Photoluminescence Enhancement by Band Alignment Engineering in MoS2/FePS3 van der Waals Heterostructures.

Single-layer semiconducting transition metal dichalcogenides (2H-TMDs) display robust excitonic photoluminescence emission, which can be improved by controlled changes to the environment and the chemical potential of the material. However, a drastic emission quench has been generally observed when TMDs are stacked in van der Waals heterostructures, which often favor the nonradiative recombination of photocarriers. Herein, we achieve an enhancement of the photoluminescence of single-layer MoS2 on top of van der Waals FePS3. The optimal energy band alignment of this heterostructure preserves light emission of MoS2 against nonradiative interlayer recombination processes and favors the charge transfer from MoS2, an n-type semiconductor, to FePS3, a p-type narrow-gap semiconductor. The strong depletion of carriers in the MoS2 layer is evidenced by a dramatic increase in the spectral weight of neutral excitons, which is strongly modulated by the thickness of the FePS3 underneath, leading to the increase of photoluminescence intensity. The present results demonstrate the potential for the rational design of van der Waals heterostructures with advanced optoelectronic properties.

Introduction

In the past decade, two-dimensional (2D) crystals have attracted the attention of a broad community of chemists, physicists, and material scientists due to their novel mechanical, electrical, and optical properties when thinned down to just a few atomic layers.[1−7] The direct gap and photoluminescent properties of single-layer 2H TMDs have facilitated their use as the active media of optoelectronic devices.[3,8−10] Recently, a growing interest in a new family of 2D compounds has emerged, namely the transition metal chalcogenophosphates, with the general formula MPX3 (where M is a transition metal, P is phosphorus, and X is a chalcogen). MPX3s have been explored in terms of their antiferromagnetic phase transition,[11−19] photo-response,[20−26] and promising applications in spintronics.[27−32]

A fascinating perspective of the field of van der Waals materials is the endless possibilities of combining and modifying their properties by stacking different types of 2D materials in heterostructures with an atomically sharp heterointerface. When two materials with different chemical potentials are brought close, charge carriers distribute across the interface until electrostatic equilibrium is reached. This will be conditioned by the relative energy band alignment between the Fermi levels, the band onsets, and the interface quality between the two materials. The study of band alignment and charge transfer across heterostructures containing single-layer semiconducting TMDs is a powerful approach to tailor their optical and electronic properties. Hence, through the proper selection of the 2D materials, it is possible to engineer the electronic and optical properties of the materials involved.

In the case of single layers of doped semiconducting TMDs, such as MoS2, charge transfer has a remarkable influence on its photoluminescence emission (PL).[8−10] Indeed, a strong enhancement of MoS2 PL of about 2 orders of magnitude due to charge transfer and dipolar interactions with the surroundings has been reported.[33] However, most of the works where these observations are reported include solution-processed functionalization methods. Several works have shown how the photoluminescence yield of these 2D materials can be strongly enhanced by molecular adsorbates[34−36] and acid treatment.[33,34] However, in heterostructures of stacked 2D materials, charge transfer seems to be less efficient across the van der Waals barrier in terms of enhancement of the PL intensity of single-layer TMDs. While an enhancement of photoluminescence has been observed in certain heterostructures with a type I band alignment, type II band arrangements usually lead to a quench of light emission.[37−39] Nevertheless, a fast and efficient photo-induced electron–hole dissociation into adjacent layers of a 2D heterostructure notably reduces the probabilities of exciton recombination in their constituent materials and, thus, causes a dramatic drop in the PL emission of these systems.[8−10,40−42] Besides charge transfer, other tuning knobs for PL modulation of single-layered materials are based on strain engineering[43−46] and the application of external back-gate electric fields.[47]

In this work, we take advantage of the strong p-type character of intrinsic FePS3 semiconductor and the optimal energy band alignment with n-type one-layer (1L) 2H-MoS2 to build vertically stacked MoS2/FePS3 heterostructures with efficient charge carrier transfer and improved light emission properties. At room temperature, the intensity of the photoluminescence of MoS2 increases, and the emission peak is blue shifted according to an increase of excitonic versus trionic recombination. Also, a remarkable increase in defect-bound exciton emission is observed at low temperatures. All these observations point to a scenario where a high proportion of the free electrons in the single-layer MoS2 is transferred to the FePS3. The efficiency of this transfer, only comparable to the adsorbates case, leads to an almost full depletion of the MoS2 layer, which is followed by the narrowing and raising of the PL emission. We show how these effects strongly depend on—and can be tuned by—the thickness of the FePS3 layer.

Experimental Results

Figure a shows one of the fabricated heterostructures consisting of a monolayer of MoS2 transferred onto a multilayer FePS3 flake (see Methods for fabrication details). The PL spectrum of the fabricated heterostructure has been measured at room temperature under a 532 nm laser excitation and compared with the PL emission of a control sample (1L MoS2 flake deposited directly onto the 300 nm SiO2/Si substrate) (Figure b). The two emission peaks corresponding to A (1.84–1.9 eV) and B (2.01–2.04 eV) excitons in 1L MoS2 are present in both PL spectra. These two emission peaks come from the recombination of electrons in the conduction band with holes in the spin–orbit split valence bands in monolayer MoS2.[2] We observe that the PL spectral shape changes depending on the material where the MoS2 monolayer lies: the PL emission associated with exciton A coming from the heterostructure is brighter and narrower with an intensity about two times higher and clearly blue shifted if compared to 1L MoS2 directly deposited on SiO2. We also observe a drop in the relative spectral weight associated with exciton B in the heterostructure spectrum when compared to the control sample (see Supporting Information Section S3). Because this signal is considerably weaker, we focus our analysis on the evolution of the A exciton peak.

Figure 1

(a) Optical microscopy image of the fabricated heterostructure onto a SiO2/Si substrate, where the single-layer MoS2 (1L) is placed on top of a multilayer FePS3 flake. The green dot in (a) indicates the zone of the heterostructure where the spectrum shown in (b,d) was taken. The scale bar in (a) corresponds to 10 μm. (b) Photoluminescence spectra taken at the 1L-MoS2/FePS3 heterostructure (green curve), which is shown in (a), and at a control sample (orange curve), 1L-MoS2, which is directly deposited on the SiO2/Si substrate. (c,d) Analysis of the photoluminescence spectral shapes for the as-prepared MoS2 monolayer and 1L MoS2/FePS3 heterostructure, respectively, assuming three peaks with Lorentzian functions: trion (X–) and neutral excitons (X0 and B).

To unveil the origin of these PL spectral changes, we decompose the PL peak coming from exciton A into two subexcitonic contributions: the neutral exciton X0 (an electron and a hole bounded) and the trion or negatively charged exciton X– (two electrons and a hole bounded).[48] For the case of as-prepared 1L MoS2 (Figure c), the contribution of the negative trion peak (X–), located at ∼1.84 eV (red curve), prevails over the PL spectral weight of the neutral exciton (X0), located at ∼1.88 eV (purple curve). This dominant recombination mediated by trions (X–) reveals a heavily n-type doped monolayer MoS2, which is consistent with previous observations.[35]

In contrast to the MoS2 monolayer on SiO2, the PL emission from the heterostructure (Figure d) is clearly dominated by the neutral exciton peak (X0) at ∼1.88 eV, due to the presence of FePS3. Considering the p-type nature of FePS3,[20,21] the experimental results suggest a strong charge transfer of electrons from the MoS2 monolayer toward the FePS3 flake, when these two are interfaced, and consequent depletion of the TMD layer. This experimental observation highly resembles the strong tunability and enhancement of the PL properties in monolayer TMDs via chemical doping.[33−36]

The equilibrium among exciton, trion, and free-electron populations in MoS2 can be viewed as a simple chemical reaction: X0 + e ↔ X–, where the rate equality of the forward and reverse reactions are described by a mass action law model.[35]

The population of the three species is then governed by a rate equation , where N and N are the number of trions (X–) and excitons (X0), respectively, while KT and ne are the rate constant for trions and the free electron density, respectively (see Supporting Information Section S4 and ref (35) for details).

The ratio between the contributions (area under the curve) of the trion (A) and exciton (A) is expected to be proportional to their respective populations in equilibrium:

Similarly, the emission ratios for the heterostructure and control should then be proportional to the respective populations in the heterostructure and control samples

This provides a first estimation of the electron depletion in the 1L-MoS2 due to charge transfer when placed on FePS3. Under this assumption, the calculated relative electron concentration, (nelcon – nelhet)/nelcon, changes proportionally to (rcon – rhet)/rcon for all the fabricated heterostructures and is within the range of ∼81–∼99%, reaching ∼95% for the specific heterostructure shown in Figure (see Supporting Information Sections S4 and S5 for further details).

Moreover, assuming the values reported in the literature for the effective masses of electrons, excitons, and trions and the trion binding energy, as well as the radiative decay rates of trions and excitons at room temperature,[35] we can obtain approximated values for actual electron densities in MoS2 in both samples (see Supporting Information Section S4 for details). Thus, the estimated electron densities of the 1L-MoS2 flake in the control sample and in the heterostructure are ∼4.8 × 1013 and ∼3.0 × 1012 cm–2, respectively. These results support our hypothesis about an efficient transfer of electrons in 1L-MoS2 toward FePS3.

While similar results have been obtained by chemical treatments or molecular physisorption on single-layer TMDs, our observation is something unique in the case of van der Waals type II heterojunctions, where typically the PL emission is strongly quenched due to spatial electron–hole separation and/or the formation of interlayer excitons.[8−10,42]

To obtain further insight into the origin of this efficient charge transfer, we determine the band onset energies for FePS3 and 1L MoS2 separately. To do this, we performed ultraviolet photoelectron spectroscopy (UPS) in bulk FePS3. The deduction of the work function for bulk FePS3 is obtained from the UPS spectrum (Figure a) as ϕ = ℏω – SEC ≈ 4.9 eV, where ℏω is the excitation energy (He I: 21.22 eV), and SEC is the energy cut-off of the secondary electron region of the spectrum obtained from a linear fit to the data[49,50] (see inset of Figure a). The work function for bulk FePS3 deduced in our work is slightly larger than two recently published works, reporting values of ∼4.7 and ∼4.17.[51,52] Nevertheless, we have also obtained a similar work function value for bulk FePS3 through Kelvin probe force microscopy (see Supporting Information Section S7).

Figure 2

(a) UPS spectrum of bulk FePS3 using He I (ℏω = 21.22 eV) as a monochromatic excitation source, where emission peaks coming from valence band (VB) states and secondary electrons (SEC) can be observed. The zero binding energy indicates the Fermi level. Inset: Zoom-in of the secondary electron cut-off (SEC). (b) Experimentally estimated band diagram of the 1L MoS2/ML FePS3 junction forming a type II heterostructure. (c) Side view of the atomic MoS2/FePS3 heterointerface and its corresponding charge transfer representation using an isovalue equal to 0.05 in the XCrySDen package.[56] The difference between the charge density and the superposition of atomic densities shows the gain (red) and depletion (blue) zones along the heterostructure, evidencing the absence of gain and depletion zones at the heterointerface. (d) Charge transfer in the heterostructure, relative to a control sample, obtained from the analysis of photoluminescence spectra as a function of the thickness of the FePS3 flake underneath.

On the other hand, electron acceptor levels in FePS3 have been postulated to arise from Fe2+ defects.[53] By fitting the conductivity as a function of temperature to an Arrhenius model for multilayer flakes of FePS3, we obtain an activation energy of ∼0.37 eV, which is in the range of the electron acceptor energies reported for bulk FePS3[53] (see Supporting Information Section S8) and UPS valence band determination (see Supporting Information Section S6). Assuming this and considering that the bandgap energy of a several-layer FePS3 flake is ∼1.23 eV, previously deduced from photo-responsivity measurements,[21] it is possible to draw a diagram of the energy band alignment for an FePS3 flake (Figure b).

Taking into account the energy values for the electron affinity and bandgap for monolayer MoS2 reported in the literature,[54] ∼4.3 and ∼1.89 eV, respectively, and considering a work function of ∼4.8 eV for exfoliated 1L MoS2 measured in ambient conditions,[55] a diagram of the energy band alignment for the 1L MoS2/FePS3 heterostructure has been built (Figure b). The justification for using work function values obtained in vacuum and in air for FePS3 and 1L MoS2, respectively, falls on the fact that the MoS2 monolayer may act as an encapsulating material for the area of FePS3 on which it is deposited. We indeed observe that the PL is quenched if the samples are not prepared under a controlled atmosphere, whereas the PL enhancement of heterostructures prepared in a controlled environment can be observed even after months of preparation.

In Figure c, the valence band maximum (VBM) of FePS3 is located above the VBM of 1L MoS2, whereas the conduction band minimum (CBM) of 1L MoS2 is below the CBM of FePS3. Therefore, for the van der Waals heterojunction, the VBM and CBM are localized on FePS3 and MoS2, respectively, confirming a type II heterointerface.The exact location of the bands for FePS3 has an estimated error of about ±0.2 eV due to the uncertainty in the determination of the UPS slope and the lack of an exact determination of dopant and free carrier densities. There is also a similar range of variation in the reported energy positions for the MoS2 levels. Even taking those uncertainties into account, the qualitative description of a type II band alignment holds. In this scenario, the observed depletion of the MoS2 layer must arise from the transfer of free electrons from the conduction band of 1L MoS2 to the available states in the FePS3 valence band. Moreover, we observe a small increase in the exciton lifetime (see Supporting Information Section S12) associated with the increase of its relative spectral weight in agreement with other works.[33,38,39] Furthermore, the fact that photoluminescence quenches in heterostructures prepared under a normal atmosphere (see Supporting Information Section S11) indicates that mechanisms requiring atomic proximity are responsible for the observed PL changes. This allows us to discard other leading mechanisms such as long-range energy transfer in our samples.

In the absence of dopants, charge transfer would be very limited by the unfavorable conditions provided by a pristine heterostructure in which both materials end up in sulfur atoms. To demonstrate this, we have carried out Hubbard-corrected DFT calculations (see computational details in Supporting Information Section S9) followed by a charge transfer Bader analysis. For simplicity, we have focused on a system formed by a bilayer MoS2/FePS3 (Figures c and S9). The Bader analysis, in agreement with the charge transfer analysis obtained from the ab initio calculations, indicates that only a small portion of the charge is transferred between the two stacked materials (see Figure c and details in Table S3) and that the charge redistribution occurs only inside each material. We conclude that for the case where FePS3 and MoS2 are intrinsic semiconductors, charge transfer between both materials is negligible. Then, we provide an estimation of the band alignment of bulk FePS3 and single-layer MoS2 using an ML slab model (see Computational Details in Methods). Work function values obtained from DFT calculations for defect-free intrinsic crystals of MoS2 and FePS3 yield a type I band alignment, regardless of the thickness of FePS3 (see Supporting Information Section S9, Figures S10 and S11). To provide a more realistic picture, which contemplates the existence of dopants, we calculate the electronic structure of MoS2 in the presence of S vacancies using a 4 × 4 × 1 supercell (see Section S10, Figure S12). This picture results in a type II band alignment between FePS3 and vacant MoS2 (see Section S10, Figure S13) and provides a closer description of the experimental results, suggesting, due to the chemical similarity, the presence of sulfur vacancies also in FePS3. These can adsorb oxygen atoms and induce oxidation of Fe2+ to Fe3+ that facilitates charge transfer at the interface.

We conclude that the strong electron acceptor character of naturally doped FePS3 combined with the natural electron doping of MoS2 are the key features, together with a favorable band alignment, that facilitate the observed charge transfer. Charge conservation requires that electron depletion in MoS2 is accompanied by a similar amount of hole depletion at FePS3. This creates a built-in potential across the junction that, in our case, acts as an energy barrier preventing the nonradiative recombination of photogenerated carriers and, thus, preserving the excitons and their photoluminescent recombination in MoS2. Furthermore, the fact that the VBM of FePS3 and the CBM of MoS2 have different momentum (see band structure calculations in Section S9), prevents the formation of interlayer excitons.

While charge transfer in the MoS2 is limited to a single layer, in the case of FePS3, hole depletion can extend over several layers of the material. Indeed, we find that the thickness of FePS3 flakes limits the charge transfer. For FePS3 flakes with thicknesses above 100 nm, we observe a PL enhancement from two to four times larger than in the control sample (see Section S5, Figure S6), and depletion of MoS2 carriers larger than 95%. These are unusually high values, both for enhancement and depletion, in the case of van der Waals heterostructures. This is illustrated by comparing the emission of several samples with different FePS3 thicknesses which reveals a clear dependence on the estimated amount of charge transferred between MoS2 and FePS3 (Figure d). Roughly speaking, we can attribute the thickness dependence to a reduced number of acceptors available in the p-doped material compared to thicker FePS3. Also, while depletion at MoS2 must necessarily occur at the single layer, the interface equilibrium at the FePS3 side can result in an extended depletion layer, which can be of interest for photovoltaic or photodetection purposes. The reported photogating effects in FePS321 could also play a role in the dynamic enhancement of the MoS2 depletion upon illumination.

To obtain more comprehensive details on the effects of charge transfer from the PL of 1L MoS2/FePS3 van der Waals heterostructure, temperature-dependent measurements have been carried out from 180 to 10 K (Figure a) in one of our heterostructures and contrasted with the low-temperature PL emission from the control sample (Figure d). In our analysis, we focus on the three more prominent PL peaks, which are labeled as D, X–, and X0 in Figure a,d, and obviate the peak related to exciton B (located at ∼2.1 eV) (see fit details in Supporting Information Section S13).

Figure 3

(a–c) Temperature evolution of photoluminescence within the range of 10–180 K in steps of 5 K in the heterostructure sample (a) PL spectra. (b) Peak energy positions extracted from a fit of the data to a multipeak model (see Supporting Information Section S13) as a function of temperature. The solid line represents the fit to a standard semiconductor model. (c) Peak areas. (d–f) Photoluminescence as a function of temperature in the control sample. (d) PL spectra. (e) Peak energy positions. (f) Peak areas.

Clearly, peak D is evident in the heterostructure in the full range of temperatures, whereas in the control sample, it starts to be more appreciable only below 80 K. This peak, moving between 1.6 and 1.8 eV depending on temperature, has been observed previously in the PL emission of single-layer MoS2 and has been attributed to the radiative recombination of excitons bounded to intragap defects formed from sulfur vacancies[57,58].

We observe that for both samples, control and heterostructure, the positions of the three peaks, D, X–, and X0, are all blue shifted as temperature diminishes (Figure b,e). This is attributed to a decreased electron-phonon interaction as well as to small changes in the bonding length.[59] To quantify the blue shifting of the PL emission in the heterostructure and control samples when decreasing temperature, a standard semiconducting bandgap model has been used (see ref (60) and Section S14).

The parameters obtained from fitting the evolution of peak energy positions with temperature to the model are summarized in Table S4 and are consistent with the previous works[36] for the case of the two excitonic peaks X– and X0. From these values, the trion binding energies for the heterostructure and control samples are similar, being ∼30 and ∼36 meV, respectively. We attribute the small difference in binding energies between the samples to the different local dielectric screening of the Coulomb interaction in the MoS2 monolayers.[61] On the other hand, the larger energy shift of peak D with varying temperature is also manifested through a higher electron–phonon coupling strength in contrast with the one obtained for the two excitonic peaks, X– and X0, in both samples (see fitted values for parameter S in Table S4).

There is also a temperature-dependent change in the relative spectral weight between X– and X0 emission peaks (Figure c). This gradual change of trion-exciton contribution is also observed in the control sample (Figure f). This observation has been previously attributed to electrons escaping their trion-bound state owing to thermal fluctuations.[62]

The spectral weight of the PL peak associated with defect-bound excitons increases significantly with decreasing temperature. This behavior has been observed in different single-layer TMDs,[4,63,64] follows an Arrhenius trend with activation energies in the order of tens of meV (see Section S15), and has been attributed to an increase of nonradiative recombination processes with temperature[4,63] or to a possible charged nature of bound excitons.[64]

More interestingly, the remarkable increase of the defect peak in the heterostructure corroborates the abovementioned scenario of electron transfer. In the work presented by Greben et al.,[63] a law of mass action is introduced to describe the equilibrium between the density of free excitons and exciton bound by defects: X0 + d → D. The rate between those densities is, in this case, governed by the density of unoccupied dopant levels in MoS2where ND and N are the density of defect-related excitons and trions, respectively, while KD and nD are the rate constants for defect-bound excitons and the concentration of unoccupied in-gap defect levels, respectively.

Similarly, the ratio between free carrier density in the heterostructure and control samples can be attributed to the proportion in spectral weight between defect and exciton emission peaks, which is directly related to their respective populations

This analysis shows that in the heterostructure and below 100 K, there are 20–25 times more unoccupied defects than in the control sample (Figure c). This is compatible with the electron depletion of MoS2, which in ref (63) is achieved by the application of an external electric field and is caused here by the acceptor character of FePS3. This was already qualitatively observable by the fact that at 180 K, a defect peak is present in the heterostructure but not in the control sample. Because of charge transfer, the photoluminescence of MoS2 in the heterostructure resembles that of a semiconductor with a lower degree of doping than in the case of the control sample (Figure a–c).

Conclusions

In summary, our study corroborates an efficient electron transfer from the n-doped MoS2 monolayer to the p-doped multilayer FePS3 flake by combining optical spectroscopy, UPS, ab initio calculations, low-temperature transport, and PL measurements. The charge transfer signatures obtained in the 2D heterostructure via PL measurements at room temperature are comparable to the ones achieved via chemical functionalization, where preservation or enhancement of the PL efficiency is accomplished. We attribute the charge transfer and the preservation of PL to the very favorable band alignment of the heterostructure. Our results suggest that the light emission properties of single-layer, n-type TMDs can be improved not only in some type I semiconductor heterostructures, but also in type II arrangements with indirect, smaller gap p-type semiconductors.

The enhancement and narrowing of the PL emission could inspire the design of future highly efficient light-emitting diodes based on band alignment engineering of heterostructures composed of atomically thin MoS2. Through a careful analysis of several heterostructures, we are able to track the dependence of the number of electrons removed from single-layer MoS2 as a function of the thickness of the FePS3 underneath. Thus, charge transfer and, consequently, PL can be easily tuned by a proper thickness selection of FePS3, enabling convenient control of optical and electrical properties of atomically thin MoS2. The singular PL tunability of the system invites us to continue exploring this 2D heterostructure as an optoelectronic material, where a meticulous study of the leading mechanisms between electron–hole recombinations and/or dissociations can have an impact on the efficiency of photodetectors, photovoltaic cells, light-emitting diodes, or electroluminescent junctions based on 2D materials.

Methods

Fabrication of Vertical Single-Layer MoS2/MultiLayer FePS3 Heterostructures

Commercially available MoS2 (SPI Supplies) and lab-grown FePS3via chemical vapor transport[65] were mechanically exfoliated onto transparent polydimethylsiloxane (PDMS) substrates. Optical microscopy, micro-reflectance, and Raman spectroscopies enabled us to identify the thickness of FePS3 and MoS2 flakes (see Supporting Information Sections S1 and S2). After identification, the selected flakes were deposited onto a 300 nm-thick SiO2/Si substrate via a deterministic, dry transfer method[66] to form vertically stacked heterostructures. The exfoliation of FePS3 flakes and the heterostructure fabrication was performed in an inert Argon atmosphere.

Photoluminescence Characterization

PL measurements at room temperature were performed using a commercial Raman microscope (Jasco NRS-5100) using an excitation line of 532 nm, with a laser spot of ∼1.5 μm diameter and a total power of 60 μW. Low-temperature micro-PL measurements were carried out using a diffraction-limited fiber in a confocal setup inserted into a pulse-tube-based closed-cycle Helium cryostat (attoDRY 2100, Attocube). A 532 nm solid-state laser was used with an irradiated laser power of approximately 100 μW at the sample.

Ultraviolet Photoelectron Spectroscopy

He I (ℏω = 21.22 eV) UPS spectra were taken on bulk FePS3 crystals. Samples were exfoliated while already mounted in the experiment chamber in order to reduce the air exposure of the surface down to a few seconds. A bias voltage of −10 V was applied to the sample in order to differentiate the secondary electron cut-off.

Computational Details

The electronic structure of MoS2/FePS3 heterostructure was calculated using the first-principles plane-wave DFT + U approach as implemented in the Quantum ESPRESSO package,[67] using a Hubbard U (on-site Coulomb repulsion) of 2.2 eV, as reported in ref (21) (see also Supporting Information Section S9 for more details). All chemical structures were fully optimized using the Broyden–Fletcher–Goldfarb–Shanno (BFGS) algorithm[68] until the forces on each atom were smaller than 1 × 10–3 Ry/au and the energy difference between two consecutive relaxation steps was less than 1 × 10–4 Ry. The Brillouin zone was sampled at least by a fine Γ-centered 4 × 4 × 1 k-point Monkhorst–Pack mesh[69] for all monolayer calculations choosing a well converged third k point according to the length of slabs. The heterostructure was set up by a 2 × 2 hexagonal supercell of single-layer FePS3, keeping the fully optimized lattice parameters from the bulk, combined with a 4 × 4 MoS2 supercell, assuming a 7.19% mismatch for the MoS2. The stacking was based on previous works with analogous materials.[70] An extended mesh of 8 × 8 × 2 k-points was necessary to determine the charge transfer between the layers and converge the charges during the Bader analysis. The work function was determined for MoS2 and FePS3 monolayers and bulk FePS3, which was simulated with slabs formed by 4 and 6 layers, being already converged in the 4-layers slab calculation. To evaluate the presence of defects in the work function of MoS2, we built up a 4 × 4 × 1 supercell to isolate a S vacancy.