Preparation of a sulfonated coal@ZVI@chitosan-acrylic acid composite and study of its removal of groundwater Cr(VI)

In this research, a new composite adsorbent (SC@ZVI@CS-AA) was designed and synthesized, and its application for the removal of Cr(VI) in groundwater was investigated. The interaction between SC@ZVI@CS-AA and Cr(VI) conformed to a pseudo-second-order model, and the adsorption process was dominated by chemisorption. The effects of material ratios, pH, temperature, SC@ZVI@CS-AA dosage, and coexisting ions on the removal of Cr(VI) were investigated. The removal efficiency of Cr(VI) by SC@ZVI@CS-AA reached 95%, and the reaction was significantly inhibited when SO42− was present. Thermodynamically, the adsorption of Cr(VI) proceeded spontaneously above 35 °C (ΔGθ < 0). According to scanning electron microscopy, X-ray photoelectron spectroscopy, Fourier transform infrared spectrometry, and synchronous thermal analysis, the removal mechanism of Cr(VI) by SC@ZVI@CS-AA was attributed to electrostatic attraction and reduction. In addition, SC@ZVI@CS-AA had good cyclic adsorption performance. Overall, the SC@ZVI@CS-AA composite showed great potential in the remediation of Cr(VI)-contaminated groundwater.


Introduction
Hexavalent chromium (Cr(VI)) is one of the most toxic heavy metals, and has been detected in environmental media affected by various industrial processes, such as electroplating, wood preservation, leather tanning, and paint production (Cao et al. 2022). If areas in which chromium pollution is generated are not protected, solid waste comprising chromium slag and other chromium-containing hazardous materials will leach chromium-containing compounds into soil and water under the influence of the external environment. Cr(VI) in groundwater has high solubility and mobility, which has led to extensive pollution (Jobby et al. 2018). According to the US Environmental Protection Agency, the total chromium concentration exceeding 0.1 mg/L can seriously damage the environment and endanger human health (Li et al. 2020).
Permeable reactive barrier (PRB) technology emerged in the 1990s as an in situ method for treating groundwater contamination (Hedin et al. 1994), and it has received widespread attention because of its wide remediation and control range, low cost, and low energy consumption (Obiri-Nyarko et al. 2014;Liu et al. 2015). Active materials are very important for Cr(VI) removal from groundwater, since PRB technology should support quick reactions with Cr(VI), avoid producing secondary pollution, and exhibit good stability, permeability, and economy (Bronstein 2005).
In recent years, scholars have performed considerable research on PRB fillers, among which zero-valent iron (ZVI) is commonly used as a reaction medium (Wilkin et al. 2010;Sun et al. 2019); it was also reported that ZVI can be applied in practical engineering applications (Henderson and Demond 2007). ZVI has the advantages of high reaction rates, a large specific surface area, and ample reducing capability. However, ZVI is prone to agglomeration, loss, and instability in practical applications, which leads to blockage of PRB systems (Hu et al. 2010;Li et al. 2020). Numerous studies have shown that modification of ZVI by using Responsible Editor: George Z. Kyzas cellulose, chitosan (CS), polyacrylic acid, and zeolite as stabilizers retains the advantages of ZVI and overcomes the disadvantages of ZVI, such as instability and loss (Tasharrofi et al. 2020;Guan et al. 2019).
CS contains a large number of reactive amino (-NH 2 ) and hydroxy (-OH) groups, which can chelate heavy metals and support ion exchange (Wu et al. 2001). CS can be used as a coating for metal particles to prevent aggregation (Zhiya et al. 2007;Gupta and Gupta 2005); moreover, it has the advantages of low price, biodegradability, and nontoxicity, making it a promising environmental material (Zhang et al. 2020). Bing et al. (2009) used CS as a stabilizer to modify ZVI and found that ZVI stabilized by CS removed 148.08 mg of chromium per gram, a level that was approximately three times higher than that of unmodified ZVI. More importantly, CS-stabilized ZVI exhibited antioxidant properties and good dispersion. However, the application of natural CS is limited in a practical sense due to its sensitivity to pH and poor mechanical stability in acidic media (Lei et al. 2016). To overcome these limitations, chemical modification of CS is needed (Zhao et al. 2013;Xiao et al. 2021). Among various modification methods, grafting copolymerization of CS with acrylic acid (AA) further improves the functionality of CS (Tao et al. 2003;Lu 2019) and compared the capacities of modified CS and unmodified CS for the adsorption of Pb + , Cd 2+ , and Cu 2+ in water. CS grafted onto AA showed better performance than CS alone for the adsorption of all three ions, and the surface of grafted CS exhibited a network structure in which the number of -NH 2 groups was approximately doubled. Therefore, CS grafted onto AA can be used to stabilize ZVI and serve as an active PRB material for the in situ remediation of Cr(VI) in groundwater.
In PRB remediation of groundwater polluted by Cr(VI), the reaction medium must have good adsorption and reducing properties in addition to permeability. Therefore, porous materials have been considered solid carriers for loading CS-stabilized ZVI and forming a composite medium, which would improve the permeability of the reaction medium (Belhaj et al. 2015). The porous carrier materials commonly used in the remediation of heavy metals in water bodies include activated carbon, attapulgite, and bentonite. The activity of activated carbon gradually decreases during the remediation process, and the material is expensive (Obiri-Nyarko et al. 2014). Attapulgite and bentonite are clay minerals with low permeability, so they are not favored for PRB technology (Pourcq et al. 2015). Sulfonated coal (SC) is the product obtained from the interaction of coal with fuming sulfuric acid or concentrated sulfuric acid (Ni 1997), it is rich in functional groups such as -OH, NH 2 , and -SO 3 H (Jamil et al. 2017), contains porous structures, and is an inexpensive and valuable resource. Acid modification of SC removes soluble metal oxides from inner pore surfaces and further opens the pores (Zheng 2008). Therefore, SC can be used as a solid carrier for ZVI stabilized by modified CS to prepare a new PRB active material for the removal of chromium.
In this study, an SC@ZVI@CS-AA composite was prepared by using biocompatible CS containing ample -NH 2 and -OH groups to improve the chemical stability and adsorption capacity of ZVI, and sulfonated coal particles rich in functional groups were used as a support. The surface morphology, structural composition, surface chemical properties, and thermal stability were analyzed by scanning electron microscopy (SEM), X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), and thermogravimetry-differential scanning calorimetry (TG-DSC). The effects of different factors, such as material proportions, pH, reaction time (t), initial concentration of Cr(VI) (C 0 ), adsorbent dosage, reaction temperature (T), and interfering ions, on the adsorption performance of Cr(VI) were assessed. Finally, the mechanism for the adsorptive removal of Cr(VI) was analyzed based on the adsorption kinetics and a thermodynamic model. This study provides a reference for the preparation of reaction media used in the remediation of Cr(VI)-contaminated groundwater with PRB technology.

Preparation of the SC@ZVI@CS-AA composite
Acid modification of SC SC with a particle size of 0.5 mm to 1.0 mm was cleaned by deionized water with a solid-liquid ratio of 1 to 30 until the supernatant was clear and no suspended matter was present. The clean sulfonated coal was filtered, spread evenly in an ordinary oven, and dried at 60 °C for 12 h (Zhan et al. 2017).
Diluted nitric acid (HNO 3 , 2 mol/L) was slowly poured onto the dried sulfonated coal, and the solid-liquid ratio was kept at 1 to 10. After stirring at 1500 rpm for 24 h, the sulfonated coal was filtered and washed to neutrality with deionized water until the supernatant was clear and no suspended matter was present. Then, the clean sulfonated coal was filtered again and dried at 60 °C for 12 h. SC modified by HNO 3 was obtained.

Preparation of the SC@ZVI@CS-AA composite
CS (1 g) was dissolved in 50 mL of glacial acetic acid (2%, v/v) to form a solution of CS with a mass concentration of 2% (Chen et al. 2006); the solution was fully dissolved by stirring for 2 h, and a transparent, yellowish colloid was formed. Next, 1 mL of AA was added to the colloid for grafting modification and stirred at a speed of 1500 revolutions per minute (rpm) for 2 h to obtain the modified CS colloid (CS-AA). SC (0.2 g) was added to CS-AA and stirred at 1500 rpm for 2 h. Then, 0.05 mL of ECH (CS:ECH = 2:1(m:v)) was added to CS-AA as a cross-linking agent (Mei et al. 2019), and stirring was continued for 3 h. Then, 0.07 g of ZVI was mixed and dispersed for 1 h with nitrogen. At the same time, 0.01 mL of sodium tripolyphosphate (CS:TPP = 10:1(m:v)) (Chen et al. 2003) was added to the solution, and then 2 mol/L NaOH was added until no new precipitate was formed. After the system was closed and aged for 24 h, it was washed with deionized water until the wash water was neutral. The clean material was dried at 60 °C for 12 h in a vacuum oven and ground to obtain the SC@ZVI@CS-AA composite. The basic preparation flow chart is shown in Fig. 1.

Characterization
The distributions of surface functional groups on SC and SC@ZVI@CS-AA were analyzed by Fourier transform infrared spectroscopy (Thermo Nicolet 6700) at wavelengths between 400 and 4000 cm −1 . The crystal structures of SC and SC@ZVI@CS-AA were analyzed by X-ray diffraction (Bruker D8 Advance) with a scanning 2θ range of 5 to 90° and a scan step of 0.02456°. The thermal stability and composition of SC and SC@ZVI@CS-AA were analyzed by a synchronous thermal analyzer using thermogravimetric analysis and differential scanning calorimetry over a temperature range of 30 to 1000 °C with a heating rate of 10 °C/min. The surface morphologies of SC and SC@ZVI@CS-AA were observed by scanning electron microscopy (ZEISS GEMINI 300) at an acceleration voltage of 3 kV, working distance of 5 mm, and multiple scanning with 500 to 20,000 scans.

Effects of material ratios, pH, dosage, and coexisting ions
To study the effects of the material composition on the adsorption of Cr(VI), 0.3, 0.5, 1, and 2 g of CS were dissolved into four different amounts (15, 25, 50, and 100 mL) of glacial acetic acid (2%, v/v) to give a CS solution with a 2% mass fraction (Chen et al. 2006). Based on the CS-AA sol preparation method described above, the corresponding volume of AA was added to obtain four CS-AA samples with different CS contents. Then, 0.2 g of SC was added to each of the four CS-AA samples to establish four different mass ratios of SC:CS (6:1, 4:1, 2:1, and 1:1). SC@ ZVI@CS-AA was prepared with the mass ratio of SC to ZVI unchanged. SC@ZVI@CS-AA was used for adsorption experiments with the following conditions: pH = 3, T = 25 °C, and C 0 = 200 mg/L. The optimal dosage of chitosan was determined according to the experimental results. The CS-AA sol was prepared with the optimal amount of CS. Five CS-AA sols with the same volume were added to 2 g of SC. Then, five SC@ZV@CS-AA samples with different SC:ZVI mass ratios (8:1, 6:1, 3:1, 1.5:1, and 1:1) were prepared by adding ZVI for adsorption experiments. Three parallel samples were used for each experiment. The effect of pH on the adsorption of Cr(VI) (10 mL, 200 mg/L) was studied with pH values ranging from 3 to 8.5, an adsorbent dose of 0.008 g, a 5 h reaction time, and a reaction temperature of 298 K. To explore the effect of adsorbent dosage on adsorption performance (200 mg/L), the pH was set to 3. Then, varying amounts of SC@ZVI@CS-AA (0.004, 0.008, 0.012, 0.016, 0.02, 0.025, and 0.03 g/L) were added, and the adsorption experiments were carried out at 25 °C for 5 h. To study the effects of coexisting ions in water on the adsorption of Cr(VI) (10 mL, 200 mg/L), 0.01 mol of seven electrolytes (NaCl, CaCl 2 , MgCl 2 , Na 2 SO 4 , NaNO 3 , NaH 2 PO 4 , and Cd(NO 3 ) 2 ·4H 2 O) were added to Cr(VI) solutions containing 0.008 g/L SC@ ZVI@CS-AA, and reactions were carried out at 25 °C for 5 h. Three parallel samples were used for each experiment.
After each adsorption experiment, a syringe (20 mL) was used to sample the mixtures in the reaction bottles. After filtration through a membrane (0.45 μm), the concentration of residual Cr(VI) in the solution was determined by diphenylcarbazide spectrophotometry (GB/T 7467-1987(GB/T 7467- , 2003. Total chromium (total Cr) was determined using a con-trAA700 continuous light source atomic spectrometer, based on linear fitting of the data to calculate its concentration.
The adsorption capacity (q e , mg/g) and removal efficiency (%) of SC@ZVI@CS-AA for chromium were calculated by Eqs. (1) and (2): where C 0 (mg/L) is the initial concentration of chromium, C e (mg/L) is the equilibrium concentration of chromium, m (g) is the dosage of adsorbent, and V (L) is the total volume of solution.

Effect of contact time and kinetics studies
To study the influence of reaction time on chromium adsorption, the Cr(VI) stock solution was diluted to three different concentrations (200, 300, and 400 mg/L), and the pH was adjusted to 3. Ten milliliters of solution was added to a polyethylene tube (50 mL) with 0.008 g of SC@ZVI@CS-AA. The reaction was carried out on a constant temperature oscillator with a rotating speed of 120 rpm and a temperature of 25 °C. The concentrations of Cr(VI) in the samples were measured at different times (10, 20, 30, 45, 60, 90, 120, 180, 240, 300, 360, and 480 min). Three parallel experiments were used for each determination. (1) To describe the mechanism for the adsorption of chromium and the associated rate of the reaction, the experimental data were fitted to pseudo-first-order, pseudo-secondorder, and intraparticle diffusion models.
The pseudo-first-order kinetic model is described by Eq. (3) (Lu et al. 2014): The pseudo-second-order kinetic model is given by Eq. (4) (Divya and Nalini 2008): The intraparticle diffusion model is described by Eq. (5) (Araújo et al. 2017): where q e (mg/g) is the adsorption capacity of SC@ZV@ CS-AA at equilibrium; q t (mg/g) is the adsorption capacity of SC@ZV@CS-AA at time t; K 1 (min −1 ), K 2 (g/mg·min), and K d (g/mg·min) are the rate constants for the pseudofirst-order, pseudo-second-order, and intraparticle diffusion models, respectively; and C (mg/g) is the intercept.

Effect of initial chromium concentration and isothermal models
To explore the effect of the initial chromium concentration on the adsorption process, temperatures of 25 °C, 35 °C, or 45 °C were used. The Cr(VI) stock solution was diluted to eight different concentrations (10, 20, 50, 100, 200, 300, 400, and 500 mg/L), and the pH was adjusted to 3. Ten milliliters of solution at each concentration was added to a polyethylene tube (50 mL) with 0.008 g of SC@ZVI@CS-AA. After reaction at 120 rpm for 5 h at the chosen temperature, samples were taken to determine the concentration of Cr(VI) in the solution. Three parallel samples were used for each experiment.
As the adsorption process reaches equilibrium at a certain temperature, the relationship between adsorbate and adsorbent can be assessed with various adsorption isotherm models (Hameed et al. 2007). To analyze the adsorption process, the Langmuir, Freundlich, and Temkin isothermal adsorption models were used to simulate the adsorption results.
The Langmuir isotherm model assumes that monolayer adsorption occurs on a homogeneous adsorbent surface (Wang et al. 2016). The Langmuir isotherm equation is described by Eq. (6): The Freundlich isotherm model is an empirical model for heterogeneous systems exhibiting multilayer adsorption. The Freundlich isotherm model is expressed by Eqs. (7) and (8)  The Temkin isotherm model assumes that the heat of adsorption for supported substrates undergoes a linear decrease . The Temkin isotherm model is expressed by Eqs. (9) and (10): where q e (mg/g) is the amount of Cr(VI) adsorbed at equilibrium per unit weight of SC@ZVI@CS; q m (mg/g) is the maximum adsorption capacity; C e (mg/L) is the equilibrium concentration of Cr(VI); C 0 (mg/L) is the initial concentration of Cr(VI); K L (L/mg) is the Langmuir constant; V (L) is the total volume of solution; m (g) is the dosage of adsorbent; K F ([mg/g][L/mg] −1/n ) and n are empirical constants for the Freundlich adsorption capacity and strength, respectively; R = 8.314 J/(mol K) is the ideal gas constant; T (K) is the absolute temperature; K T (L/g) is the equilibrium binding constant; B T (J/mol) is a Temkin constant; and b T is the Temkin constant related to the heat of adsorption.
Thermodynamic parameters include the enthalpy change ( ΔH ), entropy change ( ΔS ), and Gibbs free energy change ( ΔG ), which are evaluated by Eqs. (11), (12), and (13) (Aksu 2002); these parameters are helpful in further understanding the energy changes occurring in the adsorption process and predicting the adsorption mechanism more effectively.
where Kd (L/g) is the thermodynamic equilibrium constant; q e (mg/g) is the amount of Cr(VI) adsorbed at equilibrium per unit weight of SC@ZVI@CS; C e (mg/L) is the equilibrium concentration of Cr(VI); R = 8.314 J/(mol K) is the ideal gas constant; and T (K) is the absolute temperature. Plots are made with 1/T as the abscissa and ln q e ∕C e as the (7) lg q e = lg K F + 1 n lg C e ordinate, and ΔS and ΔH are calculated from the intercept and slope. The Gibbs free energy is given by Eq. (13):

Cyclic adsorption experiments
To explore cyclic adsorption with the adsorbent, adsorption experiments were carried out with an initial Cr(VI) concentration of 200 mg/L, adsorbent dosage of 0.008 g/L, temperature of 25 °C, pH of 3, and reaction time of 5 h. NaOH (50 mL, 0.1 mol/L) was added to SC@ZVI@CS-AA; the mixture was filtered after reaching adsorption saturation, and then the mixed liquor was desorbed in a constant temperature shaker at 25 °C for 6 h. SC@ZVI@CS-AA was filtered again and rinsed until neutral with deionized water, and the adsorption experiment was continued after vacuum drying. This sequence was repeated three times. The concentrations of Cr(VI) and total Cr in the desorption solution and adsorbed solution were determined. The desorption efficiency (%) is expressed by Eq. (14): where C 1 (mg/L) is the concentration of Cr(VI) in the desorption solution; V 1 (mL) is the volume of desorption solution; C 0 (mg/L) is the initial concentration of Cr(VI) (mg/L); and V 0 (mL) is the volume of the Cr(VI) solution.

FTIR results
The FTIR results for SC and SC@ZVI@CS-AA are shown in Fig. 2a. Changes in the functional groups on the surface of the materials after loading were limited, which was related to the rich content of chemical groups (-OH, -NH 2 , and -SO 3 H) in the sulfonated coal itself. For both SC and SC@ZVI@CS-AA, the peaks at 2919 cm −1 and 2870 cm −1 showed weak C-H stretching vibrations, and the peak at 1085 cm −1 was assigned to C-C bending vibrations. Compared with SC, the absorption peak at 3447 cm −1 for SC@ ZVI@CS-AA was broadened slightly, which may be due to the introduction of CS and to increases in the number of O-H and N-H groups on the surface of the material, which resulted in an increase in the number of hydrogen bonds (Lei et al. 2020). The peak at 1625 cm −1 was due to C = C stretching vibrations, and the C = C band of SC@ZVI@CS-AA became stronger because of the AA used in the modification process (Pbv et al. 2011). Compared with the SC data, the FTIR results for SC@ZVI@CS-AA suggested that the peak at 1321 cm −1 was due to C-N stretching vibrations and overlapping peaks representing C-H bending vibrations (Wu 2002). The above results indicated that the functional groups of AA and CS were introduced onto the surface of SC@ ZVI@CS-AA and that they provided coordination sites for the removal of Cr(VI) and enhanced the binding of Cr(VI) with the modified materials. In addition, the SC solid carrier did not change the chemical groups on the loaded materials. Figure 2b shows the XRD patterns for SC and SC@ZVI@ CS-AA. The sharp characteristic peak ( 2 = 26.9 • ) of silicon dioxide (SiO 2 ) appeared in the diffraction pattern of SC, since the SC used in the experiment was mixed with impurities such as sediment during production. In contrast, the characteristic peak of SiO 2 in the diffraction pattern of SC@ZVI@CS-AA was weak and broad, which may be due to grain refinement or the introduction of CS in the modification reaction. In addition, this result indirectly indicated the synthesis of new substances on the surface of SC. The peaks for SC@ZVI@CS-AA showed evidence of the (110) crystalline face of body-centered cubic α-Fe 2 = 44.9

XRD results
• , and a diffusion diffraction peak corresponding to the (211) face 2 = 82.6 • appeared, indicating that ZVI stabilized by CS-AA was successfully loaded onto the SC (Jin et al. 2018).  Figure 2e shows that the weight loss of SC from 30 to 600 °C was approximately 5%, indicating the superior thermal stability of CS. Figure 2d shows that the weight loss of SC@ZVI@CS-AA was mainly divided into three stages. The first stage occurred between 30 and 120 °C and was due to incomplete drying of free water or crystalline Fig. 2 The characterization results of SC and SC@ ZVI@CS-AA: a SEM results for SC (20 × magnification); b SEM results for SC@ZVI@ CS-AA (20 × magnification); c XRD results for SC and SC@ ZVI@CS-AA; d FTIR results for SC and SC@ZVI@CS-AA; e TG-DSC results for SC; and f TG-DSC results for SC@ZVI@ CS-AA water in the modified material (Hadi et al. 2013), and the weight loss for SC@ZVI@CS-AA was approximately 2%. In the second stage, a weight loss of 9% was observed between 280 and 320 °C. In this stage, the rate of weight loss was fast, and the DSC curve showed a small peak. Because the decomposition temperature of CS is 280 °C, polymers of CS were thermally decomposed and produced volatile gases and organic substances. The third stage occurred between 320 and 600 °C. Overall, the weight loss of SC@ZVI@CS-AA from 280 to 320 °C was 14%. Considering the weight lost by SC, it was speculated that the carbon-containing polymer loaded on SC@ZVI@CS-AA accounted for approximately 10.7% of the weight of the material itself. The results indicated that the SC carrier could be decomposed at a high temperature, and SC@ZVI@CS-AA maintained good thermal stability at temperatures below 280 °C.

SEM results
The SEM results for SC and SC@ZVI@CS-AA are shown in Fig. 2. From Fig. 2e, it can be seen that the surface of SC was relatively rough; it showed a lamellar structure with small holes and longitudinal cracks, which provided a large number of support sites for stable ZVI loading on CS-AA. In addition, this morphology hindered the agglomeration of ZVI. As shown in Fig. 2f, the surface of SC@ZVI@CS-AA constituted a continuous mesh structure with abundant small and large pores, which indicated that the modified ZVI (particle size of approximately 0.1-0.2 μm) was uniformly distributed on the surface of SC. This increased the dispersion and stability of ZVI and facilitated full contact between ZVI and Cr(VI), which provided abundant sites for reduction, adsorption, and the efficient removal of Cr(VI).

Effect of various component ratios of SC@ZVI@ CS-AA
The effect of different component ratios on the adsorption of Cr(VI) is shown in Fig. 3. Figure 3a shows that when the ratio of SC to CS was 2 to 1, the adsorption capacity for Cr(VI) was the largest (54.5 mg/g). When the ratio was higher, the amount of CS loaded on SC was insufficient, and there were not enough functional groups (-NH 2 and -OH) to react with Cr(VI), which limited the adsorption of Cr(VI). Moreover, CS could not stabilize ZVI due to the low CS content, which allowed ZVI agglomeration (He et al. 2007;Bhatia and Ravi 2003). When the SC:CS ratio was lower, the amount of CS present was so high that polymeric CS-AA formed on the surface of SC, reducing the porosity of the material and limiting contact between ZVI and Cr(VI) or Cr(III).
As shown in Fig. 3b, when the SC:ZVI mass ratio was 3 to 1, the adsorption capacity for Cr(VI) reached the highest value (142.42 mg/g). When the amount of ZVI loaded on SC was too high, the adsorption sites of SC were easily saturated. The more ultrafine iron powder there was, the more likely it was to form clusters. Appropriately reducing the load of ZVI was more conducive to uniform dispersion, thereby allowing ZVI to serve as a reductant.

Effect of pH
The effect of the initial pH on chromium removal is shown in Fig. 3c. pH had an important influence on the adsorption results. The removal efficiency decreased from 59.30 to 2.60% when the pH was increased from 3 to 8.5, and the Cr(VI) adsorption capacity of SC@ZVI@CS-AA decreased from 134.37 to 6.37 mg/g. Furthermore, the total Cr removal capacity of SC@ZVI@CS-AA decreased from 119.30 to 9.86 mg/g as the pH increased from 3 to 8.5. The removal capacity for Cr(VI) was higher than that for total Cr. This may be because SC@ZVI@CS-AA reduced some Cr(VI) to Cr(III), which led to a Cr(VI) concentration lower than the total Cr concentration in the solution, so the adsorption capacity of Cr(VI) was higher.
The removal of Cr(VI) by SC@ZVI@CS-AA was mainly attributed to adsorption and reduction, and pH played an important role in determining the existing form of Cr(VI) in aqueous solution and the control of charge on the surface of the adsorbent (Zhang et al. 2020). HCrO 4 − , Cr 2 O 7 2− , and CrO 4 2− are the main forms of Cr(VI) present in aqueous environments (Araghi et al. 2015). When the pH is between 1 and 6.8, Cr(VI) exists mainly as HCrO − 4 andCr 2 O 2− 7 . When the pH is higher than 6.8, Cr(VI) exists mainly in the form of Cr 2 O 2− 4 (Gan et al. 2015). Under acidic conditions, there were large amounts of HCrO − 4 andCr 2 O 2− 7 in solution, which were easily reduced (Zhou et al. 2016). Therefore, the reduction of Cr(VI) by SC@ZVI@CS-AA increased. At the same time, the functional groups (-OH, -NH 2 , -SO 3 H) on SC@ ZVI@CS-AA became protonated (-OH + , -NH 2 + , -SO 3 H + ), which enhanced the electrostatic attraction between SC@ ZVI@CS-AA and HCrO − 4 orCr 2 O 2− 7 (Fonseca et al. 2009;Zhang et al. 2012;Wasim et al. 2015).

Effect of temperature
The effects of temperature on chromium removal with different initial concentrations (100, 200, 300, 400, and 500 mg/L) are shown in Fig. 3d. At certain Cr(VI) concentrations, the adsorption capacities of Cr(VI) and total Cr both increased with increasing temperature. As the reaction temperature increased, the reactivities of functional groups (-NH 2 , -OH, and -SO 3 H) present on the adsorbent surface were enhanced, which increased the rate of metal ions passing through the boundary layer and pore channels of SC and further enhanced adsorption and reduction.
Based on the equilibrium values obtained for adsorption with different initial Cr(VI) concentrations at different temperatures, the thermodynamic parameters were calculated. The ΔH (12.30 kJ/mol) for the adsorption process was positive, indicating that adsorptionof Cr(VI) by SC@ZVI@ CS-AA was endothermic. The ΔH data illustrated that rising temperature enhanced Cr(VI) adsorption. The ΔS value (41.08 kJ/mol) confirmed that the disorder of the solid-liquid interface increased during adsorption (Tan et al. 2009), which was conducive to adsorption (Zhang et al. 2010). The value of ΔG (0.06 kJ/mol) decreased with increasing temperature and was negative above 35 °C, indicating that the adsorption process was spontaneous when the temperature was higher than 35 °C.

Effect of SC@ZVI@CS-AA dosage
The effect of SC@ZVI@CS-AA dosage on chromium removal is shown in Fig. 4a. With increasing SC@ZVI@ CS-AA dosage, the adsorption capacity of Cr(VI) showed an overall downward trend, but was higher than the adsorption capacity of total Cr. The adsorption capacity of Cr(VI) reached a maximum value (119.39 mg/g) at a dose of 0.8 g/L and then began to decrease. The removal efficiency for Cr(VI) showed an upward trend; moreover, it increased slowly with increasing dosage.
With increasing adsorbent dosage, the effect of the quantity of ZVI reductant on Cr(VI) was enhanced. At the same time, the number of active sites on the surface increased significantly, which improved mass transfer between = SC@ZVI@CS-AA and chromium-containing anions, increased the collision frequency of chromium particles (Chen et al. 2012), and further enhanced adsorption. However, due to the constant concentration of Cr(VI) in the solution, excess reaction sites provided by SC@ZVI@ CS-AA did not participate in the reaction. Therefore, the chromium adsorption capacity per unit mass of SC@ ZVI@CS-AA decreased due to the reduced utilization of SC@ZVI@CS-AA.

Effect of coexisting ions
Other anions and cations coexisting with Cr(VI) are often present in groundwater, and cause varying degrees of interference in mass transfer between SC@ZVI@CS-AA and chromium-containing anions (Cao et al. 2017). The effects of interfering ions on chromium removal are shown in Fig. 4b. After the addition of different interfering ions (Ca 2+ , Mg 2+ , Cd 2+ , Na + , Cl − , NO 3− , SO 4 2− , and H 2 PO 4 2− ), the adsorption capacities of Cr(VI) and total Cr changed differently. Inhibition by SO 2− 4 was the most significant, resulting in 58.35% and 58.06% decreases in Cr(VI) and total Cr removal, respectively, compared with the blank sample. On the one hand, binding between SO 2− 4 and -NH 2 led to competitive adsorption with chromium-containing anions. On the other hand, the distribution of charges around the adsorption sites occupied by SO 2− 4 weakened the electrostatic attraction between SC@ZVI@CS-AA and chromate (Mohan and Pittman 2006). Wang et al. (2018) also found that a variety of interfering ions exhibited effects on the adsorption process of PPy-Fe 3 O 4 /rGO, and the inhibitory effect of SO 2− 4 was the most significant. When the SO 2− 4 concentration increased from 0 to 0.1 g/L, the removal efficiency decreased by 44.4%.

Effect of contact time and kinetics studies
The effect of contact time on adsorption at three different initial Cr(VI) concentrations (200, 300 and 400 mg/L) is shown in Fig. 5a. The q e of Cr(VI) gradually increased with increasing reaction time. Most of the adsorption process occurred within the first 240 min. With various initial concentrations (200, 300, and 400 mg/L), the adsorption capacity reached 140.392, 150.544, and 165.148 mg/g within 240 min, respectively. The adsorption capacity increased by only 3.606, 3.377, and 3.236 mg/g, respectively, between 240 and 480 min.
There were a large number of unreacted adsorption sites on the surface of SC@ZVI@CS-AA during the primary stage of adsorption. With a decrease in available adsorption sites, the adsorption capacity of Cr(VI) increased slowly. Although the chemical groups (-OH, -COOH, and -NH 2 ) on SC@ZVI@CS-AA chelated Cr(III) and Fe(III) to promote reduction (Neto et al. 2019), a small portion of Cr(III) complexed with Fe(III) to coprecipitate Fe x Cr 1−x OOH or Fe x Cr 1−x (OH) 3 on the surface of ZVI, which hindered electron transfer between ZVI and Cr(VI) and prevented adsorption of Cr(VI) on SC@ZVI@CS-AA (Gong et al. 2016).
The results of fitting with the pseudo-first-order and pseudo-second-order models are shown in Fig. 5b and c, respectively. Figure 5b plots the relationship between log q e − q t and t, and q e and K 1 were calculated from the slope and intercept. Figure 3c plots the relationship between t q t and t. Similarly, q e and K 2 were calculated from the intercept and slope (Prakash et al. 2012(Prakash et al. , 2013. The related parameters are shown in Table S1. The correlation coefficients (R 2 ) for the pseudo-second-order model were higher than those for the pseudo first-order kinetic model, and were close to 1. Moreover, the q e values (149.47, 158.98, and 174.21 mg/L) calculated with the pseudo-second-order model were closer to the experimental values (143.998, 153.921, and 168.384 mg/L). Additionally, they increased with increasing concentration, indicating that the pseudosecond-order kinetic model was better suited for explaining the adsorption mechanism; that is, the adsorption rate was controlled by chemical adsorption ).
The curves for the intraparticle diffusion model and the related parameters are shown in Fig. 5d and Table S2, respectively. Figure 5c shows the obtained three-segment linear simulations with the two inflection points of t 0.5 , which indicated a multilinear response corresponding to three stages: rapid external diffusion and adsorption, internal surface and pore diffusion, and gradually approaching equilibrium. The rate for adsorption of Cr(VI) from the liquid phase to the surface of SC@ZVI@CS-AA was higher than that for pore adsorption because of the higher value of K p for the first stage of the process. The value of the parameter corresponding to the boundary layer effect (C), the C value for the third stage, was significantly higher than that for the other two stages, suggesting that the boundary effect contributed more with increasing reaction time. With less residual Cr(VI) in solution and fewer adsorption sites on the surface, the intraparticle diffusion rate decreased; thus, the reaction gradually reached equilibrium. The C value increased with increasing initial concentration of Cr(VI), which indicated that adsorption played an important role in the removal process. Moreover, intraparticle diffusion was not the only mechanism operating in the process because none of the C values passed through the origin.

Effect of initial chromium concentration and isothermal models
The effect of the initial Cr(VI) concentration on adsorption at 25 °C is shown in Fig. 6a. The adsorption capacity of Cr(VI) increased gradually with increasing initial concentration. The adsorption capacity increased rapidly below 100 mg/L but slowly within the range 100 to 500 mg/L. Conversely, the overall removal efficiency showed a downward trend. Since the adsorbent dosage was fixed, the number of adsorption-reduction sites on the surface was basically the same. In this case, SC@ZVI@CS-AA was not fully involved in the reaction at low concentrations of Cr(VI). The higher the concentration of Cr(VI) in the solution was, the more chromium-containing anions were present, which enhanced the chemical adsorption of Cr(VI) by functional groups (-NH 2 , -OH, and -SO 3 H). In addition, redox reactions occurred when the Cr(VI) and ZVI levels were increased, further improving the utilization rate of the adsorbent. The slow growth trend in adsorption capacity with initial concentrations between 100 and 500 mg/L was due to the limited number of groups carried by the material itself. Under the same conditions, when the Cr(VI) concentration was high, a substantial amount of Cr(VI) was reduced to Cr(III), which increased the amount of Cr(III)-Fe(III) hydroxides bound to ZVI and formed a passivation layer comprising Cr 0.667 Fe 0.0.333 OOH or Cr 0.667 Fe 0.333 (OH) 3 (Powell et al. 1995). This passivation layer prevented the iron core from transferring electrons. Although the removal efficiency of Cr(VI) decreased, the adsorption capacity increased. The Langmuir, Freundlich, and Temkin isotherm models for the adsorption of Cr(VI) are displayed in Fig. 6b, c, and d, respectively, and the related coefficients are listed in Table S3. As the results illustrate, the R 2 value for the Langmuir model was much closer to 1.0 and higher than those of the Freundlich and Temkin models, demonstrating that the Langmuir model was more suited for fitting the data for the adsorption of Cr(VI) by SC@ZVI@CS-AA. Therefore, the adsorption process likely involved monolayer adsorption, and the maximum adsorption capacity of Cr(VI) by SC@ZVI@CS-AA increased with increasing temperature. It is generally believed that a value of 1/n (an empirical constant of the Freundlich isotherm) less than 1 is more conducive to adsorption (Mi et al. 2012). The 1/n value of this experiment was between 0.08 and 0.01, indicating that Cr(VI) was easily adsorbed by the adsorbent. Based on the equilibrium binding constant ( K T ), SC@ZVI@CS-AA had a large equilibrium binding energy for the removal of Cr(VI), suggesting that a strong chemical reaction occurred during the adsorption process.

Cyclic adsorption by SC@ZVI@CS-AA
The experimental results for cyclic adsorption are shown in Fig. 7. As Fig. 7a shows, when SC@ZVI@CS-AA was recycled three times, the maximum decrease in Cr(VI) adsorption capacity was 4.56 mg/g, and the maximum decrease in total Cr(VI) adsorption capacity was 19.01 mg/g. When sodium hydroxide (NaOH) was used for desorption, part of the Cr(VI) did not elute from SC@ZVI@CS-AA, which reduced the number of sites available during reuse. As shown in Fig. 7b, the desorption efficiency decreased from 89.68 to 87.13% for the second cycle of SC@ZVI@CS-AA. This showed that NaOH had a good effect on the desorption of Cr(VI). The higher desorption rate was more conducive to the reuse of SC@ZVI@CS-AA and the recovery of heavy metals. This result also indicated that ion exchange and coordination occurred during the removal of Cr(VI). All of the above results illustrated that SC@ZVI@CS-AA realized cyclic adsorption of Cr(VI).

Mechanisms of Cr(VI) removal by SC@ZVI@CS-AA
In summary, the mechanism of action between SC@ZVI@ CS-AA and Cr(VI) is thought to be as shown in Fig. 8. ZVI reduces Cr(VI) to produce Fe(III) and Cr(III). These elements form coprecipitated species (Fe x Cr 1−x OOH and Fe x Cr 1−x (OH) 3 ) on the surface of ZVI at a high pH, which hinders electron transfer between ZVI and Cr(VI) and reduces the reaction activity.
CS modified by AA contains a large number of -NH 2 and -OH groups. They combine with some Fe(III) to form a stable complex, thereby inhibiting the formation of a passivation layer on the surface of ZVI so that the reduction of Cr(VI) by ZVI is promoted. Moreover, the functional groups (-OH, -NH 2 , and -SO 3 H) on SC@ZVI@CS-AA undergo protonation (-OH + , -NH 2 + , and -SO 3 H + ) under acidic conditions, making the surface of SC@ZVI@CS-AA positively charged. Under the same conditions, Cr(VI) exists as HCrO − 4 andCr 2 O 2− 7 . The combination of SC@ZVI@CS-AA and Cr(VI) results from electrostatic attraction.
SC plays a role in supporting ZVI modified by CS-AA. Moreover, the number of reactive pores and specific surface area are increased, and -SO 3 H on the surface is activated to form −SO 3 H + , which aids in the adsorption of Cr(VI).

Conclusion
(1) ZVI was stabilized by a CS-AA sol, and SC was used as a carrier to prepare SC@ZVI@CS-AA. Characterization showed that the surface of SC@ZVI@CS-AA exhibited a cross-linked network structure rich in amino and hydroxyl groups and showed good thermal stability.
(2) The pseudo-second-order model gave the best fit for the kinetics of Cr(VI) adsorption by the SC@ZVI@CS-AA composite, indicating that chemisorption occurred during the process. The Langmuir isotherm model was best suited for describing the adsorption process. Thermodynamic studies indicated that the adsorption of Cr(VI) was spontaneous when the temperature was above 35 °C. (3) Under the optimal conditions, the removal efficiency of Cr(VI) by SC@ZVI@CS-AA reached 98.5%. The adsorption capacity of Cr(VI) remained above 70% after three reuses, indicating good cyclic adsorption of Cr(VI) on SC@ZVI@CS-AA. In particular, SO 2− 4 in solution significantly inhibited the removal of Cr(VI).
(4) In an acidic environment, SC@ZVI@CS-AA showed good removal of adsorbed Cr(VI). SC@ZVI@CS-AA exhibited cyclic adsorption and good prospects for application. However, this experiment only explored the influences of the main environmental factors on the removal of Cr(VI) by SC@ZVI@CS-AA. The actual situation of chromium pollution in groundwater is more complex. According to the specific application, the influence of the permeability coefficients of the materials, the water flow rate, and the service lives of the materials should be considered fully.