Excitonic Evolution in WS 2 /MoS 2 van der Waals Heterostructures Turned by Out-of-Plane Localized Pressure

: In this study, we explore the exciton dynamics in a WS 2 /MoS 2 van der Waals (vdW) heterostructure under varying pressures by integrating a laser-confocal photoluminescence (PL) spectroscope and an atomic force microscope (AFM). For the WS 2 /MoS 2 heterostructure, the exciton emission belonging to MoS 2 is too weak to be distinguished from the PL spectra. However, upon contact with a Si probe, the emission intensity of WS 2 excitons significantly decreases from 34,234 to 6560, thereby matching the intensity level of MoS 2 . This alteration substantially facilitates the exploration of interlayer excitonic properties within the heterostructures using PL spectroscopy. Furthermore, the Si probe can apply out-of-plane localized pressure to the heterostructure. With increasing pressure, the emission intensity of the WS 2 trions decreases at a rate twice that of other excitons, and the exciton energy increases at a rate of 0.1 meV nN − 1 . These results elucidate that the WS 2 trions are particularly sensitive to the out-of-plane pressure within a WS 2 /MoS 2 vdW heterostructure.


Introduction
Layered two-dimensional (2D) materials have attracted considerable attention due to their excellent properties: unique electronic structure, tunable bandgaps, atomic-level thickness, and van der Waals (vdW) integration capability.Compared to the direct synthesis method, the mechanical transfer method for manufacturing vdW heterostructures allows for free stacking of 2D layers at any angle without restrictions from covalent bonds and lattice matching [1][2][3].Through the mechanical transfer method, different 2D materials can be combined into heterogeneous structures without any limitations in epitaxy, which greatly facilitates the assembly of high-quality vdW heterostructures and expands the exploration of their optical properties [4][5][6].vdW heterostructures have promoted the development of condensed physics, such as confirming Hofstadter butterfly topologies [7][8][9], finding thermodynamic evidence for fractional Chern insulators [10] and gate-tunable Mott insulators [11,12], and achieving unconventional superconductivity in magic-angle superlattices [13,14].Transition metal dichalcogenides (TMDCs) exhibit strong spin-orbit coupling (SOC) and high exciton binding energy, making them excellent candidates for building a vdW heterostructure platform to study novel physics.As representative TMDCs, monolayer tungsten disulfide (ML-WS 2 ) and molybdenum disulfide (ML-MoS 2 ) have emerged as promising candidates for constructing heterostructures due to their unconventional properties, including the conversion of indirect/direct bandgaps [15][16][17][18][19], valleytronics [20][21][22][23], and the large binding energies of excitons [18,24], positioning them as ideal platforms for exploring the quantum effect in condensed matter systems.The interlayer interaction is one of the most important properties of MoS 2 /WS 2 heterostructures [25][26][27].Wang et al. experimentally observed that the photoexcited holes transferring from MoS 2 to WS 2 take place within an ultra-short time: 50 fs [25].Lei et al. reported on the interlayer hole-transferring process and the formation of spatially interlayer excitons with longer lifetimes in MoS 2 /WS 2 heterostructures [26].Yu et al. determined that Raman scattering and their underlying interlayer interactions depend on the twist angle in twisted MoS 2 /WS 2 heterostructures [27].Understanding the interlayer interactions in vdW heterostructures under different conditions is crucial for optimizing their performance.
The precise tuning of bandgaps, photoluminescence energy, and quantum yield at the nanoscale greatly expands the application prospects of two-dimensional materials in micro-optoelectronic devices.Various methods have been reported for tuning the electronic structure of 2D materials, including electric field modulation [28,29], chemical modification [30,31], magnetic field modulation [32], and mechanical strain [15,33,34].It is noteworthy that the strong planar mechanical strength and small bending stiffness of 2D materials, coupled with their reversibility compared to other tuning methods, make them particularly suitable for strain engineering [35,36].For example, Castellanos-Gomez analyzed the strain tunability of the exciton properties of monolayer and few-layer MoS 2 , MoSe 2 , WS 2 , and WSe 2 by bending the substrate [37].Steele et al. employed exciton imaging to investigate the band structure of a monolayer MoS 2 under macroscopic strain and observed a strain-induced bandgap reduction and exciton funneling phenomenon [38].Although strain engineering has been extendedly explored in TMDC monolayers, multilayers, and heterostructures, most studies primarily focus on macroscopic strain, while single-point pressure is rarely studied.Due to the anisotropy of monolayer TMDCs, the regulation of excitons by macroscopic in-plane strain and out-of-plane localized strain is completely different.Therefore, a quantitative investigation into the single-point tuning of vdW heterostructures is necessary to study the exciton effect.
In this work, we develop a quantitative analysis of the excitonic evolution features in WS 2 /MoS 2 vdW heterostructures under controlled pressure conditions.ML-WS 2 and ML-MoS 2 samples are synthesized through chemical vapor transport (CVT) and the chemical vapor deposition (CVD) method, respectively, and the WS 2 /MoS 2 heterostructure is fabricated using a wet transfer technique, enabling precise control of the heterostructure composition and layer stacking.We integrate a spectroscopy measurement apparatus with an atomic force microscope (AFM) to accurately control and monitor the applied pressure.The observations reveal a diminution in PL intensity concomitant with an applied pressure across both the monolayer and ensemble heterostructure.Notably, within the heterostructure, a blue shift in the excitonic peak position under pressure articulates a modifiable electronic structure, suggesting that the trions are more sensitive than the natural excitons to the dielectric environment.This quantitative pressure tuning of the excitonic evolution offers a novel approach for modulating the electronic structure of vdW heterostructures and expands their potential applications in nanoscale light sources, nanophotonic switches, and nano-integrated excitonic circuits.

Synthesis of ML-WS 2 and ML-MoS 2
ML-MoS 2 and ML-WS 2 were both synthesized in an open-type double-temperaturezone tube furnace (MTI OTF-1200X-II, HF-Kejing, Hefei, China).Prior to the synthetic procedure, a SiO 2 /Si wafer (SiO 2 thickness is 300 nm) with dimensions of 20 mm × 20 mm was subjected to ultrasonic cleaning in ethanol, isopropanol, and deionized water for 15 min each in sequence.Afterward, the SiO 2 /Si wafer was dried with N 2 gas flow and served as the substrate for the growth of the 2D materials.For the synthesis of ML-MoS 2 , 0.5 g of S powder and 10 mg of MoO 3 powder were placed in Zone I and Zone II, respectively.The distances between the S and MoO 3 were 30 cm, and the as-cleaned SiO 2 /Si wafer was Appl.Sci.2024, 14, 2179 3 of 13 placed upside down on the quartz boat with MoO 3 powder.The system was evacuated to 10 Pa and then backfilled with N 2 to atmospheric pressure at a flow rate of 70 sccm.Zones I and II were separately heated to 350 • C and 760 • C at a rate of 10 • C min −1 and maintained at stable temperatures for 20 min to facilitate growth.Then, the system underwent natural cooling to room temperature.
To synthesize ML-WS 2 , a growth source of 20 mg of WS 2 powder (99.9% metallic base) was positioned near the central section of the tube.The pre-cleaned SiO 2 /Si wafer was then placed 2.2 cm behind the WS 2 powder.The tube was sealed and first evacuated to a pressure of 10 Pa, followed by the introduction of Ar gas until atmospheric pressure was reached.This process was repeated three times to ensure the complete removal of air.Zone I was heated to 1000 • C at a rate of 10 • C min −1 and kept for 1 h under the protection of 100 sccm of Ar gas flow to grow large ML-WS 2 triangles.Finally, the tube furnace was naturally cooled to room temperature to prevent the triangle from breaking.
The spin time and spin rate were set to first spin at 600 rpm for 10 s to completely cover the wafer and then spin at 3000 rpm for 60 s to form a PMMA thin film on the surface.The PMMA-coated wafer was then heated on a hot plate at 110 • C for 15 min to tighten the fitness between the ML-WS 2 and PMMA film, followed by immersion in a 2 M NaOH solution at 120 • C for 30 min to etch the SiO 2 layer.Secondly, the PMMA/WS 2 film floating on the surface of the NaOH solutions was picked up with a clean slide and washed in deionized water several times to remove the residual SiO 2 .The cleaned PMMA/WS 2 film was carefully transferred to the upper movable sample holder on the 2D material transfer platform with the WS 2 side facing downward, while the MoS 2 -grown SiO 2 /Si wafer was fixed on the lower heating sample holder with the WS 2 side facing upward.The five-axis positioning stage was adjusted to align the PMMA/WS 2 parallel to the MoS 2 /SiO 2 /Si and the movable sample holder was pushed down vertically until the film was fully attached to the wafer, ensuring that there were no visible bubbles.The lower heating sample holder was heated at 100 • C for 30 min to reinforce the adherence between the WS 2 and MoS 2 .Finally, the wafer was immersed in acetone, IPA, and deionized water for 10 min each to remove the PMMA and dried with N 2 gas flow at room temperature to obtain WS 2 /MoS 2 on the SiO 2 /Si wafer.

Measurement
Optical images, Raman spectra, and PL spectra of the samples were acquired using a homemade laser scanning confocal PL-Raman spectrometer (LSCPRS).A laser with a wavelength of 532 nm and a power of 6 mW was utilized as the excitation source.The spot diameter was focused to ~1 µm using a long work-distance M plan fluorite semiapochromat infinity-corrected objective (LMPLFLN 50X, Olympus, Tokyo, Japan).The exposure times were 3 and 10 s for the PL and Raman spectra, respectively.The emission intensity was recorded with a thermoelectrically cooled charge coupled device detector at −60 • C, and the wavelength was distinguished by 300 and 1800 lines per mm diffraction gratings for the PL and Raman spectra, respectively.The morphological features and height of the samples were investigated through atomic force microscopy (Ntegra, NT-MDT, Moscow, Russia) equipped with a Si probe (VIT_P/TERS, 50 N/m, NT-MDT & TipsNano, Tallinn, Estonia).The actual force constant was measured to be 45.5582N m −1 .Pressure was applied to the nanoscale contact area on the sample stage through a Si tip loaded onto the AFM instrument.The pressure force (F) was calculated using F = setpoint × 0.064 nN.The focusing position of the microscope was adjusted with reference to the tip to enable optical property measurements under different pressure conditions in contact mode.

DFT Calculation
The band structure of the WS 2 /MOS 2 HS was calculated with the DFT method using VASP 5.1.For the modeling of WS 2 /MOS 2 HS, the AB-stack structure of ML-WS 2 and ML-MoS 2 with aligned metal atoms was chosen to construct a supercell including 6 atoms.The initial interlayer distance between the two monolayers was set to 3.3 Å and optimized to 3.087 Å.The electron-ion interaction was described using ultra-soft pseudopotential plane waves, and the interaction correlation energy was treated using a generalized gradient approximation (GGA) with the Periodic-Boundary-Embedding (PBE) functional form.To prevent the interaction between the supercell layers caused by the periodic boundary condition, a vacuum layer with a thickness of 15 Å was added in the c direction.During the optimization and electron calculation processes, the internal interatomic interaction force and pressure thresholds were set to 0.03 eV Å −1 and 0.05 GPa, respectively.The k-point grid of 3 × 3 × 1, the energy convergence criterion of 10 −5 eV, and the cutoff energy of 650 eV were optimized for use.

Results and Discussion
Figure 1a,b show the detailed synthetic procedures of ML-MoS 2 and ML-WS 2 , respectively.Unlike the CVD synthesis method for ML-MoS 2 , ML-WS 2 was prepared using the CVT method for easier peeling from the substrate [39].Then, the formation of the WS 2 /MoS 2 heterostructure (WS 2 /MoS 2 HS) was achieved by transferring ML-WS 2 onto ML-MoS 2 using a wet transfer technique, as illustrated in Figure 2. Typically, the formation of two-dimensional vdW heterostructures is confirmed by integrating the results from optical imaging, AFM imaging, Raman spectra, and PL spectroscopy [25,40,41].
pressure force (F) was calculated using F = setpoint × 0.064 nN.The focusing position of the microscope was adjusted with reference to the tip to enable optical property measurements under different pressure conditions in contact mode.

DFT Calculation
The band structure of the WS2/MOS2 HS was calculated with the DFT method using VASP 5.1.For the modeling of WS2/MOS2 HS, the AB-stack structure of ML-WS2 and ML-MoS2 with aligned metal atoms was chosen to construct a supercell including 6 atoms.The initial interlayer distance between the two monolayers was set to 3.3 Å and optimized to 3.087 Å.The electron-ion interaction was described using ultra-soft pseudopotential plane waves, and the interaction correlation energy was treated using a generalized gradient approximation (GGA) with the Periodic-Boundary-Embedding (PBE) functional form.To prevent the interaction between the supercell layers caused by the periodic boundary condition, a vacuum layer with a thickness of 15 Å was added in the c direction.During the optimization and electron calculation processes, the internal interatomic interaction force and pressure thresholds were set to 0.03 eV Å −1 and 0.05 GPa, respectively.The k-point grid of 3 × 3 × 1, the energy convergence criterion of 10 −5 eV, and the cutoff energy of 650 eV were optimized for use.

Results and Discussion
Figure 1a,b show the detailed synthetic procedures of ML-MoS2 and ML-WS2, respectively.Unlike the CVD synthesis method for ML-MoS2, ML-WS2 was prepared using the CVT method for easier peeling from the substrate [39].Then, the formation of the WS2/MoS2 heterostructure (WS2/MoS2 HS) was achieved by transferring ML-WS2 onto ML-MoS2 using a wet transfer technique, as illustrated in Figure 2. Typically, the formation of two-dimensional vdW heterostructures is confirmed by integrating the results from optical imaging, AFM imaging, Raman spectra, and PL spectroscopy [25,40,41].2b, showing that numerous abnormal points with random heights are unevenly distributed throughout the stacking region.The abnormal points are caused by the formation of unavoidable nanobubbles between the two monolayers during the fabrication process, and the gap distances are approximately equal to the diameter of the bubbles.Figure 2c is the height profile along the straight line indicated in Figure 2b.The presence of two steps below 1 nm in the height curve indicates that the heterostructure is composed of two monolayers pointing to ML-WS 2 and ML-MoS 2 .The height of the second step (0.68 or 0.71 nm) is slightly higher than that of the first (0.63 nm), which is in agreement with the gap of the bilayer.It is also observed that most of the bubbles are less than 2 nm in diameter, but a small number of bubbles can reach diameters of over 10 nm.The nanobubbles enlarge the interlayer distance between the two monolayers, weakening or even eliminating interlayer interactions [42,43].Thus, nanobubbles should be avoided in subsequent measurements.Figure 2a is an optical image of the WS2/MoS2 HS on the SiO2/Si substrate, displaying twist angles for WS2 and MoS2 at approximately 60°.The AFM topography image is shown in Figure 2b, showing that numerous abnormal points with random heights are unevenly distributed throughout the stacking region.The abnormal points are caused by the formation of unavoidable nanobubbles between the two monolayers during the fabrication process, and the gap distances are approximately equal to the diameter of the bubbles.Figure 2c is the height profile along the straight line indicated in Figure 2b.The presence of two steps below 1 nm in the height curve indicates that the heterostructure is composed of two monolayers pointing to ML-WS2 and ML-MoS2.The height of the second step (0.68 or 0.71 nm) is slightly higher than that of the first (0.63 nm), which is in agreement with the gap of the bilayer.It is also observed that most of the bubbles are less than 2 nm in diameter, but a small number of bubbles can reach diameters of over 10 nm.The nanobubbles enlarge the interlayer distance between the two monolayers, weakening or even eliminating interlayer interactions [42,43].Thus, nanobubbles should be avoided in subsequent measurements.
Figure 2d presents the Raman spectra of ML-WS2, ML-MoS2, and WS2/MoS2 HS (the selected points are marked with yellow, purple, and blue crosses in Figure 2b, respectively).The characteristic peaks of WS2 located at 424 and 361 cm −1 are indexed to the outof-plane A1g mode and in-plane E 1 2g mode of the monolayer WS2 [44].The out-of-plane A1g mode and in-plane E 1 2g mode of individual MoS2 are located at 410 and 392 cm −1 , respectively.It should be noted that the Raman shifts are relatively higher than those in some previous studies [45][46][47][48][49], which is induced by the extensive oxidation reactions occurring between atmospheric oxygen and the defects in MoS2 during long exposure to air [50].The distance between the two Raman peaks is ~18 cm −1 , which is consistent with previous studies and confirms the monolayer nature of MoS2 [51].The Raman spectrum of the WS2/MoS2 region appears to contain all four peaks of the two monolayers, indicating the coexistence of MoS2 and WS2.Upon stacking, the intensities of the peaks corresponding Figure 2d presents the Raman spectra of ML-WS 2 , ML-MoS 2 , and WS 2 /MoS 2 HS (the selected points are marked with yellow, purple, and blue crosses in Figure 2b, respectively).The characteristic peaks of WS 2 located at 424 and 361 cm −1 are indexed to the out-of-plane A 1g mode and in-plane E 1 2g mode of the monolayer WS 2 [44].The out-of-plane A 1g mode and in-plane E 1 2g mode of individual MoS 2 are located at 410 and 392 cm −1 , respectively.It should be noted that the Raman shifts are relatively higher than those in some previous studies [45][46][47][48][49], which is induced by the extensive oxidation reactions occurring between atmospheric oxygen and the defects in MoS 2 during long exposure to air [50].The distance between the two Raman peaks is ~18 cm −1 , which is consistent with previous studies and confirms the monolayer nature of MoS 2 [51].The Raman spectrum of the WS 2 /MoS 2 region appears to contain all four peaks of the two monolayers, indicating the coexistence of MoS 2 and WS 2 .Upon stacking, the intensities of the peaks corresponding to the WS 2 E 1 2g , MoS 2 E 1 2g , and MoS 2 A 1g modes decrease to 1605, 250, and 270, respectively.In contrast, the peak intensity of the WS 2 A 1g model increases from 254 to 340.Previous studies have shown that the out-of-plane A 1g mode of WS 2 is extremely sensitive to the number of layers and its intensity increases with increasing layer number [52][53][54].The formation of a bilayer heterostructure can generate interlayer vdW force, enhancing the out-of-plane vibration of WS 2 .Therefore, the enhancement of the A 1g mode of WS 2 indicates that there is more than one layer of material in the test area.Combined with the morphology and height images characterized by AFM, these results corroborate the formation of the WS 2 /MoS 2 HS.
To investigate the interlayer electronic coupling in the WS 2 /MoS 2 HS, the PL spectra of ML-WS 2 , ML-MoS 2 , and WS 2 /MoS 2 HS are depicted in Figure 2e.Fitted by the Lorenz function, the PL spectrum of WS 2 exhibits two peaks at 1.972 and 2.003 eV, which can be attributed to the A-excitons and negative trions of ML-WS 2 .The PL spectrum of MoS 2 can also be divided into two peaks at 1.816 and 1.984 eV, corresponding to the A-excitons and B-excitons of ML-MoS 2 .After the formation of the heterostructure, only the exciton peak from WS 2 can be observed, possibly due to the low emission intensity of MoS 2 , making it indistinguishable in the overlapping region from the PL spectrum.It is worth noting that the energy from trions slightly red-shifts from 1.972 to 1.935 eV while the PL intensity maximum of the heterostructure decreases dramatically from 34,234 (for ML-WS 2 ) to 6560.This abrupt decrease in intensity indicates the formation of a heterojunction between WS 2 and MoS 2 accompanied by interlayer charge transfer, which can suppress exciton recombination and result in a decreased PL intensity.No distinct peak of interlayer exciton is observed, which could be attributed to the reduced intensity caused by non-parallel stacking of the monolayers [55].To study the influence of stacking order on the exciton performance, the PL spectrum of MoS 2 /WS 2 HS with MoS 2 in the upper layer is also conducted, as shown in Figure 2f.Interestingly, MoS 2 /WS 2 and WS 2 /MoS 2 exhibit similar exciton behavior, revealing that stacking order does not significantly affect the exciton behavior of the heterostructure.Given the absence of the peaks related to the interlayer exciton and MoS 2 exciton in the PL spectrum of WS 2 /MoS 2 HS, it becomes extremely challenging to analyze the excitonic properties and charge transfer in the vdW heterostructures.
To gain clearer insights into the variation in the heterostructure's excitonic characteristics, it is essential to enhance the proportion of exciton emission of MoS 2 in the PL spectrum of the heterostructure region.A feasible approach is to reduce the PL intensity of WS 2 without affecting the exciton emission of MoS 2 .Considering that WS 2 is the upper layer of the heterostructure, we propose a method for the upper WS 2 layer in the heterojunction to contact the tip of the AFM probe and divert the excited electrons in the conduction band of WS 2 using the high work function of the Si probe, thus inhibiting the emission caused by exciton recombination in WS 2 .The experimental setup is illustrated in Figure 3a, which integrates a homemade LSCPRS and an AFM.
number of layers and its intensity increases with increasing layer number [52][53][54].The formation of a bilayer heterostructure can generate interlayer vdW force, enhancing the out-of-plane vibration of WS2.Therefore, the enhancement of the A1g mode of WS2 indicates that there is more than one layer of material in the test area.Combined with the morphology and height images characterized by AFM, these results corroborate the formation of the WS2/MoS2 HS.
To investigate the interlayer electronic coupling in the WS2/MoS2 HS, the PL spectra of ML-WS2, ML-MoS2, and WS2/MoS2 HS are depicted in Figure 2e.Fitted by the Lorenz function, the PL spectrum of WS2 exhibits two peaks at 1.972 and 2.003 eV, which can be attributed to the A-excitons and negative trions of ML-WS2.The PL spectrum of MoS2 can also be divided into two peaks at 1.816 and 1.984 eV, corresponding to the A-excitons and B-excitons of ML-MoS2.After the formation of the heterostructure, only the exciton peak from WS2 can be observed, possibly due to the low emission intensity of MoS2, making it indistinguishable in the overlapping region from the PL spectrum.It is worth noting that the energy from trions slightly red-shifts from 1.972 to 1.935 eV while the PL intensity maximum of the heterostructure decreases dramatically from 34,234 (for ML-WS2) to 6560.This abrupt decrease in intensity indicates the formation of a heterojunction between WS2 and MoS2 accompanied by interlayer charge transfer, which can suppress exciton recombination and result in a decreased PL intensity.No distinct peak of interlayer exciton is observed, which could be attributed to the reduced intensity caused by non-parallel stacking of the monolayers [55].To study the influence of stacking order on the exciton performance, the PL spectrum of MoS2/WS2 HS with MoS2 in the upper layer is also conducted, as shown in Figure 2f.Interestingly, MoS2/WS2 and WS2/MoS2 exhibit similar exciton behavior, revealing that stacking order does not significantly affect the exciton behavior of the heterostructure.Given the absence of the peaks related to the interlayer exciton and MoS2 exciton in the PL spectrum of WS2/MoS2 HS, it becomes extremely challenging to analyze the excitonic properties and charge transfer in the vdW heterostructures.
To gain clearer insights into the variation in the heterostructure's excitonic characteristics, it is essential to enhance the proportion of exciton emission of MoS2 in the PL spectrum of the heterostructure region.A feasible approach is to reduce the PL intensity of WS2 without affecting the exciton emission of MoS2.Considering that WS2 is the upper layer of the heterostructure, we propose a method for the upper WS2 layer in the heterojunction to contact the tip of the AFM probe and divert the excited electrons in the conduction band of WS2 using the high work function of the Si probe, thus inhibiting the emission caused by exciton recombination in WS2.The experimental setup is illustrated in Figure 3a, which integrates a homemade LSCPRS and an AFM.To study the influence of the probe contact on the exciton properties, the PL spectra of ML-WS 2 , ML-MoS 2 , and WS 2 /MoS 2 HS are measured with and without Si probe contact (Figure 3b).To better compare the energy and shape of the emission peaks, the PL intensities of ML-WS 2 and WS 2 /MoS 2 HS without Si probe contact are multiplied by a shrinkage factor of 0.02.It is shown that upon contact with the Si probe, the PL shapes of ML-WS 2 and ML-MoS 2 remain basically unchanged.The PL intensities slightly decrease because the Si probe inhibits exciton recombination by guiding the escape pathway of the electrons.The peak positions exhibit a blue shift, which can be attributed to the local electric field generated by the contacted Si probe, thereby influencing the band structure and exciton binding energy of WS 2 and MoS 2 [56].Excitingly, with Si probe contact, the PL spectrum of the WS 2 /MoS 2 HS not only experiences a substantial decrease in intensity but also undergoes a significant change in shape.The peak of MoS 2 A-excitons is highlighted, enabling the detailed study of the excitonic properties in a vdW heterostructure using PL spectroscopy.
Moreover, an AFM equipped with a Si probe can serve as not only an electrode for electron transport but also a means to exert out-of-plane controllable pressure on 2D materials [57].Previous studies have shown that 2D materials exhibit good reversibility under a pressure of 12 GPa and a strain of 0.08%, and applying pressure to 2D material using an AFM probe does not cause damage [58,59].To investigate the pressure-related exciton characteristics, the PL spectra of ML-WS 2 , ML-MoS 2 , and WS 2 /MoS 2 HS are measured under different applied forces in the range of 0-120 nN at 20 nN intervals, as shown in Figure 4a-c, respectively.factor of 0.02.It is shown that upon contact with the Si probe, the PL shapes of ML-WS2 and ML-MoS2 remain basically unchanged.The PL intensities slightly decrease because the Si probe inhibits exciton recombination by guiding the escape pathway of the electrons.The peak positions exhibit a blue shift, which can be attributed to the local electric field generated by the contacted Si probe, thereby influencing the band structure and exciton binding energy of WS2 and MoS2 [56].Excitingly, with Si probe contact, the PL spectrum of the WS2/MoS2 HS not only experiences a substantial decrease in intensity but also undergoes a significant change in shape.The peak of MoS2 A-excitons is highlighted, enabling the detailed study of the excitonic properties in a vdW heterostructure using PL spectroscopy.
Moreover, an AFM equipped with a Si probe can serve as not only an electrode for electron transport but also a means to exert out-of-plane controllable pressure on 2D materials [57].Previous studies have shown that 2D materials exhibit good reversibility under a pressure of 12 GPa and a strain of 0.08%, and applying pressure to 2D material using an AFM probe does not cause damage [58,59].To investigate the pressure-related exciton characteristics, the PL spectra of ML-WS2, ML-MoS2, and WS2/MoS2 HS are measured under different applied forces in the range of 0-120 nN at 20 nN intervals, as shown in Figure 4a-c, respectively.The emission peak intensities of all three regions exhibit a decreasing trend with increasing pressure, which can be related to the out-of-plane pressure-induced transition from the direct to indirect bandgaps of the monolayers, leading to a reduction in quantum yield and a significant weakening of PL intensity [15,60,61].The spectral behavior in the heterostructure region is more complex than that in the monolayer region, as the PL properties of the heterostructure are influenced not only by the electronic structure of the monolayers but also by interlayer coupling [62].To visually demonstrate the relationship The emission peak intensities of all three regions exhibit a decreasing trend with increasing pressure, which can be related to the out-of-plane pressure-induced transition from the direct to indirect bandgaps of the monolayers, leading to a reduction in quantum yield and a significant weakening of PL intensity [15,60,61].The spectral behavior in the heterostructure region is more complex than that in the monolayer region, as the PL properties of the heterostructure are influenced not only by the electronic structure of the monolayers but also by interlayer coupling [62].To visually demonstrate the relationship between the PL spectra and pressure in the heterojunction region, the 2D pressure-dependent wavelength-intensity image is shown in Figure 4d.Three peaks of the heterostructure exhibit a strong pressure dependency.As the pressure increases, the intensity of the three peaks declines with varying degrees, where the peak related to WS 2 trions exhibits the most noticeable decrease.
To gain insights into the influence of pressure on the exciton characteristics of the vdW heterostructure, the pressure-dependent PL curves are fitted with the Lorentzian function.The evolution of the intensities and positions of the fitted exciton peaks of ML-WS 2 , ML-MoS 2 , and WS 2 /MoS 2 HS with increasing pressure are shown in Figure 5. Figure 5a reveals that with increasing pressure, the intensity of A excitons and trions in ML-WS 2 decreases at the same rate but the exciton energy remains almost constant (Figure 5b).For ML-MoS 2 , Appl.Sci.2024, 14, 2179 8 of 13 the intensity of A excitons decreases much faster than that of B excitons (Figure 5c), and the variation in exciton energy is consistent with that of ML-WS 2 (Figure 5d).The decline in PL intensity in the monolayer TMDCs is associated with the transfer of electrons from the K valley to its neighboring valleys during conduction.ML-WS 2 and ML-MoS 2 are both semiconductors with direct bandgaps.With increasing pressure, the conduction band exhibits an upward displacement at the K valley and a downward displacement at the Γ valley, enhancing the K−K bandgap and reducing the K−Γ bandgap [61].Consequently, the quantum yield diminishes, resulting in a reduction in PL intensity.Once the pressure surpasses a certain threshold, the optical bandgap undergoes a transition, shifting from direct to indirect.hibits the most noticeable decrease.
To gain insights into the influence of pressure on the exciton characteristics of the vdW heterostructure, the pressure-dependent PL curves are fitted with the Lorentzian function.The evolution of the intensities and positions of the fitted exciton peaks of ML-WS2, ML-MoS2, and WS2/MoS2 HS with increasing pressure are shown in Figure 5. Figure 5a reveals that with increasing pressure, the intensity of A excitons and trions in ML-WS2 decreases at the same rate but the exciton energy remains almost constant (Figure 5b).For ML-MoS2, the intensity of A excitons decreases much faster than that of B excitons (Figure 5c), and the variation in exciton energy is consistent with that of ML-WS2 (Figure 5d).The decline in PL intensity in the monolayer TMDCs is associated with the transfer of electrons from the K valley to its neighboring valleys during conduction.ML-WS2 and ML-MoS2 are both semiconductors with direct bandgaps.With increasing pressure, the conduction band exhibits an upward displacement at the K valley and a downward displacement at the Γ valley, enhancing the K−K bandgap and reducing the K−Γ bandgap [61].Consequently, the quantum yield diminishes, resulting in a reduction in PL intensity.Once the pressure surpasses a certain threshold, the optical bandgap undergoes a transition, shifting from direct to indirect.In the vdW heterostructure, the WS2 A-excitons and the MoS2 A-excitons exhibit similar decreasing trends in peak intensity under increasing pressure (Figure 5e).However, the WS2 trions experience a rapid decline with a reduction rate approximately twice that of the WS2 A-excitons.When the pressure increases to 100 nN, the intensity of WS2 trions is already quenched to a level smaller than that of the WS2 A-excitons.This is because the WS2 trions, composed of two electrons and one hole, are more significantly affected as the In the vdW heterostructure, the WS 2 A-excitons and the MoS 2 A-excitons exhibit similar decreasing trends in peak intensity under increasing pressure (Figure 5e).However, the WS 2 trions experience a rapid decline with a reduction rate approximately twice that of the WS 2 A-excitons.When the pressure increases to 100 nN, the intensity of WS 2 trions is already quenched to a level smaller than that of the WS 2 A-excitons.This is because the WS 2 trions, composed of two electrons and one hole, are more significantly affected as the electrons transitioning to the conduction band decrease due to pressure-induced band change.In contrast, the WS 2 A-excitons, composed of one electron and one hole, are less affected.Furthermore, the energy of WS 2 trions in the heterostructure also exhibits an anomalous behavior (Figure 5f).Compared to the stability of the A-excitons of WS 2 and MoS 2 , the trion energy increases at a rate of 0.1 meV nN −1 with increasing pressure.This blue-shift is attributed to the anticrossing effect of the electronic states under pressure, a phenomenon observed in III-V and II-VI semiconductor families [63][64][65][66][67][68].The more pronounced change in WS 2 trions compared to A-excitons is understood as the combined modulation of the electronic structure in the heterostructure region by interlayer coupling and strain induced by pressure.
To investigate the reversibility of pressure modulation on the WS 2 /MoS 2 HS, the applied pressure is gradually decreased from 140 to 0 nN, as shown in Figure 6.All the emission intensities of the three peaks increase with weakening pressure.The emission intensity of the WS 2 A excitons exhibits the fastest increasing rate.The emission peak positions of the A excitons from both WS 2 and MoS 2 remain almost unchanged, while the emission energy of the WS 2 trions gradually shifts towards lower energies.Once the pressure weakens to 0 N, the peak shape and intensity of the PL spectrum almost recover to their initial states.These variation trends are opposite to the case of increasing pressure, demonstrating the good reversibility of pressure modulation on the excitonic behavior of the WS 2 /MoS 2 HS.modulation of the electronic structure in the heterostructure region by interlayer coupling and strain induced by To investigate the reversibility of pressure modulation on the WS2/MoS2 HS, the applied pressure is gradually decreased from 140 to 0 nN, as shown in Figure 6.All the emission intensities of the three peaks increase with weakening pressure.The emission intensity of the WS2 A excitons exhibits the fastest increasing rate.The emission peak positions of the A excitons from both WS2 and MoS2 remain almost unchanged, while the emission energy of the WS2 trions gradually shifts towards lower energies.Once the pressure weakens to 0 N, the peak shape and intensity of the PL spectrum almost recover to their initial states.These variation trends are opposite to the case of increasing pressure, demonstrating the good reversibility of pressure modulation on the excitonic behavior of the WS2/MoS2 HS.A possible mechanism is proposed in Figure 7. Figure 7a shows the band structure of WS2/MoS2 HS calculated using the DFT method.At the K-point, both the conduction band and valence band of WS2 are higher than those of MoS2, indicating the formation of a type-II heterostructure.It is worth noting that the energy gap of the K−Γ transition is smaller than that of the K−K transition, suggesting the presence of an indirect bandgap, which may lead to a significant enhancement of non-radiative transitions.A possible mechanism is proposed in Figure 7. Figure 7a shows the band structure of WS 2 /MoS 2 HS calculated using the DFT method.At the K-point, both the conduction band and valence band of WS 2 are higher than those of MoS 2 , indicating the formation of a type-II heterostructure.It is worth noting that the energy gap of the K−Γ transition is smaller than that of the K−K transition, suggesting the presence of an indirect bandgap, which may lead to a significant enhancement of non-radiative transitions.
modulation of the electronic structure in the heterostructure region by interlayer coupling and strain induced by pressure.
To investigate the reversibility of pressure modulation on the WS2/MoS2 HS, the applied pressure is gradually decreased from 140 to 0 nN, as shown in Figure 6.All the emission intensities of the three peaks increase with weakening pressure.The emission intensity of the WS2 A excitons exhibits the fastest increasing rate.The emission peak positions of the A excitons from both WS2 and MoS2 remain almost unchanged, while the emission energy of the WS2 trions gradually shifts towards lower energies.Once the pressure weakens to 0 N, the peak shape and intensity of the PL spectrum almost recover to their initial states.These variation trends are opposite to the case of increasing pressure, demonstrating the good reversibility of pressure modulation on the excitonic behavior of the WS2/MoS2 HS.A possible mechanism is proposed in Figure 7. Figure 7a shows the band structure of WS2/MoS2 HS calculated using the DFT method.At the K-point, both the conduction band and valence band of WS2 are higher than those of MoS2, indicating the formation of a type-II heterostructure.It is worth noting that the energy gap of the K−Γ transition is smaller than that of the K−K transition, suggesting the presence of an indirect bandgap, which may lead to a significant enhancement of non-radiative transitions.Figure 7b illustrates the charge transfer in the heterostructure after contact with the Si probe.Prior to contact, the excited electrons flow from the WS 2 conduction band to the MoS 2 , while the holes flow in the opposite direction.The spatial separation of electrons and holes leads to the dissociation of excitons, suppressing exciton recombination emission.After contact with the probe, a large number of excited electrons flow from the WS 2 conduction band to the Si probe, thereby reducing the fluorescence intensity of WS 2 to a similar level as MoS 2 .This can be attributed to the fact that, when the Si probe contacts the upper WS 2 in the heterostructure region, the excited electrons in WS 2 will migrate to the Si probe due to the difference in the work function, resulting in a significant dissociation of the exciton species in WS 2 and a decrease in PL intensity.However, the lower MoS 2 layer is not in contact with the Si probe, and the influence on the PL intensity is minimal.When the excitonic recombination efficiency in WS 2 is lowered to be comparable to that in MoS 2 , the contribution of MoS 2 in the heterostructure's PL spectrum becomes prominent.When pressure is applied to the heterostructure, it causes changes in the coupling strength of the heterostructure.With increasing tip pressure, the energy of the A-excitons remains stable while the energy of the trions blue-shifts, indicating that the trions are more sensitive to the interlayer coupling strength than the A-excitons.

Conclusions
In summary, we employed a novel method to quantitatively study the exciton characteristics of WS 2 /MoS 2 vdW heterostructures through tip contact PL spectra.Upon Si probe contact, the emissions of the A-excitons and trions can be observed in the PL spectra at the heterostructure region, specifically highlighting the contribution of MoS A-excitons.The Si probe not only acts as an electrode to remove electrons but also exerts out-of-plane pressure on the micro-region of the heterostructure to quantitatively control the electronic structure and interlayer coupling.As pressure on the heterostructure increases, the PL intensity corresponding to WS 2 trions significantly decreases while the PL energy exhibits an obvious blue-shift, revealing that the trions are more sensitive to pressure than A-excitons.This work provides a valuable approach for understanding the relationship between interlayer interactions and electronic/optical properties in 2D TMDC vdW heterostructures.In future studies, more work should be conducted to further combine the local pressure induced by an AFM probe with novel physics in the TMDC heterostructures, such as interlayer excitons, moiré superlattices, and exciton condensations.

Figure
Figure 2a is an optical image of the WS 2 /MoS 2 HS on the SiO 2 /Si substrate, displaying twist angles for WS 2 and MoS 2 at approximately 60 • .The AFM topography image is shown in Figure 2b, showing that numerous abnormal points with random heights are unevenly distributed throughout the stacking region.The abnormal points are caused by the formation of unavoidable nanobubbles between the two monolayers during the fabrication process, and the gap distances are approximately equal to the diameter of the bubbles.Figure 2c is the height profile along the straight line indicated in Figure 2b.The presence of two steps below 1 nm in the height curve indicates that the heterostructure Appl.Sci.2024, 14, x FOR PEER REVIEW 5 of 13

Figure 3 .
Figure 3. (a) Illustration of the measurement.(b) PL spectra with probe contact.Figure 3. (a) Illustration of the measurement.(b) PL spectra with probe contact.

Figure 3 .
Figure 3. (a) Illustration of the measurement.(b) PL spectra with probe contact.Figure 3. (a) Illustration of the measurement.(b) PL spectra with probe contact.

Figure 6 .
Figure 6.The PL spectra of WS2/MoS2 HS under applied pressure gradually decreasing from 140 to 0 nN.

Figure 7 .
Figure 7. (a) DFT-calculated band structure of WS2/MoS2 HS.(b) Illustration of the charge transfer in the heterostructure with Si probe contact.

Figure 6 .
Figure 6.The PL spectra of WS 2 /MoS 2 HS under applied pressure gradually decreasing from 140 to 0 nN.

Figure 6 .
Figure 6.The PL spectra of WS2/MoS2 HS under applied pressure gradually decreasing from 140 to 0 nN.

Figure 7 .
Figure 7. (a) DFT-calculated band structure of WS2/MoS2 HS.(b) Illustration of the charge transfer in the heterostructure with Si probe contact.

Figure 7 .
Figure 7. (a) DFT-calculated band structure of WS 2 /MoS 2 HS.(b) Illustration of the charge transfer in the heterostructure with Si probe contact.