Arsenic Removal from Mining Effluents Using Plant-Mediated, Green-Synthesized Iron Nanoparticles

Abstract

Arsenic contamination in industrial and mining effluents has always been a serious concern. Recently, nano-sized iron particles have been proven effective in sorptive removal of arsenic, because of their unique surface characteristics. In this study, green synthesis of iron nanoparticles was performed using a mixed extract of two plant species, namely Prangos ferulacea and Teucrium polium, for the specific purpose of arsenic (III) removal from the aqueous environment. Results of UV-visible spectrometry, X-ray powder diffraction (XRD), and Fourier-transform infrared spectroscopy (FTIR) analyses confirmed the formation of iron nanoparticles from Prangos ferulacea (Pf) and Teucrium polium (Tp) extracts. The synthesized Fe nanoparticles morphology was studied via microscopy imaging. The particle size was 42 nm, as assessed by dynamic light scattering (DLS) analysis. Adsorption experiments were also designed and performed, which indicated 93.8% arsenic removal from the aqueous solution at 200 rpm agitation rate, 20 min agitation time, pH 6, initial concentration of 0.1 g/L, and adsorbent dosage of 2 g/L. Adsorption isotherm models were investigated, and the maximum uptake capacity was determined to be about 61.7 mg/g. The kinetic data were best represented by the pseudo-second kinetic model (R² = 0.99). The negative value of Gibbs free energy, the enthalpy (−7.20 kJ/mol), and the entropy (−57 J/mol·K) revealed the spontaneous and exothermic nature of the adsorption process. Moreover, the small quantity of the activation energy confirmed the physical mechanism of arsenic adsorption onto iron nanoparticles and that the process is not temperature sensitive.

1. Introduction

Arsenic is known as a carcinogenic element that may enter # aquatic environments through either natural processes, such as weathering, biological activities, and volcanic emissions, or anthropogenic practices, including mining and metallurgical activities or glass and ceramic manufacture [1]. With an average concentration of 6 mg/kg, arsenic is the twentieth most abundant element within the earth’s crust, being present in over 2000 mineral compounds [2]. It may exist in the form of arsenate (60%), arsenite (20%), sulfide, and sulfur-containing salts (20%). Inorganic arsenic in natural waters is usually found in the form of arsenate oxyanions (As (V)) or arsenite (As (III)). Arsenite is 25–60 times more toxic than arsenate, more mobile in aqueous environments, and more hazardous to the environment. Long-term consumption of arsenic-contaminated water is a threat to human health [3]. Among the chronic effects of exposure to arsenic are the spread of cancers, skin diseases, and infertility. According to World Health Organization (WHO) and United States Environmental Protection Agency (USEPA), the maximum arsenic concentration limit in drinking water is 10 μg/L [4].

Several methods have been proposed for removal of arsenic from aqueous solutions, each having their unique limitations and benefits. The available strategies generally include the oxidation-based [5,6], coagulation, settlement, and filtration methods [7,8]; and membrane [9], ion-exchange [10], and adsorption processes [11,12]. Most arsenic removal methodologies are non-effective for arsenite ions, which are predominantly non-charged at pH < 9.2 [5]. The precipitation, adsorption, and ion-exchange methods are, therefore, less applicable for removal of trivalent arsenic. Accordingly, the optimum arsenic removal methods are usually composed of two individual stages: initial oxidation of arsenite to arsenate, and subsequent removal of arsenate [13].

In spite of the availability of biological and chemical methods (chemical oxidation, electrochemical acidification) to remove arsenic from the aqueous solutions, the efficiency of some natural and agro-based sorbents, such as banana peel [14], rice husk [15], red mud [16], clay minerals [17], and polymeric materials [18], has been investigated. Among the other sorbents tested, metallic nanoparticles, such as titanium, zirconium, and iron, have shown higher sorption efficiencies towards arsenic ions. In recent years, iron nanoparticles (FeNPs) have attracted much attention in the field of organic and inorganic contaminant removal. Successful results have been obtained in the utilization of FeNPs for arsenic decontamination [19].

The fate and contamination of nanoparticles synthesized through chemical processes and their unwanted toxic by-products are of concern. However, through the production of bio-based methods, safe production of nanomaterials has considerably increased. The conventional chemical methods in the fabrication of nanoparticles are being substituted with green chemistry methodology. Unlike physical and chemical methods, this novel method does not rely on high input energy, pressure, temperature, or utilization of chemical reagents.

Among the biological methods, the utilization of plant extracts seems to be more feasible, as the metal ions are mitigated quickly. Nanoparticles are synthesized within a few minutes or hours, considering the plant type and phytochemical concentration, while in microbial methods, considerably longer processing time is required [20]. Therefore, utilization of plants and their derivatives in the fabrication of nanomaterials is prioritized, in which green synthesis of metal nanoparticles is facilitated by using various plant organs, such as leaves, stems, roots, and seeds [21]. Plant-derived polyphenolic and flavonoid compounds, L-ascorbic acid (vitamin C), proteins, caffeine, theophylline, and antioxidants accelerate the synthesis of nanoparticles [22]. Herein, some of the mentioned substances act as reducing agents, while some others act as stabilizers and dispersants [23].

To the best of our knowledge, the plant-mediated synthesis of FeNPs has not been reported with the Prangos ferulacea and Teucrium polium extracts. This selection was based on the abundance, low-cost, and availability of the corresponding plant species in the area of study, the novelty of their incorporation, and their chemical specifications. The phenolic content of Prangos ferulacea and Teucrium polium is 65 [24] and 150 mg [25], respectively. These plants act as both reducing and capping agents.

The study aims to prepare iron nanoparticles (FeNPs) using an eco-friendly approach in one step by reduction of FeSO₄ with polyphenols from Prangos ferulacea and Teucrium polium. The resulting product is non-toxic, eco-accommodating, and cost-effective. Hence, it can be used as an economical alternative for FeNP production on a large-scale. Also, the synthesized nanoparticles were fully characterized, and consequently, the sorptive behavior of FeNPs was assessed through equilibrium and kinetic adsorption tests for arsenic removal from mining effluents, which has not yet been reported. Adsorption efficiency was investigated in different experimental conditions and was compared with that of conventional sorbents.

2. Material and Methods

2.1. Synthesis of Iron Nanoparticles

The collected plant samples were thoroughly rinsed with deionized water, air dried, and pulverized. Next, 15 g of plant powder was agitated with 250 mL deionized water for 1 h at 80 °C [26] using a magnetic hotplate stirrer (HSD330, MTops, Gyeonggi-do, Korea). The final extract was filtered using a vacuum pump (PLATINUM, Aurora, IL, USA). Iron sulfate (0.1 M) and sulfuric acid (0.1 M) were added into the solution to minimize and reduce Fe²⁺ oxidation. The extract solution was mixed with 0.1 M iron sulfate solution at a ratio of 2:1, while agitating the mixture for 30 min at room temperature. The nitrogen gas was injected into the mixture because the dissolved oxygen had been removed. Immediately after mixing, the solution changed color from light brown to dark black. This phenomenon is attributed to the formation of iron nanoparticles based on previous studies [20]. The nanoparticles were carefully filtered and dried at 50 °C for 24 h in a vacuum oven.

2.2. Characterization of Iron Nanoparticles

UV-visible spectrometry (Shimadzu UV2550, Kyoto, Japan) detected nanoparticles within 190–700 nm absorbance spectra. Mineral phases in the specimens were detected using X-ray powder diffraction (XRD) analysis (Philips pw1730, Amsterdam, The Netherlands). Field emission scanning electron microscopy (FESEM) (BEL SORP mini, Osaka, Japan), transmission electron microscopy (TEM) under an accelerating voltage of 100 kV (Zeiss EM 10C), and dynamic light scattering (DLS) (ZEN 3600, Malvern, Worcestershire, UK) were applied to study the size and morphology of the material. The surface functional group responsible for reduction, covering, and stabilization of the reduced nanoparticles was identified by interpreting the Fourier-transform infrared spectroscopy (FTIR) spectrum (BRAIC WQF-510, Beijing, China). The pH of zero points of charge (pH_ZPC) of the sorbent was estimated by preparing several samples at different pHs and detecting their zeta potentials using a Nano ZS ZEN3600 device (Malvern, Worcestershire, UK).

2.3. Adsorption Experiments

Batch adsorption experiments were designed and performed using 25-mL flasks containing 20 mL aqueous arsenic solution with a dose of adsorbent added, and then agitated by an orbital shaker at room temperature (25 ± 1 °C). According to

Table 1. Summary of the design of adsorption experiments.

	Parameters
	pH	Time (min)	Agitation Rate (rpm)	Adsorbent Dose (g/L)	Initial Arsenic Concentration (mg/L)
Series 1	4, 5, 6, 7, 8,9	15	150	2	75
Series 2	6	5, 15, 20, 40, 60, 80, 100, 120	150	2	75
Series 3	6	20	100, 150, 200, 250	2	75
Series 4	6	20	200	0.5, 1, 1.5, 2, 2.5, 3	75
Series 5	6	20	200	2	25, 50, 75, 100, 150

, the effects of 5 different parameters on arsenic removal were assessed: pH, time, agitation rate, adsorbent dosage, and initial concentration (test series 1–5, respectively). Factors were changed one at a time, while the others were kept constant. The pH adjustments were made by adding proper amounts of HCl (37%, c) and 0.1 M NaOH (BDH chemicals Co, Hull, UK) to the solution. Samples were withdrawn at predetermined time intervals, filtered, and then analyzed for their arsenic content by AAS (ContrAA700, Jena, Germany). Each experiment was repeated three times, and the mean value was reported. The standard solution was analyzed to achieve stability and accuracy.

The arsenic removal at a time (t) was then calculated according to Equation (1). The amount of arsenic adsorbed at time t, q_t (mg/g), was calculated by Equation (2).

$$Removal (\%)=\frac{(C_0-C_t) }{C_0}\ast 100$$

(1)

$$q_t=\raisebox{1ex}{$\left(C_0-C_t\right)V$}\!\left/ \!\raisebox{-1ex}{$W$}\right.$$

(2)

where C₀ is the initial arsenic concentration, and C_t is the concentration at time (t). V is the volume of the arsenic solution (L), and m is the mass of adsorbent (g).

Test series five was also used to study the equilibrium isotherm model.

Table 2. Isotherms and their linearized expressions.

Isotherms	Equations	Linear Expression	Plot	Parameters
Freundlich	$q_e=Ln{K}_F{\left(C_e\right)}^{1/n}$	$ln{q}_e=Ln{K}_F+\frac{1}{n}ln{C}_e$	$ln{q}_e vs. ln{C}_e$	KF = exp(intercept) n = (slope)−1
Langmuir	$q_e=\raisebox{1ex}{$\left(q_m{K}_L{C}_e\right)$}\!\left/ \!\raisebox{-1ex}{$\left(1+K_L{C}_e\right)$}\right.$	Type(I) $\frac{C_e}{q_e}=\left(\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$K_L{q}_m$}\right.\right)+\left(\raisebox{1ex}{$C_e$}\!\left/ \!\raisebox{-1ex}{$q_m$}\right.\right)$	$\left(\raisebox{1ex}{$C_e$}\!\left/ \!\raisebox{-1ex}{$q_e$}\right.\right) vs. C_e$	qm = (slope)−1 KL = slope/intercept
Temkin	$q_e=q_m\ln \left(K_T{C}_e\right)$	$q_e=q_{m}ln{K}_T+q_{m}ln{C}_e$	$q_e vs. ln{C}_e$	qm = exp (intercept/slope) KL = slope
Redlich-Peterson	$q_e=\raisebox{1ex}{$A_{RP}{C}_e$}\!\left/ \!\raisebox{-1ex}{$1+B_{RP}{C}_e^g$}\right.$	$\ln \left[\left(\raisebox{1ex}{$A_{RP}{C}_e$}\!\left/ \!\raisebox{-1ex}{$q_e$}\right.\right)-1\right]=gln{C}_e+ln{B}_{RP}$	$\ln \left[\left(\raisebox{1ex}{$A_{RP}{C}_e$}\!\left/ \!\raisebox{-1ex}{$q_e$}\right.\right)-1\right] vs. ln{C}_e$	G = slope BRP = exp (intercept) ARP *
Koble–Corrigan	$q_e=\raisebox{1ex}{$A_{KC}{C}_e^p$}\!\left/ \!\raisebox{-1ex}{$1+B_{KC}{C}_e^P$}\right.$	$\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$q_e$}\right.=\raisebox{1ex}{$\left(\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$A_{KC}{C}_e^P$}\right.\right)$}\!\left/ \!\raisebox{-1ex}{$\left(\raisebox{1ex}{$B_{KC}$}\!\left/ \!\raisebox{-1ex}{$A_{KC}$}\right.\right)$}\right.$	$\raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$q_e$}\right. vs. \raisebox{1ex}{$1$}\!\left/ \!\raisebox{-1ex}{$C_e^P$}\right.$	AKC= (slope)−1 BKC = intercept/slope P *

Note: * optimized using a trial and error method.

lists the details of the isotherm models and their parameters selected for this study.

To study the kinetics model, the sorption experiments were performed at an initial arsenic concentration of 75 mg/L, sorbent dosage of 0.04 g/L, pH 6, and agitation rate of 150 rpm. Samples were withdrawn at predetermined time intervals (5, 15, 20, 40, 60, 80, 100, and 120 min) during agitation. The kinetic data were analyzed using well-known kinetic models, as given in

Table 3. Kinetic models and their linearized expressions.

Kinetic Model	Equations	Linear Expression	Plot	Parameters
Zero order	qt = qe − k0t	qt = qe − k0t	qt vs. t	k0 = −slope qe = intercept
First order	qe = qtexp (k1t)	Ln(qe/qt) = k1t	ln(qt) vs. t	k1 = −slope qe = exp (intercept)
Pseudo-first order	qt = qe[1 − exp (k1pt)]	$Ln(q_e-q_t)=ln{q}_e-k_{1}t$	$\ln (q_e-q_t) vs. t$	k1 = −slope qe = exp (intercept)
Pseudo-second order	qt = k2pqe2t/(1 + qek2pt)	$\frac{t}{q_t}=\frac{1}{k_{2p}{q}_e^2}+\frac{t}{q_e}$	$\frac{t}{q_t} vs. t$	k2p = (slope)2/intercept qe = (slope)−1
Elovich	$q_t=\frac{1}{\beta }\ln \left(\alpha \beta t\right)$	$q_t=\frac{1}{\beta }\ln \left(\alpha \beta \right)+\frac{1}{\beta }lnt$	$q_t vs. lnt$	Β = (slope)−1 α = (slope) exp (intercept/slope)
Exponential function	$q_t=k{t}^v$	$Ln{q}_t=lnk+vlnt$	$Ln{q}_t vs. lnt$	V = slope K = exp(intercept)

.

The best-fit isotherm and kinetic models were determined according to the coefficient of determination (R² value).

Moreover, to study the thermodynamics of the adsorption, experiments were performed at an initial arsenic concentration of 100 ppm, pH 6, adsorbent dosage of 2 g/L, agitation rate of 200 rpm, and five different temperatures (20, 40, 50, 60, and 70 °C) for 20 min.

3. Results and Discussion

3.1. Characterization of the Synthesized Nanoparticles

As mentioned earlier, the color of the solution changed during nanoparticle formation. The color change in the mixture of extracts after their addition to the iron solution indicates the formation of the nanoparticle [27].

Figure 1 shows the absorption spectra in the case of mixed extracts, iron sulfate, and the synthesized nanoparticles. The spectrum showed a peak at 240 nm, which indicates the presence of Fe(II) ion. The resonance wavelength for iron nanoparticles lies in the range of 400–500 nm. A peak range of 390–450 nm was attributed to the zero-valent iron state, which is in full agreement with the results of other reports using various plants such as goua, papaya, and neem leaf extracts [28].

After 10 min, the reaction was finished, and the peak disappeared, representing the complete reduction of Fe(II) ions to FeNPs. According to the UV-Vis analysis, the presence of the iron nanoparticle was confirmed, and further characterization studies focused on the synthesized FeNPs by mixture of extracts.

The FTIR spectra showed several peaks at different regions, which indicated the biological nature of the iron complexes. In Figure 2a (before the adsorption process), the band at 3407 cm⁻¹ was attributed to hydroxyl (OH) groups. The band at 1625 cm⁻¹ represented the C=C aromatic ring stretching vibration and confirmed the existence of poly-phenolic compounds (aromatic cycles) in the extracts. The absorption band in the C–H region showed the polyols, and vibrations at 1100 cm⁻¹ were attributed to the C–O bond. The peak at 590 cm⁻¹ was assigned to the Fe–O bond [29].

Overall, the peaks confirm the presence of protein, phenolic, and amide groups in the extract mixture, which are responsible for the reduction of salt precursors to nanoparticles. Flavonoids are thermodynamically able to reduce the Fe²⁺ to Fe⁰ because of their lower redox potential (0.23 < E < 0.75 V). Therefore, the Fe–O bond appeared at 590 cm⁻¹, confirming the presence of Fe nanoparticles. Figure 2b shows the shift of peaks at 3407, 2925, 1625, 1100, and 590 cm⁻¹ to new positions at 3419, 2900, 1621, 1089, and 572 cm⁻¹ (after the adsorption process), respectively. These changes in the vibration peaks indicate the higher sorption intensity, and confirm the arsenic adsorption onto iron nanoparticles.

The mechanism of nanoparticle production via biomolecules is challenging. However, the researchers found that phenolic compounds, such as flavonoids, derived from a plant may control the reduction of ferrous cations, and also act as capping agents [30].

The XRD results (Figure 3) confirmed the crystalline form of produced FeNPs, with the peaks at (2θ) 44.38° and 34.38° corresponding to the zero valent iron (Fe⁰) and magnetite (Fe₃O₄), respectively. Also, the detected peak at the 2-Theta of 27.13° is related to iron hydroxide (FeOOH). The polyphenolic compound is observed as a sharp peak at 17.58° [31]. This observation is significant for nanoparticles, which mainly consist of maghemite and hydroxide partially oxidized on their surface.

The morphology and structure of the synthesized nanoparticles were studied using field emission scanning electron microscopy (FESEM). The results of FESEM imaging at two different magnifications of 1 μm and 200 nm are given in Figure 4a. The particles were mostly spherical, with an average size of 30–70 nm. The size range was in agreement with many recent studies (50–100 nm). Okoth et al. reported an average size of 1–2 µm for synthesized iron nanoparticles from banana peel extract [32], while aqueous sorghum bran produced smaller particles (200 nm) [33]. Wang et al. reported that the average size distribution obtained from a mixture of green tea and eucalyptus was about 100 nm [34].

TEM and SEM analysis demonstrated the surface morphology of biosynthesized FeNPs. The synthesized particles have a uniform spherical structure, with sizes ranging between 35 and 56 nm (Figure 4b). The DLS analysis showed an average particle diameter of 42 nm. It is interesting to note that some of the nanoparticles have irregular shapes because of agglomeration during their preparation for SEM analysis. As a result, most FeNPs were spherical. However, TEM images were compatible with SEM analysis and confirmed the distribution of iron nanoparticle sizes. Similar results using other plant extracts have been reported in the synthesis of Fe⁰/Fe₂O₃/Fe₃O₄ nanoparticles (

Table 4. Size and morphology of Fe⁰/Fe₂O₃/Fe₃O₄ nanoparticles synthesized by various plant extracts [35].

Plants	Part Used	Size and Morphology
Green tea	Leaf	40–60 nm, Amorphous
Tea	Tea powder	40–50 nm, Spherical
Pomegranate	Leaf	100–200 nm
Azadirachta indica	Leaf	50–100 nm, Spherical
Melaleuca nesophila and Rosemarinus officinalis	Leaf	50–80 nm, Spherical
Grapes	Leaf	15–100 nm, Quasi-spherical
Tridax procumbens	Leaf	80–100 nm, Crystalline, irregular sphere shapes
Eucalyptus tereticornis	Leaf	40–60 nm, Cubic

).

3.2. Optimization of Arsenic Sorption onto Iron Nanoparticles

The influence of the main operating parameters, such as pH, contact time, stirring speed, initial arsenic concentration, adsorbent dosage, and temperature, on arsenic removal from contaminated water was investigated.

The pH plays a role in wastewater treatment processes. Therefore, arsenic removal at different acidity levels was investigated in the range of 4 to 9. Figure 5 illustrates the effects of different initial pHs on arsenic (III) removal efficiency. As can be seen, arsenic removal gradually increased up to pH 6, and then it leveled off, with no further changes for the higher pH values tested. Accordingly, the maximum arsenic removal was found to occur at neutral pH, and therefore, this pH was selected for the rest of the experiments. Low adsorption efficiency at pH 4 was attributed to the competition between As (III) and the protons for the available sorption sites over the sorbent. At pH values higher than 9.2, the negative charge of H₂AsO₃⁻ leads to electrostatic repulsion [36]. The maximum arsenic removal capacity in the previous studies was also reported at about pH 6–8 [37]. Negative zeta potential at this range of pH confirmed the physical adsorption of arsenic on the adsorbent.

The effect of contact time on arsenic removal efficiency is illustrated in Figure 6. According to the results, 48.62% of the arsenic was removed during the initial adsorption stage (0–5 min), and as time passed, the arsenic content of the solution decreased at a higher rate so that after 20 min, 93.18% removal was achieved with no specific further change afterward. Thus, to ensure the achievement of equilibrium, the contact time was fixed at 20 min for the rest of the batch experiments. Generally, the rate of adsorption was high at first, and then it experienced a gradual increase until the equilibrium occurred. This fact may relate to a large number of the vacant surface site at an initial stage, and after a few minutes, the occupation of the rest is difficult due to repulsive forces between the solute molecules in solid and bulk phases [38].

The agitation rate influences the solute distribution bulk and the thickness of the Nernst boundary film. As can be seen in Figure 7, arsenic uptake increased as the agitation rate was raised to 200 rpm because of lower boundary-layer resistance. Similar results were reported by many researchers [38]. The negative effect of higher stirring speed on adsorption (250 rpm) can be attributed to excessive turbulence of the solution.

The effect of adsorbent dosage on removal efficiency was shown in Figure 8. The efficiency of As adsorption considerably increased with the increase of the amount of adsorbent as the maximum static uptake of arsenic (about 93.8%) was achieved for an adsorbent-to-solute ratio of 2 g/L. The reason may be attributed to the fact that the exchangeable sites or surface area are more available for the solute at the higher dosage of the adsorbent [39]. However, a further increase in adsorbent dosage (>2.0 g/L) did not affect the As adsorption due to the completion of adsorption of all arsenic. Thus, 2.0 g/L of iron nanoparticle in contaminated water was chosen for further experiments. The results of this section are in agreement with previous findings by Bazrafshan et al. [40] and Yao et al. [39].

Moreover, Figure 9 illustrates the effect of an initial arsenic concentration on removal efficiency. At an initial concentration of 25 mg/L, 99.92% arsenic removal is achieved by increasing the initial arsenic content, the efficiency is substantially decreased to 64.72%. This may be due to the less availability of the exchangeable sites or mitigation of nanoparticles effective surface area for arsenic capture at higher adsorbate concentration. The maximum capacity of iron nanoparticle for arsenic adsorption was achieved about 61.7 mg/g, which is considerable regarding adsorbent.

3.3. Arsenic Adsorption Isotherm

The process of adsorption is generally studied through adsorption isotherm curve. This graph shows the distribution of molecules between liquid and solid phase at the equilibrium state, which is considered as a fundamental factor in determining the sorption capacity [41]. To assess the sorption behavior of the iron nanoparticles, some well-known isotherms, namely Langmuir, Fruendlich, Temkin, Redlich–Peterson, and Koble–Corrigan, were considered to describe the phenomena. The results of As adsorption on adsorbents were analyzed by using the linear form of the models. The parameters associated with the adsorption behavior were calculated from the slope, and intercept of the plot. In each isotherm, adsorption capacity increased by raising the equilibrium concentration of arsenic. Isotherms models and their linearized expressions for the arsenic sorption are given in

Table 5. Arsenic adsorption modeling through different isotherms and their respective parameters.

Isotherms	Parameters	R2
Freundlich	KF = 25.81 mg g−1 (L mg−1)1/n n = 3.75	0.9901
Langmuir	qm = 64.51 mg g−1 KL = 0.49 L mg−1	0.9849
Temkin	qm = 8.0041 mg g−1 KL = 45.02 L mg−1	0.917
Redlich-Peterson	g = 1.361 BRP = 32.52 (L mg−1) ARP = 2 mg g−1 (L mg−1)	0.9987
Koble–Corrigan	AKC = 0.052 mg g−1 (L mg−1) BKC = 0.012 (L mg−1)P P = 0.2	0.9974

. The results of adsorption isotherms showed that there exist high correlation coefficients between the isotherms and experimental data. By comparing the coefficient values, the Redlich-Peterson isotherm had the highest correlation coefficient (R² = 0.9987), although only slightly (Figure 10).

3.4. Adsorption Kinetics

By understanding the kinetic sorption model, the mechanisms of the adsorption process are determined, such as mass transfer or chemical reaction. Several kinetic models, such as zero-order, first-order, second-order, pseudo second-order, Elovich, and exponential functions, were used to test the experimental data. The kinetic models, their linear expression form, and the model parameters are summarized in Table 3. The correlation coefficient (R²) was used for the evaluation of the best kinetic model (

Table 6. Kinetic parameters for arsenic removal using synthesized iron nanoparticles.

Isotherms	Parameters	R2
Zero-order	K0 = −0.0731 mg/g h qe = 41.784 mg/g	0.29
First-order	K1 = −0.00181/h qe = 41.09 mg/g	0.28
Pseudo-first-order	K1 = 0.003 1/h qe = 18.81 mg/g	0.32
Pseudo-second-order	K2 = 0.017 g/mg h qe = 48.87 mg/g	0.99
Elovich	β = 0.24 g/mg α = 478.1 mg/g h	0.63
Exponential function	V = 0.1031 K = 31.31 mg/g h	0.61

). Among all models, the pseudo-second order kinetic model had a good correlation coefficient, and it fit best with experimental data (Figure 11).

3.5. Thermodynamic Studies

Thermodynamic parameters, such as Gibbs energy (ΔG), enthalpy (ΔH), and entropy (ΔS), may assist in the determination of the possibility, spontaneity, or exothermic or endothermic states of a reaction, which are the critical aspects of an adsorption process. To calculate the thermodynamic parameters, some experiments were carried out at various temperatures ranging 20–70 °C. The amount of arsenic uptake from the solution was decreased from 90% to 73%, respectively (Figure 12). Enthalpy and entropy values were calculated according to the van’t Hoff equation:

$$ln{k}_d=-\left(\frac{\Delta H}{RT}\right)+\left(\frac{\Delta S}{R}\right)$$

(3)

where ΔH and ΔS are the changes in enthalpy and entropy of the adsorption process, respectively. T is the temperature (°K), and k_d is the distribution coefficient, which can be calculated by:

$$K_d=\left(C_0-Ce\right)V/CeM$$

(4)

where C₀ and C_e are the initial and equilibrium arsenic concentrations, respectively. V is the solution volume and M is the mass of the adsorbent. By using the van’t Hoff equation and plotting lnK_d against 1/T, the values of ΔH and ΔS may be estimated through the slope and intercept of the curve (Figure 11). Moreover, Equation (5) was used to calculate the Gibbs free energy (ΔG).

ΔG = ΔH − TΔS

(5)

The calculated values of the thermodynamic parameters for the arsenic sorption process are listed in

Table 7. Thermodynamic parameters of arsenic adsorption onto the synthesized iron nanoparticles.

Adsorbent	Initial Concentration (mg/L)	ΔH (kJ/mol)	ΔS (J/mol·k)	T (°K)	ΔG (kJ/mol)
Iron nanoparticles	100	−20.7	−57.3	293	−3898.3
				313	−2751.9
				323	−2178.6
				333	−1605.4
				343	−1032.2

. The negative ΔH value indicates the exothermic nature of arsenic sorption onto iron nanoparticles and that the lower temperature favors the reaction (Figure 13). The negative ΔS value also shows the adsorption of arsenic ions from the solution onto the iron nanoparticle. Finally, the negative ΔG confirms the spontaneous nature of the adsorption and exergonic reaction at all applied temperatures.

3.6. The Activation Energy of Adsorption

For the evaluation of activation energy, some experiments were performed at the initial arsenic concentration of 100 mg/L and at three different temperatures, 20, 40, and 60 °C. Samples were taken at predetermined time intervals (5, 10, and 20 min) and analyzed for arsenic content by the authors. Figure 14 shows the kinetics of arsenic adsorption onto iron nanoparticles during 20 min of contact time. As can be seen, higher temperatures declined the sorption rate. The activation energy is calculated by the Arrhenieus equation:

$$ln{k}_l=lnA-\left(\raisebox{1ex}{$E_a$}\!\left/ \!\raisebox{-1ex}{$RT$}\right.\right)$$

(6)

where k_l is the pseudo-second kinetic constant, which can be calculated from Figure 12, A is a temperature-independent coefficient (min. g⁻¹. mm⁻¹), E_a is the activation energy (kJ·mol⁻¹), R is the universal gas constant, and T is the temperature (°K).

E_a can be geometrically estimated by plotting the variations of lnk_l against 1000/T (Figure 15). According to Figure 16, the E_a for the arsenic sorption process is obtained as −17 kJ/mol. This negligible amount of E_a indicates the physical nature of the adsorption process. Therefore, the reaction is not temperature-sensitive, which confirmed that the system was controlled by the diffusion process [42].

3.7. Arsenic Removal from Real Mining Effluents

Real arsenic-containing effluent samples were collected from Meiduk (Kerman province, Iran) and Ghale Zari (South Khorasan province, Iran) copper processing plants and analyzed for their arsenic contents. The arsenic concentration was 69 µg/L and 43 µg/L for Meiduk and Ghale Zari samples, respectively.

For each sample, an adsorption experiment was performed using 0.04 g adsorbent dosage at optimum conditions, according to the previous tests (pH 6, agitation rate of 200 rpm, 20 min contact time) at ambient temperature. Figure 16 demonstrates the arsenic removal efficiencies in two cases. As can be seen, more than 70% of arsenic can be removed by iron nanoparticles in both cases. Therefore, the adsorbent was successfully confirmed as an efficient option to be considered for mine wastewater treatment plants.

3.8. Comparison of Arsenic Sorption Capacity for Different Adsorbents

Table 8. The comparison of arsenic sorption capacity for different adsorbents.

No.	Adsorbent	Uptake Capacity (mg/g)	Ref.
1	Granular ferric hydroxide (GFH)	1.1	[43]
2	a-FeOOH nanoparticles	76	[43]
3	Ultra-fine δ-FeOOH	37.3	[43]
4	Hematite	0.2	[44]
5	Magnetite–maghemite nanoparticles	3.71	[43]
6	Red mud	0.94	[44]
7	Fe3O4 nanoparticles	46.06	[43]
8	Alumina	25	[43]
9	Activated carbon	3.15	[43]
10	γ-Fe2O3 nanoparticles	2.9	[43]
11	Fe3O4–γ-Fe2O3 nanoparticles	3.71	[43]
12	Nano zero valent iron (NZVI)	1.8–2.0	[43]
13	NZVI/montmorillonite	59.9	[43]
14	Fe–Al layered double hydroxide	40.6	[43]
15	Nano-iron/oyster shell	0.9	[43]
16	NZVI-synthetized by mint	86.53	[44]
17	Iron oxide coated fungal biomass	5.4	[43]
18	Modified iron oxide	0.73	[44]
19	PVA–alginate-supported iron oxide nanoparticles (IONPs)	40.3	[35]
20	chitosan-alginate-supported IONPs	34.4	[35]
21	Iron nano particle/Pf-Tp	61.7	Present study

draws a comparison with the synthesized iron nanoparticles and other reported adsorbents in terms of arsenic adsorption capacity (mg/g). As can be seen, the proposed sorbent is 20 times higher than activated carbon and alumina, which are the conventional materials for arsenic removal. Iron nanoparticles synthesized by Prangos ferulacea and Teucrium polium leaf extracts still demonstrate a remarkable uptake capacity and may be a promising arsenic removal option to be used in various industrial applications. Note that the other iron-based nano adsorbents, which are illustrated in Table 8, are all chemically produced.

4. Conclusions

Arsenic is one of the toxic and dangerous pollutants in the industrial and mineral processing of wastewater. The purpose of this study is to synthesize biological nanoparticles using the extracts of Prangos ferulacea and Teucrium polium herbs, and consequently, the application of produced FeNPs for arsenic (III) removal from aqueous solutions. The results of this study can be drawn as follows:

• Iron nanoparticles were synthesized via green chemistry methods. XRD, UV-Vis, and FTIR analyses confirmed the successful formation of iron nanoparticles using Prangos ferulacea and Teucrium polium leaf extracts, and the morphology was determined as spherical particles of average 40 nm diameter by SEM and TEM analysis.
• The influence of parameters such as contact time, pH, adsorbent dosage, initial arsenic concentration, temperature, and agitation rate was investigated. Results showed that 93.9% of arsenic was eliminated at a stirring speed of 200 rpm, pH 6, 2 g/L nanoparticle dosage, and initial concentration of 0.075 g/L within 20 min.
• The synthesized nanoparticles were able to absorb 61.7 mg/g arsenic from the solution at 298 °K, which is a considerable capacity in comparison with most of the available sorbents. This was also confirmed through post-sorption FTIR analysis.
• The adsorption behavior of the synthesized nanoparticles is represented by different isotherm models.
• Thermodynamic adsorption investigations showed the negative values for enthalpy and entropy of the adsorption process, which revealed its exothermic and spontaneous nature. Moreover, the activation energy for arsenic removal was obtained as −17 kJ/mol, which indicated the physical sorption mechanism and that the process is not temperature-sensitive.
• Arsenic adsorption tests performed over the real industrial effluent samples collected from two different mines confirmed that the proposed adsorbent is capable of removing more than 70% of arsenic contamination in real cases.
• The comparison of arsenic sorption capacity shows that iron nanoparticles synthesized by Prangos ferulacea and Teucrium polium leaf extracts still demonstrate a remarkable uptake capacity and may be a promising arsenic removal option to be used in various industrial applications.