Investigation of Band Alignment for Hybrid 2D-MoS2/3D-β-Ga2O3 Heterojunctions with Nitridation

Hybrid heterojunctions based on two-dimensional (2D) and conventional three-dimensional (3D) materials provide a promising way toward nanoelectronic devices with engineered features. In this work, we investigated the band alignment of a mixed-dimensional heterojunction composed of transferred MoS2 on β-Ga2O3(\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ 2- $$\end{document}2-01) with and without nitridation. The conduction and valence band offsets for unnitrided 2D-MoS2/3D-β-Ga2O3 heterojunction were determined to be respectively 0.43 ± 0.1 and 2.87 ± 0.1 eV. For the nitrided heterojunction, the conduction and valence band offsets were deduced to 0.68 ± 0.1 and 2.62 ± 0.1 eV, respectively. The modified band alignment could result from the dipole formed by charge transfer across the heterojunction interface. The effect of nitridation on the band alignments between group III oxides and transition metal dichalcogenides will supply feasible technical routes for designing their heterojunction-based electronic and optoelectronic devices.


Background
Beta-gallium oxide (β-Ga 2 O 3 ) has attracted considerable interests due to its superior material properties [1,2]. With ultra-wide bandgap (4.6-4.9 eV), the theoretical breakdown electric field (E C ) is estimated to be around 8 MV/cm [3,4]. Combined with its high relative dielectric constant (ε) and electron mobility (μ), the Baliga's figure of merit ( εμE 3 C ) is triple that of GaN or SiC, reducing the conduction loss significantly [1]. In addition, the availability of large bulk single crystals synthesized via melt-growth and epitaxial techniques delivers significant advantages for industrial applications [5,6]. By far, β-Ga 2 O 3 has been well demonstrated in a wide range of electronic applications, including light-emitting diodes, gas sensors, photodetectors, as well as field-effect transistors [7][8][9][10]. Very recently, hybrid heterojunctions, i.e., the integration of two-dimensional (2D) materials with three-dimensional (3D) materials, are of particular interest due to the complementary properties of their material systems [11].
To date, diverse 2D layered materials have been stacked on wide bandgap semiconductors to construct hybrid heterojunctions for novel applications with varying functionalities, such as MoS 2 /GaN, WSe 2 /GaN, MoS 2 /SiC, and so on [12][13][14][15]. Structurally, the MoS 2 crystal is composed of a Mo atomic layer sandwiched between two sulfur layers, forming a two-dimensional hexagonal trilayer which is bonded to its neighboring layers by weak van der Waals forces [16,17]. Unlike graphene with a zero bandgap, the thickness-dependent modulation of bandgaps motivated the exploration of MoS 2 in optical and electrical devices [18,19]. Based on the physics of MoS 2 , the density of states of few-layer MoS 2 is three orders of magnitude higher than that of single-layer (SL) MoS 2 , leading to high drive currents in the ballistic limit. In this context, few-layer MoS 2 may deliver significant advantages for transistor applications than SL MoS2 [18]. Thus, the integration of MoS 2 with β-Ga 2 O 3 is of great interest for combining respective merits of both the established 2D and 3D materials. And the optical and electrical properties for hybrid heterojunctions are inherently dominated by the interfacial energy band alignment. Consequently, it is quite desirable to have tunable band alignments for improving the performance of heterojunction based devices. In this work, we investigated the band alignment of 2D-MoS 2 /3D-β-Ga 2 O 3 heterojunctions with and without nitridation treatment via X-ray photoelectron spectroscopy (XPS) characterizations and first principles calculations.

Methods
The SiO 2 /Si substrate was ultrasonicated with acetone and visopropanol for each 10 min, respectively, followed by rinsing in deionized water and drying with N 2 . Fewlayer MoS 2 films were grown on the SiO 2 /Si substrate by chemical vapor deposition (CVD) using precursors of MoO 3 (0.08 mg, 99%, Alfa Aesar) and S powder (1 g, 99%) [20,21]. The MoO 3 and S powder were placed into two separate crucibles with a SiO 2 /Si substrate in the quartz tube, as shown in Fig. 1a. During the growth process, the quartz tube was held at 800°C for MoS 2 film growth within 5 min. Figure 1b displays the optical microscopic image of uniform MoS 2 film on SiO 2 /Si substrate. After the growth of MoS 2 film, it would be transferred to β-Ga 2 O 3 (Tamura Corporation, Japan) substrate via PMMA-assisted method, [22] as sketched in Fig. 1c. During the transfer process, PMMA was first spin-coated on as-grown MoS 2 film as a supporting layer, and then the samples were immersed in KOH solution for etching away the SiO 2 layer. Subsequently, the PMMA layer with MoS 2 film would float on the solution, after which the sample would be rinsed in deionized water for 1 min to remove the residual K + and

Results and Discussions
Raman spectroscopy was employed to investigate the quality of few-layer MoS 2 film as well as to check relevant layer numbers. The Raman spectra of MoS 2 film before and after transfer are presented in Fig. 2, which was characterized by RENISHAW inVia Raman spectroscopy. Two characteristic Raman modes could be observed around 381.91 cm −1 and 405.84 cm −1 , corresponding to the in-plane ( E 1 2g ) mode and out-ofplane (A 1g ) mode, respectively [23,24]. Compared with as-grown MoS 2 film, there is almost no Raman shift in E 1 2g and A 1g modes after transfer process, indicative of undamaged MoS 2 after transfer process. The peak at 412.99 cm −1 after transfer process stems from the β-Ga 2 O 3 substrate, in consistent with previous reports [25]. The frequency difference between E 1 2g and A 1g mode was deduced to be about 23.93 cm −1 , designating four layers of few-layer MoS 2 film [26]. Further, as shown in the inset of Fig. 2, the thickness of MoS 2 film was verified to be 3 nm approximately (around four layers) by high-resolution transmission electron microscope (HRTEM), which is in good agreement with our Raman spectra. It can be seen from Fig. 3a that a high intensity peak of N 1 s was detected from the nitride β-Ga 2 O 3 substrate, suggesting the presence of nitrogen. Figure 3b shows the SIMS profiles of MoS 2 /β-Ga 2 O 3 heterojunction with nitridation, where the signals of main components represented by Mo, N, and Ga are plotted against depth. It is observed that the N peak is located at the MoS 2 /β-Ga 2 O 3 interface, and the N spreading into β-Ga 2 O 3 substrate could be contributed by the N injection into the underlying layer during plasma treatment or primary beam bombardments. The higher Ga profile in the MoS 2 layer than β-Ga 2 O 3 substrate probably stems from the different ion yield in the different material matrix [27]. Moreover, the tail of Mo in β-Ga 2 O 3 could be ascribed to the diffusion or depth resolution problem, which is caused by primary beam bombardment [28].
To obtain the band alignments of MoS 2 /β-Ga 2 O 3 heterojunctions, XPS measurements with a step of 0.05 eV were carried out on VG ESCALAB 220i-XL system with a monochromatic Al Kα X-ray source (hν = 1486.6 eV). The constant pass energy was set at 20 eV. Additionally, the standard C 1 s (284.8 eV) was used for binding energy (BE) calibration [29]. To evaluate the valence band offset (VBO) at the MoS 2 / β-Ga 2 O 3 interface, Mo 3d and Ga 3d core levels  Figure 4a shows the XPS narrow scan of Mo 3d and valence band spectra from few-layer MoS 2 [30]. The binding energy difference (BED) between CLs of Mo 3d 5/2 and valence band maximum (VBM) for MoS 2 was calculated to be 228.59 ± 0.1 eV. As shown in Fig. 4b, the BE of Ga 3d CL and VBM from few-layer β-Ga 2 O 3 were deduced to be 20.25 ± 0.05 and 3.23 ± 0.05 eV, respectively. The corresponding BED was determined to 17.02 ± 0.1 eV, which is well consistent with that reported by Sun et al. [31]. Figure 4c depicts the measured XPS spectra of Mo 3d and Ga 3d CLs for MoS 2 /β-Ga 2 O 3 heterojunctions with/without nitridation. It is noted that the Mo 3d 5/2 CL shifted from 228.95 ± 0.05 eV for the unnitrided heterojunction toward 229.60 ± 0.05 eV for the nitrided heterojunction while Ga 3d CL shifted from 20.25 ± 0.05 to 20.65 ± 0.05 eV. Based on Kraut' method, [32] the valence band offset (VBO, ΔE V ) of few-layer MoS 2 /β-Ga 2 O 3 heterojunctions was calculated according to the following equation, Ga 3d ) is the binding energy difference between Mo 3d 5/2 and Ga 3d CLs for MoS 2 /β-Ga 2 O 3 heterojunctions. Hence, the ΔE V of MoS 2 on β-Ga 2 O 3 substrate with and without N 2 plasma treatment was calculated to be 2.62±0.1 and 2.87 ± 0.1 eV, respectively. Figure 4d shows the O 1 s CL energy loss spectra of β-Ga 2 O 3 substrates with and without nitridation. It is noted that the bandgap keeps unchanged after nitridation treatment with a value of 4.70 ± 0.1 eV. Thus, the conduction band offset can be extracted as follows, where E Ga 2 O 3 g and E MoS 2 g are the bandgaps of β-Ga 2 O 3 and few-layer MoS 2 , respectively. The bandgap of 1.4 ± 0.1 eV for few-layer MoS 2 was used in this work. 34 According to Eq. (2), the ΔE C between MoS 2 and β-Ga 2 O 3 with and without nitridation were deduced to be 0.68 ± 0.1 and 0.43 ± 0.1 eV, respectively. The calculated band diagrams for heterojunctions without/with nitridation are shown in Fig.  5(a) and 5(b), respectively.
Next, the electronic structures of nitrided and unnitrided heterojunctions were further examined through the Vienna ab initio simulation package (VASP) based on density functional theory (DFT) [33][34][35]. The generalized gradient approximation (GGA) of Perdew-Burke-Ernzerhof (PBE) parameterization was adopted for exchange-correlation function [36,37]. We used the DFT-D3 dispersion corrections approach to describe the long-distance van der Waals (vdW) interactions [38][39][40]. The projector augmented wave (PAW) pseudopotential method was used to describe the core-valence interaction with a kinetic energy cutoff of 650 eV for plane wave expansion. We employ a 4 × 4 × 1 G-centered kmesh for structural relaxation of the unit cell, with the smallest spacing between k-points of 0.04 Å −1 , which is precise enough by the convergence test with respect to the number of k points. The convergence thresholds are set to 10 −4 eV for energy differences of the system and 10 −2 eV Å −1 for Hellman-Feynman force. In order to eliminate artificial interactions between two adjacent atomic layers, the thickness of the vacuum layer is set to functionals [41]. The mixing ratio is 25% for the shortrange HF exchange. The screening parameter is 0.2 Å −1 .
The MoS 2 /β-Ga 2 O 3 heterojunctions were constructed as shown in Fig. 6a. The universal binding energy relation (UBER) method, which provides a Fig. 4 a XPS spectra of Mo 3d CL and valence band from few-layer MoS 2 . b XPS spectra of Ga 3d CL and valence band from β-Ga 2 O 3 substrate. c XPS spectra of Mo 3d and Ga 3d CLs for fabricated MoS 2 /β-Ga 2 O 3 heterojunction with/without surface nitridation. d XPS spectra of O 1 s CL energy loss of β-Ga 2 O 3 substrate with/without surface nitridation simple universal form for the relationship between binding energy and atomic separation, [42] was applied to determine the energetically stable structure before electronic structure calculation. Various interlayer distances were considered and the surface adhesion energy W ad for the heterojunctions are shown below, where A is the interface area, E Ga 2 O 3 , E MoS 2 , and E Ga 2 O 3 =MoS 2 are the total energies of β-Ga 2 O 3 , mono-layer MoS 2 and the MoS 2 /β-Ga 2 O 3 heterojunction, respectively. Once the W ad reaches a maximum, the optimal interlayer distance will be obtained. After structure optimizations, a nitrogen atom is substitutionally doped in the original MoS 2 /β-Ga 2 O 3 heterojunction, as shown in Fig. 6b. The concentration of nitrogen in DFT calculation is around 4.17%, which is close to that (3.61%) in experiments. The electronic structures for both nitrided and unnitrided MoS 2 /β-Ga 2 O 3 heterojunctions were calculated as illustrated in Fig. 6c and d. It was seen that mid-gap states were introduced, which may enhance the charge transfer across the MoS 2 /β-Ga 2 O 3 interface, and the resulting interface dipole contributed to the measured binding  using the HSE06 method to further confirm the above conclusion, and find that the corrected ΔE C are 0.87 and 1.08 eV for undoped-and doped-β-Ga 2 O 3 /MoS 2 heterojunctions respectively.

Conclusions
In conclusion, respective MoS 2 film has been transferred onto unnitrided and nitride β-Ga 2 O 3 for constructing MoS 2 /β-Ga 2 O 3 heterojunctions. Raman spectroscopy was used to investigate the quality of transferred MoS 2 film, and SIMS study was performed to probe the elemental depth profiles of the MoS 2 /β-Ga 2 O 3 heterojunction with nitridation. The VBOs were determined to be 2.62 ± 0.1 eV for nitrided MoS 2 /β-Ga 2 O 3 heterojunction and 2.87 ± 0.1 eV for unnitrided heterojunction by XPS, respectively. The resultant CBOs were deduced to be 0.68 ± 0.1 and 0.43 ± 0.1 eV, which was in the same trends with the DFT calculations. These findings demonstrated that the band offsets can be modified via surface nitridation process. This study offers glorious perspectives on the implementation of designed electronic devices based on 2D/3D vertical heterojunctions.