Heavy Metals Nanofiltration Using Nanotube and Electric Field by Molecular Dynamics

Laboratory of Preparation and Computation of Nanomaterials (LPCN), Federal University of Pará, C. P. 479, 66075-110 Belém, PA, Brazil Pos-Graduate Program in Engineering of Natural Resources of the Amazon, ITEC, Federal University of Pará, C. P. 2626, 66.050540 Belém, PA, Brazil Universidade Federal do Sul e Sudeste do Pará, Campus de Marabá. FL 17, QD 04, LT Especial, 68505080 Marabá, PA, Brazil Pos-Graduation Program in Chemical Engineering, ITEC, Federal University of Pará, C. P. 479, 66075-900 Belém, PA, Brazil Science Faculty, ICEN, Federal University of Pará, C. P. 479, 66075-900 Belém, PA, Brazil


Introduction
The term heavy metals (HM) is used to classify a group of highly reactive molecules, toxic to the environment and human life [1]. Even though it is a consolidated term, it has been questioned for bringing together elements with different chemical properties and reactivity [2]. Among the most harmful and frequent elements found in the environment above the permitted limits are copper, mercury, chromium, lead, manganese, cadmium, nickel, zinc, and iron [3].
Many of the HMs are naturally present in the environment, associated with other minerals and in rocks, but the distribution of HM can be altered by forces of nature such as floods and soil erosion; however, anthropic activities are mainly responsible by increasing the levels of these pollutants in the soil [4,5], water [6], and atmosphere [7]. The urbanization and industrialization are the main causes of the increase of the concentration of HM [8] and reached a stage in which society and the scientific community are no longer able to control and predict its real impact [9,10]. The effluents generated by industries, mining, agriculture, hospitals, and laboratories must receive treatment on the places where they are generated before disposal in sewage networks to eliminate the risk of contamination of water and soil [11,12]. Contaminated effluent water can be absorbed by plants and animals, causing poisoning at all levels of the food chain [13]. The soil and sediments favor the adsorption and permanence of metals, while water is a transition environment and facilitates part of the chemical reactions, transport, and diffusion of metals [14][15][16].
Nanotube-based water purification devices have the potential to transform the desalination and demineralization field through their ability to remove salts and HM [17] without significantly affecting the fast flow of water molecules [18]. In the environment, the high reactivity of HM ions facilitates the formation of new compounds with organic molecules or metals and thus promote their persistence in soils and water [19]. There are new compounds with HM basis that have been used as food for many species, and because they are foreign to metabolism, they are not expelled out. When man feeds on contaminated vegetables and animals or consumes contaminated water by them, the HM accumulates in the body and will cause serious diseases through progressive bioaccumulation [20,21].
Molecular dynamics (MD) allows the nanoscale computational study of macrosystems, analyzing the interactions between atoms as a function of time [22,23], and has contributed to the development of new technologies for protection and conservation of the environment. Thus, through the integration of the equations of motion described by Newtonian mechanics [24], the movements of atoms, which occur at different time scales, are continuously registered considering the intermolecular and intramolecular forces that act on them in their trajectories [25].
The selective transport of HM ions in aqueous media has been analyzed in studies of MD simulations [26,27] and has potential use in effluent nanofiltration (NF). Simulation studies of the MD of transporting solid molecules in water using carbon nanotubes (CNT) and boron nitride nanotubes (BNNT) have contributed in several areas of knowledge, such as medicine and current processes desalination [28,29]. In recent years, research with BNNT has revealed the superiority of some properties over CNT, such as the extremely high resistance to torsion [30], the selectivity for ion separation [31,32], and the conductivity and thermal stability at high temperatures. Such advantages lead to the hypothesis that BNNT is an adequate substitute for CNTs, but studies on the biocompatibility for the use of BNNT have not been conclusive [33].
Previous simulations revealed that the adsorption capacity of electrically charged solid particles is improved by using the nanotube surface modification or by applying an external electric field (EF) [34][35][36]. EF increases the reactivity of HM inside the nanotube and increases the adsorption capacity of these ions to CNT and BNNT [37,38]. The application of electric cap in BNNT for NF of pollutants [39] and desalination [40] is also a reality. Studies with BNNT are more recent than studies with CNT and even with better performance on CNT; more research is needed to better evaluate the possible biocompatibility of this nanostructure [41,42].
In this work, we present a computational modeling study analyzing the MD of HM filtration in water, in two simulation phases: in CNTs and BNNT, both under the effect of constant EF, longitudinal, and external to the single-walled type nanotubes, where we have the system close to the real model in Figure 1(a). Also, we have our computational model in Figure 1(b). In this study, it was observed that the effect caused by the rotation of the metal molecules around the NT produces an evanescent field [43]. This work aimed to use CNT and single-walled BNNT to make an NC that removes HM ions from the water, through the action of a uniform EF. The MD simulation allowed the understanding of the interaction properties of HM for the: Pb 2+ , Hg 2+ , Fe 2+ , Zn 2+ , and Cd 2+ ions in water, and it was observed that both nanotubes are effective in the NC of these HM in water under the electric field 10 -8 au.

Literature Review
The contamination of living beings by HM is increasing and harmful to the metabolism of all species in the world [44]. The poorest countries have more pollution due to higher HM due to the lack of basic sanitation and the disposal of untreated effluents in the environment. [45].
Among the 35 known types of metals, 23 elements have an atomic radius between 63.546 Å and 200.590 Å, with a density greater than 4.0 g/cm 3 , chemical criteria to be classified as HM. In this category, in addition to chemical properties, these metals have a gloss, excellent heat conductivity, and high melting and boiling points [46] This group of inorganic elements, in any concentration, is highly toxic   Journal of Nanomaterials and other HMs need control to not exceed the minimum concentrations allowed in the environment and living organisms [44]. Part of HM are toxic to the human body metabolism such as lead, considered a major environmental threat, since it is found in the air, water, dust, and soil [47] and in several products used by man, without the population knowing the risks of its manipulation and the correct disposal [48]. There is still no safe level of exposure to lead, an extremely toxic heavy metal, and the biggest victims of poisoning are developing children and pregnant women [49]. Children poisoning is irreversible, chronic and cumulative, impairs cognitive impairment, and school and work performance. [50].
Some HM participates in human metabolism in minimal concentrations and is called trace elements. In this class are copper, cobalt, iron, nickel, magnesium, molybdenum, chromium, selenium, manganese, and zinc. These metals must not exceed the plasma concentration limit to avoid the risk of acute or chronic intoxication [51].
The daily requirement for copper ranges from 50 to 120 mg, with 80 mg being the ideal average for a 70 kg adult. Some tissues require more copper, such as the liver, brain, spleen, bone and skeletal muscle [52], the liver, and spleen being considered reserve organs. Minimal concentrations of copper are essential for the metabolic functions of copperdependent enzymes cuproenzymes, such as cytochrome C oxidase, cytosolic superoxide dismutase, lysyl oxidase, tyrosinase, ceruloplasmin, and dopamine β-hydroxylase [20]. These enzymes catalyze physiological reactions related to oxidative phosphorylation, inactivation of free radicals, collagen and elastin biosynthesis, melanin formation, blood clotting, iron metabolism, and catecholamine synthesis [53], zinc activates immune responses in children and adolescents, and vanadium contributes to regulating insulin activity in glucose metabolism; however, these metals must be within the established limits to be beneficial to health [54].
Excessive HM in the body causes endocrine interference, mental disorders, damage blood cells, and respiratory, hepatic, and renal diseases that compromise the vitality of all systems [55,56]. The progressive accumulation of HM results in neuromuscular slowness that progresses as a degenerative process that simulates states of dementia or the symptoms of Parkinson's and Alzheimer's disease [57,58]. In the long run, the accumulation of MP or its compounds destroys nucleic acids, causes mutations, behavioral disorders (autism, attention deficit and hyperactivity, aggressiveness), heart disease, kidney disease, and cancer and interferes with the reproductive process [51].
Keeping the environment balanced concerning the toxicity of HM, it is necessary to seek the chemical safety of the environment, achieved by carrying out processes to ensure the safety of human health and the environment [52]. The WHO International Chemical Safety Program (ICSP) [44] establishes scientific bases for the management of chemical products and strengthens safety standards for natural elements, as well as covering all situations of exposure, considering the natural occurrence in the environment until extraction or synthesis, industrial production, use, reuse, and disposal [59].
The use of single-walled CNT or multiple-walled to filter HM ions in aqueous solutions is efficient in the NF processes of other pollutants [60][61][62][63]. Among the sensors used to detect HM [35,36], some have high sensitivity to identify the kind and level of contamination, based on a threedimensional hybrid electrode system in CNT [64]. After HM or its compounds are detected, separation methods will be chosen, such as the use of chemical precipitation [65], reverse osmosis [31], filtration [45], adsorption on activated carbon [66], the use of adsorbents such as aluminosilicates [67,68], or processes of oxy-reduction [69].
CNT and single-walled BNNT are nanostructures used for the NC of HM ions in aqueous solutions in MD studies and have shown good results [30]; however, the performance of the BNNT has surpassed the CNTs in some properties. BNNTs have all the properties of NTC and also excellent chemical inertness and superior thermal stability [70]. Mixed with polymers, BNNTs can be used to reinforce synthetic fibers, exceeding the resistance of the same fiber by 35% when reinforced with CNT; however, many studies are still needed to better understand the stresses at the interface of polymers with these two kinds of nanotubes [71]. In medicine, the use of CNTs is well established as a biomaterial, whereas, for BNNT, biocompatibility has not yet been confirmed [33].
2.1. Nanofiltration. NF membranes used to recycle wastewater, require high flow and energy rates during the NF process [66]. Reverse osmosis is the ideal separation method for ion removal [67], but it does not remove monovalent ions. NF can distinguish molecules based on their size or valence [72] and is often used to remove bi-or trivalent ions from water because they represent part of "water solids." The main disadvantage of membrane NF lies in the blocking of pores (initiated with organic molecules that incorporate other molecules and prevent NF), a process called fouling [73,74]. This block reduces the flow of matter and reduces the life of the membrane, requiring mechanical or chemical cleaning [75,76]. The separation process combined with the addition of an adjustable EF to the NF membrane was first investigated and denoted as electron filtration (EFT) [77] and brings together two driving forces: pressure and EF used mainly for the separation of charged molecules or particles electrically [78]. EFT works by trapping contaminants loaded in the membrane pores with the help of EF. The flow of electroosmosis generated by the superimposed EF also contributes to the improved permeate flow [60]. The ionic currents through the membranes depend on the movement of the ions through the pores of the membrane, a process called permeation. The way water is transported in nanotubes can be predicted with the use of EF, which influences the electrical dipoles in water molecules giving them a direction [79].
In order to understand the process of simulating the NC of water inside the nanotubes, it is necessary to understand that the nanotube must have enough diameter to allow a greater flow of water molecules, which also facilitates the transport of the solid molecules of the simulation, without friction on the walls of the CNT [80]. The direction taken by the water molecules inside the CNTs is a phenomenon induced by the application of external and longitudinal EF to the NT that increases the speed of the flow internally (since it acts on a system that also has a charge) and drives the molecules of water in the direction of the positive pole with an electrical force acting on the HM ions due to the EF where the electrostatic attraction would normally direct them, due to the electronegativity of oxygen [81].
The complexity of the interactions of hydrogen bonds in water [82] and the impossibility of carrying out real experiments with water in high-intensity electric fields [83,84] are some of the reasons that require the advancement of simulation studies in the field of MD.
In MD simulations, the stoichiometry of hydrogen bonds and molecular orientations undergo changes due to increased temperatures or increased intensities of the electric fields [85], and these hydrogen bonds can stabilize molecules whose electrical dipoles are oriented perpendicularly to the direction of the EF [86]. Molecular systems present movements that occur at different time scales, considering both intermolecular and intramolecular forces. The effect caused by the rotation of the molecules around the nanotubes produces an evanescent field [87], such an effect occurs on the walls of the nanotube. EF does not propagate as an electromagnetic wave, but its activity is spatially concentrated on the walls of the nanotube, also functioning as a "force trap" or valve. The trajectories of ions in nanotubes can be quite complex, and the effect of applied EF can trap electrically charged ions [88].

Materials and Methods.
In order to perform the computer simulations for this study, the both single-walled armchair structures of the CNT and BNNT were first modeled, because this structure has a better capacity to conduct energy. The CNT was modeled with 1064 atoms, 91.052 Å in length and 10.081 Å in diameter; the BNNT received modeling with 1162 atoms, measured in length 91,985 Å and the diameter measured 12.691 Å. It is hard to build two nanotubes of different nature with the same diameter and length. We compared the results in a very general way of the NF of the two nanotubes due to the percentage difference of 20.6% in these diameters.
HM ions and water molecules were modeled in the Gaussian 09 software. Then, the conformational analysis was performed in the Hyperchem 7.5 software when the studied molecules go through different conformations to minimize the energy of each molecule. Interactive cycles in energy minimization were applied individually, for each molecular structure [89]. The convergence criterion of the energy gradient was 0.01 kcal/(mol Å).
We use the methodology of molecular mechanics together with the processes and parameterizations of the force fields and molecule connections. The system consists of 100 water molecules for each ion present, 80 simulations were performed, 40 molecules of water for NTC, and 40 molecules of water for BNNT. The intermolecular forces have been described in terms of potential energy, kinetic, and thermal capacity functions, as well as lengths and angles of hydrogen bonds and nonbinding interactions. After the energy minimization, the conformational analysis [90] of the molecules was carried out, where a systematic search was made for the values of the dihedral angles of all the rotatable connections [91], to explore the conformational space of the molecules and so find a lower energy arrangement.
Once more stable molecular structures were obtained, MD calculations were made, based on the MM+force field, calculations of the interaction energies, and Newton's equations for motion. The equations were used to predict the position and velocity of all atoms at each time interval. Thus, MD was carried out with the system at an approximate temperature of 300 K, varying slightly during the simulation but controlled by the NPV ensemble [91,92]. EF was applied longitudinally to the outer surface of the two kinds of nanotubes and with constant intensities in a vacuum, so that the nanotubes remain rigid during all the simulations, while the water molecules and ions relaxed.
The duration of each simulation was 100 ps, and they were carried out under the action of eight uniform electric fields between 10 -1 au and 10 -8 au (1 au = 5:14 × 109 V/cm). The study allows the calculation of some physical properties such as kinetic energy (EKIN), potential energy (EPOT), and total energy (ETOT). The HM ions were Hg 2+ , Fe 2+ , Zn 2+ , Cd 2+ , and Pb 2+ . The ions of each heavy metal were placed, individually with the water molecules, inside the nanotubes. Each set of molecules was subjected to sixteen simulation processes, eight in the NTC and eight in the BNNT.
The simulation was performed using the similar MD methodology proposed by Yang et al. [77] and Neto et al. [79]. The MD method calculated the trajectory of the molecules when exposed to the electric fields of EF.

Results and Discussion
The MD simulation of HM NF in nanotubes of different molecular conformations made it possible to evaluate the influence of external EF on CNT and BNNT as well as possible interactions between HM ions, the inner surface of nanotubes and water molecules. Figure 1 demonstrates the NF process under the action of an EF that results in the retention of HM ions inside the nanotube, and the passage and reflux of water molecules, in the experiment, with high kinetic energy. The flow of water molecules is different from the molecules that pass through the interior and those that return towards the EF due to internal collisions with other water molecules, HM ions, and the nanotube wall. The favorable direction of the flow of water molecules under the EF is towards the positive pole of the nanotube due to the electronegativity of the oxygen atoms; however, the EF also increases the intermolecular interactions of the water, and thus the NF tends to be attracted to positive poles also contributes to this effect.
EF has an electrical force effect on the transport of water molecules in CNT and BNNT, but the influence of the size, chemical structure, and electrostatic potential of the molecules that makes up the system generates internal charges when the EF is applied. Nanotube simulations for studying the transport of solid molecules in water report the importance of the internal electrical charge in the total energy of the simulated system. The effect of trapping heavy metal ions is due to the intensity of the EF applied in interaction with 4 Journal of Nanomaterials the intermolecular and intramolecular forces of the water molecules and HM ions. These forces are more intense in metal ions, good conductors of energy; in the water molecules closest to the nanotube walls, there is an evanescent effect, which temporarily retains these molecules next to the nanotubes [34,35]. The effect of EF on NF in CNT and BNNT concerning the kinetic energy of HM ions is recorded in Figure 2(a) for CNT and Figure 2(b) for BNNT. The action of the electric field in the nanotube and molecules in the system during the simulation time was able to trap ions and few water molecules, causing most of them to be expelled from inside the nanotube with high values of velocities and kinetic energy. In general, in both nanotubes, the increase in kinetic energy was proportional to the intensity of the EF, but in both nanotubes for molecules with lower weight and atomic density (iron and zinc), they showed greater variations in kinetic energy and in potential energy. However, there are small variations in the linear growth of the kinetic energy, as in the case of the zinc 2+ ion, due to the constant molecular rearrangement inside the nanotubes. Two molecules with higher weight and molecular density showed little variations in kinetic energy (lead and mercury). The kinetic energy of the metal ion (which has an electric charge) increased directly proportional to the EF, because it follows the classic electric force (electric force = electric charge × EF).
The in situ temperature of the system was evaluated because the electric fields influence the thermal capacity of the molecules even when the NPT ensemble is applied, as they are variations of little intensity and are generally controlled by the system. These changes occur in the initial picoseconds of the simulation time. The in situ temperature of the HM ion simulation is displaying in Figure 3(a) for CNT and Figure 3(b) for BNNT. The heating capacity is divided by the amount of substance, mass, or volume; therefore, this value does not depend on the size or extension of the molecules and does not remain constant. Other system variables such as temperature and pressure were kept constant in this experiment. The thermal capacity varies with the application of EF, even though this field is uniform, as it causes an increase in the kinetic energy of the molecules that increases the total energy of the system. Thus, the increase in EF is a decisive factor in the systems in situ temperature. In CNTs, cadmium, iron, and lead had the highest temperatures in situ, followed by lead; these ions are the most electronegative in BNNT, lead, cadmium, and zinc.
Evaluating the transport of water molecules in CNT and BNNT, under the influence of the electric field, a sequence of 16 simulations with the same boundary conditions, without containing the HM ions, was performed, but without the presence of the metal ions ( Figure 4). It seeks to know the percentage of water molecules filtered in the CNT and BNNT, to understand whether the HM molecules under the action of EF and the inner walls of the nanotubes influence the transport of confined water. We tried to evaluate the percentage of water molecules filtered under the influence of EF and van der Waals forces.
The dipole moment of water molecules allows the intensity of the EF to influence its flow in a direction [80,93]. The alignment of the water molecules induced by EF will cause the preexisting hydrogen bonds to become unstable and break. The balance between hydrogen bonds and van der Waals forces will give rise to dispersed molecular clusters, because EF decreases the lengths of the molecular bonds between oxygen and hydrogen, causing changes in the stability of water molecules, vibrational frequencies, and dissociation energy that influenced the percentage of molecules in both nanotubes. There was a better yield of NF in BNNT because even though both are hydrophobic, the chemical constitution of BNNT increases their degree of hydrophobicity, which  Journal of Nanomaterials contributed to the expulsion of more water than in CNT [90]. The influence of EF on water dynamics, generating flows inversely proportional to the increase in EF, is well established in these conditions, as well as that the propelling effect of water in the transport of molecules inside the nanotube is related to size, chemical structure, and electrostatic potential of the molecules to be filtered.
The NF of the HM molecules in Figure 5(a) for CNT and Figure 5(b) for BNNT to confirm the influence of EF and the kind of nanotube structures (BNNT and CNT) used to remove water from HM ions; however, the results showed that both BNNT and CNTs are effective in retaining HM ions. The difference between them is made by the greater hydrophobicity and thermal conductivity of the BNNT. The EF effect on HM at 10 -8 au is enough for filtration. In addition to implying lower energy consumption in systems that aim to separate persisting HM in wastewater after conventional NF, both nanotubes perform this separation; however, in the evaluated parameters, there was a better efficiency of NF of BNNT about CNTs, but the difference is not a cause of replacement of CNT by BNNT because both are effective. The nanotube option should also take into account the costs of an NF system budget.
The trapping of HM ions in modified NC membranes composed of CNTs was the subject of previous studies and concluded that the same metal ions examined were almost 100% trapped in NBBT [62]; in this study, there was a better performance of the Cd 2+ ion [5]. The CNTs, due to its properties that keep it neutral and temporarily charged with the use of NF, showed better trapping in the Pb 2+ ion, in the highest concentration examined. The metals caused a minimal decline in temperature flow and thermal capacity due to the incrustation of the membrane in the order of gravity.
The number of filtered water molecules was lower for CNT than BNNT due to the intermolecular forces present [94]. Both nanotubes have hydrophobic potential; however, BNNT has adsorption properties superior to CNT due to its greater transfer of electrons on the wall. The adsorption energy of HM in BNNT is directly related with the hydrophobicity of its molecules [95].
The effect of BNNT nanofiltration on Pb 2+ ion is not good, because Pb is the heaviest and has little effect of the applied electrical force (electrical force = electric charge × EF). The removal effect is better when the electric field (EF) is 10 -8 au, because the evanescence effect is smaller. Also, the greater the EF, the greater the evanescent effect is. The   Journal of Nanomaterials evanescence effect produces the axial spin movement of all water molecules and HM inside the nanotube, thus decreasing the nanofiltration of water molecules and increase the kinetic energy to perform the axial turn inside the nanotube. The effect of EF on HM NF at 10 -8 au EF is the best for HM filtration in this EF range. Recent studies on confined water bring illuminating results on the influence of hydrogen bonds and the changes in this behavior in the presence of external electrical charges [96]. It is interesting to mention that hydrogen bonds account for a large part of the interactions of the water molecules. The hydrogen bonds in confined water, under EF, are different from the bonds established in free water because in the latter the bonds are stable and occur in a shorter distance between the atoms [89,91] displayed in Figure 6.
The confined water MD is characterized by flexible hydrogen bonds [97] are unstable bonds (they quickly rupture and redox). The effects of the NT surface and water confinement requires that the MD simulations be carried out on NT with a diameter between 10 Å to 12.2 Å to facilitate the diffusion of water and friction on the internal surface of the NT [98]. Confined water needs minimal space to alternate the two geometric shapes of its molecular arrangements in closed and open chains [99].
Pure water is a poor conductor of electricity, but its free ions are attracted or attract other molecules. The EF, depending    Journal of Nanomaterials on the intensity, can cause electrolysis [100,101]. Previous studies have shown that applying metallic electrodes to water, even at low intensity, causes interference in the orientation of water molecules and in the positioning of their atoms, which can be attracted or repelled. Similar orientations can occur on the surface of minerals containing alternating positive and negative charges, under the effect of electric fields.
Due to the partial covalence of hydrogen bonds in water, HM ions are easily distributed among their agglomerates under the action of EF increasing the intermolecular forces [102,103]. Both the carbon nanotube and the boron nitride have achieved higher levels of water flow compared to carbon nanotubes and are therefore expected to provide a more efficient water purification device.

Conclusion
The behavior of ionic compounds solubilized in water was analyzed, interacting with CNTs and BNNT under the action of EF. The trapping of HM ions in the NF process was effective in both nanotubes, but the intermolecular forces that act in the system allowed a different filtration flow, through filtered ions. The highest percentage of water molecules sequentially filtered in the NF of the ions was for CNT Pb 2+ , Cd 2+ , Hg 2+ , Zn 2+ , and Fe 2+ . In the case of BNNT, they were Cd 2+ , Fe 2+ , Hg 2+ , Zn 2+ , and Pb 2+ . In this simulation, BNNT demonstrated thermal and chemical stability in EF applications superior to that of CNT. Differences were found in the thermodynamic properties of CNT and BNNT, such as kinetic and potential energy, in addition to the increase in the thermal capacity of the system, where BNNT showed greater stability in all parameters than CNT. The evanescent effect was observed close to the inner walls of CNT and BNNT, resulting from the interactions of the intensity of the EF and the internal electrical charges of the molecules. The evanescence effect produces the axial spin movement of all water molecules and HM inside the nanotube, thus decreasing the nanofiltration of water molecules and increase the kinetic energy to perform the axial turn inside the nanotube. This effect temporarily retains water molecules close to the nanotubes. Ion capture by CNT is related to the effect of an evanescent attractive potential. Also, the greater the EF, the greater the evanescent effect is. Thus, probably, the nanofiltration would possibly get a better value to small EF. The results of this study can contribute as a theoretical basis for the development of an NF membrane system for the removal of HM ions from contaminated waters.

Data Availability
The data used to support this study are included within the article.

Conflicts of Interest
The authors declare that there is no conflict of interest regarding the publication of this paper. Table 1: Numerical values of the CNT kinetic energy (Kcal/mol) (Figure 2(a)). Table 2: Numerical values of the BNNT kinetic energy (Kcal/mol) (Figure 2(b)). Table 3: CNT in situ temperature (K) table referring to numerical values of the Figure 3(a). Table 4: BNNT in situ temperature (K) numerical values of the Figure 3(b). Table 5: Expulsion values of the NTC of the Figure 5(a). Table 6: Expulsion values of the BNNT to the Figure 5(b). Table 7: Percentage of water molecules expelled without adding ions ( Figure 4). (Supplementary Materials)