Tunable Electronic and Magnetic Properties of 3d Transition Metal Atom-Intercalated Transition Metal Dichalcogenides: A Density Functional Theory Study

Abstract

Currently, intercalation has become an effective way to modify the fundamental properties of two-dimensional (2D) van der Waals (vdW) materials. Using density functional theory, we systematically investigated the structures and electronic and magnetic properties of bilayer transition metal dichalcogenides (TMDs) intercalated with 3d TM atoms (TM = Sc–Ni), TM@BL_MS₂ (M = Mo, V). Our results demonstrate that all the studied TM@BL_MS₂s are of high stability, with large binding energies and high diffusion barriers of TM atoms. Interestingly, most TM@BL_MoS₂s and TM@BL_VS₂s are found to be stable ferromagnets. Among them, TM@BL_MoS₂s (TM = Sc, Ti, Fe, Co) are ferromagnetic metals, TM@BL_MoS₂ (TM = V, Cr) and TM@BL_VS₂ (TM = Sc, V) are ferromagnetic half-metals, and the remaining systems are found to be ferromagnetic semiconductors. Exceptions are found for Ni@BL_MoS₂ and Cr@BL_VS₂, which are nonmagnetic semiconductors and ferrimagnetic half-metals, respectively. Further investigations reveal that the electromagnetic properties of TM@BL_MoS₂ are significantly influenced by the concentration of intercalated TM atoms. Our study demonstrates that TM atom intercalation is an effective approach for manipulating the electromagnetic properties of two-dimensional materials, facilitating their potential applications in spintronic devices.

1. Introduction

Since the discovery of graphene [1,2], various two-dimensional (2D) materials with finite thickness have attracted wide attention due to their versatile properties and potential applications in various fields [3,4]. Particularly, the 2D candidates with rich magnetic properties are regarded as a good platform for the development of spintronic devices [2,3,4,5,6], which facilitates low energy consumption and fast device operation. Experimentally, a few 2D magnetic materials have been produced, including ferromagnetic (FM) transition metal trihalide monolayer (e.g., CrI₃, VI₃) [7,8], Cr₂Ge₂Te₆ monolayer [9], Fe₃GeTe₂ monolayer [10], and antiferromagnetic (AFM) MnPS₃ monolayer [11,12,13]. Inspired by the above progress, considerable theoretical efforts have been made on the development of 2D magnetic candidates, such as transition metal (TM) borides [14,15,16], TM phosphides [17,18,19], TM disulfides [20,21,22], etc.

Despite the theoretical and experimental achievements, the members of known 2D FM or antiferromagnetic (AFM) candidates are still limited. Moreover, to advance the development of 2D spintronics devices, two major challenges should be solved: (i) robust magnetic orders with high transition temperature and (ii) the feasibility of fabrication experimentally, for example, many predicted 2D FM or antiferromagnetic (AFM) candidates so far face challenges in preparation. Therefore, finding a feasible way to obtain 2D magnetic materials is still a great challenge. It is found that introducing magnetism to grown non-magnetic 2D materials is an effective way to solve the challenge (ii) mentioned above. Various modification strategies have been developed for nonmagnetic materials in terms of tuning their electronic and magnetic properties, such as atom doping or substitution [23], introducing vacancies [24], external strains [25], etc. Unfortunately, these methods usually bring significant distortion to the structures.

Alternatively, the intercalation of magnetic species into the vdW gap of 2D materials provides a promising way to explore novel 2D magnets without destroying the host lattices [26,27,28,29,30]. Taking 2D transition-metal dichalcogenides (TMDs) as an example, the TM atom-intercalated vdW TMDs with different stoichiometries have been produced by electrochemical intercalations, which has been identified to be an effective way to modulate the electronic and magnetic properties [31,32]. For example, Guzman et al. presented a comprehensive investigation into the structural evolution of CVD-grown V_1+xSe₂ nanoplates. They found the VSe₂-to-V₅Se₈ phase and V₃Se₄-to-V₅Se₈ structural transformation under 300 °C and 250 °C annealing [33]. Zhao et al. extended the ferromagnetic order in tantalum-intercalated TaS(Se)_y, such as Ta₉S₁₆, Ta₇S₁₂, Ta₁₀S₁₆, and Ta₈Se₁₂ [34]. Besides, FM orders were observed in ultra-thin films of V₅X₈ (X = S, Se) [35] and V₅Se₈/NbSe₂ 2D-heterostructures [36]. In addition, a number of different Cr_1+δTe₂ phases were found to have a broad range of magnetic orders and novel magnetic phases [37,38]. Using density functional theory calculation, Kumar et al. found that the late TM-intercalated WSe₂ exhibited substantial magnetic moments and pronounced ferromagnetic order [27]. Guo et al. found that filling either Cr or I atoms into the van der Waals gap of stacked and twisted CrI₃ bilayers can induce the double exchange effect and significantly strengthen the interlayer ferromagnetic coupling [39]. Correspondence to the experimental progress, the theoretical understanding of the electronic and magnetic properties of these intercalated systems is still very limited.

In this work, by using the density functional theory, we systematically studied the structural, electronic, and magnetic properties of 3d TM atom-intercalated bilayers MS₂ (M = Mo, V), TM@BL_MS₂ (TM = Sc–Ni). Our results show that all the systems are very stable with large binding energies. Except for Ni@BL_MoS₂ and Cr@BL_VS₂, all the other TM@BL_MoS₂s and TM@BL_VS₂s are ferromagnetic (half-)metals and semiconductors. Ni@BL_MoS₂ is a nonmagnetic semiconductor, and Cr@BL_VS₂ is a ferrimagnetic half-metal. Most importantly, the electronic and magnetic properties of these TM@BL_MoS₂s are largely influenced by the concentration of TM atoms.

2. Results and Discussion

2.1. Structure and Stability of TM@BL_MS₂ (TM = Sc–Ni, M = Mo, V)

The space groups for H-MoS₂ and H-VS₂ monolayers are the same, P$\overline{6}$m2, with lattice constants of 3.16 Å and 3.17 Å, respectively. The MoS₂ monolayer is a nonmagnetic semiconductor [40,41,42], and the H-VS₂ monolayer is predicted to be a FM semiconductor with a magnetic moment of 1.0 µB per V atom [43,44,45,46,47]. Three types of stacking configurations were considered for MoS₂ or VS₂ bilayers: (i) AA-stacking, in which the atoms from the top MS₂ layer sit exactly atop those in the bottom layer [see Figure 1a,f and Figure S1a,f]; (ii) AB-stacking, where both layers rotate by 30 degrees relative to each other, with the M atoms from the top layer sitting atop those in the bottom layer, and the S atoms from one layer located on the hexagonal hole site of the adjacent layer [see Figure 1b,g and Figure S1b,g]; (iii) AB′-stacking, which is similar to AB-stacking but with reversed interlayer configurations of the M/S atoms [see Figure 1c,h and Figure S1c,h]. Our results show that the favored stacking structures for bilayer MoS₂ and VS₂ are the AB-stacking ones, which are 0.60 eV/0.56 eV and 0.55 eV/0.55 eV lower in energy than their AA- and AB′-stacking counterparts, respectively. In the following study, we only focus on the AB-stacked MoS₂ bilayer and VS₂ bilayer, whose interlayer spacing is 3.19 Å and 3.04 Å, respectively. In order to study the most stable configurations of TM@BL_Mo(V)S₂s, two configurations with different intercalation sites for TM atoms are considered: (i) Conf_I: in which the intercalated TM atom sits below the Mo/V atom (A site) in the top layer of Mo(V)S₂ [see Figure 1d,i and Figure S1d,i]; (ii) Conf_II: in which the TM atom is located below the S atom (B site) of the upper layer of Mo(V)S₂ [see Figure 1e,j and Figure S1e,j]. Testing calculations show that the TM@BL_MoS₂s with Conf_I intercalation are the favored ones, which are 1.3 eV and 0.9 eV lower than that with B Site intercalating for Ti@BL_MoS₂ and V@BL_MoS₂, respectively. Therefore, we will only use Conf_I for further exploration, and the optimized structures are shown in Figure 2 and Figure S2 in the Supporting Information (SI). The detailed structural information is summarized in

Table 1. The interlayer spacing (d_S-S), the distance between the TM atoms and the two-side MS₂ layer (d_TM-S), magnetic moment (μ) per unit cell, binding energy (E_b), the Bader charges transferred from TM to adjacent MS₂ layers (ΔQ), and ground state (GS) of TM@BL_MS₂ systems.

Metal	dS-S$/Å$	dTM-S$/Å$	μ (μB)	Eb (eV)	ΔQ (e−)	GS
TM@BL_MoS2
Sc	3.46	1.73	2.78	−6.50	1.48	M
Ti	3.34	1.67	2.45	−5.19	1.28	M
V	3.28	1.64	3.44	−6.80	1.10	HM
Cr	3.37	1.69	5.93	−5.66	0.92	HM
Mn	3.37	1.68	5.00	−5.82	0.89	SC
Fe	3.17	1.58	2.29	−5.42	0.73	M
Co	3.19	1.60	2.92	−7.11	0.60	M
Ni	3.16	1.09	0	−7.21	0.32	SC
TM@BL_VS2
Sc	3.37	1.69	36.86	−6.67	1.53	HM
Ti	3.31	1.66	37.65	−6.69	1.41	SC
V	3.16	1.58	38.94	−6.51	1.37	HM
Cr	3.03	1.51	33.79	−4.23	1.26	HM
Mn	3.05	1.52	40.62	−4.63	1.18	SC
Fe	3.09	1.55	39.71	−5.22	1.04	SC
Co	3.01	1.50	36.69	−5.78	0.65	SC
Ni	3.05	1.53	37.72	−5.45	0.61	SC

.

Except for Ni@BL_MoS₂, all the TM atoms prefer to occupy the center of the vertical interlayer space of both TM@BL_MoS₂s and TM@BL_VS₂s (see Figure 2 and Figure S2), in which the TM atoms are distanced from the two-side MS₂ layer by about 1.09 Å~1.73 Å and 1.50 Å~1.69 Å for TM@BL_MoS₂s and TM@BL_VS₂s, respectively. Compared with free-standing Mo(V)S₂ bilayers, the intercalation of TM atoms enlarges the interlayer spacing to 3.16 Å~3.46 Å and 3.05 Å~3.37 Å in the case of TM@BL_MoS₂s and TM@BL_VS₂s, respectively (see Table 1). To quantitatively characterize the structural stability of these TM@BL_Mo(V)S₂ systems, we calculate the binding energies (E_bs) of the TM atom to the Mo(V)S₂ bilayers based on the following equation:

$$E_{\mathrm{b}}=E_{\mathrm{Total}}-E_{\mathrm{TM}}-2E_{\mathrm{M}{\mathrm{S}}_2}$$

(1)

where E_Total, E_TM, and E_MS2 represent the total energy of the TM@BL_MS₂, the isolated transition metal atom, and the free-standing Mo(V)S₂ bilayers, respectively. As shown in Table 1 and Figure 3a, the calculated E_bs are around −7.21~−5.19 eV and −6.69~−4.23 eV for TM@BL_MoS₂s and TM@BL_VS₂s, respectively, which are comparable to those of TM-intercalated graphene/TMD heterostructures [48]. Such large negative E_b values indicate that the intercalation of TM atom in Mo(V)S₂ bilayers is energetically favorable. In the case of TM@BL_MoS₂, the E_b of Ni@BL_MoS₂ system is the lowest (−7.21 eV), and that of Ti@BL_MoS₂ is the highest (−5.19 eV). For the TM@BL_VS₂ system, the E_b of Ti intercalation system is the lowest (−6.69 eV), and that of Cr intercalation system is the highest (−4.23 eV).

In order to clarify the bonding characteristics of these systems, the charge density differences (CDDs) of TM@BL_MS₂s (TM = Ti, V, Cr) are given in Figure S3, which are calculated based on the following equation:

$$∆\rho ={\rho }_{\left[{\mathrm{T}\mathrm{M}@\mathrm{B}\mathrm{L}\_\mathrm{M}\mathrm{S}}_2\right]}-{\rho }_{\left[\mathrm{T}\mathrm{M}\right]}-{\rho }_{\left[{\mathrm{M}\mathrm{S}}_2\left(\mathrm{T}\right)\right]}-{\rho }_{\left[{\mathrm{M}\mathrm{S}}_2\left(\mathrm{L}\right)\right]}$$

(2)

where ρ_{[TM@BL_MS2]}, ρ_[TM], ρ_[MS2(T)], ρ_[MS2(L)] represent the charge density of the whole system, the TM atom, the top MS₂ layer, and the low MS₂ layer, respectively. Clearly, the embedding of TM atoms resulted in charge redistribution in the Mo(V)S₂ sublayers. It was found that the charge densities were reduced at the center of TM atoms and accumulated between the TM-S bonds, indicating the obvious covalent bonding characteristics in the systems. Quantitatively, the Bader charges of the TM atoms were calculated (see Table 1). It can be seen that the amount of charge transfer gradually decreased from Sc to Ni in both TM@BL_MoS₂ and TM@BL_VS₂ systems (see Figure 3b). Moreover, the diffusion of Ti atom in Ti@BL_MoS₂ along the MoS₂ interface was examined (see Figure S4), and the large diffusion barrier (1.4 eV) confirms the difficulty of clustering of Ti atoms.

2.2. Electronic and Magnetic Properties of TM@BL_MS₂ (TM = Sc–Ni, M = Mo, V)

The total magnetic moments of both TM@BL_MoS₂s and TM@BL_VS₂s are summarized in Table 1 and Figure 3c. Interestingly, the intercalation of TM atoms in MoS₂ bilayers introduces magnetism into the systems. Except for Ni@BL_MoS₂, all the other TM@BL_MoS₂s (TM = Sc–Co) exhibit ferromagnetism. Among them, Cr@BL_MoS₂ has the largest magnetic moment (5.93 μB), and Fe@BL_MoS₂ displays the smallest one (2.29 μB) per unit cell (see Table 1). Such induced ferromagnetisms in nonmagnetic MoS₂ bilayers are similar to those of TM-intercalated TMDs [27] and other 2D materials [28,48]. The partial density of states (PDOS) and spin density plots are shown in Figure 4. The magnetic moments of these TM@BL_MoS₂ systems are mainly contributed by the 3d electrons from TM atoms. In addition, spin polarization of some Mo atoms close to the TM atoms occurs, as shown in the spin density plots in Figure 4. Taking Cr@BL_MoS₂ as an example, five d orbitals (d_xy, d_x2–_y2, d_xz, d_yz, d_z2) from the spin-up channel of the Cr atom contribute 5.0 μB magnetic moment, and one d orbital from the spin-up channel of the two Mo atoms above/below the Cr atom contributes 1.0 μB magnetic moment. As a result, the total magnetic moment of this system is about 6.0 μB. Similar analysis can also be applied to other systems.

As for TM@BL_VS₂s, most systems are found to be ferromagnetic, with ferromagnetic coupling between V atoms and intercalated TM atoms (see Figure S5). The magnetic moments range from 36.69 μB to 40.62 μB per unit cell. Among them, Co@BL_VS₂ has the smallest magnetic moment (36.69 μB), and Mn@BL_VS₂ has the largest magnetic moment (40.62 μB). Exceptions are found for Sc@BL_VS₂ and Cr@BL_VS₂. For the former, zero magnetic moment is found for the Sc atom, and for the latter, it is ferromagnetic, with the Cr atom displaying opposite magnetic moments to those of the V atoms.

Compared with the semiconducting properties of free-standing MoS₂ monolayers, TM@BL_MoS₂s (TM = Sc, Ti, Fe, Co) are changed to be ferromagnetic metals (see Figure 4 and Figure S6 and Table 1), in which the conduction band bottom passes through the Fermi level. V@BL_MoS₂ and Cr@BL_MoS₂ are ferromagnetic half-metals, in which one spin channel shows conducting behavior, and the contrast spin channel shows semiconducting properties. Besides, Mn@BL_MoS₂ and Ni@BL_MoS₂ are ferromagnetic semiconductor and non-magnetic semiconductor with band gaps of about 0.74 eV and 0.99 eV, respectively, which are significantly reduced compared to those of the MoS₂ monolayer (1.8 eV) [41]. Moreover, the embedded TM atom produces impurity energy bands in the gap of MoS₂, leading to the Fermi level of the energy band shifting upwards moderately. For TM@BL_VS₂s, as shown in Figure S7, the systems with TM = Sc, V, Cr become ferromagnetic or ferrimagnetic half-metals. In the case of TM@BL_VS₂s with TM = Ti, Mn, Fe, Co, Ni, they are all ferromagnetic semiconductors with a band gap of about 0.52, 0.46, 0.53, 0.19, and 0.52 eV, respectively.

2.3. Effect of TM Intercalation Concentration on the TM BL_MS₂ (TM = V, Cr, Mn, Fe, M = Mo)

In order to determine the influence of TM intercalation ratios on the structures and electronic properties of these systems, the structures intercalated with two or more TM atoms (TM = V, Cr, Mn, Fe), i.e., 2TM@BL_MoS₂ and TMC@BL_MoS₂, were considered, in which two adjacent TM atoms inserted in the 4 × 4 supercell and one chain intercalated TM atoms inserted in the 4 × 1 supercell were considered. The optimized structures of 2TM@BL_MoS₂s and TMC@BL_MoS₂s are shown in Figure 5. Similar to the TM@BL_MoS₂ systems, the TM atoms from the 2TM@BL_MoS₂ and TMC@BL_MoS₂ systems were firmly stabilized in the middle of the space of the MoS₂ bilayer. Moreover, for the 2TM@BL_MoS₂ systems, the interlayer distances were around 3.44 Å, 3.55 Å, 3.50 Å, and 3.24 Å, and the distances between TM atoms and the two-side MoS₂ layer were 1.72 Å, 1.77 Å, 1.75 Å, and 1.62 Å in the systems with TM = V, Cr, Mn, Fe, respectively (see

Table 2. The interlayer spacing (d_S-S), the distance between the TM atoms and the two-side MoS₂ layer (d_TM-S), magnetic moment (μ) per unit cell, binding energy (E_b) per TM atom, the Bader charges transferred from TM to adjacent MoS₂ layers (ΔQ), and ground states (GS) of 2TM@BL_MoS₂s and TMC@BL_MoS₂s.

Metal	${\mathit{d}}_{\mathbf{S}-\mathbf{S}}/Å$	${\mathit{d}}_{\mathbf{TM}-\mathbf{S}}/Å$	μ (μB)	Eb (eV)	ΔQ (e−)	GS
2TM@BL_MoS2
2V	3.44	1.72	7.85	−4.21	1.13	HM
2Cr	3.55	1.77	11.89	−3.79	0.98	M
2Mn	3.50	1.75	10.00	−3.45	0.95	M
2Fe	3.23	1.62	6.14	−4.18	0.73	M
TMC@BL_MoS2
VC	3.14	1.57	3.03	−3.39	1.09	M
CrC	4.55	2.27	5.14	−2.35	0.64	M
MnC	3.43	1.71	3.61	−1.95	0.95	M
FeC	3.54	1.77	2.44	−3.04	0.73	M

). As for TMC@BL_MoS₂s, the interlayer distances were around 3.14 Å, 4.56 Å, 3.43 Å, and 3.54 Å for the structures inserted with chains of V, Cr, Mn, and Fe, respectively, and the distances between TM atoms and adjacent MoS₂ layers were around 1.57 Å, 2.27 Å, 1.71 Å, and 1.77 Å, respectively (see Table 2). The binding energies of these 2TM@BL_MoS₂ and TMC@BL_MoS₂ systems were in the range of −4.21~−3.45 eV and −3.39~−1.95 eV, respectively, in which the systems with Mn atoms exhibited the lowest E_bs, while those with the V atoms showed the largest E_bs. Similar to those TM@BL_MoS₂ systems, the charges transferred from TM to adjacent MoS₂ layers per unit cell were in the range of 0.73~1.13e and 0.73~1.09e for the 2TM@BL_MoS₂ and TMC@BL_MoS₂ systems, respectively (see Figure S8 and Table 2).

The PDOS, spin density diagram, and band structures of these 2TM@BL_MoS₂s and TMC@BL_MoS₂s are shown in Figure 6 and Figure S9. It was found that all 2TM@BL_MoS₂s and TMC@BL_MoS₂s are FM metals except for 2V@BL_MoS₂, which is FM half-metal. For 2TM@BL_MoS₂s, the magnetic moments were 7.85 μB, 11.89 μB, 10 μB, and 6.14 μB per unit cell for the systems with TM = V, Cr, Mn, Fe, respectively, nearly double those of their TM@BL_MoS₂ counterparts (see Table 2). As for TMC@BL_MoS₂s, the magnetic moments were 3.03 μB, 5.17 μB, 3.61 μB, and 2.44 μB per unit cell for the ones with TM = V, Cr, Mn, Fe, respectively. Similar to TM@BL_MoS₂s, the magnetic moments of the above systems were mainly contributed by the inserted 3d TM atoms (see Figure 6). However, the magnetic moments per unit cell of TMC@BL_MoS₂s differed from those of TM@BL_MoS₂s, indicating that the magnetic properties of these systems are sensitive to the intercalated TM concentrations. Moreover, with the intercalated TM atom increasing, the electronic states around the Fermi level increased, leading to the metallic character in them. Specifically, the number of impurity bands from TM atoms in 2TM@BL_MoS₂s increased and almost doubled those in TM@BL_MoS₂s (see Figure S9). In contrast, in TMC@BL_MoS₂s, the electronic bands around the Fermi level were no longer flat and became much more dispersed, indicating that the hybridization between TM_d and MoS₂_p orbitals was largely strengthened. Therefore, it is very effective to control the electronic and magnetic properties of these systems by varying the TM ratios.

3. Methods

All the calculations were performed using the Vienna ab initio simulation package (VASP) [49,50,51]. The exchange–correlation potentials were treated by the generalized gradient approximation (GGA) of optimized Perdew–Burke–Ernzerhof (PBE) [52]. The interactions between the ion core and the valence electrons were modeled with projector augmented wave (PAW) potentials [53], and the DFT-D2 method was used to account for van der Waals (vdW) interaction [54]. A vacuum layer of 20 Å was added in the z direction to eliminate the interaction of interlayer caused by periodic boundary conditions. In order to consider the Coulomb interaction and exchange interactions on TM electrons, the GGA+U method with U_eff = 4.0 eV was adopted according to the previous literature, which has been proven to give very close results for these systems [55,56,57,58]. A plane-wave basis set with a kinetic cutoff energy of 500 eV was employed. The 4 × 4 supercells with the lattice constants of a = b = 12.76 Å and a = b = 12.72 Å for bilayer MoS₂ and VS₂ were applied, respectively. The Brillouin zone (BZ) was sampled using 3 × 3 × 1 and 5 × 5 × 1 gamma-centered Monkhorst–Pack grids for the calculation of structural relaxation and electronic structures, respectively. The criteria for energy and atomic force convergence were set to 10⁻⁵ eV per unit cell and 0.01 eV Å⁻¹, respectively.

4. Conclusions

In summary, the structural, electronic, and magnetic properties of 3d TM atom (TM = Ti–Ni)-intercalated Mo(V)S₂ bilayers, TM@BL_ Mo(V)S₂, are explored using density functional theory methods. All the studied systems are thermal dynamically stable with large binding energies, ranging from −4.23 to −7.21eV. The intercalation of TM atoms introduces rich electronic and magnetic properties to Mo(V)S₂ bilayers. Except for Ni@BL_MoS₂, which is a nonmagnetic semiconductor, most studied systems are found to display robust magnetic properties. TM@BL_MoS₂s (TM = Sc, Ti, Fe, Co) are ferromagnetic metals, TM@BL_MoS₂ (TM = V, Cr) and TM@BL_VS₂ (TM = Sc, V, Cr) are ferromagnetic or ferrimagnetic half-metals, and the remaining systems are found to be ferromagnetic semiconductors. Furthermore, the electronic and magnetic properties of these TM@BL_MoS₂s are largely influenced by the concentration of TM atoms. Our study proposes that the intercalation of TM atoms is a feasible way to tune the properties of 2D materials and promote their potential applications in electronic and spin electronic devices.