Shedding Light on Miniaturized Dialysis Using MXene 2D Materials: A Computational Chemistry Approach.

Materials science can pave the way toward developing novel devices at the service of human life. In recent years, computational materials engineering has been promising in predicting material performance prior to the experiments. Herein, this capability has been carefully employed to tackle severe problems associated with kidney diseases through proposing novel nanolayers to adsorb urea and accordingly causing the wearable artificial kidney (WAK) to be viable. The two-dimensional metal carbide and nitride (MXene) nanosheets can leverage the performance of various devices since they are highly tunable along with fascinating surface chemistry properties. In this study, molecular dynamics (MD) simulations were exploited to investigate the interactions between urea and different MXene nanosheets. To this end, detailed analyses were performed that clarify the suitability of these nanostructures in urea adsorption. The atomistic simulations were carried out on Mn2C, Cd2C, Cu2C, Ti2C, W2C, Ta2C, and urea to determine the most appropriate urea-removing adsorbent. It was found that Cd2C was more efficient followed by Mn2C, which can be effectively exploited in WAK devices at the service of human health.

Introduction

Kidney disorder is one of the fatal diseases worldwide, with 1.7 million deaths per year.[1] Among kidney diseases, end-stage kidney disease (ESKD) is the most threatening by affecting 3 million patients annually. Although kidney transplantation is the most promising way to treat ESKD, due to the long waiting time for finding a compatible kidney and the high risk of kidney failure, researchers are about to use substitutions.[2] Hemodialysis (HD) and peritoneal dialysis (PD) are the most common treatments for kidney patients. HD, which contains pumping the blood out of the body and pumping it back after toxin filtration with an artificial kidney,[3] can neither remove the toxins properly from the blood nor is applicable to continuous removal of waste solute and excess water off the blood. Furthermore, the HD device’s large weight (in comparison to 175 g of the human kidney) causes difficulties in patients’ mobility and quality of social life.[3,4] PD is based on sending the fluid through a catheter to the belly to remove waste solute and accumulated toxins from the body.[2] However, the PD is an off-hospital method for kidney disease treatment. Its efficiency for blood purification is less than that of HD, and its period of treatment is limited to 30–40 min each time.[5,6] Besides, PD is not a permanent treatment and is replaced gradually by HD. Another potential treatment for kidney diseases is a wearable artificial kidney (WAK).[2] This method has been introduced in 1960s on the basis of the wearable artificial kidney that is regenerable due to the consumption of a certain amount of dialysate solution and provision of a closed loop as the human kidney does.[7] As PD, the WAK performs dialysis outside the hospital.[2] Despite both HD and PD, it provides an almost permanent and continuously clear bloodstream,[7] called continuous renal replacement (CRT).[8] An applicable WAK should be less than 2 kg to be worn in the body and also covers the requirement of a light power source with the ability to supply enough energy for an extended period without abruption as well as a sorbent system with the ability to regenerate and detoxify the bloodstream without the need for fresh dialysate to be injected in the systems. Other features such as appropriate additives, fluid removal system, and proper safety will promote the rate and efficiency of clearness of blood and consequently the quality-of-life factors, like social contribution and sleep pattern for the patients.[5,6,9] Recently, in times of the COVID-19 pandemic, HD and PD treatments are not that applicable due to hospitalization consideration and acute kidney injuries.[10]

One of the most challenging points in accordance with the dialysate is the ability to remove the waste solutes and toxic compounds. These compounds contain ions like phosphate and extra potassium, sodium, heavy metals complexes, and urea, resulting from the metabolism of proteins and other nitrogen-containing compounds. Based on the previous study, the most effective solution for ion removal from blood is activated carbon (AC) compounds that charcoal has shown to be efficient in this regard.[11] Moreover, due to time extension in the CRT method, phosphate detoxification improves. In addition to structural properties like the lack of surface charge at the physiological pH of the human body, no structural reactivity, and slow hydrolysis in water compared to other compounds, the volume of urea production is approximately 240 mmol/day, making it that problematic.[12] Conventional methods used two main protocols to remove urea: first, enzymatic hydrolysis, resulting in the outgassing of carbon dioxide and ammonia, and second electro-oxidation that suffers from slow kinetics and incomplete urea removal. Research shows that max phase and specifically metal carbide and nitride (MXene) sorbent can play a vital role in decreasing the concentration of urea in the artificial kidney by up to 99%, which helps regeneration of the kidney.[13] In this regard, the presence of a polar charged nitrogen, which intensifies the interaction of MXene and urea,[14] has been found to have an eminent effect on controlling and modifying urea concentration in blood.[15]

MXenes inheriting both metal and ceramic characteristics[11] show elevated electrical and thermal conductivity, higher ductility, lower friction coefficient, and improved thermal stability compared to common ceramics on the one hand and higher density, rigidity, Young’ modulus, thermal resistance; outstanding strain, creep performance, and vibration; and higher thermal resistance than those of regular metals on the other hand.[12,14−17] These led to their extraordinary performance in the fields of energy, electricals, optics, electronics, magnetism, molecular adsorption, catalysis, environment, sensing, as well as biomedical including drug delivery, phototherapy, diagnostic imaging, biosensing, tissue engineering, and cancer therapy.[18−20] MXenes can be represented as M2XT, M3X2T, M4X3T, M5X4T, M2M′X2T, and M2M′X3T,[21][21] where T shows surface termination. In this regard, fluoride ions are a possible termination due to etching, in addition to −OH, =O, and other oxygen-containing groups as a result of aqueous media.[22] The electrical, biological, and optical properties are highly dependent on surface functionalization.

The T intensifies the hydrophilic nature of MXene. Along with its high hydrophilic nature, abundant surface area makes MXene a potential option for urea adsorption. Different metal atoms in the two-dimensional (2D) MXene structure are an effective parameter in the adsorption of various compounds from aqueous media due to their capacity to bond with a certain molecule. Metal termination is necessary to reveal the supreme electronic and biological properties along with environmental applications, as electrochemical double-layer capacitors,[23] water treatment agents,[24] and kidney detoxifiers.[13] The 2D MXenes have strong interflake bonds and weak interlayer bonds. Due to the weak van der Waals (vdW) interlayer bonds, MXene can interact with external ions and molecules, resulting in the termination of the metal groups. For either thoroughly exfoliating the 2D MXene and making a MXene nanosheet or increasing the interlayer spacing,[25] intercalants are necessary. Intercalation and delamination provide accessibility to the active sites due to enhancement of the pillaring effect, within which the trapping and swelling of the penetrated molecules or ions occur. Thus, through an electrochemical process, the reaction with −OH, =O, and −F groups determinate them. The presence of intercalants results in increasing the interlayer spacing or, in other words, changing the c lattice parameter.[24,26]

In this study, an attempt has been made to thoroughly investigate the application of emerging 2D structures in urea removal. To this end, computational materials science tools have been employed to screen the novel MXene nanostructures in urea removal and blood purification. The molecular dynamics (MD) simulations were carried out on Cd-, Cu-, Mn-, Ti-, and W-based MXenes, and their viability in blood purification was accessed. In this study, we simulated different MXene structures to get a better response in this regard; as both the production and implementation of these 2D materials are expensive and sensitive, researching for different possible materials for getting a better response is necessary. However, doing laboratory research is a primary need to conduct animal study and clinical research; thus, molecular dynamics simulation can be exploited to get a deep insight into the intermolecular interactions to properly choose the most effective structures. This, in turn, prevents cumbersome and time-consuming laboratory studies. This work can enable the use of emerging nanostructures in newly developed wearable devices and accordingly improve the human’s life quality.

Results and Discussion

Stability

The structures of Ti2C, Cd2C, Mn2C, Cu2C, W2C, and Ta2C had suitable minimization energies after performing the optimization process. The total energies of all MXene structures after optimizations are reported in Figure , indicating the stability of all considered MXene nanosheets. To clarify the viability of an emerging nanostructure in the WAK devices, it is essential to investigate the interactions between the adsorbate and urea in detail. This enables us to determine the most applicable nanostructures. In this regard, applications of the MXene nanostructures in urea removal have been studied thoroughly in terms of the adsorption energy and other factors, by which the most effective MXene type can be detected. In the following, results obtained from each analysis will be discussed.

Figure 1

Total energy of the proposed MXene structures; the reported value for Ti2C is from the Material project site.

Energy and Gibbs Free Energy of Urea and MXene Nanosheet Reaction

The energy of intermolecular interaction can be considered to determine the adsorption energy of investigated atoms as a function of time. The more negative the energy (mmpsa), the stronger the binding of the considered MXenes and urea.[27] Accordingly, the MXene nanosheet leading to the most negative adsorption energy value is regarded as the most effective adsorbent in the WAK devices. Furthermore, formation of strong bonds between MXene with urea helps the urea to be absorbed from the body fluid and, as a result, detoxifies the body from the wastes. As Figure shows, the electrostatic interaction energies between urea and all of the investigated structures were about −2 mmpsa, which is neglectable. Moreover, Figure a indicates that van der Waals (vdW) energy interaction between urea and Cd2C was in the lowest level with an average of −214.104 mmpsa followed by that of Mn2C. Based on Figure b–e, the energy levels were higher for Mn2C, Cu2C, Ti2C, and W2C with approximate levels of −190.489, −187.9, −163.238, and −158.24 mmpsa as a function of time. The highest vdW energy level belonged to Ta2C with −41.330 mmpsa, as shown in Figure f. The observed difference can be referred to the differences between these structures’ vacant orbitals and their tendency to whether form van der Waals bonds with urea via complexing or stay in the aqueous media. The chemical complex of urea and Cd2C is more potent than other structures; thus, the energy level is significantly more appropriate than other structures. Actually, energy analysis is the most appropriate index to compare the performance of various adsorbents. As discussed hitherto, the Cd2C MXene is the most favorable material for urea removal in terms of the adsorption energy, which needs to be verified by other analyses.

Figure 2

Urea interaction electrostatic and van der Waals energy with (a) Cd2C, (b) Mn2C, (c) Cu2C, (d) Ti2C, (e) W2C, and (f) Ta2C and (g) Gibbs free energy for Cd2C, Mn2C, Cu2C, Ti2C, W2C, and Ta2C.

Gibbs free energy of the reaction between urea and the above-mentioned MXenes is also (Figure g) an indicator of the probability of the reaction. The lower the Gibbs free energy, the more probable the reaction. As it is shown, the least Gibbs free energy belongs to the Cd2C–urea reaction with a value of −42.32 kJ. Moreover, based on this diagram, the second value belongs to Mn2C with a sharp increase in the Gibbs free energy value to −25.81 kJ. Consequently, it can be predicted according to the energy diagrams that the highest value of Gibbs free energy belongs to Ta2C with a value of −5.02 kJ.

Analysis of Radius Distribution Function (RDF) of Urea and MXene Nanosheets

The radius distribution function (RDF) shows the density of a specific matter at a distance (r) around a molecule, computed through eq .Here, dnr shows the number of particles in a particular shell of thickness and ρ is the local density. The calculation is based on the RDF that helps us to define the coordination number and many other parameters. As the distance increases, the probability of bonding also decreases, and as a result, the coordination is less likely. Figure demonstrates the RDF for the considered MXenes. As the results show, Cd2C had a higher g(r) peak of 10.99 at 0.474 Å, and as a result, it had a brighter coordination shell, which illustrates the higher urea density in a shorter distance and, accordingly, a higher chance for bonding and adsorption was well provided. The g(r) peak intensity decreased for Mn2C, Cu2C, Ti2C, W2C, and Ta2C and appeared more distant.

Figure 3

RDF of urea with W2C, Ti2C, Mn2C, Cu2C, Ta2C, and Cd2C.

Root-Mean-Square Deviation (RMSD) of Urea and MXene Nanosheets

The root-mean-square deviation (RMSD) implies the flexibility of an atom to depart from a tagged structure or, in other words, the distance from a target that can be whether an organic molecule and protein or even a single atom. For the transformation of the hard-ball model, the RMSD can be defined as belowIn eq , X is the position of atom i and X is the position of atom j. Although the above-mentioned equation is a function of time and in a particular Δt(a), the RMDS may differ from this value in Δt(b).[11] The whole picture shows the strength and tendency of a bond to form or break. Furthermore, if Δt → 0, thenwhere a is a constant and is approximately equal to the lattice parameter of formation.[28]

Figure shows the average distance between a tagged urea molecule and a part of the MXene structure. According to what has been shown, the RMSD value for Ta2C fluctuated between 1.15 × 10–5 and 7.13417 Å, which is a greater value compared to other structures; this illustrates the weaker bonds between urea and Ta2C and more flexibility in the particle movements as a result of thermal fluctuations. On the other hand, W2C fluctuates between 2.19 × 10–5 and 7.0448 Å with an increasing slope. For Cu2C, Ti2C, and Mn2C, the RMSD values fluctuated in a more limited range with average values of 4.8559, 4.8379, and 4.6849 Å, respectively. Cd2C on the other hand fluctuated in the most restricted range of 4.96 × 10–5 to 5.53991 Å with an average of 4.430084 Å. The finding shows that Cd2C forms more resistant bonds with urea due to the limited fluctuation range, which proves that urea and Cd2C bonds are the firmest and thus are more applicable in the WAKs.[29]

Figure 4

RMSD values of urea with W2C, Ti2C, Mn2C, Cu2C, and Cd2C MXenes.

Root-Mean-Square Fluctuation (RMSF) of Urea and MXene Nanosheets

Root-mean-square fluctuation (RMSF) is the displacement in the position of a single atom or a molecular structure of a reference atom or structure. In GROMACS, it is computed with the Poisson–Boltzmann equation (eq ), which goes as follows, where M is the number of the frames, r(t) is atom i of the complex at time k, and i is the reference atom or structure.Figure shows the RMSF value for the considered MXene structures. The observation shows that the highest residue belongs to the Ta2C structure with an average of 0.05689 nm. Moreover, significant fluctuations are observed for Ta2C, which, all in all, implies loose bonds such as bend, turn, and coil that introduce high flexibility, which causes the whole complex to be unstable. W2C with 0.04856 nm is at the second position of instability and Ti2C with a 0.01666 nm peak in the residue is the third probable unstable complex in the presence of urea. Cu2C and Mn2C with less fluctuation and negligible peaks are more stable than the mentioned structure. However, Cd2C is formed with the least flexibility. This is due to the least fluctuations with an average of 0.01286 nm between the reviewed structures. In Cd2C, π–π interactions are more durable, thus making the structure most stable.

Figure 5

RMSF values of urea with (a) Ta2C, (b) W2C, (c) Ti2C, (d) Cu2C, (e) Mn2C, and (f) W2C.

Number of Hydrogen Bonds between Urea and MXene Nanosheets

Hydrogen bonds (H-bonds) are nonbinding interactions occurring through electrostatic interactions. In molecular dynamics simulations, the number of H-bonds is determined by the LINCS algorithm. Figure shows the number of intermolecular hydrogen bonds for the investigated structures. Since an average of 13.27 hydrogen bonds were found for Cd2C, it may interact with urea. Consequently, the most abundant values for hydrogen bonds of 8.18 and 2.19 belong to Cu2C and Mn2C. For the rest of the structures, this value is approximately zero. Since the bonding of D–H–A and their angles are of great importance in the bonding, no hydrogen bonding might be due to the lack of any bond between O and N or O/N as well as the probability of the unsuitable direction that makes urea and MXene bonding impossible. This results in weak C–H and π interactions.

Figure 6

Number of hydrogen bonds between urea and (a) Cd2C, (b) Cu2C, (c) Mn2C, (d) Ti2C, and (e) W2C, and (f) Ta2C.

Due to the high polarity of urea and its hydrophilic nature and the presence of hydrogen and oxygen atoms in the aqueous solution, aqueous media acts as a cross-linking agent for interacting MXene and urea. As the solvent-accessible surface area (SASA) of all of the studied structures is significant, the chance for cross-linking agent action is present. The formation of partially negative charges on the MXene surface due to MXene hydration may enhance the interactions. Based on the cross-linking bond capacity, the probability of bond formation varies in the studied structures.[30]

Density of Urea and MXene Nanosheets

In GROMACS, the density is based on the partial density from a reference point. In other words, the determination of the reactant density from a certain point of the reference structure is calculated by three-dimensional multi-Gaussian relations of atomic density. Based on this fact, as a reaction completes or the concentration of the reactant decreases, the density function decreases as well. Figure shows the density of urea around the mentioned MXenes. Figure a reveals that the reaction of urea and Mn2C showed slight fluctuation, which indicates that homogeneous distribution of urea could be found around the Mn2C structure with an average density of 18.93 kg/m3. However, Cd2C (Figure b) with an average density of 18.57 kg/m3 had better urea adsorption due to the uniformity of its structure compared to Mn2C. Based on the studied structures’ molecular dynamics simulation results, the average densities for Mn2C, Cd2C, Ti2C, Ta2C, Cu2C, and W2C are 18.93, 18.57, 18.47, 17.59, 16.34, and 16.34 kg/m3, respectively. Thus, urea adsorption is less likely compared to the other studied structures (Mn2C and Cd2C). The ups and downs in the Ta2C, W2C, and Cu2C diagrams (Figure d–f) show a heterogeneous distribution of urea around the named structures, which is not desirable. The findings proved that Mn2C and Cd2C owing to the homogeneous distribution of urea around their structures and high intensity values are the most promising structures to interact with urea for effective urea removal from organs.

Figure 7

Density fluctuations of urea around (a) Mn2C, (c) Cd2C, (f) Ti2C, (g) Ta2C, (h) W2C, and (i) Cu2C and fluctuation of urea in (b) Mn2C and (d, e) Cd2C.

Radius of Gyration of Urea and MXene Nanosheets

The radius of gyration is an indicator of the compactness of a structure and its change as a function of time.[12] In Figure , the radius of gyration is shown for different modeled MXene structures. Since the radius of gyration for Cu2C was 2.84861 Å, it can be stated that Cu2C was less compact in comparison to Cd2C, Mn2C, Ta2C, W2C, and Ti2C with R(g) values of 2.64254, 2.38891, 2.24154, 2.11727, and 1.94982 Å, respectively. So, the adsorption of urea by Ti2C may be the strongest due to the tight bonds, while the weakest adsorption was achieved for Cu2C for which loose bonds are anticipated. Moreover, a slight decrease in the bond distance in Cd2C from 2.939321 to 2.3972 Å demonstrates that the adsorption of urea and Cd2C increased with time, while other structures were stable since their radii of gyration were not fluctuating with time sharply.

Figure 8

Radii of gyration for Mn2C, Cu2C, Cd2C, Ti2C, Ta2C, and W2C.

Adsorption Capacity

Solvent-Accessible Surface Area (SASA) of Urea and MXene Nanosheets

The term solvent-accessible surface area (SASA) refers to the capacity of a material in an aqueous media to be surrounded by water molecules. The SASA is calculated through eq where i is a specific atom indictor and Δσ is its atomic solvation parameter and A shows its SASA. As the SASA increases, the active sites of the adsorbate increase as a function of higher capacity; moreover, the free energy transfer increases and adsorption of nonpolar molecules is enhanced, and this means that the detoxification of the body fluid will be affected as well. Figure demonstrates that the SASA value for Cd2C was about 97.7964 nm2 with an adsorption percentage of 60%. An improvement was observed in the SASA for other structures, especially for Ta2C, which reached 111.5022 nm2, while its adsorption percentage falls to 6.25%. The SASA values for W2C, Ti2C, Cu2C, and Mn2C were 109.0914, 108.7972, 103.6181, and 101.2825 nm2, respectively. The observed differences are due to the size of the molecules. As smaller molecules provide a higher surface area, the SASA is the highest for Ta2C with the Ta ion radius of 31 pm and the lowest for Mn2C with 97 pm radius. Although the SASA value reveals the ability of an adsorbent to be accessible to the adsorbate molecules, it is not an absolute index to compare the MXenes and it just verified the acceptable performance of all MXenes in providing enough accessible area to remove urea and detoxify the blood. However, comparing the adsorption percentage in Figure c, it is found that Cd2C shows the best performance.

Figure 9

(a) Solvent-accessible surface area (SASA) per time for W2C, Ti2C, Mn2C, Cu2C, Ta2C, and Cd2C; (b) adsorption of urea on MXene structures Mn2C, Ta2C, and Cd2C; and (c) adsorption percentages of W2C, Ti2C, Mn2C, Cu2C, Ta2C, and Cd2C.

Evaluation of Urea Adsorption Rate

In this section, the urea adsorption rate by MXenes was investigated. To do this, the amount of urea adsorbed by MXenes during the simulation is shown in Figure . The intensity of urea adsorption versus time is shown in Figure , so the rate of urea adsorption can be compared by investigating the slope of these graphs. According to the obtained results, Cd2C and Mn2C diagrams had the highest slope. This indicates a higher urea adsorption rate by these two structures. On the other hand, since the slope of the graphs decreased with time, it is inferred that the intensity of urea adsorption by MXenes decreased with time. The decrease in the urea adsorption rate may be due to decreasing urea concentration in solution with time as well as limited MXenes’s surface (Table ).

Figure 10

Urea adsorption rate of MXenes during the simulations.

Table 1

Cd2C–Urea Characteristic Behavior at 150 °C

analysis		0–10 ns	10–20 ns	20–30 ns
average Gibbs free energy (kJ/mol)		–24.404
average energy (kJ/mol)	electrostatic	1.295	1.591	2.685
	vdW	20.178	34.185	77.541
	total	21.473	35.776	81.226
maximum RDF		5.265	2.892	1.01
average RMSD (nm)		0.5156	1.4915	1.9296
average RMSF (nm)		0.0156	0.5161	0.8226
release %		15	80	100
hydrogen bond number		15	10	2
area (nm2)		114.214	131.984	145.255
average gyration (nm)		2.31	3.11	3.43

Refurnishing the Surface of MXene after Urea Adsorption

To verify the application of MXenes in urea removal and blood purification, the reusability of the MXene adsorbents must be studied. In the literature, some methods have been investigated to regenerate the MXenes used in the removal of various contaminants. Mechanochemical (MC) etching on Ti3C2T for methylene blue removal from water proved to be useful, which could be regenerated using Al powder as a regeneration agent.[31] Another research showed that MXene has a great absorption-desorption capacity while using 0.1 M NaOH as an effective regeneration agent. Moreover, the absorption capacity of MXene was more than 80% after six cycles.[15] Moreover, another method to refurnish the spent MXene surface is the thermal desorption method. In this regard, the active surface sites will be regenerated as the temperature increases. To investigate the effect of thermal desorption, a simulation has been conducted at 150 °C for occupied Cd2C to clarify the rate of urea desorption (Figure ). Moreover, the energy value, RDF, RMSD, RMSF, area, number of hydrogen bonds, and the radius of gyration (Table ) have been calculated to support the given information in Figure . As can be detected, the urea desorption cycle can be completed in 30 ns.

Figure 11

Urea desorption rate as a function of time from the Cd2C surface at 150 °C.

Table 2

Comparison of Simulation and Experimental Results of Urea Adsorption at Different Concentrations

		equilibrium concentration (mg/dL)
		50	150	250
adsorbed urea (mg/g of adsorbent)	Ti3C2Tx (reference experimental)	11	15	16
	Ti3C2Tx (simulation)	9	11	13

The average energy value for urea desorption shows that the reaction (urea release) is thermodynamically desirable (due to the negative value for urea desorption). Moreover, decreasing the number of particles around a tagged particle (RDF), the number of hydrogen bonds, and the compactness of the structure (radius of gyration) and increasing the flexibility of a structure (RMSD), the instability of the residue (RMSF), and solvent-accessible surface area (SASA) are all indicators of thermal desorption of urea from the Cd2C surface.

Validation

A comparison between experimental (reported in the literature) and simulated urea adsorption is shown in Figure a. The results are in accordance with the experimental reference data. In the report by Meng et al.,[13] urea adsorption on Ti3C2T was investigated. The volume of urea solution used in the experiment was equal to 6 mL, along with 0.155 g of MXene. The urea concentration in the simulation was determined according to the laboratory conditions and was in proportion to the volume of the simulation boxes (which is equal to 1000 nm3). Moreover, simulations have been performed according to the experimental conditions (37 °C and different laboratory concentrations). Validation of diagrams in Figure e from the experimental work is reported in Figure b and validation of diagrams in Figure from the experimental work is reported in Table . A good agreement of the experimental data with the data obtained from the simulation is shown in Figure b and Table . The amount of urea adsorbed by Ti3C2T in the simulation was lower than the experimental adsorption. This is due to the limited simulation time. On the other hand, urea adsorption by Cd2C has been investigated using molecular dynamics simulations. Figure b shows the concentration of urea in solution at the beginning of the simulation and after adsorption by Cd2C. The results show high urea adsorption by Cd2C and the effectiveness of this structure in eliminating urea.

Figure 12

(a) Comparison between experimental and simulated urea adsorption. (b) Comparison between simulation and experimental urea concentration in the solution at the beginning and after adsorption by MXenes.

Conclusions

In this study, through performing MD simulations on a set of emerging 2D nanostructures, the viability of employing MXenes in purifying the blood from urea was accessed. To this end, different MXene types including Mn2C, Cu2C, Cd2C, Ti2C, Ta2C, and W2C were considered for the urea removal process. Applicability of nanosheets in detoxifying the blood through removing urea was investigated from stability and adsorption capacity perspectives, which include different parameters like adsorption energy, solvent-accessible surface area, RDF, RMSD, RMSF, and radius of gyration. The finding attested to the fact that all considered MXenes have acceptable performance in urea removal and being exploited in the WAK devices. In this regard, Cd2C possessed the best performance with respect to adsorption and stability followed by Mn2C. It was clarified that Cd2C is a suitable option due to the lowest energy level of −214.04 mmpsa compared to other systems. Moreover, the least flexibility as well as homogeneous distribution of urea around the system, high probability of bonds formation, and reasonable compactness, which have been driven from RMSD, RMSF, number of H-bonds, density, gyration, and SASA of 0.86871 Å, 0.01286 nm, 13.27, 18.57 kg/m3, 2.64 Å, and 97.7964 nm2, respectively, proves that using Cd2C will benefit us from different aspects. This work concentrates that MD simulations can be effectively utilized to develop nanomaterials with direct applications in enhancing the human’s quality of life.

Computational Methods

Theory

Molecular dynamics simulation is a computational method to direct laboratory studies at a microscopic scale based on general physics rules and is widely used in different fields of science, from biology to materials science.[32] Two families of simulations are accessible in this regard, molecular dynamics (MD) simulation and Monte Carlo (MC). MD is more favorable due to the time dependency which gives results on rheological properties,[32] diffusion of particles,[33] turbulence,[34] transport coefficient, etc.[35] To determine the positions of atoms, it considers individual particle’s motion and interatomic interactions[33] on a scale of time. This results in an investigation of the numerous possible conditions of atoms binding, folding, and conformational changes.[36]

Classic MD simulation has been constructed based on the idea of no bond formation or breaking,[37] where Newton’s equation of motion based on the interaction potential for the periodic behavior of a one-dimensional inharmonic chain (Fermi–Pasta–Ulam) and three-dimensional hard-sphere model (Alder–Wainwright) (eq ) is applied.[36] Ab initio molecular dynamics (AIMD), density functional molecular dynamics (DFMD), and such[38] simulations have been developed since that classic MD does not consider the complex chemical formation as a matter of time; besides, it also ignores the effect of geometrical importance on developing different structure models.[38] Quantum mechanics/molecular mechanics (QM/MM) simulations, which are constructed on the computation of the wave equation of Schrödinger, contain local density approximation (LDA) of the Hartree–Fock and Kohn–Shom LDA theory[39] (eq ). This theory involves numerous numerical stability checks in each step on a scale as short as a few femtoseconds[38] using the perturbation theory.[36] Rules of motion are applied to define each atom’s position as a function of time in each step. Consequently, they are updated repeatedly based on the computation of interaction between phonons[38] for a long range of structures (eq ).[40]where M is the atomic mass, R is the degree of freedom, and E is the Ehrenfest potential.where H is the Hamiltonian operator of the system and the initial state of φ°.[40]where ω is an indicator of the unperturbed harmonic frequency, Δ is the frequency shift, and r is the frequency linewidth.where n is the phonon occupation number and δn is the fluctuation of n. However, the whole story is based on the approximation of motions and actions on an atomic scale. It is a powerful method due to position calculation as a function of time and careful control of atoms’ possible conditions based on the structural equilibrium and fluctuations.[41]

Simulation Details

In this study, GROMACS software was used for molecular dynamics simulations and data analyses.[36,42] To this end, a 32-core X5670 CPU and a 1080 Ti graphics card with a Ubuntu 18.04.1 operating system were used[43] and all of the molecular structures were modeled by Avogadro software. MXene adsorbent structures were designed via Avogadro software and the most stable state was reached using Gaussian software. The structure of the Ti2C nanosheet was downloaded from the Materials Project website with ID mp-10721, for which the energy stability of the Ti2C structure has been reported.

Validation

In this study, we validated our simulation results by comparing them with an experimental study by Meng et al.[13] For this purpose, three simulations for three MXene molecules (Ti3C2T, Mo2TiC2T, and Ti2CT) have been performed in a 10 × 10 × 10 nanocube for 300 ns with GROMACS 2020.1 in the all-atom optimized potential for liquid simulations (OPLS-AA) force field. MXene structures were downloaded from the Materials Project website and changed based on the reference. The changes in the structures were applied using Avogadro software. Gaussian software has been used for atomic charge calculation based on the electrostatic potential (ESP). For each MD simulation, a certain MXene nanosheet and ten urea molecules were simulated, and the mass ratio of the absorbed urea per studied MXene has been calculated.