentage of each experiment was selected as a response and calculated with Equation (6):

$$\mathrm{As}\left(\mathrm{V}\right)\mathrm{removal}\mathrm{percentage}={\displaystyle \frac{C_0-C_f}{C_0}}\times 100$$

(6)

where C₀ is the initial concentration (0.265 mg/L) and C_f is the final concentration, at time t, of As(V), expressed in mg/L. The temperature and working volume were 25 °C and 200 mL, respectively. A second-order polynomial was used to study the percentage of As(V) removal and identify the most relevant terms of the model. The complete quadratic model is presented in Equation (7).

$$Y={\beta }_0+{{\displaystyle \sum }}_{i=1}^k{\beta }_j{X}_i+{{\displaystyle \sum }}_{i=1}^k{\beta }_{ii}{X}_i^2+{{\displaystyle \sum }}_{i=1}^k{\beta }_j{X}_i+{{\displaystyle \sum }}_{i,j=1\left(i\ne j\right)}^k{\beta }_{ii}{X}_i{X}_j+\epsilon$$

(7)

where Y is the response, β₀ is the constant coefficient, β_i, and β_ii are the linear and quadratic coefficients (respectively) of the independent factor X_i, β_ij is the coefficient of interaction between the independent factors X_i and X_j, and, finally, ε represents the model error.

4. Conclusions

This work addresses the development of a new hybrid sorbent, M-Q, to remove As(V) in groundwater. The synthesis of Q, M, and M-Q spheres was carried out successfully, and XRD was used to characterize them. The resulting spheres of sorbents demonstrated good mechanical resistance when used in batches.

Three different factors, the pH, the time of contact, and the mass of each sorbent (Q and M-Q), were evaluated in order to optimize the response, namely the As(V) removal percentage, through experimental designs, applying a CCD response surface method (multivariate). This methodology has the advantage of providing global information, considering interactions between the factors, and requiring a small number of experiments. The removal of As(V) using only Q spheres as a sorbent was inefficient, considering that it only eliminated 30% under the optimal conditions proposed by the experimental design (pH 6.1, sorbent mass 0.5 g, and time 120 min). Incorporating M into the Q spheres generated a hybrid sorbent with a sorption capacity of 80.1%. Beyond the possible mechanisms that may occur for the capture of As(V), it was found that the (optimal) response increased by 2.7 times when M was incorporated.

Other situations that must be highlighted here are the following: (1) this work was obtained on groundwater, which has several interferent species that could diminish the removal process due to the competition of oxoanions, while one of our previous works was realized with distilled water, and (2) the initial As(V) concentration in this work (0.265 mg/L) is a more accurate reflection of the real As(V) concentration found in groundwater in Santa Fe province and is smaller than the one used in our last work. Under these conditions, the driving force of the removal process is lower, but the removal percentage is more significant, indicating the high affinity of the hybrid sorbent for this contaminant.

It was also demonstrated that magnetite remains unaltered for at least 6 months in aerobic conditions when incorporated into chitosan spheres, as could be seen during the XRD.

Another advantage of this hybrid sorbent is its capability to be separated due to its magnetic properties, as was probed during this work in our laboratory.

It is possible to propose that the increased uptake for As(V) in the hybrid material is due to a combination of electrostatic interactions due to the opposite charges between the sorbent surface and As(V) anions and, possibly, the later formation of Fe-O-As compounds on the sorbent surface, as was found in our previous work.

Considering that the remediation process must be beneficial with regard to the cost of synthesis of the sorbent, process performance, and eco-compatibility, this synthesized hybrid sorbent appears to be a sustainable and economical option to be used in regions with scarce economic resources as an appropriate treatment for drinking water. This is a promising development that could significantly improve the accessibility of clean drinking water in resource-constrained regions.

Additionally, this study used a method that does not require much equipment to detect As(V); measurements can be performed in the area in question without transporting the samples to large cities, thus avoiding contamination. The silver diethyldithiocarbamate technique is accepted for detection and quantification and only requires a few reagents, a fume extraction hood, and a photocolorimeter to read at 530 nm. As long as the appropriate work protocol is respected, this methodology is sensitive and reproducible enough to be used in compliance with the current regulations.