Experimental and DFT study of Cu(II) removed by Na-montmorillonite Open Access

In recent years, with the advancement of science and technology, the content of heavy metals (e.g. Cu(II), As(III), Pb(II) and Hg(I)) in the water environment has gradually increased, posing a serious threat to the stability of the Earth's ecosystem and the health of organisms (Turan & Ozgonenel 2013; Chen et al. 2015; Behera et al. 2021). Cu(II) in water mainly comes from industrial activities such as copper mining, brass manufacturing, electronics processing and non-ferrous metal smelting (Maskawat Marjub et al. 2019). The World Health Organization (WHO) recommends that the maximum quantity of Cu(II) in drinking water should be 0.2 mg · L⁻¹, while the European Communities Directive sets it at 0.1 mg · L⁻¹. Drinking excess copper-containing water (at a concentration greater than 6 mg · L⁻¹) can cause vomiting, stomach cramps, skin irritation, nausea and anemia, while excess Cu(II) can be deposited in the kidneys, liver, brain and other human organs in enrichment, leading to Wilson's symptoms, hypertension, atherosclerosis and coronary heart disease, among many other adverse consequences (Paksamut & Boonsong 2018; Ameh & Sayes 2019; Pesavento et al. 2019; Chu et al. 2020). Therefore, the removal of Cu(II) is an important and urgent topic in heavy metal wastewater treatment technology.

The current methods for the removal of Cu(II) include ionic redox (de Carvalho et al. 2013), chemical precipitation (Fedje & Strömvall 2019), electrode-position (Schmidt et al. 2021), ion exchange (Antuganova et al. 2016; Groenendijk & van Wunnik 2021; Yousefzadeh et al. 2022), thermal deposition (Liu et al. 2018), biological methods (Szewczuk-Karpisz et al. 2021), liquid–liquid extraction (Molaei et al. 2018), membrane separation (Sharma et al. 2020), reverse osmosis (Abdullah et al. 2020; Hamid et al. 2020) and adsorption (Zhu et al. 2019; Giusto & Pissetti 2021; Xue et al. 2021). Among these methods, adsorption is efficient, simple, economical and widely used. Environmentally friendly and inexpensive adsorbent materials are a hotspot for current applications. Among them, montmorillonite has gained international attention as a potential adsorbent due to its large specific surface area, high cation exchange capacity, low-cost and no secondary environmental pollution (Zhu et al. 2019; Xue et al. 2021). Montmorillonite is a common clay mineral consisting of a 2:1 cationic layered mineral with aluminum octahedral sheets (Xiao et al. 2020). Isomorphic substitutions in the tetrahedral or octahedral sheet cause a negative charge on the montmorillonite lamellae that are neutralized by cations in the interlayer space. Hence, the cationic interchangeability and electrostatic attraction of the montmorillonite are both important, which facilitates Cu(II) removal (Chen et al. 2015).

For the past few years, the adsorption performance of montmorillonite on Cu(II) has been studied by adsorption experiments. Zhou et al. (2012) investigated the adsorption performance of montmorillonite on a single heavy metal Cu(II) in simulated metallurgical wastewater, with a removal rate (RR) of 99.9% at a Cu(II) concentration of 50 mg · L⁻¹. Turan & Ozgonenel (2013) investigated the removal efficiency of Cu(II) from industrial leachate by biosorption of montmorillonite and maximized the removal efficiency of Cu(II), achieving the optimization of the levels of the analyzed factors, resulting in a RR of 80.7% for Cu(II) with raw montmorillonite, which could be increased to 88.91% for the modified montmorillonite. Chen et al. (2015) used Na-Mt and Ca-Mt to remove Pb(II), Cu(II), Co(II), Cd(II), Zn(II) Ag(I), Hg(I) and Cr(VI) from aqueous solutions, finding that ion exchange was the main adsorption mechanism and Na-Mt was more effective for heavy metal adsorption than Ca-Mt and proved to be a potentially useful material for Pb(II), Cu(II), Co(II), Cd(II) and Zn(II) removal from aqueous solution. Atasoy & Bilgic (2017) studied the adsorption of Cu(II) by montmorillonite to explore the effectiveness of montmorillonite in environmental applications. The results showed that the adsorption of Cu(II) by montmorillonite reached 8.136 mg · g⁻¹ by the Langmuir model. Montmorillonite feedstock had the potential to be used effectively in the production of low-cost adsorbents for the removal of copper from wastewater. Vezentsev et al. (2022) studied the adsorption of Cu(II) ions on bentonite (the main component is montmorillonite) as a function of the contact time and temperature of the medium. The adsorption of Cu(II) on bentonite clay increased as the temperature increased from 293 to 333 K. At the same time, the shape of the adsorption isotherms remains unchanged. Recently, adsorption experiments have been used to examine the characteristics of montmorillonite for the adsorption of Cu(II). However, the adsorption process is quite intricate that experimental approaches cannot fully explain the mechanism of Cu(II) adsorption. If the energy changes and kinetic parameters are investigated further from a quantum mechanical and molecular dynamics (MD) standpoint using molecular simulation techniques, insight into the microscopic mechanisms of Cu(II) adsorption on montmorillonite can be gained, allowing for a reasonable understanding and prediction of montmorillonite adsorption and modification. The density of states, electron transfer numbers, bonding locations and ionic ligand structures were investigated using density functional theory (DFT) and MD methods to fully understand the adsorption of heavy metal ions by minerals (Greathouse & Cygan 2005; Zhao & He 2014; Xing et al. 2015; Kong & Wang 2016). Therefore, it is still feasible to study the mechanism of Cu(II) adsorption by montmorillonite using a combination of experimental and simulation methods.

In this study, sodium-based montmorillonite (Na-Mt) was used as the adsorbent material and characterized by X-ray powder diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), Brunner–Emmet–Teller (BET), scanning electron microscopy (SEM) and energy-dispersive spectroscopy (EDS). The adsorption of Cu(II) by Na-Mt in an aqueous solution was investigated under batch experimental conditions. The effects of adsorption time, initial Cu(II) concentration, temperature, initial pH, adsorbent dosage and coexisting cations on the adsorption effect were studied. In addition, cyclic regeneration experiments were carried out to investigate the locking ability of Na-Mt on heavy metal Cu(II). Finally, the DMol³ module of Material Studio software based on the DFT was used to calculate the adsorption sites, energy changes and Hirshfeld population of montmorillonite. The Forcite module was applied to investigate the distribution characteristics, self-diffusion coefficients and concentration distribution of montmorillonite at the atomic level. The possible adsorption configuration and adsorption mechanism were given and analyzed.

The sodium-based montmorillonite (Na-Mt) (PGW type) was produced from the Nanocor Company, USA. PGW, it has an ion exchange capacity of 145 mEq/100g, an aspect ratio of 200–400, a specific gravity of 2.6, a moisture content of 12% and an average particle size of 16–22 μm. For NVE, the microcanonical, or NVE, ensemble represents a system of a fixed number of particles (N), in a fixed volume (V), and with a fixed total energy (E). For NVT, application of classical Legendre transforms allows alternative ensembles to be derived, an example being the canonical, or NVT, ensemble, where a system can exchange heat with the environment and so remain at constant temperature (T). Copper chloride (CuCl₂) and cobalt chloride (CoCl₂) were purchased from Adamas-Beta. Magnesium chloride (MgCl₂) and calcium chloride (CaCl₂) were obtained from Macklin. Sodium nitrate (NaNO₃) and sodium chloride (NaCl) were purchased from the Shanghai Titan Company. Hydrochloric acid (HCl) and sodium hydroxide (NaOH) were obtained from Tianjin Chemical Reagent Company. The reagents used in this study were all of the analytical grade and were not further purified.

The adsorption process is favorable when R_L is between 0 and 1 and unfavorable when R_L > 1. R_L = 1 means irreversible adsorption.

Five adsorption–desorption experiments were performed to evaluate the recoverability of the Na-Mt material. For each adsorption–desorption cycle, the dosage of Na-Mt (2 g · L⁻¹) and the initial concentration of Cu(II) (100 mg · L⁻¹) were kept constant and the contact time was 80 min, then the used Na-Mt material was immersed in 1 M NaCl solution. After each adsorption–desorption cycle, the Na-Mt material was filtrated, washed with deionized water and dried at 105 °C for the next adsorption.

The simulations (DMol³ and Forcite modules) were completed using Materials Studio software. The DFT calculation was implemented in the DMol³ program using the generalized gradient approximation (GGA) with the Perdew-Burke-Ernzerhof (PBE) functional (Perdew et al. 1996). A Grimme dispersion correction was added to the DFT calculation (Grimme 2006). In addition, spin unrestricted and formal spin as initial were also used in all calculations of the DMol³ module. The basic set of atomic orbitals used in the calculation was double numerical plus polarization (DNP). The core electrons were treated with effective core potentials (ECP), which provide no special treatment of cores. The number of integration points that used to integrate the wavefunction in reciprocal space was 3 × 2 × 1 (Kong & Wang 2016). The self-consistent field (SCF) convergence was set to 1 × 10⁻⁶ eV per atom. The effect of water on the adsorption was studied with the aid of the conductor-like screening model (COSMO) as implemented in DMol³ program, where the solvent was treated as a dielectric continuum. Based on the above settings, the configuration optimization and related energy (ΔE_ads) calculation of the montmorillonite unit cell model established were performed. In addition, the Hirshfeld population was analyzed by means of an optimized Cu(II) with montmorillonite model and a Na(I) with montmorillonite model.

In this study, the initial structure of montmorillonite was modeled by referring to literature (Viani et al. 2002). The parameters of the unit cell were a = 5.180 Å, b = 8.980 Å, c = 15.000 Å, α = β = γ = 90°. The cleavage of the (001) crystal plane was performed on the montmorillonite unit cell and a vacuum of 15 Å was added on the surface along the c-axis to minimize the artificial interference from the adjacent layer. Moreover, the charged nature of the montmorillonite laminate endowed its surface model with a negative charge and all the DFT and MD calculations for adsorption were executed on this plane.

MD involves the stepwise integration of Newton's equations from a given starting point. It is the most natural method of performing equilibrium statistical-mechanical calculations through a simulation. MD simulation calculations were executed with periodic periphery conditions in three dimensions. The supercell model of montmorillonite 4a × 2b × 1c was constructed on the basis of the unit cell, with x- and y-axis lengths of 20.72 and 17.96 Å, respectively (Peng et al. 2018; Chu et al. 2019; Zhu et al. 2019). This provided an appropriate montmorillonite surface for MD studies of Cu(II) adsorption. In order to obtain a realistic model of a clay mineral in contact with an aqueous solution, 480 water molecules and 20 Cu(II) were added to the interlayer of montmorillonite supercell during the simulation process. Firstly, the structure of the completed simulation box was optimized and annealed using the Forcite module and then the MD simulation was performed. Anneal algorithm for structural relaxation was used in microcanonical ensemble synthesis (NVE) with an equilibrium time of 500 ps (Zhu et al. 2019). For NVE, the microcanonical, or NVE, ensemble represents a system of a fixed number of particles (N), in a fixed volume (V), and with a fixed total energy (E). The final balanced structures of the NVE ensemble simulations were regarded as the initial structures of the canonical (NVT) ensemble simulations (Zhu et al. 2019). For NVT, application of classical Legendre transforms allows alternative ensembles to be derived, an example being the canonical, or NVT, ensemble, where a system can exchange heat with the environment and so remain at constant temperature (T). All the dynamic time step was 1 fs, the NVT simulation time was 1,000 ps and the number of dynamic steps was 1 × 10⁶. UFF (Universal force field) was applied and the Ewald method and the atom-based method were used to calculate the long-range electrostatic interaction and the van der Waals interaction, respectively (Zhu et al. 2019). The experiments (T = 298 K, 1 × 10⁵ Pa) were carried out using a Nose thermostat with the pressure regulated by a Berendsen constant pressure device and all atoms were relaxed (Zhu et al. 2019). In addition to the MD simulation of Cu(II) in an aqueous solution in the presence of montmorillonite, the kinetic behavior of Cu(II) in an aqueous solution was simulated and analyzed. The concentration distribution of Cu(II) adsorbed by montmorillonite was analyzed. Then, the mean square displacement (MSD) curve was drawn and the diffusion coefficient of Cu(II) was calculated. Finally, the radial distribution function of Cu(II) adsorption on montmorillonite was analyzed.

Based on the results of the analysis of the adsorption kinetics in Section 3.2.1, the value of the activation energy E_a can be + derived by selecting the natural logarithm of the rate constant of the pseudo-second-order kinetic equation to plot against 1/T, as depicted in Supplementary Figure S1 (Chen et al. 2021b; Tseng et al. 2022). The adsorption activation energy (E_a) was calculated to be 37.08 kJ · mol⁻¹. It was reported that an E_a of 40 kJ · mol⁻¹ was the critical value to distinguish physical adsorption (<40 kJ · mol⁻¹) and chemical adsorption (>40 kJ · mol⁻¹), thus the adsorption process of Cu(II) by Na-Mt was dominated by physical adsorption (Chen et al. 2021a).

The effect of the initial Cu(II) concentration on the adsorption of Cu(II) by Na-Mt at three temperature conditions (25, 35 and 45 °C) is presented in Figure 6(f). It can be noticed that as the equilibrium concentration of Cu(II) increased, the equilibrium adsorption capacity exhibited a trend of steep increase first followed by a gentle increase. To figure out the relationship between the adsorbent and the adsorbate, adsorption isotherm models, including the Langmuir model, Freundlich model and Langmuir–Freundlich model (Equations (3)–(5) were used for modeling these adsorption isotherm data. The fitting results are shown in Figure 6–6(i) and the relevant parameters are listed in Table 3. The correlation coefficient (R²) of the Langmuir isotherm model (R² = 0.974 at 25 °C) was greater than that of the Freundlich isotherm model (R² = 0.969 at 25 °C) and Langmuir–Freundlich model (R² = 0.969 at 25 °C). From the result, the Langmuir isotherm model can fit the adsorption process of Cu(II) by Na-Mt, reflecting that the adsorption process of Cu(II) by Na-Mt is mainly monolayer adsorption. The Q_L values fitted to the Langmuir equation showed an increase from 60.32 to 113.75 mg · g⁻¹ with increasing temperature, illustrating the significant effect of raising the temperature on the adsorption process. The separation factor R_L was 0–1 in the range of 25–45 °C, suggesting that the adsorption process is favorable at all three temperatures.

The thermodynamic parameters ΔG^T, ΔH^T and ΔS^T (Equations (7)–(9) are investigated to analyze the driving forces of adsorption or to determine whether the adsorption process is spontaneous. As shown in Supplementary Figure S2, ΔH^T and ΔS^T were obtained from the slope and intercept of van't Hoff plots of ln vs. 1/T. The thermodynamic parameters for the adsorption are listed in Table 4. The ΔH^T and ΔS^T values were positive at three temperature conditions (25, 35 and 45 °C), indicating that the adsorption of Cu(II) by Na-Mt was an endothermic and entropy-increasing process. The values of ΔG^T were negative at three temperature conditions (25, 35 and 45 °C), implying that the adsorption process of Cu(II) by Na-Mt was a spontaneous process. And the absolute values of ΔG^T increased with increasing temperature, indicating that increasing the temperature is beneficial for the adsorption to proceed to the right, thus the adsorption capacity of Na-Mt on Cu(II) is increased. Furthermore, the removal of Cu(II) from aqueous solutions by Na-Mt is not a simple adsorption process, but also affected by solvent effect and ion exchange, which cause a different phenomenon from that shown by normal adsorption processes (ΔH < 0 and ΔS < 0).

In practice, Cu(II) is not the only cation present in water, so four cations (Na(I), Ca(II), Co(II) and Mg(II)) were selected as representatives to study the effect of coexisting cations on the adsorption capacity of Cu(II). The effect of different cations on the adsorption capacity of Cu(II) with different concentrations are shown in Figure 7(b). The adsorption capacity of Cu(II) by Na-Mt decreased as the concentration of coexisting cations increased. The order of influence on the adsorption capacity of Cu(II) by Na-Mt was Mg(II) > Co(II) > Ca(II) > Na(I). This is because in ion exchange, the higher the valence of the ion and the smaller the radius, the more likely the ion exchange will occur. If the adsorption process of montmorillonite materials follows the electrostatic principle, the higher the value of Z/r (charge/radius) the greater the affinity, then the affinity of the multivalent ion is greater than that of the monovalent ion (Ucarli et al. 2020). The magnitude of Z/r is Mg(II) (2/72) > Co(II) (2/74.5) > Ca(II) (2/100) > Na(I) (1/102) (Marcus 1993), which corresponds to the order in the diagram.

The effect of pH on the adsorbent is a significant factor in adsorption experiments. Therefore, the adsorption capacity of Cu(II) at pH 3–8 was studied. The effect of pH on adsorption was investigated by adjusting the initial pH with 0.1 M NaOH solution and 0.1 M HCl solution. The adsorption result is shown in Figure 7(c) and the adsorption capacity of Cu(II) gradually increased from 21.22 (pH = 3) to 49.80 mg · g⁻¹ (pH = 8) with increasing pH. When the pH in the solution is low, a large amount of H⁺ is present in the solution and the H⁺ and Cu(II) ions will compete with each other for the adsorption sites on Na-Mt. A part of the H⁺ will occupy the adsorption sites on Na-Mt, weakening the adsorption of Cu(II); when the pH in the solution is further increased, the presence of a large amount of OH⁻ in the solution makes it easy for Cu(II) to precipitate Cu(OH)₂ or Cu(OH)⁺ complexes on the Na-Mt surface, resulting in an increase in adsorption capacity. This is also in agreement with the trend of similar previous studies (Han et al. 2018).

In practical applications, the adsorbent can be used repeatedly for the purpose of reusing to save material, which is the key to economic attractiveness, so we tested the recycled performance of the adsorbed Na-Mt. The reusability experiments of Cu(II) by Na-Mt were carried out and five adsorption–desorption cycles were repeated, as depicted in Figure 7(d). The results showed that the removal efficiency of Cu(II) decreases from 95 to 67% with increasing times of the reuse after five cycles. This indicated that Na-Mt adsorbent is quite stable and feasible for the repeated application of heavy metals such as Cu(II).

The adsorption of Na(I) or Cu(II) on montmorillonite are mainly an electrostatic attraction and this interaction can lead to the transfer of electrons between the interlayer cations and the laminate. By analyzing the Hirshfeld charges of Al, Si, O and interlayer cations in montmorillonite before and after adsorption, the charge transfer that occurred when the montmorillonite interacted with cations was verified. Therefore, Hirshfeld population analysis was used to calculate the charges of the atoms after adsorption on montmorillonite and the results are shown in Supplementary Table S2. It was found that the charge transfer amount of Cu(II) was more than that of Na(I), indicating that the electrostatic attraction between Cu(II) and montmorillonite laminate may be stronger than the electrostatic attraction between Na(I) and montmorillonite laminate. This is consistent with the conclusion that the interaction between Cu(II) and the laminate is stronger for adsorption energy studies.

The adsorption kinetics study is based on a comparison of the diffusion coefficients of Cu(II) in the pure aqueous phase system and in the aqueous phase system in the presence of montmorillonite to accurately evaluate the adsorption properties of montmorillonite. The MSD of them were computed through the 1,000 ps simulations and the results are presented in Supplementary Figure S4. The diffusion coefficients of Cu(II) in the aqueous phase system and the aqueous solution system containing montmorillonite were 0.13 × 10⁻⁸m² · s⁻¹ and 0.85 × 10⁻¹⁰ m² · s⁻¹, respectively. As expected, the diffusion coefficient of Cu(II) in the aqueous system of montmorillonite was significantly lower than that in the aqueous phase system, indicating strong adsorption of Cu(II) by montmorillonite (Atasoy & Bilgic 2017; Chu et al. 2020).

The interaction relationship and microscopic distribution of ions can be deduced by the radial distribution function. The average location of the individual molecules in a solvent is expressed in terms of a radial distribution function, g(r). This function relates the probability of finding another molecule at a particular distance r from each molecule. The interaction between Cu(II) and montmorillonite was studied by using radial distribution function. The radial distribution function results shown in Supplementary Figure S5 are calculated according to the MD simulation trajectory. The first sharp peak was associated with Cu(II) and was located at a distance of 2.25 Å from the surface of the montmorillonite laminate. This indicated that at a distance of 2–3 Å the interaction between Cu(II) and the surface of the montmorillonite laminate is strong and a substantial amount of Cu(II) ions are adsorbed on the surface of the montmorillonite laminate. This is also consistent with the experimental data of SEM-EDS and XRD.

In this study, the adsorption characteristics and mechanism of Cu(II) on Na-Mt were investigated by means of adsorption experiment and molecular simulation. The characterization of the samples by XRD and EDS showed that Cu(II) was successfully adsorbed on the Na-Mt. The adsorption capacity of Na-Mt on Cu(II) was 35.23 mg · g⁻¹ at 80 min. The adsorption process was in accordance with the pseudo-second-order kinetic and the Langmuir isotherm models. The activation energy (E_a) (37.08 kJ · mol⁻¹) indicates that the adsorption process was dominated by physical adsorption. The adsorption was a spontaneous, endothermic and entropy-increasing process. The order of influence of coexisting cations on the adsorption capacity of Cu(II) is Mg(II) > Co(II) > Ca(II) > Na(I). The variation in pH has a greater effect on the adsorption capacity of Cu(II) by Na-Mt. After five cycles, Na-Mt still exhibited a high Cu(II) RR of 67%. The simulations demonstrated that there were no significant differences in the adsorption energy of Cu(II) at the four adsorption sites on the surface of montmorillonite(001). The Hirshfeld population analysis found that the electrostatic attraction between Cu(II) and montmorillonite laminate may be stronger compared with Na(I). The diffusion coefficients of Cu(II) in the aqueous phase system and the aqueous solution system containing montmorillonite were 0.13 × 10⁻⁸ and 0.85 × 10⁻¹⁰ · m² s⁻¹, respectively. A considerable amount of Cu(II) ions were adsorbed at a distance of 0.26 and 2.25 Å from the surface of the montmorillonite laminate. These results and methods are important guidelines for the research and application of Cu(II) adsorbents.

The works were supported by the Major Science and Technology Project of Xinjiang Bingtuan (2020AA004), the Major Science and Technology Project of Bashi Shihezi City (2020ZD02), the National Natural Science Foundation of China (NSFC Grant No. 21968029), the Major Science and Technology Project of Xinjiang Bingtuan (2017AA007), the Applied Basic Research Program of Xinjiang Bingtuan (2015AG004) and Science & Technology Research and Achievement Transformation Project of Xinjiang Bingtuan (2015AD027).

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

The authors declare there is no conflict.