Molecular Dynamics Simulation of Natural Organic Matter–TiO 2 Nanoparticle Interaction in Aqueous Environment: Effects of Ca 2+ and Na + Ions

Abstract

This work aimed to explore the dynamic properties and mechanisms between natural organic matter and engineered nanomaterials (NOM–ENMs) interactions in aqueous solutions with the existence of monovalent and divalent cations. Due to the limitation of experimental method characterizing the interactions at the atomic level, the molecular dynamics simulation method was used in the study. Comparative interaction processes between NOM and TiO₂ nanoparticles in the aqueous solutions with or without Ca²⁺ and Na⁺ ions were simulated. Results showed that the Temple-Northeastern-Birmingham (TNB) encircled the TiO₂ nanoparticles and formed a horseshoe-style binding by two hydrogen bonds in the absence of cations. However, in the presence of cations, strong attraction between the mediating divalent cations and the negatively charged COO⁻ groups induced the O_COO− atoms to move away from the surface, which resulted in a hinge-style binding. The less TNB fragment encircling the nTiO₂ might decrease the steric and electrostatic repulsion between TiO₂ nanoparticles coated by humic acid. That is, the Ca²⁺ might destabilize the TiO₂ nanoparticles in the presence of NOM. The mediating effect of Na⁺ cation on NOM–ENMs interaction is insignificant, whereas the adsorbed Na⁺ on the surface of the nTiO₂ might change the surface charge and potentially affect the environmental behavior of the nTiO₂. This study facilitates to understand the fate and potential toxicity of ENMs in the aquatic environments.

Introduction

Interaction between natural organic matter (NOM) and engineered nanomaterials (ENMs) plays an important role in the fate and transformation of ENMs in the natural aquatic environment (Pettibone et al., 2008; Ottofuelling et al., 2011; Liu et al., 2013; Philippe and Schaumann, 2014; Huangfu et al., 2015; Dong et al., 2016; Li et al., 2016), the fouling process of nano-sized membrane in water supply treatment (Li and Elimelech, 2004; Ahn et al., 2008b; Kim et al., 2009), and the analysis of nanomaterial toxicity for the aquatic organism (Edgington et al., 2010; Mileyevabiebesheimer et al., 2010; Wang et al., 2011; Yang et al., 2013, 2014). Especially for addressing the behavior of nanoparticles in aqueous environment, plenty of studies have been carried out (Chen and Elimelech, 2006; Zhang et al., 2008; Domingos et al., 2009; French et al., 2009; Keller et al., 2010; Bian et al., 2011; Lanphere et al., 2014; Zhu et al., 2014). Many studies have ascertained that the commercially available nanoparticles, such as TiO₂, ZnO, SiO₂, are able to quickly aggregate as micrometer-size particles in aqueous solutions (Keller et al., 2010). This phenomenon may be beneficial for mitigating the risk that nanoparticles threaten human health. However, the existence of NOM could stabilize these nanoparticles for several hours in the lake water and the wastewater samples (Keller et al., 2010), so the study of the interactions of ENMs with NOM can provide insights into understanding their fate and toxic effect.

In natural waters, the NOMs coexist with ions. Studies have shown that humic substance can stabilize colloidal particles through steric and electrostatic repulsion in the presence of monovalent cations (Thio et al., 2011). However, some of divalent cations, such as Ca²⁺, Cu²⁺, and Pb²⁺, could diminish the stabilization effect of NOM (Ottofuelling et al., 2011). The destabilization effect of the Ca²⁺ ions on the stability of TiO₂ nanoparticles was presumably caused by the electrical double layer compression and ion bridge effect between nanomaterials coated by humic acid (Zhang et al., 2009; Liu et al., 2013). In addition, it was also attributed to the complexation between Ca²⁺ and the humic substance. However, the microscopic mechanism of the destabilization is still unclear. Thus, a molecular thermodynamic view of NOM–ENMs interactions in aqueous solutions with the presence of monovalent and divalent is needed.

Since the interaction process between NOM and nanomaterials in the presence of cations conducts very fast and involves many atomic-level details, the current experimental observation and measurement methods are limited to adequately address the microscopic interactions and the dynamic processes at short time intervals (e.g., femtosecond). Molecular dynamic (MD) simulations provide a possible way that explores and understands the microstructure and the dynamics at the atomic level. MD has been applied to investigating the structural, dynamic, and energetic characterization of metal ions-NOM complexation and NOM aggregation in aqueous solutions (Xu et al., 2006; Kalinichev and Kirkpatrick, 2007; Ahn et al., 2008a; Iskrenova-Tchoukova et al., 2010; Kalinichev et al., 2011). The results indicate that the type of cations affects the complexation style and the form of NOM aggregation. In the presence of Na⁺ and Mg²⁺ ions (Ahn et al., 2008a), the NOM molecules aggregate spontaneously through hydrogen bonding between their functional groups, and resulting in the formation of uniformly shaped NOM clusters. However, the Ca²⁺ ions can strongly bind with two different NOM molecules by co-complexing the carboxylate groups (Kalinichev et al., 2011), which forms linear or branched structures. The aggregation of the NOM molecules could equilibrate much faster with more stable structures compared with other ions, such as Mg²⁺ or Na⁺, which was proved by the previous experimental observations (Brigante et al., 2007; Nebbioso and Piccolo, 2009). The MD method has proved as an excellent alternative to study the interaction process involving NOM molecules and cations. However, to the best of our knowledge, a molecular thermodynamics view of the interaction between the NOM and nanoparticle, as well as the effect of cations, are not reported yet.

To reveal the interaction mechanism and the effect of cations on the interaction process between NOM and TiO₂ nanoparticles at the atomic level, we attempted to simulate the interaction process between NOM and TiO₂ nanoparticles in aqueous solutions with the presence of Ca²⁺ and Na⁺ by MD simulation. By comparing the system without cations in terms of the NOM structural and functional properties and adsorption configuration, the interaction mechanism and mediation of cations during the interaction of NOM–ENMs were addressed in a nano-analysis point of view.

Simulation Details

Model system description

Two model systems (i.e., with or without ions), which contained the TiO₂ nanoparticles (131 TiO₂ units), NOM, and a large number of water molecules (9,957), were used to study NOM–nTiO₂ interaction in aqueous solutions. The TiO₂ nanoparticles were generated by removing atoms from a TiO₂ crystal that were constructed using the lattice parameters for rutile (Diebold, 2003; Koparde and Cummings, 2005). A simplified neutral surface was used to present the system under pH_pzc condition. The excess surface atoms (Ti and O ions) were removed to maintain a zero surface charge in the model system. Both the “Steelink” and “Temple-Northeastern-Birmingham” (TNB) models can be used to describe the atomic structures of humic acid monomers of NOM. Since the Suwannee River NOM (SRNOM) was often used as a typical representative of NOM in experiments (Roger et al., 2010), the TNB model, in which the C/O/H ratio was closer to the SRNOM (Xu et al., 2006), was selected to describe the atomic structures of NOM in the simulations. Moreover, the results of the theoretical studies using the TNB model in the large-scale MD simulations have demonstrated a good agreement with the available experimental observations (Kalinichev and Kirkpatrick, 2007; Iskrenova-Tchoukova et al., 2010). As shown in Fig. 1, the TNB model (753 Da) had three carboxylic groups, three carbonyl groups, two phenolic groups, two amine groups, and four other R–OH alcohol groups. For the pH_pzc conditions (pH_pzc ∼6.5) (Romero et al., 2013), it is reasonable to assume that the three carboxylic groups of NOM fragment are completely deprotonated (pK_a values between 4 and 5), whereas the hydroxyl groups and amine groups are often protonated (pK_a values of ∼9) (Kalinichev, 2013). The TNB molecule has a net charge of −3 e, which was compensated by adding a Ca²⁺ ion and a Na⁺ ion. As reported in a previous study, Ca²⁺ is the predominate cation in groundwater and river water samples, where its concentration is in the range of 4–10 mM (Keller et al., 2010). In this study, the concentrations of Ca²⁺ and Na⁺ ions were both 4.27 mM, and both were in the reasonable range. Figure 2 shows the initial conformations of the NOM–nTiO₂in the absence or presence of cations. The water molecules were not displayed in the snapshot.

FIG. 1.

TNB model. First-COO⁻, second-COO⁻ and third-COO⁻ groups are illustrated respectively. TNB, Temple-Northeastern-Birmingham.

FIG. 2.

Natural organic matter and TiO₂ nanoparticles model systems: (a) without cations; and (b) with cations.

The system was hydrated by water molecules in a cubic MD simulation box with the periodic boundary conditions along the x, y, and z directions with a dimension of 73 Å. The TiO₂ nanoparticles and TNB were located at the center of the model system. The nearest distance from the groups of TNB to the surface atom of the TiO₂ nanoparticles was 1 nm at the initial state.

Force field and simulation methods

The Matsui and Akaogi force field is suitable for representing various titanium dioxide crystals (Matsui and Akaogi, 1991; Bandura and Kubicki, 2003; Naicker et al., 2005). Water adsorption on TiO₂ nanoparticles has been well studied using the Matsui and Akaogi force field (Koparde and Cummings, 2007). In this study, the Matsui and Akaogi force field was thus applied to describe the TiO₂ nanoparticle. The Van der Waals interactions were expressed by the Buckingham function. The site–site interaction could be expressed as below: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} U ( { { \rm r } _ { ij } } ) = { A_ { ij } } \exp \bigg( - { \frac { { r_ { ij } } } { { \rho _ { ij } } } } \bigg) - { \frac { { C_ { ij } } } { r_ { ij } ^6 } } \tag { 1 } \end{align*} \end{document}

where U(r_ij) was the interaction energy between atom i and j. The A_ij, ρ_ij and C_ij parameters are listed in Table 1. The interactions between the Ti atoms and the water oxygen atoms (Ti–O_W) were modeled using the ab initio-derived interaction parameters reported by Bandura and Kuibicki (2003) in Table 1. The charges of Ti and O ions were +2.196 e and −1.098 e, respectively.

Table 1.

Interaction Parameters for Matsui and Akaogi Force Field (Matsui and Akaogi, 1991)

Interaction	A_ij, kcal/mol	ρ_ij, Å	C_ij, kcal/(mol · Å⁶)
Ti–Ti	717647.4	0.154	121.0676
Ti–O	391049.1	0.194	290.3317
O–O	271716.3	0.234	696.8883
Ti–OH₂O^a	28593.02	0.265	148.000

Interaction between the water oxygen and the Ti atom was approximated by the Buckingham function.

Předota et al. (2004) has verified that the nondissociating, rigid, nonpolarizable model is computationally efficient and yields the realistic structural results both for bulk solutions and at solid/solution interfaces down to subatomic length scales. The rigid and nonpolarizable simple point charge-extended (SPC/E) model, where partial charge was assigned to the O_W (−0.8476 e) and two H_W atoms (+0.4238 e), was used to describe the water molecules in this study. This force field applied the Lennard-Jones (LJ) potential to the oxygen site (Berendsen et al., 1987). The Van der Waals interactions between the oxygen atoms in rutile TiO₂ nanoparticles and water molecules were treated as the O–O LJ interactions of the SPC/E water model (Předota et al., 2004). The expression of the potential was written as: \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} \begin{align*} U ( { r_ { ij } } ) = { \varepsilon _ { ij } } \left[ { { { \bigg( { \frac { { \sigma _ { ij } } } { { r_ { ij } } } } \bigg) } ^ { 12 } } - 2 { { \bigg( { \frac { { \sigma _ { ij } } } { { r_ { ij } } } } \bigg) } ^6 } } \right] \tag { 2 } \end{align*} \end{document}

where ɛ_ij was the depth of the potential well and σ_ij was the finite distance at which the interparticle potential was zero. The ɛ_ij and σ_ij parameters are listed in Table 2. The intramolecular interactions of atoms in the NOM molecules and dissolved ions in our simulations are described by the consistent valency force field (Kitson and Hagler, 1988).

Table 2.

Simple Point Charge-Extended Potential Parameters (Berendsen et al., 1987)

Interaction	ɛ_ij, kcal/mol	ρ_ij, Å
O_W–O_W	0.15539	3.5532
O_W–H	0.000
H–H	0.000

SPC/E, simple point charge-extended.

After minimizing the energy of the model system, MD simulations were conducted by using the large-scale atomic/molecular massively parallel simulator (LAMMPS) program package at a constant temperature of 300 K with an integration time step of 1 fs in the constant-temperature, constant-volume (NVT) ensemble using a Nose–Hoover thermostat (Plimpton, 1995). The cutoff distance of R_c = 12.0 Å was applied for all real-space interactions. The long-range electrostatic interactions were treated by the particle–particle particle–mesh solver, where the short-range interactions were truncated at 12 Å. Both simulations were conducted for 5.5 ns.

Results and Discussion

NOM–ENMs interaction process

First, Fig. 3 shows the variations of temperature and the total energy of the system in the presence or absence of cations. The temperature exhibited fluctuations because the velocities were rescaled at only certain intervals. However, the average temperature remained constant during the whole simulation. The total energy was also constant, except a jump due to the adsorption of TNB on the nTiO₂.

FIG. 3.

Variations of system temperature and total energy during adsorption process in presence or absence of cations.

To interpret the dynamic interaction process between NOM and TiO₂ nanoparticle, the nearest neighbor (NN) distance was calculated. The NN distance was defined as the shortest distance between the atoms belonging to TNB model and TiO₂ nanoparticles, respectively. The variation of the NN distance between TNB and TiO₂ nanoparticles during the whole simulation period is shown in Fig. 4.

FIG. 4.

Variation of NN distance between TNB and TiO₂ nanoparticles. NN, nearest neighbor.

The initial distance between TNB and TiO₂ nanoparticles was 10.8 Å. In the presence of cations, the distance first increased and then decreased. At 1 ns, the distance reduced to 1.93 Å and remains stable during the rest of the simulation. The variation of the NN distance implied first that the TNB moved away from the TiO₂ nanoparticles, and then moved to the vicinity of the TiO₂ nanoparticle and interacted with it. In the absence of ions, the distance quickly decreased to 2.0 Å at 0.1 ns and also slightly fluctuated during the remaining simulation time. It showed that the TNB quickly approached the nanoparticles during the simulation. The results indicated that the interacting process was delayed by the cations. Furthermore, the short distance between TiO₂ nanoparticles and TNB could be addressed as direct interactions between them.

To understand the interaction mechanism between NOM and TiO₂ nanoparticles, the interaction conformations between the TNB molecules and the TiO₂ nanoparticles in aqueous solutions were analyzed. In the presence of the cations, the terminal H atom (H_phenol) in the benzene ring of TNB was attracted by the O_surf atom in the TiO₂ nanoparticles when the TNB molecule moved to the vicinity of the TiO₂ nanoparticles at 1 ns, forming an hydrogen bond interaction. Evidences from ultraviolet and infrared spectroscopy also indicate that phenol groups are responsible for the attachment of humic acid to the surface (Yang et al., 2009; Erhayem and Sohn, 2014). The hydrogen bond looked like a hinge existing between the TNB molecule and the TiO₂ nanoparticles. The other terminal group of TNB persistently rotated, but did not form any linking structure with bonds during the simulation. In the absence of cations, the TNB molecule moved quickly to the vicinity of the TiO₂ nanoparticles within 100 ps. The hydrogen bonds were formed between the terminal H (H_COH) atoms of the TNB molecule and the O_surf atoms of the TiO₂ nanoparticles at 0.2 ns. One of the terminal groups in the TNB molecules still oscillated to find the favor energy point. At 1.5 ns, one of terminal H atoms was also attracted to the O_surf atoms of the TiO₂ nanoparticles, forming another hydrogen bond. Therefore, the TNB molecule was adsorbed on the surface of TiO₂ nanoparticles with a horseshoe style through two hydrogen bonds until the end of the simulation run. The variation of bond lengths of these hydrogen bonds between TNB and TiO₂ nanoparticles is shown in Fig. 5. All the bond lengths were smaller than 3.4 Å, which indicates the formation of hydrogen bond (Song et al., 2010; Stachowicz and Korchowiec, 2013). Due to the continuous self-regulation of the TNB, the distance of H_COH−O_TiO2 was larger than 3.4 Å casually. The interaction conformation of TNB on the surface of the TiO₂ nanoparticles in aqueous solutions at 5 ns is presented in Fig. 6. Due to less sites interacted and a hinge style formed between TNB and nTiO₂ in the presence of cations, it can be inferred that less TNB fragment surrounded the nTiO₂ in comparison to the conformations without the cations, thus the steric repulsion between TiO₂ nanoparticles coated by humic acid will be decreased. That is, the cations may destabilize the TiO₂ nanoparticles in the presence of NOM.

FIG. 5.

Variation of bond lengths of hydrogen bonds between TNB and TiO₂ nanoparticles: (a) in presence cations; (b) absence of cations.

FIG. 6.

Interaction conformation of TNB on surface of the TiO₂ nanoparticles in aqueous solutions at 5 ns: (a) in the presence cations; (b) absence of cations.

As shown in Fig. 6, more COO⁻ groups encircled the TiO₂ nanoparticles in the absence of cations. To quantitatively describe the interaction between the COO⁻ group of TNB and TiO₂ nanoparticles, the NN distance was calculated. As shown in Fig. 7, all the distances between carboxyl groups and TiO₂ nanoparticles changed slightly after 2 ns. In the absence of cations, the NN distances were about 6, 3, and 5 Å for the first COO⁻, second COO⁻, and third COO⁻ groups, respectively, whereas the NN distances were 8, 13, and 5.5 Å for the first COO⁻, second COO⁻, and third COO⁻ groups in the presence of cations. The results indicated that all the carboxyl groups surrounded the TiO₂ nanoparticles in the absence of cations. However, in the presence of cations, the carboxyl groups were far from the surface of the TiO₂ nanoparticles. It indicated that in the presence of cations, the surface of the nTiO₂ coated by humic acid had less negative charge due to COO⁻ groups far from the surface of nTiO₂. Thus, the electrostatic repulsion between nTiO₂ reduced. The less electrostatic repulsion might be another reason for the destabilization of the TiO₂ nanoparticles coated by humic acid in the presence of cations.

FIG. 7.

Variation of NN distance between COO⁻ and TiO₂ nanoparticles: (a) first COO⁻, (b) second COO⁻, (c) third COO⁻.

In addition, the interaction energy between the nTiO₂ and TNB during the adsorption process was calculated. The interaction between the nTiO₂ was attractive when the interaction energy value was negative, and vice versa. As shown in Fig. 8, the interaction energy changed slightly before 2.5 ns and then dramatically decreased to −90 kcal/mol. The sudden change of the interaction energy arose from the adjustment of the interaction conformations. The interaction conformation in Fig. 8 indicates that the benzene ring near the formed hydrogen bond moved to cover above on the surface of the TiO₂ nanoparticle after 2.5 ns.

FIG. 8.

Interaction energy between the nTiO₂ and TNB during the adsorption process.

In general, the results suggested that the dynamic interaction process and the interaction conformation were mediated by the cations. Moreover, the mediated mechanisms for monovalent and divalent cations were investigated in the next section.

Metal cations mediating interaction between NOM–ENMs

The bulk properties of cations

In aqueous solutions, the Ca²⁺ and Na⁺ ions might interact with the water molecules and form a water shell. As shown in Fig. 9, the density profiles of the water molecules around the nanoparticles were calculated to illustrate the distribution of water layers. The peaks corresponding to water oxygen were observed at 2.35 and 2.45 Å for Na⁺ and Ca²⁺ ions, respectively. In addition, other peaks were observed at 5.05 and 4.85 Å. Similar to the water oxygen, the hydrogen profiles of Na⁺ and Ca²⁺ ions were observed at 3.05 and 5.45 Å, 3.05 and 5.35 Å, respectively. The results indicated that there were two hydration shells surrounding the Na⁺ and Ca²⁺ ions. These results were in an excellent agreement with the previous studies (Kalinichev and Kirkpatrick, 2007; Wu et al., 2012).

FIG. 9.

Water density profiles of cations.

Metal cations mediating interaction between NOM–ENMs

Variations of NN distance between the cations and TiO₂ nanoparticle or TNB molecule are presented in Fig. 10. The distance between the Na⁺ ion and TNB quickly reduced to 6 Å. It indicated that the Na⁺ ion moved quickly to the vicinity of TNB at the initial simulation due to electrostatic attraction for the terminal O atoms of TNB. The distance between Na⁺ ion and TiO₂ nanoparticles also reduced to 2.5 Å at about 1 ns. Due to the strong attraction between TNB and TiO₂ nanoparticles as described in Fig. 4, it suggested that the Na⁺ ion moved along with the TNB to the vicinity of the TiO₂ nanoparticle. Since more negative O atoms existed on the surface of TiO₂ nanoparticles, the Na⁺ ion tended to desorb from the TNB due to the stronger attraction from the TiO₂ nanoparticles and keep stable adsorption during the simulation time. The distance was kept at about 2.5 Å.

FIG. 10.

Variation of NN distances between cations and TiO₂ nanoparticle, TNB molecule.

Distance between the Ca²⁺ ion and TNB rapidly reduced to 2.30 Å within 10 ps and remained stable during the rest of simulation time. The distance between the Ca²⁺ ion and TiO₂ nanoparticles changed significantly at the initial stage, and it persisted at 5.2 Å after 2 ns. The variation of NN distance between Ca²⁺ ion and TNB indicated that the Ca²⁺ ion quickly moved to the vicinity of TNB and formed a stable configuration with the TNB. The distance between Ca²⁺ ion and TiO₂ nanoparticles indicated that there is no direct interaction between them during the whole simulation.

Adsorption configuration and the variation of the NN distance between the Na⁺ ion and O_surf atoms of TiO₂ nanoparticle are presented in Fig. 11. Water molecules within a distance of 6 Å from Na⁺ ion are also shown in Fig. 11. First, the inner-sphere and outer-sphere concepts were introduced to characterize the adsorption mode of ions (Wu et al., 2013). An ion in the outer-sphere mode is adsorbed at the surface through its solvation shell, whereas the inner-sphere adsorption mode is defined that the first H₂O on the nonhydroxyl surface substitute water molecules from the solvation shell of ion, or no water molecules exist between the nonhydroxyl surface and ion. No water molecules existed between the Na⁺ ion and the TiO₂ nanoparticles. It suggested that the Na⁺ ion was adsorbed on the surface of TiO₂ nanoparticles in an inner-sphere style. This was also proved by the distance between the Na⁺ ion and O_surf atoms. From the discussion of the bulk properties of ion, two water layers existed around the Na⁺ ion. The direct attraction between the Na⁺ ion and the TiO₂ nanoparticles suggested that the TiO₂ nanoparticle had broken through the water layers around the Na⁺ ion and TiO₂ nanoparticle. It implied that the absorbed Na⁺ might change the surface charge of the nTiO₂, and thus affect the following environmental behavior of the TiO₂ nanoparticle.

FIG. 11.

Adsorption configuration and variation of NN distance between the Na⁺ ions and O_surf atoms of TiO₂ nanoparticle.

Adsorption configuration and the variation of the NN distance between the Ca²⁺ ion and the O_−COO atoms of TNB are shown in Fig. 12, in which the water molecules within the distance of 6 Å from the Ca²⁺ ion were included. Due to the strong electrostatic attraction, the Ca²⁺ ion and two COO⁻ groups formed the tetradentate coordinated contact ion pairs. It showed that the Ca²⁺ ion interacted with the TNB with the style of the inner sphere (Iskrenova-Tchoukova et al., 2010). According to the above discussion of the bulk properties of ions, two water layers existed around the Ca²⁺ ion. It suggested that the water layers around the Ca²⁺ ion and TNB could be broken through by the Ca²⁺ ion because of the direct attraction between the Ca²⁺ ion and the TNB.

FIG. 12.

Adsorption configuration and variation of NN distance between the Ca²⁺ ion and O_−COO atoms of TNB molecule.

In general, the Na⁺ ion mainly mediated with the TiO₂ nanoparticles by changing the surface charge of the nTiO₂ to potentially influence the following environmental behavior of the nTiO₂, while the Ca²⁺ ion mediated with the TNB due to the strong attraction for the two COO⁻ groups, to play an important role in the NOM–TiO₂adsorption process.

Discussion on Mediating Mechanism

Based on the above results, the difference of Na⁺ and Ca²⁺ ions in mediating mechanisms during the NOM–TiO₂adsorption process was discussed. The adsorption configuration in the absence of ions showed that the carboxyl groups of TNB encircled the TiO₂ nanoparticles through a strong electrostatic attraction. The presence of Ca²⁺ decreased the adsorption affinity and changed the adsorption configuration between NOM and TiO₂ nanoparticles due to a strong attraction with carboxyl groups of NOM, whereas the effect of Na⁺ ion is insignificant in mediating the adsorption process.

There were three structurally distinct carboxylic groups in the TNB molecular model of NOM used in our simulations, which were negatively charged. Both the Na⁺ and Ca²⁺ ions could be associated with these carboxylic groups through electrostatic attraction. The Ca²⁺ ion was observed to associate with two carboxylic groups and form tetradentate coordinated contact ion pairs in terms of relatively stable coordination geometries. The Ca²⁺ ion was coordinated with the two oxygen atoms of the carboxylic group simultaneously and stayed approximately equidistant from both of them predominately near the bisector plane orthogonal to the plane of the carboxylic group. The water molecules in the first coordination shell of Ca²⁺ were partially replaced by the carboxylate oxygen. The results suggested that Ca²⁺ ion was more effective and stable than Na⁺ ion in mediating the interaction between TNB and TiO₂ nanoparticles. It suggested that the mediation strength was related to the charge of the cation.

Kalinichev et al. (2011) reported a similar view on the metal ion–NOM complexation. They suggested that the relative strength of the metal ion–NOM complexation could be interpreted in terms of simple electrostatic considerations, which was described by the charge/radius (z/R) ratio of the cation. Ca²⁺ is almost the same size as Na⁺ (R: Na⁺, 0.97 ± 0.06; Ca²⁺, 1.03 ± 0.05) (Marcus, 1983), but interacts more strongly with the carboxylate groups in NOM due to its greater z/R (Na⁺, \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$ \frac { 1 } { { 0.97\; \pm\; 0.06 } } $$ \end{document} ; Ca²⁺, \documentclass{aastex}\usepackage{amsbsy}\usepackage{amsfonts}\usepackage{amssymb}\usepackage{bm}\usepackage{mathrsfs}\usepackage{pifont}\usepackage{stmaryrd}\usepackage{textcomp}\usepackage{portland, xspace}\usepackage{amsmath, amsxtra}\usepackage{upgreek}\pagestyle{empty}\DeclareMathSizes{10}{9}{7}{6}\begin{document} $$ \frac { 2 } { { 1.03\; \pm\; 0.05 } } $$ \end{document} ) (Kalinichev and Kirkpatrick, 2007). In contrast, for the ions with the same charge, the larger ions showed a stronger association with NOM since water molecules were loosely bonded in their hydration shells. Wu et al. (2012) reported the different mediation behaviors for monovalent and divalent cations in the study of Asp adsorption on the TiO₂ plates. In the NaCl, KCl, and RbCl solutions, if the number of stable hydrogen bonds between the O_COO− atoms and the H_OH atoms is large enough to trap the Asp side chain on the surface, the mediating cation is not essential to the Asp/rutile binding and can leave freely. The divalent cations show a completely different mediation behavior. The strong attraction of the mediating divalent cations for the negatively charged Asp side chain causes the O_COO− atoms to move away from the surface.

Although no mediation interaction between Ca²⁺ and the TiO₂ nanoparticle was observed in our study, the mediation interaction between Na⁺ and the TiO₂ nanoparticle was obvious. The absorbed Na⁺ on the TiO₂ nanoparticle by an inner-sphere style might change the surface charge of the TiO₂ nanoparticle. Thomson et al. (1996) also pointed out that the surface charge of the mineral can be neutralized in the presence of sufficient Na⁺, resulting in much more positively charged surfaces as good adsorbents for the negatively charged adsorbates. Divalent cations, in particular Ca²⁺, showed the capacity for effectively neutralizing negative surface charge of the mineral surface. The mediating effect between Ca²⁺ and TiO₂ nanoparticle that was neglected in our study might attribute to the strong attraction with carboxyl groups of NOM and the less number of Ca²⁺ ions in the model system. Further research was expected on the mediation interaction between massive cations and TiO₂ nanoparticles.

Conclusion

The MD study of the dynamic properties and the mechanisms of NOM–ENMs interactions in the existence of Na⁺ and Ca²⁺ ions in aqueous solutions provided a new quantitative insight into the dynamics and structure of NOM with TiO₂ nanoparticles and the effects of metal cations on the interaction between NOM and TiO₂ nanoparticles. The simulation results showed that the TNB combined with the TiO₂ nanoparticles formed a horseshoe style in the absence of cations. When the cations were present, the strong attraction of the mediating Ca²⁺ ion to the negatively charged COO⁻ group led to the O_COO− atoms to move away from the surface, which resulted in a hinge-style binding. The mediation strength was related to the charge of cation. The Ca²⁺ ion was more effective and stable than the Na⁺ ion in mediating the interaction between TNB and TiO₂ nanoparticles. The Ca²⁺ ion interacted with the TNB with the style of an inner sphere. The strong electrostatic attraction between Ca²⁺ ion and two COO⁻ groups formed tetradentate coordinated contact ion pairs. The Na⁺ adsorbed on the TiO₂ nanoparticle by an inner-sphere style might change the surface charge of the TiO₂ nanoparticle, thus, affecting its following environmental behavior. Our results agreed with the previous experimental findings and could substantially explain the effect of Ca²⁺ on the stability of aqueous nanoparticles in the presence of NOM. Also, this mechanism was able to address the complexing action of Ca²⁺ with NOM with respect to the fouling of the nanofiltration membranes in water purification (Chen et al., 2006; Zhang et al., 2009; Tian et al., 2013).

Footnotes

Acknowledgments

This work was financially supported by the Funds from State Key Laboratory of Urban Water Resource and Environment (Harbin Institute of Technology) (No. 2017TS03), the Creative Research Groups of China (51121062), and the National Natural Science Foundation of China (51278147).

Author Disclosure Statement

No competing financial interests exist.

References

Ahn

, Kalinichev

, and Clark

(2008a). Effects of dissolved Ca²⁺, Mg²⁺, and Na⁺ ions on the supramolecular aggregation of natural organic matter in aqueous solutions. AGU Fall Meeting Abstracts p. 2154.

Ahn

W.Y.

, Kalinichev

A.G.

, and Clark

M.M.

(2008b). Effects of background cations on the fouling of polyethersulfone membranes by natural organic matter: Experimental and molecular modeling study. J. Membr. Sci., 309, 128.

Bandura

A.V.

, and Kubicki

J.D.

(2003). Derivation of force field parameters for TiO₂−H₂O systems from ab initio calculations. J. Phys. Chem. B, 107, 11072.

Berendsen

H.J.C.

, Grigera

J.R.

, and Straatsma

T.P.

(1987). The missing term in effective pair potentials. J. Phys. Chem., 91, 6269.

Bian

S.-W.

, Mudunkotuwa

I.A.

, Rupasinghe

, and Grassian

V.H.

(2011). Aggregation and dissolution of 4 nm ZnO nanoparticles in aqueous environments: Influence of pH, ionic strength, size, and adsorption of humic acid. Langmuir, 27, 6059.

Brigante

, Zanini

, and Avena

(2007). On the dissolution kinetics of humic acid particles: Effects of pH, temperature and Ca²⁺ concentration. Colloids Surf. A Physicochem. Eng. Asp., 294, 64.

Chen

K.L.

, and Elimelech

(2006). Aggregation and deposition kinetics of fullerene (C60) nanoparticles. Langmuir, 22, 10994.

Chen

K.L.

, Mylon

S.E.

, and Elimelech

(2006). Aggregation kinetics of alginate-coated hematite nanoparticles in monovalent and divalent electrolytes. Environ. Sci. Technol., 40, 1516.

Diebold

(2003). The surface science of titanium dioxide. Surf. Sci. Rep., 48, 53.

10.

Domingos

R.F.

, Tufenkji

, and Wilkinson

K.J.

(2009). Aggregation of titanium dioxide nanoparticles: Role of a fulvic acid. Environ. Sci. Technol., 43, 1282.

11.

Dong

Y.N.

, Li

, Huang

, Wang

, and Li

(2016). Coagulation and dissolution of zinc oxide nanoparticles in the presence of humic acid under different pH values. Environ. Eng. Sci., 33, 347.

12.

Edgington

A.J.

, Roberts

A.P.

, Taylor

L.M.

, Alloy

M.M.

, Reppert

, Rao

A.M.

, Mao

, and Klaine

S.J.

(2010). The influence of natural organic matter on the toxicity of multiwalled carbon nanotubes. Environ. Toxicol. Chem., 29, 2511.

13.

Erhayem

, and Sohn

(2014). Stability studies for titanium dioxide nanoparticles upon adsorption of Suwannee river humic and fulvic acids and natural organic matter. Sci. Total Environ., 468–469, 249.

14.

French

R.A.

, Jacobson

A.R.

, Kim

, Isley

S.L.

, Penn

R.L.

, and Baveye

P.C.

(2009). Influence of ionic strength, pH, and cation valence on aggregation kinetics of titanium dioxide nanoparticles. Environ. Sci. Technol., 43, 1354.

15.

Huangfu

, Wang

, Liu

, Lu

, Zhang

, Cheng

, Jiang

, and Ma

(2015). Effects of humic acid and surfactants on the aggregation kinetics of manganese dioxide colloids. Front. Environ. Sci. Eng., 9, 105.

16.

Iskrenova-Tchoukova

, Kalinichev

A.G.

, and Kirkpatrick

R.J.

(2010). Metal cation complexation with natural organic matter in aqueous solutions: Molecular dynamics simulations and potentials of mean force. Langmuir, 26, 15909.

17.

Kalinichev

(2013). Molecular models of natural organic matter and its colloidal aggregation in aqueous solutions: Challenges and opportunities for computer simulations. Pure Appl. Chem., 85, 149.

18.

Kalinichev

A.G.

, Iskrenova-Tchoukova

, Ahn

W.-Y.

, Clark

M.M.

, and Kirkpatrick

R.J.

(2011). Effects of Ca²⁺ on supramolecular aggregation of natural organic matter in aqueous solutions: A comparison of molecular modeling approaches. Geoderma, 169, 27.

19.

Kalinichev

, and Kirkpatrick

R.J.

(2007). Molecular dynamics simulation of cationic complexation with natural organic matter. Eur. J. Soil Sci., 58, 909.

20.

Keller

A.A.

, Wang

, Zhou

, Lenihan

H.S.

, Cherr

, Cardinale

B.J.

, Miller

, and Ji

(2010). Stability and aggregation of metal oxide nanoparticles in natural aqueous matrices. Environ. Sci. Technol., 44, 1962.

21.

Kim

, Shan

, Davies

S.H.R.

, Baumann

M.J.

, Masten

S.J.

, and Tarabara

V.V.

(2009). Interactions of aqueous NOM with nanoscale TiO₂: Implications for ceramic membrane filtration-ozonation hybrid process. Environ. Sci. Technol., 43, 5488.

22.

Kitson

D.H.

, and Hagler

A.T.

(1988). Catalysis of a rotational transition in a peptide by crystal forces. Biochemistry, 27, 7176.

23.

Koparde

V.N.

, and Cummings

P.T.

(2005). Molecular dynamics simulation of titanium dioxide nanoparticle sintering. J. Phys. Chem. B, 109, 24280.

24.

Koparde

V.N.

, and Cummings

P.T.

(2007). Molecular dynamics study of water adsorption on TiO₂ nanoparticles. J. Phys. Chem. C, 111, 6920.

25.

Lanphere

J.D.

, Rogers

, Luth

, Bolster

C.H.

, and Walker

S.L.

(2014). Stability and transport of graphene oxide nanoparticles in groundwater and surface water. Environ. Eng. Sci., 31, 350.

26.

, and Elimelech

(2004). Organic fouling and chemical cleaning of nanofiltration membranes: Measurements and mechanisms. Environ. Sci. Technol., 38, 4683.

27.

, Ma

, Wallis

L.K.

, Etterson

M.A.

, Riley

, Hoff

D.J.

, and Diamond

S.A.

(2016). Impact of natural organic matter on particle behavior and phototoxicity of titanium dioxide nanoparticles. Sci. Total Environ., 542, 324.

28.

Liu

, Sun

, Borthwick

A.G.L.

, and Ni

(2013). Comparison on aggregation and sedimentation of titanium dioxide, titanate nanotubes and titanate nanotubes-TiO₂: Influence of pH, ionic strength and natural organic matter. Colloids Surf. Physicochem. Eng. Asp., 434, 319.

29.

Marcus

(1983). Ionic radii in aqueous solutions. J. Sol. Chem., 12, 271.

30.

Matsui

, and Akaogi

(1991). Molecular dynamics simulation of the structural and physical properties of the four polymorphs of TiO₂. Mol. Simul., 6, 239.

31.

Mileyevabiebesheimer

O.N.

, Zaky

, and Gruden

C.L.

(2010). Assessing the impact of titanium dioxide and zinc oxide nanoparticles on bacteria using a fluorescent-based cell membrane integrity assay. Environ. Eng. Sci., 27, 329.

32.

Naicker

P.K.

, Cummings

P.T.

, Zhang

, and Banfield

J.F.

(2005). Characterization of titanium dioxide nanoparticles using molecular dynamics simulations. J. Phys. Chem. B, 109, 15243.

33.

Nebbioso

, and Piccolo

(2009). Molecular rigidity and diffusivity of Al³⁺ and Ca²⁺ humates as revealed by NMR spectroscopy. Environ. Sci. Technol., 43, 2417.

34.

Ottofuelling

, Von Der Kammer

, and Hofmann

(2011). Commercial titanium dioxide nanoparticles in both natural and synthetic water: Comprehensive multidimensional testing and prediction of aggregation behavior. Environ. Sci. Technol., 45, 10045.

35.

Pettibone

J.M.

, Cwiertny

D.M.

, Scherer

, and Grassian

V.H.

(2008). Adsorption of organic acids on TiO₂ nanoparticles: Effects of pH, nanoparticle size, and nanoparticle aggregation. Langmuir, 24, 6659.

36.

Philippe

, and Schaumann

G.E.

(2014). Interactions of dissolved organic matter with natural and engineered inorganic colloids: A review. Environ. Sci. Technol., 48, 8946.

37.

Plimpton

(1995). Fast parallel algorithms for short-range molecular dynamics. J. Comput. Phys., 117, 1.

38.

Předota

, Bandura

A.V.

, Cummings

P.T.

, Kubicki

J.D.

, Wesolowski

D.J.

, Chialvo

A.A.

, and Machesky

M.L.

(2004). Electric double layer at the rutile (110) surface. 1. Structure of surfaces and interfacial water from molecular dynamics by use of ab initio potentials. J. Phys. Chem. B, 108, 12049.

39.

Roger

G.M.

, Durand-Vidal

, Bernard

, Mériguet

, Altmann

, and Turq

(2010). Characterization of humic substances and polyacrylic acid: A high precision conductimetry study. Colloids Surf. A Physicochem. Eng. Asp., 356, 51.

40.

Romero

, Marco

, Giménez

, and Esplugas

(2013). Adsorption and photocatalytic decomposition of the β-blocker metoprolol in aqueous titanium dioxide suspensions: Kinetics, intermediates, and degradation pathways. Int. J. Photoenergy, 2013, 1.

41.

Song

D.-P.

, Chen

M.-J.

, Liang

Y.-C.

, Bai

Q.-S.

, Chen

J.-X.

, and Zheng

X.-F.

(2010). Adsorption of tripeptide RGD on rutile TiO₂ nanotopography surface in aqueous solution. Acta Biomater. 6, 684.

42.

Stachowicz

, and Korchowiec

(2013). Bond detectors for molecular dynamics simulations, part I: Hydrogen bonds. J. Comput. Chem., 34, 2261.

43.

Thio

B.J.R.

, Zhou

, and Keller

A.A.

(2011). Influence of natural organic matter on the aggregation and deposition of titanium dioxide nanoparticles. J. Hazard. Mater., 189, 556.

44.

Thomson

N.H.

, Kasas

, Hansma

H.G.

, and Hansma

P.K.

(1996). Reversible binding of DNA to mica for AFM imaging. Langmuir, 12, 5905.

45.

Tian

J.Y.

, Ernst

, Cui

, and Jekel

(2013). Effect of different cations on UF membrane fouling by NOM fractions. Chem. Eng. J., 223, 547.

46.

Wang

, Li

, Zhao

, and Xing

(2011). Toxicity and internalization of CuO nanoparticles to prokaryotic alga Microcystis aeruginosa as affected by dissolved organic matter. Environ. Sci. Technol., 45, 6032.

47.

, Chen

, Skelton

A.A.

, Cummings

P.T.

, and Zheng

(2013). Adsorption of arginine-glycine-aspartate tripeptide onto negatively charged rutile (110) mediated by cations: The effect of surface hydroxylation. ACS Appl. Mater. Interfaces, 5, 2567.

48.

, Skelton

A.A.

, Chen

, Vlček

, and Cummings

P.T.

(2012). Modeling the interaction between integrin-binding peptide (RGD) and rutile surface: The effect of cation mediation on Asp adsorption. Langmuir, 28, 2799.

49.

, Kalinichev

A.G.

, and James Kirkpatrick

(2006). 133Cs and 35Cl NMR spectroscopy and molecular dynamics modeling of Cs⁺ and Cl⁻ complexation with natural organic matter. Geochim. Cosmochim. Acta, 70, 4319.

50.

Yang

, Lin

, and Xing

(2009). Interactions of humic acid with nanosized inorganic oxides. Langmuir, 25, 3571.

51.

Yang

S.P.

, Bar-Ilan

, Peterson

R.E.

, Heideman

, Hamers

R.J.

, and Pedersen

J.A.

(2013). Influence of humic acid on titanium dioxide nanoparticle toxicity to developing zebrafish. Environ. Sci. Technol., 47, 4718.

52.

Yang

, Jiang

, Hsu-Kim

, Badireddy

A.R.

, Dykstra

, Wiesner

, Hinton

D.E.

, and Meyer

J.N.

(2014). Silver nanoparticle behavior, uptake, and toxicity in Caenorhabditis elegans: Effects of natural organic matter. Environ. Sci. Technol., 48, 3486.

53.

Zhang

, Chen

, Westerhoff

, and Crittenden

(2009). Impact of natural organic matter and divalent cations on the stability of aqueous nanoparticles. Water Res. 43, 4249.

54.

Zhang

, Chen

, Westerhoff

, Hristovski

, and Crittenden

J.C.

(2008). Stability of commercial metal oxide nanoparticles in water. Water Res. 42, 2204.

55.

Zhu

, Wang

, Keller

A.A.

, Wang

, and Li

(2014). The effect of humic acid on the aggregation of titanium dioxide nanoparticles under different pH and ionic strengths. Sci. Total Environ., 487, 375.