Mechanism of adsorption affinity and capacity of Mg(OH)₂ to uranyl revealed by molecular dynamics simulation

Abstract

Because of its remarkably high adsorption affinity to uranyl ions, Mg(OH)₂ can effectively extract trace-level uranyl and has been exploited for the treatment of field water samples. In this work, we used molecular dynamics simulation to systematically study the dynamics, energetics and structure aspects of uranyl adsorption on the Mg(OH)₂ (001) surface. The approach of the uranyl cation causes the redistribution of surface OH groups and the emergence of a negatively charged surface region, which accommodates the adsorption of uranyl. The adsorption stability of uranyl is largely attributed to the coordination interaction with surface OH groups, and the calculated adsorption free energy is in quantitative agreement with experimental results. On the other hand, the adsorbed uranyl affects the orientation of surrounding OH groups, which may hinder the additional uranyl adsorption to the adjacent region and limit the adsorption capacity. The estimation of monolayer surface coverage is also well consistent with the experiments. Taken together, our results reveal the mechanisms of both adsorption affinity and capacity of Mg(OH)₂. As suggested by this work, comprehensive studies about uranyl adsorption can provide insight into the adsorption properties and should be helpful for the further development of uranyl adsorbents.

---

Introduction

One of the challenges of the nuclear power industry is the removal of toxic and radioactive uranium from industry wastewater. Uranium disposed into the environment can lead to radionuclide contamination of groundwater systems, and eventually reaches the top of the food chain, becoming a risk factor for human health.¹ The dominant and mobile form of dissolved uranium in contaminated groundwater system is uranyl cation (UO₂²⁺). However, the uranyl concentration in contaminated groundwater sample is at trace level (around 100 μg L⁻¹),² which hampers the uranyl removal. A variety of techniques have been developed to realize efficient removal of uranyl from low concentration samples, e.g., ion exchange, membrane based separation, solvent extraction, co-precipitation, and adsorption.^3–7 Adsorption is an effective and economic purification method for removal, recovery and recycling of uranyl from wastewater. In past decades, the development of adsorbents toward effective extraction and enrichment of uranyl, and the studies about underlying mechanisms have attained intensive interests.^8–10

A growing body of experimental works has been conducted to study the uranyl adsorption on various adsorbents, e.g., clay minerals,^11–15 layered double hydroxides (LDHs),^8,9 carbon,^16,17 and so on. In many cases, efforts have focused on the thermodynamic properties of uranyl adsorption, wherein the influences of experimental parameters, including pH, temperature, and ion concentration were also discussed. In addition, extended X-ray absorption fine-structure spectroscopy (EXAFS)¹⁸ and time-resolved laser-induced fluorescence spectroscopy (TRLFS)¹⁹ were recently exploited to provide information regarding the species and structure of adsorbed uranyl after equilibrium, while detailed information about the interaction between uranyl and adsorbent is still lacked.

Molecular dynamics simulation has been widely used to study the interaction between uranyl and mineral surfaces including montmorillonite (001),^20–24 kaolinite (001),^25,26 orthoclase (001),^27,28 calcite (101̄4),²⁹ quartz (001)/(010),^30,31 rutile TiO₂ (110),^32,33 and goethite (110).³⁴ These studies revealed the influence of porous structures and surface functional groups to uranyl adsorption, and provided atomic structure information about adsorbed uranyl. Comprehensive understanding about the adsorption properties requires full description of uranyl adsorption behavior. However, to our knowledge, limited attempts have been made to study the process of uranyl adsorption.

As a nontoxic and environmentally friendly mineral, Mg(OH)₂ has been widely used for the extraction of heavy metal ions.^35–39 More importantly, Mg(OH)₂ has a higher Langmuir constant (i.e., b value, reflecting the affinity and selectivity of adsorbent) toward uranyl than many other minerals including hematite, diatomite, sepiolite, silica gels.^35,38 Hence, Mg(OH)₂ has been successfully exploited to extract trace uranyl (μg L⁻¹ to mg L⁻¹) from field water samples and the extraction behavior follows monolayer adsorption mechanism, interestingly, the monolayer adsorption capacity is found to be relatively low (6.92 mg g⁻¹).³⁵ Such relatively low adsorption capacity seemingly contradicts to the high affinity of materials. Even though the uranyl sorption amount dramatically increases when Mg(OH)₂ exposed to concentrated solution (e.g., hundreds of mg L⁻¹), the mechanism deviates from monolayer adsorption model of trace uranyl extraction.

In this work, we used molecular dynamics simulation to systematically study the dynamics, energetics and structure aspects of uranyl adsorption on the Mg(OH)₂ (001) surface. The approaching of uranyl cation causes the redistribution of surface OH groups and the emergence of negatively charged surface region which accommodates the adsorption of uranyl. Besides, the adsorption stability of uranyl is largely attributed to the coordination interaction with surface OH groups. The estimated adsorption free energy of uranyl (−26.08 kJ mol⁻¹) suggests remarkably high adsorption affinity of Mg(OH)₂ to uranyl. Moreover, we found that adsorbed uranyl can affect the orientation of surrounding OH groups, which may hinder the additional uranyl adsorption to the adjacent region and limit the monolayer adsorption capacity. According to this model, we estimated the monolayer surface coverage of Mg(OH)₂ (001) surface, and the result is quantitatively consistent with previous experiment.^35,38 In short, our results reveal both mechanisms about high adsorption affinity and relatively low adsorption capacity of Mg(OH)₂. Systematic studies about the adsorption behavior of uranyl and the underlying mechanisms are highly demanded to facilitate further development of adsorbent.

Systems and methods

Crystal structure of Mg(OH)₂ at ambient conditions is brucite (trigonal, P3̄m1). The Mg(OH)₂ supercell was constructed according to the crystallographic parameters reported by Kazimirov et al.⁴⁰ The supercell consists of 14 × 14 × 4 crystallographic unit cells, resulting in 784 Mg atoms and 1568 OH groups. The dimension of supercell is 44.08 × 38.18 × 19.08 ų. And the supercell is terminated by the (001) plane, the cleavage plane of brucite, with all OH groups remaining intact as in our previous study.⁴¹ The bottom layer of the supercell was restrained to its initial position. The average position of surface hydroxyl oxygen atoms was defined as zero in the following analysis. An aqueous region with the thickness of 90 Å was placed above the substrate. As indicated by previous work, the species of uranyl is affected by the solution pH.^42–44 Free uranyl ion is considered as the dominant species in the solution under acidic condition,⁴⁴ while the uranyl tends to form complexes with carbonate when the pH increases.^42,43 Accordingly, both kinds of initial configurations were considered in this work. We first studied the situation where the unrayl and carbonate were initially placed apart (i.e., free-uranyl system). And the case wherein the uranyl and carbonate form pair before adsorption was also studied (i.e., uranyl–carbonate-pair system). In both systems, the initial position of uranyl is at least 5 Å above the upmost surface. We also studied the impact of ionic strength to the uranyl adsorption by considering the adsorption in the solution of 0.6 M NaCl (the salinity of seawater). Requisite numbers of sodium and chloride ions were added to the free-uranyl system.

After initial conjugate gradient energy minimization, we performed 10 ns production run in NVT (constant number of atoms, volume, and temperature) ensemble at 300 K. A Nosé–Hoover thermostat with a relaxation time of 0.2 ps was used for temperature control. Five independent simulations were performed with different initial configurations, and these simulations share substantially similar final configurations of uranyl adsorption. A representative adsorption process in the free-uranyl system is chosen to be studied in detail.

Umbrella sampling algorithm⁴⁵ was exploited to construct the potential of mean force (PMF) for uranyl desorption. We first used steered molecular dynamics (SMD) to pull the steadily adsorbed uranyl away from the surface, generating a series of conformations. Fifteen systems were then constructed with different positions of uranyl, i.e., z = 1.6, 1.9, 2.2, 2.8, 3.0, 3.2, 3.4, 3.6, 5.6, 7.6, 9.6, 11.6, 13.6, 15.6, 17.6 Å. For each system, 20 ns NVT simulation at 300 K was performed and the last 15 ns trajectory was collected for analysis. During the simulations, the uranyl uranium atom was restrained to the desired position by a harmonic potential of 1000 kJ mol⁻¹ nm⁻². And the weighted histogram analysis method (WHAM)⁴⁶ was used to extract PMF profile.

A time step of 1 fs was used in all simulations. A typical 12 Å cutoff distance was used in the calculation of short-range electrostatic interaction as well as van der Waals interaction. Long-range electrostatic interactions were computed using the particle mesh Ewald (PME) method (a Fourier grid spacing of 1.2 Å was adopted, and a cubic B-spline interpolation was used). All simulations were carried out with Gromacs (v4.6.5) program.⁴⁷ The potential parameters for Mg(OH)₂, sodium and chloride ions were taken from the CLAYFF force field.⁴⁸ The flexible version of the simple point charge (SPC) water model was used.^49,50 The recently improved version of Guilbaud and Wipff model was employed for uranyl.⁵¹ And the carbonate parameters of Greathouse and Cygan were adopted.^24,31 All of these parameters were listed in Table S1 of the ESI.†

Results and discussion

The configurations before and after the adsorption of uranyl in the free-uranyl system are shown in Fig. 1a and b. Due to the electrostatic attraction between uranyl cation and surface OH groups, the uranyl cation quickly approached onto Mg(OH)₂ (001) surface. The approaching of uranyl caused the redistribution of surface OH groups: the positions of two OH groups underwent fluctuation. At t ≈ 700 ps, the uranyl inserted to the surface defect caused by the fluctuation of these two OH groups which then served as the uranyl adsorption site. As illustrated in Fig. 1b, the adsorbed uranyl inserted into Mg(OH)₂ (001) surface with the [O=U=O]²⁺ axis tilted. And the surrounding OH groups coordinated to the adsorbed uranyl. Similar phenomena can be found in the other four independent simulations (Fig. S1 of the ESI†). Such redistribution is also observed in the uranyl–carbonate-pair system (Fig. S2 of the ESI†). Taken together, the redistribution of surface OH groups serves as the key step of uranyl adsorption. And the movement of the OH group in the free-uranyl system was then studied in detail.

Fig. 1 (a) Initial configuration and (b) the configuration after uranyl adsorption. (c) Time evolution of the position of the fluctuating OH group. (d) Representative snapshots for the surface OH group underwent fluctuation (highlighted in purple) at t = 30, 150, 700 ps (top, middle, bottom). Water molecules are not represented.

The position of this OH group and the representative snapshots are shown in Fig. 1c and d. As mentioned above, there was another surface OH group underwent position fluctuation in the meantime (Fig. S3 of the ESI†). After the adsorption of uranyl (at t = 700 ps), such fluctuation diminished. The adsorbed uranyl can then form stable coordination with neighboring OH groups. Similar configuration of adsorbed uranyl can be found in other four independent simulations (Fig. S1†). It should be noted that the fluctuation of surface OH groups during ion adsorption is observed experimentally in other mineral adsorbents.^52,53 Moreover, surface defects are considered to dominate the uranyl adsorption in some mineral adsorbents. For example, Yang et al. invoked two silanol defects serving as the uranyl adsorption site on the kaolinite (001) surface.²⁵ Li et al. also found the defect sites essentially promote the adsorption stabilities of metal ions on kaolinite.²⁶

As mentioned above, the adsorption of uranyl on Mg(OH)₂ (001) surface is accompanied with the fluctuation of surface OH groups. Besides, adsorbed uranyl can form stable coordination with OH groups. Given their importance to uranyl adsorption, we further studied the underlying mechanism about the position fluctuation of the OH group. Firstly we studied the possible impacts of uranyl's interaction. The total interaction energy between the uranyl cation and the OH group substantially remained steady (Fig. S4 of the ESI†), exhibiting a limited relationship with hydroxyl dissociation. Interestingly, we found the position fluctuation can be attributed to the repulsion between the inserted uranyl oxygen and the OH group characterized by the interaction energy between the inserted uranyl oxygen and the OH group as well as the distance between the inserted oxygen and the hydroxyl oxygen (Fig. 2a and b). At t = 700 ps, the distance between two oxygen atoms sharply decreased to 2.7 Å and the interaction energy increased up to 214 kJ mol⁻¹. Such suddenly enhanced repulsion may cause the movement of OH group.

Fig. 2 Comparison of the fluctuation of OH group (red) with the interaction between the uranyl and this OH group (black): (a) the distance between the inserted oxygen of uranyl and hydroxyl oxygen and (b) interaction energy between the inserted oxygen and the OH group.

The distribution and orientation of adsorbed uranyl in the free-uranyl system was then investigated. The relative density profile of uranyl peaks at z = 1.6 Å (Fig. 3a), suggesting the stable adsorption of uranyl because of the coordination interaction with OH groups. On the other hand, the carbonate locates farther from the surface than the uranyl due to the electrostatic repulsion with surface OH groups, and its interaction with surface is very limited. It should be mentioned the stable adsorption of uranyl is also in line with the formation of alkaline earth uranyl complex on Mg(OH)₂ (001) surface found in our previous experiments.⁵⁴ Moreover, the uranyl is almost immobilized in the adsorption site, considering that the diffusion constant is around 1.22 × 10⁻¹⁰ cm² s⁻¹, 5 orders of magnitude slower than aqueous uranyl (0.865 × 10⁻⁵ cm² s⁻¹) (Fig. S5 of the ESI†). The orientation of adsorbed uranyl is characterized as the angle between [O=U=O]²⁺ axis and the surface normal, i.e., φ_OUO (Fig. 3b). The orientation of adsorbed uranyl has a narrow distribution centers at ∼47°. Similar phenomena about the tilting adsorption of uranyl can be found in other minerals. For example, uranyl can adsorb on montmorillonite (001) surface with a tilting angle of ∼45°.⁵⁵

Fig. 3 (a) Relative density profiles for uranyl (U atom, solid line), carbonate (C atom, dashed line) and water (O atom, dotted line). (b) The orientation distribution of adsorbed uranyl. Inset: the schematic illustrating how the angle is measured. (c) PMF profile for the desorption of uranyl from surface. (d) Snapshot about the coordination of adsorbed uranyl with OH groups (the OH groups underwent fluctuation are highlighted in purple).

The adsorption free energy has been calculated to evaluate the stability of adsorption. More specifically, the potential of mean force (PMF) profile corresponding to uranyl desorption in the free-uranyl system was constructed by umbrella sampling (Fig. 3c). The PMF profile is featured by four minima at z = 1.60, 2.72, 4.13, 7.02 Å, corresponding to the adsorbed uranyl and three intermediate states during the desorption (Fig. S6 of the ESI†). The coordination number of OH groups decreases from 4 (z = 1.60 Å) to 2 (z = 2.72 Å), 1 (z = 4.13 Å). At z = 7.02 Å, the interaction between uranyl cation and OH groups is mediated by water molecules. In the region of z > 15 Å, the interaction with Mg(OH)₂ (001) surface is negligible. The free energy of adsorption is then calculated as ΔG = G(z = 1.6 Å) − G(z > 15 Å). The estimation of adsorption free energy is −26.08 kJ mol⁻¹, in quantitative agreement with the experimental result based on temperature dependence of adsorption capacity (∼−28 kJ mol⁻¹).^35,38 In comparison, previous simulation studies indicated the free energy of uranyl adsorption on quartz (010) and montmorillonite (001) surfaces are −6.5 and −4.2 kJ mol⁻¹, respectively.^24,30 The estimated adsorption free energy supports the experimental observation that the adsorption affinity of Mg(OH)₂ to uranyl is better than many other mineral adsorbents as previously reported.^35,38 Again, such stable adsorption of uranyl is largely attributed to the coordination interaction with OH groups. There are up to four OH groups can form coordination to the adsorbed uranyl (Fig. 3d).

Mg(OH)₂ (001) surface is commonly recognized as a hydrophilic and electrostatically neutral surface. In our previous experiments, clean Mg(OH)₂ surface was found to carry a small amount of positive charges (∼4.7 mV). After the adsorption of monolayer uranyl, the surface of Mg(OH)₂ become more positively charged (∼26.0 mV).⁵⁴ To further illustrate the impact of adsorbed uranyl to the surface properties of Mg(OH)₂ and compared with the experimental observation, we calculated the electrostatic potentials of the Mg(OH)₂ (001) surface before and after the redistribution of surface OH groups caused by uranyl. As illustrated in Fig. 4, the redistribution results in a local area of negative charge which should facilitate the uranyl adsorption, while the other region become more positively charged. Moreover, the adsorption of positively charged uranyl further increases the positive charge distribution on Mg(OH)₂ surface. Similar phenomena can be found in the uranyl–carbonate-pair system: the approaching uranyl can also directly interact with surface and cause the redistribution of surface OH groups leading to the negative charge site (Fig. S7 of the ESI†). Besides, the carbonate' interaction with surface is also limited due to the electrostatic repulsion with surface OH groups.

Fig. 4 The electrostatic potentials for the Mg(OH)₂ (001) surface (a) before and (b) after the uranyl adsorption, calculated via the APBS software. Positively and negatively charged regions are colored in blue and red, respectively.

Besides the distribution of surface OH groups, the impact of uranyl adsorption to the orientation of surface OH groups was also evaluated. We investigated the dipole orientation of surface OH groups, φ_OH, i.e., the angle between the OH dipole and the surface normal. The orientation of OH group of brucite has been successfully studied by molecular dynamics simulations.^41,56 In this work we aim to investigate the impact of adsorbed uranyl on the OH orientation. According to the surface mesh, the OH groups with different separation from the adsorbed uranyl were chosen and their orientations were calculated separately (Fig. 5a). For OH groups far away from the adsorbed uranyl (d_xy > 7.9 Å), the orientation distribution peaks at ∼20°, i.e., surface OH groups largely point upward. For the OH groups within 7.1 Å from the adsorbed uranyl, there are some OH groups lay nearly parallel to the surface (with the peak around 75°) and point away from uranyl, while other OH groups adopt more restricted upward orientation (Fig. 5b). It is interesting to note that the OH orientation in the region of 7.1 Å < d_xy < 7.9 Å is modestly affected, and the orientation distribution is similar to the case of d_xy > 7.9 Å. In short, the adsorbed uranyl has a profound effect on the orientation of surface OH groups within 7.1 Å, and a modest effect on the OH groups with d_xy = 7.1–7.9 Å (see also Fig. S8† for schematic diagram). The additional adsorption of uranyl should avoid the overlapping of susceptible regions, and two surface distributions of uranyl were proposed accordingly (Fig. S8†). In the first distribution, modestly affected OH groups are not shared by the adjacent uranyl, while in the second distribution the uranyl share these OH groups. The estimated surface density of adsorbed uranyl is 0.47/0.58 nm⁻², respectively. Taken the OH surface density of 11.65 nm⁻², our finding suggests the surface region containing 25/20 OH sites accommodates only one uranyl. The simulation result is in quantitative agreement with the experimental calculation based on adsorption capacity (one out of every 19 OH sites was occupied).³⁵ In other words, only a small portion of possible adsorption sites can be occupied by uranyl. The distributions of dipole orientation of surface OH groups exhibit similar features in the uranyl–carbonate-pair system (Fig. S9 of the ESI†). It should be noted that previous experiments about other minerals (e.g., montmorillonite and kaolinite) also reported there were limited surface sites can be occupied by uranyl.¹³ Our results, for the first time, suggest a molecular mechanism about the limitation of mineral's monolayer adsorption capacity.

Fig. 5 (a) Distributions of dipole orientation of surface OH groups, φ_OH, with different in-plane separation (d_xy) from the adsorbed uranyl. (b) The scheme for the susceptible regions of the adsorbed uranyl (d_xy = 7.9 Å, solid circle; d_xy = 7.1 Å, dotted circle).

We also investigated the impact of ionic strength by studying the uranyl adsorption in the solution of 0.6 M NaCl (Fig. S10 in the ESI†). The uranyl cation can also insert into the surface with a tilting adsorption configuration (Fig. S10c and d†). Besides, the redistribution of surface OH groups still serves as the key step of uranyl adsorption. As reflected in the electrostatic potentials of the surface after the uranyl adsorption, the negative charge site for uranyl adsorption still emerges (Fig. S11 of the ESI†). In short, the ionic strength does not appear to change the adsorption behavior of uranyl on Mg(OH)₂ (001) surface. We also evaluated the impact of uranyl adsorption to the orientation of surface OH groups. It is interesting to note that the susceptible region increase to ∼8.7 Å (Fig. S12 of the ESI†): larger than the free-uranyl system (∼7.9 Å). This may decrease surface coverage of uranyl, leading to the lower adsorption ability of uranyl. Similarly, the uranyl adsorption equilibrium constant was found to decrease as the ionic strength of the aqueous solution increased in previous simulation study about uranyl adsorption on montmorillonite.²⁴ It has been widely proposed that the competitive adsorption of other cations will suppress the adsorption of uranyl in high ionic strength. Our results suggest the decreased adsorption capacity may be also related to the impact to surface OH groups.

To better illustrate the mechanism of uranyl adsorption as well as the impact on the surface structure of Mg(OH)₂ and discriminate from the generic cations, we analyzed the position and orientation of OH groups in the vicinity of sodium ion in the case of 0.6 M NaCl system. The chosen sodium ion is more than 1.5 nm from the adsorbed uranyl. The fluctuation of three nearest-neighbor OH groups is negligible during the adsorption of sodium ion (Fig. S13 in the ESI†). This further confirms that the considerable fluctuation of surface OH groups during uranyl adsorption is largely attributed to the electrostatic repulsion with negatively charged uranyl oxygen, rather than the attraction to the cation. Besides, sodium ion cannot cause negative charge site on the surface of Mg(OH)₂ (Fig. S11 in the ESI†). On the other hand, the adsorbed sodium ion can also affect the orientation of neighboring OH groups, while the susceptible region is considerably smaller: d_xy < 4.3 Å (Fig. S14 in the ESI†). Taken together, while the upper limit of cation adsorption capacity may share similar mechanism, the adsorption behavior of uranyl on Mg(OH)₂ (001) surface is unique.

Conclusions

In summary, we provided a detailed description of the dynamics, energetics and structure of uranyl adsorption on the Mg(OH)₂ (001) surface. The redistribution of surface OH groups serves as the key step of uranyl adsorption which creates a negative charge site to accommodate uranyl. The redistribution is caused by the repulsive interaction between the incoming uranyl oxygen and OH groups. The adsorption stability of uranyl is largely attributed to the coordination interaction of surface OH group to the uranyl. And the estimated adsorption free energy is in quantitative agreement with the experimental result. On the other hand, the adsorption of uranyl can affect the orientation of surrounding OH groups, which may hinder the additional uranyl adsorption to the adjacent region and limit the monolayer adsorption capacity. The estimated monolayer surface coverage is well consistent with the experiment observation. Given that the limited monolayer surface coverage was also reported in other mineral adsorbents, our finding should be helpful to the studies of these adsorbents. Taken together, our results reveal the mechanism about remarkable performance of Mg(OH)₂ to extract low concentration uranyl and explain the seemingly contradictive observations of relatively low capacity. The computational studies about the underlying mechanisms of adsorption properties may provide guidance to the further development of adsorbent to realize the complete removal of low concentration uranyl from field water samples.

Acknowledgements

Financial support was provided by the 100-talent project of Chinese Academy of Science, the National Basic Research Program of China (2013CB933704, 2010CB933501), the Knowledge Innovation Program of the Chinese Academy of Sciences (KJCX2-YW-N50), the Outstanding Youth Fund (21125730), the National Natural Science Foundation of China (21273240, 21477129), and the Special Program for Applied Research on Super Computation of the NSFC-Guangdong Joint Fund (the second phase).

References

1. C. R. Preetha, J. M. Gladis, T. P. Rao and G. Venkateswaran, Environ. Sci. Technol., 2006, 40, 3070–3074.
2. R. A. Zielinski, D. T. Chafin, E. R. Banta and B. J. Szabo, Environ. Geol., 1997, 32, 124–136.
3. A. P. Kryvoruchko, L. Y. Yurlova, I. D. Atamanenko and B. Y. Kornilovich, Desalination, 2004, 162, 229–236.
4. B. Gu, Y.-K. Ku and P. M. Jardine, Environ. Sci. Technol., 2004, 38, 3184–3188.
5. M. C. Duff, J. U. Coughlin and D. B. Hunter, Geochim. Cosmochim. Acta, 2002, 66, 3533–3547.
6. Y. Shen, X. Tan, L. Wang and W. Wu, Sep. Purif. Technol., 2011, 78, 298–302.
7. J. N. Malin, J. G. Holland, S. A. Saslow and F. M. Geiger, J. Phys. Chem. C, 2011, 115, 13353–13360.
8. S. Ma, L. Huang, L. Ma, Y. Shim, S. M. Islam, P. Wang, L. D. Zhao, S. Wang, G. Sun, X. Yang and M. G. Kanatzidis, J. Am. Chem. Soc., 2015, 137, 3670–3677.
9. M. J. Manos and M. G. Kanatzidis, J. Am. Chem. Soc., 2012, 134, 16441–16446.
10. L. Zhou, M. Bosscher, C. Zhang, S. Ozcubukcu, L. Zhang, W. Zhang, C. J. Li, J. Liu, M. P. Jensen, L. Lai and C. He, Nat. Chem., 2014, 6, 236–241.
11. S. Bachmaf and B. J. Merkel, Environ. Earth Sci., 2010, 63, 925–934.
12. M. Schindler, C. A. Legrand and M. F. Hochella, Geochim. Cosmochim. Acta, 2015, 153, 15–36.
13. B. Campos, J. Aguilar-Carrillo, M. Algarra, M. A. Gonçalves, E. Rodríguez-Castellón, J. C. G. Esteves da Silva and I. Bobos, Appl. Clay Sci., 2013, 85, 53–63.
14. H. Geckeis, J. Lutzenkirchen, R. Polly, T. Rabung and M. Schmidt, Chem. Rev., 2013, 113, 1016–1062.
15. M. O. Barnett, P. M. Jardine, S. C. Brooks and H. M. Selim, Soil Sci. Soc. Am. J., 2000, 64, 908.
16. J. Górka, R. T. Mayes, L. Baggetto, G. M. Veith and S. Dai, J. Mater. Chem. A, 2013, 1, 3016.
17. Y. Sun, S. Yang, Y. Chen, C. Ding, W. Cheng and X. Wang, Environ. Sci. Technol., 2015, 49, 4255–4262.
18. R. J. Reeder, M. Nugent, G. M. Lamble, C. D. Tait and D. E. Morris, Environ. Sci. Technol., 2000, 34, 638–644.
19. E. Veilly, J. Roques, M. C. Jodin-Caumon, B. Humbert, R. Drot and E. Simoni, J. Chem. Phys., 2008, 129, 244704.
20. O. F. Zaidan, J. A. Greathouse and R. T. Pabalan, Clays Clay Miner., 2003, 51, 372–381.
21. J. A. Greathouse and R. T. Cygan, Environ. Sci. Technol., 2006, 40, 3865–3871.
22. X. Y. Liu, L. H. Wang, Z. Zheng, M. L. Kang, C. Li and C. L. Liu, J. Hazard. Mater., 2013, 244–245, 21–28.
23. W. Yang and A. Zaoui, J. Hazard. Mater., 2013, 261, 224–234.
24. J. A. Greathouse and R. T. Cygan, Phys. Chem. Chem. Phys., 2005, 7, 3580–3586.
25. W. Yang and A. Zaoui, Appl. Clay Sci., 2013, 80–81, 98–106.
26. X. Li, H. Li and G. Yang, Sci. Rep., 2015, 5, 14377.
27. S. Kerisit and C. Liu, Environ. Sci. Technol., 2012, 46, 1632–1640.
28. S. Kerisit and C. Liu, Environ. Sci. Technol., 2014, 48, 3899–3907.
29. S. Doudou, D. J. Vaughan, F. R. Livens and N. A. Burton, Environ. Sci. Technol., 2012, 46, 7587–7594.
30. J. F. Boily and K. M. Rosso, Phys. Chem. Chem. Phys., 2011, 13, 7845–7851.
31. J. A. Greathouse, R. J. O'Brien, G. Bemis and R. T. Pabalan, J. Phys. Chem. B, 2002, 106, 1646–1655.
32. H. Perron, J. Roques, C. Domain, R. Drot, E. Simoni and H. Catalette, Inorg. Chem., 2008, 47, 10991–10997.
33. K. Sebbari, J. Roques, E. Simoni, C. Domain, H. Perron and H. Catalette, Surf. Sci., 2012, 606, 1135–1141.
34. H. M. Steele, K. Wright and I. H. Hillier, Geochim. Cosmochim. Acta, 2002, 66, 1305–1310.
35. Z. Chen, Z. Zhuang, Q. Cao, X. Pan, X. Guan and Z. Lin, ACS Appl. Mater. Interfaces, 2014, 6, 1301–1305.
36. C. Li, Z. Zhuang, F. Huang, Z. Wu, Y. Hong and Z. Lin, ACS Appl. Mater. Interfaces, 2013, 5, 9719–9725.
37. W. Liu, F. Huang, Y. Liao, J. Zhang, G. Ren, Z. Zhuang, J. Zhen, Z. Lin and C. Wang, Angew. Chem., Int. Ed. Engl., 2008, 47, 5619–5622.
38. Q. Cao, F. Huang, Z. Zhuang and Z. Lin, Nanoscale, 2012, 4, 2423–2430.
39. H. Yan, J. Bai, X. Chen, J. Wang, H. Zhang, Q. Liu, M. Zhang and L. Liu, RSC Adv., 2013, 3, 23278.
40. V. Y. Kazimirov, M. B. Smirnov, L. Bourgeois, L. Guerlou-Demourgues, L. Servant, A. M. Balagurov, I. Natkaniec, N. R. Khasanova and E. V. Antipov, Solid State Ionics, 2010, 181, 1764–1770.
41. X. Ou, J. Li and Z. Lin, J. Phys. Chem. C, 2014, 118, 29887–29895.
42. C. Moulin, I. Laszak, V. Moulin and C. Tondre, J. Appl. Spectrosc., 1998, 52, 528–535.
43. M. C. Duff and C. Amrhein, Soil Sci. Soc. Am. J., 1996, 60, 1393.
44. A. Kilincarslan and S. Akyil, J. Radioanal. Nucl. Chem., 2005, 264, 541–548.
45. G. M. Torrie and J. P. Valleau, J. Comput. Phys., 1977, 23, 187–199.
46. S. Kumar, J. M. Rosenberg, D. Bouzida, R. H. Swendsen and P. A. Kollman, J. Comput. Chem., 1992, 13, 1011–1021.
47. B. Hess, C. Kutzner, D. van der Spoel and E. Lindahl, J. Chem. Theory Comput., 2008, 4, 435–447.
48. R. T. Cygan, J. J. Liang and A. G. Kalinichev, J. Phys. Chem. B, 2004, 108, 1255–1266.
49. H. J. C. Berendsen, J. P. M. Postma, W. F. Gunsteren and J. Hermans, in Intermolecular Forces, ed. B. Pullman and D. Reidel, Amsterdam, 1981, vol. 14, ch. 21, pp. 331–342.
50. O. Teleman, B. Jönsson and S. Engström, Mol. Phys., 1987, 60, 193–203.
51. S. Kerisit and C. Liu, J. Phys. Chem. A, 2013, 117, 6421–6432.
52. M. P. Asta, J. Cama, M. Martinez and J. Gimenez, J. Hazard. Mater., 2009, 171, 965–972.
53. P.-Y. Wu, Y. Jia, Y.-P. Jiang, Q.-Y. Zhang, S.-S. Zhou, F. Fang and D.-Y. Peng, J. Environ. Chem. Eng., 2014, 2, 2312–2320.
54. Z. Zhuang, X. Ou, J. Li, Y. Zhou, Z. Zhang, S. Dong and Z. Lin, ACS Sustainable Chem. Eng., 2016, 4, 801–809.
55. J. A. Greathouse, H. R. Stellalevinsohn, M. A. Denecke, A. Bauer and R. T. Pabalan, Clays Clay Miner., 2005, 53, 278–286.
56. T. R. Zeitler, J. A. Greathouse, J. D. Gale and R. T. Cygan, J. Phys. Chem. C, 2014, 118, 7946–7953.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra00384b

---