Adsorption of nitrate and nitrite from aqueous solution by magnetic Mg/Fe hydrotalcite

Since the 1960s, with the development of industry and agriculture, several countries have experienced different degrees of groundwater pollution from nitrate (Galloway et al. 2008). Nitrates in high concentrations and their reduction to nitrites result in an adverse effect on the environment, and could potentially cause human health problems such as blue baby syndrome in infants and stomach cancer in adults (Majumdar & Gupta 2000; Kim-Shapiro et al. 2005; Boehm 2019). Excess of nitrate in drinking water can also lead to various types of cancer in humans (Aliaskari & Schäfer 2020).

Several methods for removal of nitrate or nitrite from water have been applied, such as reverse osmosis (Berkani et al. 2019), catalytic reduction (Han et al. 2019; Elias et al. 2020), biological denitrification (Mulholland et al. 2008; Liu et al. 2021a), ion exchange and electrodialysis (Wiercik et al. 2020; Zeng et al. 2020). However, biological processes are easily affected by temperature, and the effluent needs further treatment, such as disinfection (Chen et al. 2009; Zeng et al. 2020). Reverse osmosis and electrodialysis are relatively expensive, and merely displace nitrate or nitrite into the concentrated waste brine, causing disposal problems (Samatya et al. 2006). Ion exchange needs high energy or a high or expensive dose of the reagent (Zeng et al. 2020). In recent decades, adsorption methods have received great attention in the removal of nitrate or nitrite from water, due to their simplicity, sludge-free operation, ease of handling and the availability of various adsorbents (Song et al. 2016; Nassar et al. 2020). So far, clay materials such as hydrotalcites, with the advantages of low cost, abundant sources and easy preparation, have gained great attention from researchers. Hydrotalcite comprises up and down parallel layers and a large internal space for exchange of anions from water (Jung et al. 2020). Carbonate ions and crystal water are present in the interlayer (Saifullah & Hussein 2015). Upper and lower surfaces typically include metal oxide and metal hydroxide (Xia et al. 2020). The layer structure exhibits positive charge, and the internal anion exhibits negative charge, eventually rendering the hydrotalcite electrically neutral (Li et al. 2020). Hydrotalcite possesses a unique microporous structure, tunable denaturation, memory effect of calcination, interlayer anion-exchange ability, and a high degree of order (Ogata et al. 2018; Cheng et al. 2021). In a study reported previously by our group, calcined Mg/Al hydrotalcite possessed a high adsorption capacity for nitrate (34.36 mg N/g) and nitrite (37.17 mg N/g) (Wan et al. 2012). However, issues related to the effective separation and recovery of hydrotalcite from solution still need to be resolved. The preparation of magnetic hydrotalcite has been developed to resolve the above-mentioned problem, and it has been applied to adsorb toxic anions or compounds such as methyl orange (Deng et al. 2016), phosphate (Sun et al. 2013) and arsenic (Toledo et al. 2010). However, as far as we know, research on the adsorption properties of magnetic hydrotalcite for nitrate and nitrite is still limited.

Considering that Al can damage the human body and has negative effects on health, in this study, magnetic Mg/Fe hydrotalcite calcined material (M-CHT) was synthesized through the co-precipitation and calcination method, and then was used to remove nitrate and nitrite from water. The adsorption properties (including kinetics and isotherm) for nitrate and nitrite over M-CHT under batch conditions were investigated. Moreover, the adsorption mechanism was also analyzed based on characterization (XRD, XPS, FTIR and VSM).

FeCl₃·6H₂O, MgCl₂·6H₂O, FeCl₂·4H₂O, NaNO₂, NaNO₃, NaOH and Na₂CO₃ were all of analytical grade and purchased from Kemiou Chemical Reagent Co., Ltd (Tianjin, China). The solutions used in all experiments were prepared using ultrapure water of 18.25 MΩ. The 20% ammonia used was in the form of ammonium solution.

M-HT was synthesized by the co-precipitation method. First, a magnetic matrix solution was prepared by dissolution of FeCl₂·4H₂O (0.24 mol/L Fe²⁺) and FeCl₃·6H₂O (0.48 mol/L Fe³⁺) in 100 mL deionized (DI) water. Under the conditions of a controlled temperature of 45 ± 1 °C and vigorous stirring, 20% ammonia solution was added dropwise into the above solution to adjust the pH at 11 ± 1. The resulting precipitate was aged at 45 ± 1 °C for 30 min. The as-obtained oily black precipitate was centrifuged and washed with the deionized water several times until the solution pH was neutral. The obtained substance was stored in a 500 mL conical flask containing 100 mL of deionized water for further use.

Next, MgCl₂·6H₂O (1.2 mol/L) and FeCl₃·6H₂O (0.4 mol/L) were dissolved in 200 mL of deionized water (solution A). Then, solution B containing a mixture of 25.60 g of NaOH (3.2 mol/L) and 4.24 g of Na₂CO₃ (0.2 mol/L) was prepared. The two solutions (A and B) were simultaneously added dropwise into 100 mL of the as-prepared magnetic matrix water under vigorous stirring. The temperature and pH were maintained constant at 40 ± 1 °C and 10 ± 1, respectively. The resulting slurry was stirred for 2 h and added into a thermostatic water bath at 65 ± 1 °C for ∼18 h. The resulting product was centrifuged and washed with the deionized water several times until the electrical conductivity of the supernatant was less than 300 μs/cm. Then, the product was dried at 70 °C and sieved with 100 mesh to obtain the powder, which was marked as M-HT. M-HT was subjected to calcination at 500 °C for 5 h, and sieved with 100 mesh to obtain the final product, which was marked as M-CHT.

Adsorption equilibrium studies were carried out by utilizing a constant mass (0.10 g) of M-CHT with 100 mL of the nitrate or nitrite solution. Nitrate or nitrite concentrations were 5, 10, 15, 20, 30, 45, and 60 mg N/L. M-CHT with the nitrate or nitrite solution was placed in a temperature-controlled orbital shaker with a stirring speed of 150 rpm. The pH of the mixture was not adjusted to avoid the effect of other anions. After shaking the flasks for 24 h, the solution sample was filtered by a 0.45 μm membrane. The adsorption capacity of M-CHT toward nitrate or nitrite at equilibrium is denoted as q_e (mg N/g).

Nitrate and nitrite concentrations were measured by Hitachi U-3010 spectrophotometer. XRD data were recorded in a 2θ range of 5° to 80° on a D8 Advance diffractometer using Cu Kα radiation. X-ray photoelectron spectroscopy (XPS) was measured on a Thermo Scientific Escalab 250SXi. Transmission electron microscopy (TEM) was measured on a Tecnai G2F30 microscope. Fourier transform infrared (FTIR) spectra were measured through a WQF-510 spectrometer.

As seen in Figure 1, in the first 500min, M-CHT rapidly adsorbed nitrate and nitrite; then, the adsorption rate became sluggish, and the adsorption saturation time was 750min. A majority of active adsorption sites were available for nitrite or nitrate in the first 500min, while after that, the active adsorption sites on M-CHT were gradually saturated. With the contact time increasing, the amount of adsorbed nitrate and nitrite increased, and at 308K, almost 80% of the nitrate and nitrite was removed in 750min.

Table 1 shows the adsorption kinetics results which were fitted by the first-order, pseudo-second-order, and intraparticle diffusion models. According to the values of the correlation coefficient with nitrate and nitrite, higher R² values (≥0.97) were fitted by the first-order and pseudo-second-order models. Specifically, the first-order model is well fitted, indicating that nitrate and nitrite adsorption rates are controlled by diffusion. In addition, the adsorption data were fitted with the pseudo-second-order model, and adsorption was concluded to be chemical adsorption (Hu et al. 2016). In addition, the initial adsorption rate can be calculated by from the pseudo-second-order model. The adsorption rate of nitrate increased from 209.33 to 404.99 mg/(g·min), while that of nitrite increased from 164.56 to 476.23 mg/(g·min). It can be explained that improving the solution temperature could facilitate the initial adsorption rate of nitrate and nitrite. The adsorption capacities of M-CHT fitted by the first-order model for nitrate at 298, 303 and 308 K were respectively 8.37, 8.11 and 8.78 mg N/g. The corresponding values for nitrite were 7.92, 8.21, and 8.59 mg N/g, indicating that the increase of solution temperature can promote the adsorption of nitrate or nitrite on M-CHT, which is consistent with the above result.

The adsorption isotherm is an important way to describe the maximum adsorption capacity of M-CHT for nitrate or nitrite. To discuss the effect of different temperatures, the adsorption equilibrium of nitrate and nitrite on M-CHT was investigated at 298, 303 and 308 K. As shown in Figure 2, the equilibrium concentration of nitrate or nitrite increased, and the equilibrium adsorption capacity was also increased. With the temperature increasing from 298 to 308 K, adsorption equilibrium was achieved more rapidly. The adsorption capacities of nitrate and nitrite were high at high temperatures.

Table 2 summarizes the adsorption isotherm parameters for nitrate and nitrite. The Langmuir isotherm model afforded the better fitting results with R² values (≥0.97), which was greater than those obtained by the Freundlich model, indicating that nitrate and nitrite are uniform on the adsorbent surface and adsorption may be monolayer adsorption (Wan et al. 2012; Rodrigues et al. 2019). Meanwhile, at 298, 303 and 308 K, the maximum adsorption capacities of M-CHT for nitrate were 7.89, 13.00, and 14.28 mg N/g, respectively. The corresponding values for nitrite were 10.60, 12.40 and 16.90 mg N/g. The adsorption capacities of nitrate and nitrite at different temperatures increased, ranking in the order of 298 K < 303 K < 308 K. High temperature was favorable for adsorption; this tendency was in agreement with that reported by Rodrigues et al. (2019). At 308 K, M-CHT exhibited highest adsorption capacity, and the Langmuir parameters for nitrate adsorption were Q₀ = 14.28 mg N/g, and K_L = 1.39 L/mg; the corresponding values for nitrite were Q₀ = 16.90 mg N/g, and K_L = 0.59 L/mg.

Figure 3 and Table 4 present the Arrhenius equations, considering the satisfactory correlation coefficient values of 0.9934 and 0.9298. The adsorption processes have an adsorption activation energy value of 35.92 kJ/mol for nitrate and that of 67.12 kJ/mol for nitrite onto M-CHT. When the E_a value is lower than 40 kJ/mol, the adsorption type can be considered as a physical adsorption process. When the E_a value is greater than 40 kJ/mol, it suggests chemical adsorption (Bagheri et al. 2015). Herein, the E_a value of nitrite adsorption was greater than 40 kJ/mol, indicating the feasibility of the adsorption process being predominantly chemical in nature. The E_a value of nitrate adsorption was lower than 40 kJ/mol, which might be physical adsorption.

Hydrotalcite-like material is an alkali compound, and the solution pH profoundly affects its adsorption performance. Figure 4 shows the results of adsorption capacities at equilibrium (q_e) under different initial pH values. M-CHT exhibited a high nitrate adsorption capacity at initial pH range of 3.36–8.45, as well as high nitrite adsorption capacity at initial pH range of 3.57–9.4, indicating that M-CHT has a high adsorption capacity toward nitrate and nitrite in a wide range of solution pH. The maximum nitrate and nitrite adsorption capacities were 9.91 mg N/g at pH 6.33 and 14.24 mg N/g at pH 6.38, respectively.

When initial pH was >4.0, the final pH after adsorption exceeded 10.5, suggesting that M-CHT is a strongly alkaline material. After adsorption, pH increased possibly due to the release of OH⁻ from hydrotalcite. At the same time, at initial pH values of 4–10, these trends were not significant, indicating that hydrotalcite exhibits a certain buffering effect on the change of the solution pH; hence, within a certain range of pH, the effect of pH on the adsorption capacity of nitrate and nitrite by CHT is not extremely significant, and the adoption range is wide (Ahmed et al. 2020). With the decrease in the solution pH to 2.50, the nitrate and nitrite adsorption capacities decreased to 6.41 mg N/g and 10.12 mg N/g, respectively, with a corresponding decrease in the final solution pH to 10.09 and 10.04. Thus, a strongly acidic environment reduces the stability of the laminate structure of materials (Ferreira et al. 2006), thus decreasing adsorption capacity for anions. With the increase in the pH, the competitive adsorption of nitrate or nitrite by a high number of OH⁻ in the solution increased, leading to the decreased adsorption of the nitrate or nitrite.

Typically, anions such as F⁻, Cl⁻, ClO₄⁻, SO₄²⁻, CO₃²⁻, and PO₄³⁻ are present in nitrate- and nitrite-contaminated water, which can compete with nitrate or nitrite for adsorption sites on materials (Gierak & łazarska 2017). As shown in Figure 5, in the control group (no coexisting ions), the removal efficiencies of nitrate and nitrite by M-CHT were 46.34% and 68.56%, respectively. From the general trend observed in the figure, the adsorption capacity of nitrate and nitrite significantly decreased in the presence of coexisting anions. The order of influence is PO₄³⁻ > CO₃²⁻ > F⁻ > SO₄²⁻ > Cl⁻ > ClO₄⁻. The adsorption ability of M-CHT for nitrite or nitrate from the solution was mainly dependent on the electrical affinity of its positive surface. In the presence of PO₄³⁻ in the solution, the adsorption capacity was significantly decreased. After the calcination of M-CHT, the interlayer water or interlayer anions were lost, and the material surface exhibited a positive charge. The higher the coexisting anion valence, the poorer the adsorption of nitrite or nitrate by M-CHT (Li et al. 2016). Thus, PO₄³⁻ is the most competitive anion. In addition, anion radius affected the adsorption capacity. Compared with Cl⁻, the anion F⁻ with smaller radius has the more adverse effect on the adsorption of nitrate or nitrate.

XRD was employed to investigate sample structure. Figure 6 shows the XRD patterns of M-HT, M-CHT, and M-CHT-A. M-CHT-A-NO₃⁻ and M-CHT-A-NO₂⁻ represent the calcined hydrotalcite with adsorbed nitrate and nitrite, respectively. As shown in Figure 6, before calcination, M-HT exhibited a typical HT-CO₃²⁻ structure with sharp, symmetric (003), (006), (110) and (113) reflections, as well as wide, symmetric (012), (015) and (018) reflections, revealing the characteristics of the hydrotalcite-like compounds (Wang et al. 2018; Kang et al. 2020). A peak observed at a 2θ of 30.34° was indexed to Fe₃O₄ (JCPDS 26-1136), indicating that the magnetic substrate had been successfully loaded on the hydrotalcite (Zhang et al. 2013). After calcination at 500 °C for 4 h, all of the characteristic reflections (003, 006, 110, 113, 012, 015 and 018) of M-HT disappeared, suggesting the destruction of the layer structure of the M-CHT. Meanwhile, a mixed oxide of Mg(Fe)O with peaks at 43° and 62° (M-CHT) was formed (Yang et al. 2012). The peak indexed as Fe₃O₄ can still be clearly observed at 30.34°. Hence, M-CHT should be a mixed metal oxide, and the calcination treatment did not destroy the structure of Fe₃O₄. After adsorption of nitrate or nitrite on M-CHT (M-CHT-A-NO₃⁻, M-CHT-A-NO₂⁻), the layered structure was reconstructed, which is indicative of the adsorption of nitrate or nitrite on the positive layer and formation of a negative layer.

The interlayer spacing was calculated by using the basal spacing (d003) minus the width of the brucite-like layer (Wan et al. 2012). Herein, the internal spaces of CHT and M-CHT were 0.293 nm and 0.299 nm, respectively, indicating that the addition of the magnetic matrix does not affect the internal space of M-HT.

To analyze the surface composition and elemental states of M-CHT, XPS was adopted. All elements were marked in the full spectrum map (Figure 7(a)). The typical binding energies were observed at 56.92 eV, 301.38 eV, 532.80 eV, and 727.12 eV, corresponding to Mg2p, C1s, O1s, and Fe2p, respectively, which are consistent with the main constituent elements of M-CHT.

As presented in Figure 7(b), the binding energies at 712.3 eV and 725.6 eV were assigned as characteristic of Fe³⁺, and the binding energies at 710.6 eV and 723.9 eV were assigned as characteristic of Fe²⁺ (Liu et al. 2020a). In addition, the binding energy at 719.06 eV is the common satellite peak of both Fe³⁺ and Fe²⁺. As calculated by the peak areas from XPS, the peak area ratios of Fe²⁺ and Fe³⁺ were 35.96% and 64.04%, respectively. The actual Fe²⁺/Fe³⁺ value of 0.56 is basically consistent with the theoretical value of 0.5, indicating that Fe₃O₄ is doped in hydrotalcite (Yan et al. 2015; Liu et al. 2020a).

Figure 8 shows FTIR spectra of M-HT, M-CHT and M-CHT-A. The wide band at 3,460 cm⁻¹ corresponds to the -OH bending vibration from hydroxyl groups and interlayer water (Yan et al. 2015; Shi et al. 2020). A weak band at 1,649 cm⁻¹ was assigned as the bending vibration of the interlayer water (Abdelkader et al. 2011; Liu et al. 2020a). The peak at 1,364 cm⁻¹ is the vibrational peak of CO₃²⁻ (Saiah et al. 2009). The band between 400 cm⁻¹ and 800 cm⁻¹ is attributed to the stretching bands of the magnesium iron skeleton (Liu et al. 2020a). The above results indicate that M-HT exhibits characteristics of hydrotalcite, comprising interlayer water and carbonate, and the introduction of the magnetic substrate does not change its properties.

After calcination, the characteristic peaks of hydrotalcite at 3,460 cm⁻¹ and 1,649 cm⁻¹ became weak or disappeared, mainly due to the collapse of the lamellar structure, disappearance of functional groups such as OH⁻, CO₃²⁻, and H₂O at high temperature, and conversion of the sample to a mixed oxide (Wan et al. 2012). After the adsorption of nitrate or nitrite, new peaks were observed at 1,384 cm⁻¹ and 1,271 cm⁻¹, corresponding to nitrate and nitrite (Ogata et al. 2018), indicating the successful adsorption of nitrate and nitrite on M-CHT. Moreover, for all samples, the peak at 582 cm⁻¹ corresponds to the Fe-O stretching vibration (Liu et al. 2020b, 2021b), which is a feature of Fe₃O₄ (Pandi & Viswanathan 2016). It indicated that Fe₃O₄ was successfully loaded on the hydrotalcite, and the calcination and adsorption process does not affect the structure of Fe₃O₄.

XRD, XPS and FTIR results indicated that Fe₃O₄ was successfully loaded on the hydrotalcite matrix. Figure 9 shows the magnetic hysteresis curve of M-CHT at room temperature. M-CHT exhibited a magnetization of 9.15 emu/g. The coercive force and remanence were close to zero, indicating that M-CHT is a superparamagnetic material (Xu & Wang 2012; Shen et al. 2019). The inset image in Figure 9 shows the result of magnetic separation of M-CHT after 5 min, which indicates that M-CHT can be easily separated and recovered.

In this study, M-CHT synthesized by the co-precipitation and calcination method exhibited high adsorption capacity for nitrate and nitrite from contaminated water under a wide range of pH (initial pH ranged from 3 to 9). The adsorption kinetics and isotherm of nitrate and nitrite can be described with the first-order, pseudo-second-order model and Langmuir model, respectively. In the presence of coexisting anions, the removal efficiency of nitrate and nitrite over M-CHT decreased in the order of PO₄³⁻ > CO₃²⁻ > F⁻ > SO₄²⁻ > Cl⁻ > ClO₄⁻. XRD and FTIR analysis revealed that M-CHT can recover its original layered structure after the adsorption of nitrate or nitrite. Meanwhile, XRD and XPS analysis confirmed that Fe₃O₄ was successfully loaded on hydrotalcite, and did not affect the hydrotalcite structure. M-CHT is a magnetic material and it can be easily recycled using a magnet. Thus, M-CHT exhibits great prospects for application in wastewater purification.

This work was financially supported by National Natural Science Foundation of China (51878251), Key Scientific and Technological Research Project in Henan Province (192102210170, 14B430005, 182102210398, 172102310137), State Key Laboratory of Materials Processing and Die & Mould Technology, Huazhong University of Science and Technology (P2019-004).

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

All relevant data are included in the paper or its Supplementary Information.