1 Introduction
Solar power has been recognised as an essential green energy resource to replace the conventional fossil energy, which attracts considerable attention for advanced applications, involving solar photovoltaics, solar heating system and solar thermoelectric devices [
1-
4]. Among them, photovoltaics have been developed as a mature technology that converts energy from the sunlight into a flow of electrons by collecting excited electrons or holes at counter electrodes. Currently, the highest power conversion efficiency of the photovoltaic solar system reaches 25% at an organometallic stabilizing perovskite [
5]. However, the costs, stability and toxicity of a perovskite-based devices still limit its possibility for practical large-scale production [
6]. Therefore, it is still of great significance to design a stable non-perovskite system with suitable bandgap and large charge carrier separation for advanced photovoltaic applications.
Two-dimensional (2D) semiconducting materials exhibits great potentials on photovoltaic applications, due to its large specific surface area, highly-efficient light absorption, ultrahigh charge carrier transportation [
7]. Significantly, these atomically thin materials possess excellent flexibility in terms of tuning their physicochemical properties [
8-
10]. Among them, 2D transition metal dichalcogenides (TMDs) with a sandwiched MX
2 (M = transition metals and X = S, Se and/or Te) stoichiometry possesses diverse electronic/optical characters [
11]. Especially, semiconducting 2H-TMDs, such as MoX
2 and WX
2 nanosheets, with stimuli-controllable electronic structures has shown great potentials on photoelectrochemical solar energy conversion [
12]. Furthermore, the development of techniques on heterostructure engineering remarkably enhance the efficiency of light absorption and excited charge carrier separation by forming the proper band alignment [
13-
15]. The photogenerated electrons or holes can accumulate at different components of heterostructures due to the induced built-in electric fields in type-II or z-scheme paths [
16-
18]. Thus, the construction of 2D heterostructures with proper band alignment can efficiently improve the performance of photovoltaic/optoelectronic nanodevices.
Since the in-plane electric polarization can lead to the emergence of a spontaneous photovoltaic effect in WSe
2/black phosphorus heterostructure [
19], it is highly expected to achieve the polarization-controlled photovoltaic performance by constructing the polarized heterostructure [
20]. Recently, 2D Janus TMDs (MXY, X ≠ Y) with asymmetric geometry have been successfully fabricated by using the chemical vapor deposition (CVD) approach, attracting considerable attention for its intriguing optoelectronic features [
21-
23]. An inherent electric polarization is generated from the structural asymmetry, which is beneficial to the separation of excited electron−hole pairs by limiting charge carrier recombination [
24]. The carrier mobility of photogenerated electrons and holes could be completely different in Janus TMDs [
25-
28]. Besides, the lattice parameters of MXY generally keep similar to the MX
2 containing the same transition metals [
29]. Therefore, MXY is a promising materials platform to build polarized heterostructures with MX
2 monolayers. The superior photovoltaic/optoelectronic performance is also expected to be achieved in MXY/MX
2 heterostructures.
2D Janus MoSSe and MoTe
2 possess excellent stability and appropriate bandgap of around 2.0 eV [
21,
30]. However, their optoelectronic application is still limited by the short lifetime excited charge carriers. To improve the photo response character and prolong the lifetime of excitons, in this work, we constructed a heterostructure based on the Janus MoSSe and MoTe
2 monolayers and systematically investigated its polarization-controlled optoelectronic properties. The negative formation energy indicates the thermal stability of the constructed heterostructures (MoSSe−MoTe
2 and MoSeS−MoTe
2). Electronically, the bandgaps of MoSSe−MoTe
2 and MoSeS−MoTe
2 heterostructures are remarkably reduced to 0.71 and 0.03 eV, respectively, due to the formed z-scheme band alignment between the monolayers. Notably, the vertical polarization induced by MoSSe monolayer will further trigger the interlayer charge redistribution, suggesting the polarization-controlled charge transfer and built-in electric fields. Besides, the light absorption is also improved after forming the heterostructure, especially in the visible-light region. The proper band alignment and polarization-controllable feature of MoSSe/MoTe
2 heterostructures provide a promising platform for advanced optoelectronic nanodevices.
2 Computational methods
All calculations were performed based on density functional theory (DFT) method, as implemented in Vienna Ab-initio Simulation Package (VASP) [
31-
33]. The projected augmented wave (PAW) [
34] method was used to describe the ion−electron correlation, while the Perdew−Burke−Ernzerhof generalized gradient approximation (PBE-GGA) [
35] functional was adopted. The energy cut-off was set to be 500 eV, and the first Brillouin zone (BZ) was sampled by a 21 × 21 × 1 Gamma-centred k-mesh [
36]. To prevent the adjacent interaction between the periodic layers, a vacuum thickness of 20 Å was used. Since the bandgap would be underestimated by the PBE-GGA functional, the Heyd−Scuseria−Ernzerhof hybrid functional (HSE06) was applied to calculate the electronic properties. Besides, the DFT-D3 method was adopted to correct the long-range van der Waals interaction [
37]. Moreover, all structures were fully relaxed until the energy and force on each atom converge to 10
−6 eV and 0.001 eV/Å.
The optical absorption can be calculated by
where
ε1(
ω) and
ε2(
ω) represent the real and imaginary parts of the complex dielectric function, respectively, which was obtained by using the vaspkit code [
38].
3 Results and discussion
The optimized MoSSe and MoTe
2 monolayers are in the space group of
P3
m1 with the lattice parameters of 3.23 and 3.52 Å, respectively, close to previous experimental researches [
39,
40]. Accordingly, the MoSSe/MoTe
2 heterostructure is constructed between the unit cell of each layer with the lattice mismatch of around 8%. As shown in Fig.1, two heterostructures with reversed electric polarization of MoSSe layer are presented as MoSSe−MoTe
2 and MoSeS−MoTe
2, forming the Se/Te and S/Te interface, respectively. The lattice parameter of MoSSe/MoTe
2 heterostructure is calculated to be 3.36 Å, suggesting a tiny tensile (compression) in the MoSSe (MoTe
2) layer. Besides, the interlayer distances of MoSSe−MoTe
2 and MoSeS−MoTe
2 are 3.31 Å and 3.18 Å (see Tab.1), indicating the polarization-controlled interlayer interaction. To evaluate the thermal stability of heterostructure, the formation energy (
Ef) is computed by,
=
, where
EH,
and
EMoSSe represent total energies of the heterostructure, MoTe
2 and MoSSe monolayer, respectively. As shown in Tab.1,
Ef is calculated to be −8.92 meV for the MoSSe−MoTe
2 heterostructure and −6.77 meV for the MoSeS−MoTe
2, indicating that the formation of the heterostructure is energetically favourite.
To understand the behaviour of the photoelectric response of constructed polarized heterostructure, we next investigated their electronic characters. All the band structures are calculated based on the hybrid functional (HSE06) to obtain a more accurate bandgap. As shown in Fig.2, MoSSe and MoTe
2 monolayers are direct semiconductors (at K point) with the bandgap of 2.10 and 1.60 eV, respectively, close to the experimental values (Tab.1) [
39,
40]. Then, we calculated the electrostatic potential of the separated layers along the
z direction [Fig.2(c)]. Due to the polarized structure, MoSSe monolayer has two work functions, namely −6.40 eV for S surface and −5.63 eV for Se surface. Besides, the work function for MoTe
2 monolayer is calculated to be −5.11 eV, slightly higher than those of MoSSe monolayer. When forming the heterostructure, different work functions will have impact on the interfacial carrier redistribution and built-in electric field, thus, manipulate its electronic properties. Fig.2(d) presents the potentials of valence band maximum (VBM) and conduction band minimum (CBM) of MoSSe and MoTe
2 monolayers. The gap between VBM of MoTe
2 layer and CBM of MoSSe layer narrows for S/Te interface of the MoSeS−MoTe
2 heterostructure [0.03 eV as detailed in Fig. S1 of the Electronic Supplementary Materials (ESM)] in comparison with Se/Te interface of the MoSSe−MoTe
2 (0.71 eV), due to the large difference on interlayer electrostatic potentials of S−Te and Se−Te of the polarized heterostructure. This is also confirmed by the orbital-resolved band structure and band decomposed charge density of the heterostructures, as given in Fig.2(e) and (f). In both MoSSe−MoTe
2 and MoSeS−MoTe
2 heterostructures, the valence band (VB) is mainly contributed by Mo orbitals of MoTe
2 layer, while the conduction band (CB) originates from the Mo orbitals of MoSSe layer. The location of VBM and CBM in MoSSe/MoTe
2 heterostructure meets the requirement of type-II or z-scheme band alignment, which remains to be further investigated by considering the interlayer charge carrier redistribution.
The charge density difference (Δ
ρ) can be given by
, where
,
and
are calculated charge densities of heterostructure, MoTe
2 and MoSSe monolayers, respectively. As plotted in Fig.3, the electrons are mainly depleted from the MoTe
2 layer, while accumulated at Se or S atomic layer in both heterostructures, indicating an interlayer electron transfer from Te surface to S or Se surface. Bader analysis confirms that 0.03
e and 0.05
e electrons transfer occurs at the interface of MoSSe−MoTe
2 and MoSeS−MoTe
2 heterostructures, respectively. Obviously, the interlayer charge distribution of a polarized heterostructure determines by both the electric polarization of the MoSSe and the electronegativity between the surface atoms, which also has impacts on the transportation of the excited electrons and holes [
25]. Since the electronegativity of Te atom is lower than that of S or Se atom, the direction of the electron transfer in such a heterostructure cannot be changed by the reversal of the electric polarization of MoSSe.
The interlayer electron transfer in MoSSe/MoTe
2 heterostructure gives rise to an electrostatic difference at the interface (Fig. S2 of the ESM), suggesting a built-in electric field pointing from the MoTe
2 layer to MoSSe layer. The effective electric potential difference of the heterostructure is given in Fig.4 by subtracting the potential of MoSSe and MoTe
2 layers from the MoSSe/MoTe
2 heterostructure. Accordingly, the built-in electric field (
Ebuilt-in) at the interface can be obtained by
, where d
V2−1 is the potential difference of two interfacial layers, and d
z represents the interlayer distance. Therefore, the calculated
Ebuilt-in is 0.11 and 0.16 V/Å for MoSSe−MoTe
2 and MoSeS−MoTe
2 heterostructures, respectively, relying on the polarization-manipulated interlayer interaction [
41,
42]. Such built-in electric field will also have impact on the transport of photogenerated electrons and holes. Consequently, excited holes of the MoTe
2 layer are allowed to transfer at the interface, as is the transport of excited electrons of the MoSSe layer, as illustrated in Fig.4(c). It indicates the charge carrier recombination at both Te/Se and Te/S interface. However, excited electrons of the MoTe
2 and holes of the MoSSe layer are forbidden to transport at the interface. Thus, the depletion of excited holes (electrons) of the MoTe
2 (MoSSe) layer benefits to prolong the lifetime of its excited electrons (holes), achieving the separation of the photogenerated charge carriers. This mechanism of photogenerated charge carrier separation in MoSSe/MoTe
2 heterostructures belongs to the direct z-scheme band alignment.
Finally, excellent light absorption is also critical for highly efficient optoelectronic applications, and here we investigated the optical properties of the MoSSe/MoTe2 heterostructure. The optical absorption spectra (from 300 nm to 1200 nm) of MoSSe and MoTe2 monolayers as well as MoSSe−MoTe2 and MoSeS−MoTe2 heterostructures are shown in Fig.5. Clearly, the absorption performance of the MoTe2 monolayer is better than that of MoSSe layer corresponding to their bandgap. After forming the MoSSe/MoTe2 heterostructure, the absorption strength significantly improved compared to the separated layers, especially at the visible light region (400−750 nm), due to the interfacial interaction. The peaks observed at 450, 600 and 700 nm of MoSSe/MoTe2 heterostructure correspond well with the absorption peaks of single-layer MoSSe and MoTe2, respectively, indicating a red or blue shift for their absorption peaks after forming the heterostructure. Such a shift is mainly due to the regulation of the bandgap caused by interlayer charge redistribution when forming the polarized heterostructure. Notably, a new absorption peak appears at around 900 nm in the MoSeS−MoTe2 heterostructure, ascribing to the interlayer optical transition. These indicate an enhanced efficiency of solar light absorption via heterostructure engineering.
4 Conclusions
In conclusion, by using the first-principles calculation, we investigated a new z-scheme 2D polarized heterostructure (MoSSe/MoTe2) with intriguing electronic and optical properties, rendering them potential materials platform for advanced optoelectronic applications. The existence of inherent electric polarization in the Janus MoSSe monolayer endows the heterostructure with the polarization-controlled optoelectronic response. The bandgap of the MoSSe−MoTe2 is 0.71 eV, while it changes to 0.03 eV by reversing the polarization. Meanwhile, the interlayer electron transfer, built-in electric field, and light absorption are also various with the different polarization. Most importantly, the MoSSe/MoTe2 heterostructure possesses z-scheme band alignment, mainly due to the contribution of its VB and CB and the direction of built-in electric field. The excited holes in the MoTe2 layer can easily recombine with the excited electrons in the MoSSe layer, as a result, the photogenerated holes and electrons will remain in the MoSSe and MoTe2 layer, respectively. Furthermore, the light absorption is also improved by forming the heterostructure compared to the separated layers. The improved performance on photogenerated carrier separation and light absorption makes the polarized MoSSe/MoTe2 heterostructure a promising candidate for optoelectronic nanodevices.