Hybrid G/BN@2H-MoS2 Nanomaterial Composites: Structural, Electronic and Molecular Adsorption Properties

Hybrid structures often possess superior properties to those of their component materials. This arises from changes in the structural or physical properties of the new materials. Here, we investigate the structural, electronic, and gas-adsorption properties of hybrid structures made from graphene/hexagonal boron nitride and 2H-molybdenum disulfide (G/BN@MoS2) monolayers. We consider hybrid systems in which the G/BN patch is at the Mo plane (model I) and the S plane (model II). We find that the implanted hexagon of G or BN in MoS2 alters its electronic properties: G@MoS2 (I,II) are metallic, while BN@MoS2 (I) is an n-type conducting and BN@MoS2 (II) is semiconducting. We study the molecular adsorption of some diatomic gases (H2, OH, N2, NO, CO), triatomic gases (CO2, NO2, H2S, SO2), and polyatomic gases (COOH, CH4, and NH3) on our hybrid structures while considering multiple initial adsorption sites. Our results suggest that the hybrid systems may be suitable materials for some applications: G@MOS2 (I) for oxygen reduction reactions, BN@MoS2 (I,II) for NH3-based hydrogen production, and G@MoS2 (I) and BN@MoS2 (I,II) for filtration of No, Co, SO2, H2S, and NO2.

Transition metal dichalcogenides (TMDs, e.g., MoS 2 ) are 2D layered structures with many applications in electronics [1]. Due to an intrinsic band gap ranging from 0.4-3.1 eV [12,13], they have recently received much attention. Furthermore, TMDs monolayers can possess high mobility for charge carriers at room temperature, which makes them attractive materials for optoelectronics and energy-harvesting applications [14].
Molybdenum disulfide monolayers, TMD-(MoS 2 ), have two phases 2H-MoS 2 (two layers per hexagonal unit cell) or 1T-MoS 2 (one layer per trigonal unit cell). Several experimental techniques can change the physical and chemical properties of the MoS 2 structure: doping [15], surface functionalization by metal atoms [16], ion bombardment [17], and defect formation [18]. Density functional theory (DFT) is widely used to investigate the effect of substitutional doping with nonmetal, halogen, and transition atoms [19,20] on the electronic and magnetic properties of 2H-MoS 2 . The ferromagnetic behavior of Co- [21] and Fe- [22] doped MoS 2 monolayer were demonstrated. Furthermore, it is found that the substitutional doping of MoS 2 can enhance its electrochemical catalytic response [23].

Computational Methods
All calculations are performed using density functional theory (DFT) on the basis of the projector augmented wave method (quantum espresso package) [41]. First, the energies and wave functions are calculated within the generalized gradient approximation (the Perdew-Burke-Ernzerhof exchange-correlation functional) [42]. The cell-volume and ionic position relaxations of all structures are carried out until all the atomic forces on each ion are less than 10 −4 eV/Å. A vacuum region of~16 Å is used to avoid the interaction between the layers in the z-direction. We use norm-conserving pseudopotentials with a 50 Ry energy cutoff and a 9 × 9 × 1 k-point grid. The valence electron configurations 4s 2 4p 6 4d 5 5s 1 for Mo and 3s 2 3p 4 for S atoms are used to calculate their potentials. The Van der Waals correction is considered [43]. A 6 × 6 × 1 G/BN@MoS 2 supercell is created by embedding a patch of graphene/BN into MoS 2 monolayer. Löwdin charges are used to calculate the charge transfer between the monolayers and molecules. Spin-polarized calculations show that all the considered heterostructures in this study are nonmagnetic. The stability of the considered hybrid structures is estimated by calculating their formation energies (E form ) using the following equation: where E hyb and E MoS (n Mo E Mo + n S E S ) are the total energy of the hybrid sheet and the total energy of the constituent atoms of the sheet, respectively, with a total number of the atoms in the hybrid structure (n hyb ). The energy of the patch is E pat = n C E C for the G-patch and E pat = n B E B + n N E N for the h-BN-patch, where n x is the total number of the x atom (x = Mo, S, C, B, N). The adsorption energy (E ads ) of a molecule on a sheet is calculated by: where E sheet+molecule , E sheeat and E molecule are the total energies of the sheet with the adsorbed molecule, the sheet, and the isolated molecule, respectively.

Structural and Electronic Properties of Hybrid Monolayers
To investigate the effect of G and BN patches on the electronic properties of the MoS 2 sheet, we first calculate the structural parameters and electronic properties of the pristine MoS 2 sheet to establish a reference. The lattice constant of the optimized structure is 3.22 Å with an S-Mo bond length of 2.42 Å, which agrees well with the corresponding experimental values of 3.18 Å and 2.41 Å (2.41 Å) [44]. Now, we investigate the effect of G-patch on the structural properties of the MoS 2 sheet. Interface models with different edge terminations (Mo or S) have been considered [45]. The electronic properties are found to be strongly dependent on the termination, which can be correlated with the existence of polar C-Mo bonds or defects caused by the C-S bonds at the interface. The two relaxed G@2H-MoS 2 structures are shown in (Figure 1b,c). The first hybrid structure (Figure 1b), built by removing one Mo atom and six S atoms has the G-patch connected to Mo atoms. The other hybrid structure is created when the G-patch is formed in the S layer, after the removal of 3 S atoms, giving a total number of 111 atoms in the supercell (Figure 1c). The patch's boundary constitutes the main defect in the MoS 2 monolayer, which may change the physical and/or chemical properties of MoS 2 . Structural relaxation yields a local symmetric distortion in the Mo sites surrounding the carbon atoms.
where E sheet+molecule , E sheeat and E molecule are the total energies of the sheet with the molecule, the sheet, and the isolated molecule, respectively.

Structural and Electronic Properties of Hybrid Monolayers
To investigate the effect of G and BN patches on the electronic properties of sheet, we first calculate the structural parameters and electronic properties of th MoS2 sheet to establish a reference. The lattice constant of the optimized structu Å with an S-Mo bond length of 2.42 Å, which agrees well with the correspondin mental values of 3.18 Å and 2.41 Å (2.41 Å) [44]. Now, we investigate the effect of G-patch on the structural properties of sheet. Interface models with different edge terminations (Mo or S) have been co [45]. The electronic properties are found to be strongly dependent on the term which can be correlated with the existence of polar C-Mo bonds or defects caus C-S bonds at the interface. The two relaxed G@2H-MoS2 structures are shown i 1b,c). The first hybrid structure (Figure 1b), built by removing one Mo atom atoms has the G-patch connected to Mo atoms. The other hybrid structure is crea the G-patch is formed in the S layer, after the removal of 3 S atoms, giving a tota of 111 atoms in the supercell (Figure 1c). The patch's boundary constitutes the ma in the MoS2 monolayer, which may change the physical and/or chemical pro MoS2. Structural relaxation yields a local symmetric distortion in the Mo sites sur the carbon atoms.   The Mo-Mo distance decreases from 3.22 Å to 3.02 Å with no significant change in the Mo-S bonds (~2.41 Å) for hybrid and pristine cases. The C-C bond length in the hybrid structures is 1.42 Å (1.47 Å) for the first (second) model, which is close to the corresponding value in pristine graphene (1.42 Å). The average C-Mo and C-S lengths are 2.13 Å and 1.79 Å, respectively, which are smaller than the S-Mo of 2.41 Å due to the small size of the C atom compared to the S atom.
Turning to the first h-BN@MoS 2 hybrid structure (Figure 1d), we see that it is also slightly buckled with~0.35 Å, especially inside the BN patch. The average bond length for Mo-Mo decreases from 3.21 Å to 3.05 Å. The bond length of B-N in the structure is 1.45 Å, which is the pristine h-BN value [5]. We find the optimized bond lengths for N-Mo and B-Mo bonds are 2.14 Å and 2.09 Å, respectively. These are smaller than S-Mo (2.41 Å) due to the small size of the B and N atoms compared to the S atom. The second hybrid structure ( Figure 1e) has an optimized B-N bond length of 1.47 Å, while the N-Mo and B-Mo bond lengths are 2.21 Å and 2.27 Å, respectively-larger than N-Mo and B-Mo bonds of the first configuration. We notice that no buckling occurs, which causes the Mo-Mo bond length to be 3.21 Å. We calculate the formation energy E form of the hybrid structures using Equation (1) and find the pristine energy to be −7.09 eV, while the energies of the hybrid structures range between −7.14 and −7.17 eV, reflecting the stability of the studied hybrid structures. Now, we discuss the electronic properties of the hybrid structures. To establish a reference, we first discuss the properties of the pristine MoS 2 . The DOS of pristine MoS 2 is shown in (Figure 2a). The band gap is 1.73 eV, which is in good agreement with the reported experimental (theoretical) value of 1.80 eV (1.74 eV) [44,46,47]. The electronic states near the top of the valence band and the bottom of the conduction band are mainly composed of Mo states which agree with the previous literature [48].

Molecular Adsorptions
Now, we discuss the adsorption properties of the hybrid sheets compared to the pristine MoS2. We consider the following gases: diatomic (OH, NO, CO, N2, H2), triatomic

Molecular Adsorptions
Now, we discuss the adsorption properties of the hybrid sheets compared to the pristine MoS 2 . We consider the following gases: diatomic (OH, NO, CO, N 2 , H 2 ), triatomic (NO 2 , H 2 S, SO 2 , CO 2 ), tetratomic (COOH, NH 3 ), and polyatomic (CH 4 ). The relaxed structures are used with distinct starting sites for the adsorption. For the pristine MoS 2 , we place the gas on the top of the hollow site (H P ), (Figure 3a), which has been shown to be the most favorable site [49].  We started the relaxation from multiple initial positions close to the hexagon cen of the G/BN patch, as well as on the pore edge (including the Mo and S atoms at the ed Figure 3). We find that for some molecules, there are multiple final positions. The adso tion energies of various final locations are discussed below. The adsorption energy E ad calculated using Equation (2).
In Figure 4, we show the adsorption results of four molecules. From left to right, t subfigures show the adsorption energy (first column), the charge transfer between t sheet and the adsorbent (second column), and the shortest distance between the adsorbe atom and the sheet atom (third column). Each subfigure considers nine cases: the pristi and four hybrid systems, each with two starting locations for the adsorbent. The starti and ending locations are shown on the adsorption energy subfigures above the bar rep senting the adsorbent, while the shortest distance between the adsorbent atom and t sheet atom is shown in the corresponding bar in the distance subfigure. We started the relaxation from multiple initial positions close to the hexagon center of the G/BN patch, as well as on the pore edge (including the Mo and S atoms at the edge, Figure 3). We find that for some molecules, there are multiple final positions. The adsorption energies of various final locations are discussed below. The adsorption energy E ads is calculated using Equation (2).
In Figure 4, we show the adsorption results of four molecules. From left to right, the subfigures show the adsorption energy (first column), the charge transfer between the sheet and the adsorbent (second column), and the shortest distance between the adsorbent atom and the sheet atom (third column). Each subfigure considers nine cases: the pristine, and four hybrid systems, each with two starting locations for the adsorbent. The starting and ending locations are shown on the adsorption energy subfigures above the bar representing the adsorbent, while the shortest distance between the adsorbent atom and the sheet atom is shown in the corresponding bar in the distance subfigure.  We first report the bond lengths of the isolated adsorbents. These are 0.75 Å and 0.98 Å for H 2 and OH, respectively, which are similar to experimental values of 0.74 Å [50] and 0.97 Å [51], respectively. After structural optimization, the bond lengths of both molecules do not change for the pristine and the hybrid structures. The bond lengths of the isolated N 2 , NO, and CO are 1.09 Å, 1.16 Å, and 1.14 Å, respectively, similar to experimental bond lengths (1.10 Å, 1.15 Å, and 1.13 Å [51,52], respectively). The bond lengths of For H 2, we find the largest adsorption energy, 0.24 eV at C1, is on the G@MoS 2 (II) sheet. All other hybrid sheets adsorb the H 2 molecule by approximately 0.08 eV. Our results are very close to the previous published results for pristine MoS 2 (0.06 eV) [53] and pristine G (0.08 eV) [54].
For the OH molecule, it is physisorbed on the pristine MoS 2 sheet with 1.71 eV, in agreement with published results [55].  (Figure 4k,n), and at a distance of 2.14/3.37 Å for (Mo s -N m )/(C s -C m ) bonds, (Figure 4l,m). The NO is also chemisorbed on BN@MoS 2 (I) at (H BN , H MoS2 ) with energy of 2.29 eV, (Figure 4j), a charge transfer of −0.1e, (Figure 4k), and at a distance of 2.14 Å (Mo s -N m ) (Figure 4l). The pristine and hybrid structures thus adsorb NO more strongly than CO. The adsorption of NO/CO on the hybrid structures is generally stronger than that of graphene (0.03/0.01 eV [60]) and h-BN (0.03/0.02) eV [61]. Adsorption of NO/CO on the hybrid systems is also superior to that on MoS 2 doped Au, Pd, Pt, and Ni (1.62/1.38 eV for NO/CO [62]), which indicates that our hybrid systems may be considered for NO/CO filtration. Summarizing our results for the diatomic molecules, the hybrid structures (especially G@MoS 2 (I)) are suitable for the physisorption and chemisorption for N 2 , NO, and CO. However, all the considered structures adsorb the H 2 very weakly. For charge transfer, OH and H 2 act as acceptors, while N 2 and CO act as donors.
We now move to triatomic gases and begin our discussion by H 2 S. Our SO2 is chemisorbed by BN@MoS2(I) at (HMoS2, B1) with an energy of 2.71 eV (Figure 5d), a charge transfer of -0.57e (Figure 5e), and a distance of 1.38 Å (Bs-Om) (Figure 5f). The largest physisorption for H2S is on BN@MoS2 (I) at (HBN, HBN) with energy, charge transfer, and distance of 1.96 eV, -0.29e and 1.58 Å (Bs-Om), respectively, (Figure 5d-f).  The last triatomic gas we study is CO 2 . Our calculated C-O bond length and O-C-O angle for the free gas are 1.18 Å and 180 • (the corresponding literature values are 1.16 Å and 180 • .00 [65]). There is no significant change in the bond length and the angle when we add the CO 2 to the different studied systems. The molecule is weakly physisorbed on all structures, with an energy of a fractional eV. The adsorption energy ranges from 0.15 to 0.38 eV, which is better than the corresponding values (0.14 to 0.25 eV) obtained for 2H-@MoS2 nanosheets, nanotubes, and nanopores [66].
Turning to the last gases group, polyatomic gases (COOH, NH 3 , and CH 4 ), the carboxyl molecule COOH (Figure 6a for BN@MoS 2 (I), at the two different final positions for every hybrid structure. However, the bond lengths of COOH did not change significantly. We also notice that the molecule is closest to the sheet with a Cs-Cm distance of 1.52 Å for the G@MoS 2 (II) system (Figure 6c). In all cases, the charge transfer with the sheets is very weak. Most importantly, COOH is chemically adsorbed by G@MoS 2 (II) at (H GMo , C1) with an energy of 3.65 eV without any charge transfer (indicating covalent bond). All other hybrid sheets physically adsorb the COOH gas with an average energy of~1.2 eV. Our calculations show that the adsorption energy of COOH reaches 1.75 eV on our hybrid structures compared to 1.6 on hybrid 1T-@2H-MoS 2 monolayer which may be utilized in many applications such as the decomposition of organic dyes [34].
Regarding the isolated NH 3 and CH 4 gases, the bond length and angles are 107 • and 1.02 Å for NH 3 [68], and 109.47 • and 1.10 Å for CH 4 [68]. When they are adsorbed on the considered sheets, the bond lengths and the characteristic angles of these molecules do not significantly change compared to the corresponding isolated cases. The largest physisorption for NH 3 (Figure 6d) is on BN@MoS 2 (I) with1.80 eV at (Mo, H BN ) with a distance of 1.58 Å (B s -N m ) and a charge transfer of 0.50e (Figure 6e,f). BN@MoS 2 (I) improves the adsorption energy of NH 3 compared to pristine G and h-BN (0.03 eV [60,61]). For CH 4 , the most significant adsorption energy is 2.66 eV for BN@2H-MoS 2 (II) at (Mo, B1) and the closest distance between the molecule and the sheet is 2.71 Å (C s -H m ) with a charge transfer of −0.04e (Figure 6g-i). All considered monolayers adsorb them with energy less than 2.00 eV (physisorption). To summarize, the adsorption energies of COOH are larger than those of NH 3 and CH 4 of the ability to redistribute the charge over the length of COOH. We also notice that NH 3 acts as a donor, which matches some experimental results of NH 3 adsorption on a G sheet [69].
Although we have considered hybrid structures with a small G/BN patch, we expect our results to be applicable to structures with bigger patches of various shapes. This is because the adsorption properties largely depend on the structural defects at the border of the G/BN patch rather than the inner part of the patch. The adsorption capacity will vary with the G/BN concentration but will roughly depend on the square root of the concentration.

Conclusions
In this work, first principle calculations are employed to study the structure, electronic, and molecular adsorption properties of graphene/hexagonal boron nitride@2Hmolybdenum disulfide (G/BN@MoS2) monolayers. We consider systems where the G/BN patch is at the Mo plane (model I) and the S plane (model II). The G@MoS2 systems are metallic, while the BN@MoS2 (I) is n-type semiconducting, and BN@MoS2 (II) is semiconducting. Compared to the pristine G, BN, and MoS2, the hybrid systems have higher adsorption energies for the considered gases (diatomic gases: H2, OH, N2, NO, CO, triatomic gases: CO2, NO2, H2S, SO2, and polyatomic gases: COOH, CH4, and NH3). OH is physisorbed on G@MoS2 (I,II), which can be used for oxygen reduction reactions. NH3 is physisorbed on BN@MoS2 (I,II), making them a suitable material for NH3-based hydrogen production. H2, CO2, and CH4 are weakly physisorbed on all hybrid structures. We also find that chemisorption occurs for: NO and CO on G@MoS2 (I), SO2 on BN@MoS2 (I), and H2S and NO2 on BN@MoS2 (II), which makes these hybrid systems suitable for use as filter materials for these toxic gases.

Conclusions
In this work, first principle calculations are employed to study the structure, electronic, and molecular adsorption properties of graphene/hexagonal boron nitride@2Hmolybdenum disulfide (G/BN@MoS 2 ) monolayers. We consider systems where the G/BN patch is at the Mo plane (model I) and the S plane (model II). The G@MoS 2 systems are metallic, while the BN@MoS 2 (I) is n-type semiconducting, and BN@MoS 2 (II) is semiconducting. Compared to the pristine G, BN, and MoS 2 , the hybrid systems have higher adsorption energies for the considered gases (diatomic gases: H 2 , OH, N 2 , NO, CO, triatomic gases: CO 2 , NO 2 , H 2 S, SO 2 , and polyatomic gases: COOH, CH4, and NH 3 ). OH is physisorbed on G@MoS 2 (I,II), which can be used for oxygen reduction reactions. NH 3 is physisorbed on BN@MoS 2 (I,II), making them a suitable material for NH 3 -based hydrogen production. H 2 , CO 2 , and CH 4 are weakly physisorbed on all hybrid structures. We also find that chemisorption occurs for: NO and CO on G@MoS 2 (I), SO 2 on BN@MoS 2 (I), and H 2 S and NO 2 on BN@MoS 2 (II), which makes these hybrid systems suitable for use as filter materials for these toxic gases.