# Fileset

[Nisi_2025_2D_Mater._12_015023.pdf](https://mdr.nims.go.jp/filesets/a8955828-f028-4df1-99ac-141ff385c093/download)

## Creator

K Nisi, [J C Thomas](https://orcid.org/0000-0002-2151-7725), [S Levashov](https://orcid.org/0009-0001-4244-8554), E Mitterreiter, T Taniguchi, [K Watanabe](https://orcid.org/0000-0003-3701-8119), S Aloni, [T R Kuykendall](https://orcid.org/0000-0003-1362-3285), [J Eichhorn](https://orcid.org/0000-0003-2413-6079), [A W Holleitner](https://orcid.org/0000-0002-8314-4397), [A Weber-Bargioni](https://orcid.org/0000-0003-2986-1819), [C Kastl](https://orcid.org/0000-0001-5309-618X)

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Scanning probe spectroscopy of sulfur vacancies and MoS<sub>2</sub> monolayers in side-contacted van der Waals heterostructures](https://mdr.nims.go.jp/datasets/1e200014-934f-4a7c-8f8a-0e423555b58d)

## Fulltext

Scanning probe spectroscopy of sulfur vacancies and MoS2 monolayers in side-contacted van der Waals heterostructures2D Materials     PAPER • OPEN ACCESSScanning probe spectroscopy of sulfur vacanciesand MoS2 monolayers in side-contacted van derWaals heterostructuresTo cite this article: K Nisi et al 2025 2D Mater. 12 015023 View the article online for updates and enhancements.You may also likeAdsorption and epitaxial growth of smallorganic semiconductors on hexagonalboron nitrideM Kratzer, A Matkovic and C Teichert-Direct growth of hBN/Grapheneheterostructure via surface deposition andsegregation for independent thicknessregulationWenyu Liu, Xiuting Li, Yushu Wang et al.-Electrical spin injection, transport, anddetection in graphene-hexagonal boronnitride van der Waals heterostructures:progress and perspectivesM Gurram, S Omar and B J van Wees-This content was downloaded from IP address 144.213.253.16 on 01/01/2025 at 02:47https://doi.org/10.1088/2053-1583/ada046/article/10.1088/1361-6463/ab29cb/article/10.1088/1361-6463/ab29cb/article/10.1088/1361-6463/ab29cb/article/10.1088/1361-6528/ac8994/article/10.1088/1361-6528/ac8994/article/10.1088/1361-6528/ac8994/article/10.1088/1361-6528/ac8994/article/10.1088/2053-1583/aac34d/article/10.1088/2053-1583/aac34d/article/10.1088/2053-1583/aac34d/article/10.1088/2053-1583/aac34d2D Mater. 12 (2025) 015023 https://doi.org/10.1088/2053-1583/ada046OPEN ACCESSRECEIVED28 August 2024REVISED11 December 2024ACCEPTED FOR PUBLICATION17 December 2024PUBLISHED30 December 2024Original content fromthis work may be usedunder the terms of theCreative CommonsAttribution 4.0 licence.Any further distributionof this work mustmaintain attribution tothe author(s) and the titleof the work, journalcitation and DOI.PAPERScanning probe spectroscopy of sulfur vacancies and MoS2monolayers in side-contacted van der Waals heterostructuresK Nisi1,2,3, J C Thomas3, S Levashov1, E Mitterreiter1,2, T Taniguchi4, K Watanabe5, S Aloni3,T R Kuykendall3, J Eichhorn1, A WHolleitner1,2, A Weber-Bargioni3 and C Kastl1,2,∗1 Walter Schottky Institute and Physics Department, Technical University of Munich, Am Coulombwall 4a, 85748 Garching, Germany2 Munich Center for Quantum Science and Technology (MCQST), Schellingstraße 4, 80799 Munich, Germany3 Molecular Foundry, Lawrence Berkeley National Laboratory, Berkeley, CA 94720, United States of America4 Research Center for Materials Nanoarchitectonics, National Institute for Materials Science, 1-1 Namiki, Tsukuba 305-0044, Japan5 Research Center for Electronic and Optical Materials, National Institute for Materials Science, 1-1 Namiki, Tsukuba 305-0044, Japan∗ Author to whom any correspondence should be addressed.E-mail: christoph.kastl@wsi.tum.deKeywords: TMDC, scanning tunneling microscopy, conductive atomic force microscopy, defects, van der Waals heterostructureSupplementary material for this article is available onlineAbstractWe investigate the interplay between vertical tunneling and lateral transport phenomena inelectrically contacted van der Waals heterostructures made from monolayer MoS2, hBN, andgraphene. We compare data taken by low-temperature scanning tunneling spectroscopy to resultsfrom room-temperature conductive atomic force spectroscopy on monolayer MoS2 with sulfurvacancies and with varying hBN layers. We show that for thick hBN barrier layers, where tunnelingcurrents into the conductive substrate are suppressed, a side-contact still enables addressing thedefect states in the scanning tunneling microscopy via the lateral current flow. Few-layer hBNrealizes an intermediate regime in which the competition between vertical tunneling and lateraltransport needs to be considered. The latter is relevant for device structures with both a thintunneling barrier and a side-contact to the semiconducting layers.Two-dimensional (2D) van der Waals heterostruc-tures possess promising characteristics for next-generation (opto)electronic devices [1, 2]. Verticalheterostructures, with van derWaals layers assembledto form atomically well-defined, ultra-thin tunnel-ing devices, can leverage the particular advantagesof the 2D materials platform, overcoming funda-mental limits of conventional device architectures[3] and providing new functionality for applicationsin quantum technologies [4, 5]. These 2D hetero-structures typically consist of atomically thin highly-conductive electrodes (e.g. graphene), insulating bar-riers (e.g. few-layer hBN), and monolayer semicon-ductors (e.g. MoS2, WS2, MoSe2 orWSe2). The wide-gap insulator hBN has proven to be an excellent tun-neling barrier in vertical device structures [6, 7]. Itslow defect density [6] and high dielectric breakdownstrength [8, 9] allow fabricating barriers which areonly a few angstroms thick and can be used to finetune non-equilibrium tunneling currents, electron-hole interactions, and equilibrium charge transferbetween two neighboring layers [6, 10].Moreover, the 2D van der Waals heterostruc-tures allow incorporating various defects as single-photon sources directly integrated into the devicearchitecture [11]. For example, vacancy defects [12]in monolayer transition metal dichalcogenides canbe implanted into tunneling circuits [13] with nano-meter precision using ion beam patterning [14, 15].For an atomic defect to be an effective two-level sys-tem, it is crucial to reduce the degeneracies of itsground and excited states, which can be achievedin defect complexes with lowered symmetries orin charged defects [16, 17]. In this context, recentscanning tunneling microscopy (STM) studies onsulfur vacancies in monolayer MoS2 have directlyresolved the spontaneous symmetry lowering of thevacancy’s defect orbitals upon charging, known as© 2024 The Author(s). Published by IOP Publishing Ltdhttps://doi.org/10.1088/2053-1583/ada046https://crossmark.crossref.org/dialog/?doi=10.1088/2053-1583/ada046&domain=pdf&date_stamp=2024-12-30https://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/https://orcid.org/0000-0002-2151-7725https://orcid.org/0009-0001-4244-8554https://orcid.org/0000-0003-3701-8119https://orcid.org/0000-0003-1362-3285https://orcid.org/0000-0003-2413-6079https://orcid.org/0000-0002-8314-4397https://orcid.org/0000-0003-2986-1819https://orcid.org/0000-0001-5309-618Xmailto:christoph.kastl@wsi.tum.dehttp://doi.org/10.1088/2053-1583/ada0462D Mater. 12 (2025) 015023 K Nisi et althe Jahn–Teller distortion [18–20]. Both experimentand theory demonstrate significant energetic correc-tions (10 s of meV) at the single particle level [18–22]. However, so far, these STM studies only con-sidered defects in MoS2 layers which are supporteddirectly on a graphene [18–20] or gold substrate [23].In that case, the (semi)-metallic substrate not onlyinduces substantial charge transfer into the mono-layer semiconductor [24], but it also quenches theoptical emission from the relatively long-lived, loc-alized defect excitons [25], while emission from thefast free exciton is still detectable. Therefore, a directcorrelation between the single particle defect orbitalstypically resolved by STM of transition metal dichal-cogenides on graphene and the excitonic defect statesrevealed by optical spectroscopy of MoS2 on insu-lating hBN remains challenging. An insulating hBNbarrier can potentially provide the necessary decoup-ling of the semiconductingMoS2 layer and thus of thepoint defects from the metallic graphene. Ultimately,such a heterostructure facilitates resolving localizedexciton transition at the atomic scale via tip-inducedluminescence [26–28]. The latter approach has beensuccessfully applied to molecules on metal surfaces,where a thin insulating barrier, typically few layers ofsingle crystal salt, provides the necessary decouplingfrom the substrate [29].In this work, we present scanning probe meas-urements of MoS2/hBN heterostructures with sul-fur vacancy defects. We employ scanning tunnelingspectroscopy (STS) at 4 K to gain insights into theelectronic properties of the heterostructures. A chal-lenge in such experiments is the identification oftheir respective locations on the same heterostructureusing the limited scan range and limited coarse posi-tioning accuracy available in the STM. Therefore, weutilize conductive atomic force microscopy (cAFM)at room temperature as a complementary tool tocharacterize the charge transport for different het-erostructure configurations with varying hBN thick-ness on the same sample. To characterize also thickdecoupling barriers, where vertical tunneling is sup-pressed, we rely on side-contacted heterostructures,where the lateral current flow enables electronic spec-troscopy of the localized states.1. ResultsThin layers of hBN (down to amonolayer) andmono-layers (ML) of MoS2 were prepared via micromech-anical cleavage and placed onto an epitaxially-growngraphene on (6H)-SiC substrate via a stampingprocedure (see Methods). The layers were stackedto create different regions of interest within onesample. Specifically, we investigated the followingheterostructure configurations: i) hBN on epitaxialgraphene (EG), ii)MoS2 onhBNonEG, and iii)MoS2on EG. Importantly, the MoS2 layer was contacteddirectly to the graphene substrate via a side-contact toensure a conductive pathway even for thick hBN bar-riers, where direct vertical tunneling is suppressed. Inother studies, similar side-contacts were achieved viagold or nanopatterned graphene electrodes [30, 31].The heterointerfaces were probed by scanning probemicroscopy and spectroscopy with the bias appliedbetween the EG and the tip (figure 1(a)). We stud-ied the properties, in particular the charging state andin-gap states, of single sulfur vacancy defects (indic-ated exemplarily by the black circles in figure 1(a)) inthe MoS2 monolayer for the different layer configur-ations. Figure 1(b) shows a helium ion microscopeimage of such a heterostructure sample. Due to thehigh material contrast of He-ion microscopy, all lay-ers are visible enabling the assessment of the exactposition of each material, including microcracks inthe layers. The different materials are labeled accord-ingly. For this particular sample, the ML MoS2 isdivided in three regions (on thick hBN, on ML hBN,and directly on EG/SiC).In our STMmeasurements, the coarse position ofthe sample with respect to the tip was identified viaan optical access port. Gold contacts and other bulkcrystals in close vicinity to the heterostructure servedas optical reference markers for the approach of thetip. To identify the different locations on the sample,we relied on STS and STM images. Hereby, distinctivefeatures such as lattice constants and defects supportSTS measurements to enable the assignment of eacharea (figure S1). Figures 1(c) and (d) depict repres-entative low-temperature STS measurements for dif-ferent areas on the heterostructure. The dI/dV curveswere recorded with amodulation voltage AM= 5mVand a current setpoint of Iset = 150 pA. For eachconfiguration, we recorded spectra on multiple loc-ations and determined averaged values for the bandonsets.For MoS2 directly on graphene (figure 1(c)), theonsets of the tunneling current appear at+0.64 V and−1.76 V corresponding to the conduction band min-imum (CBM) and valence bandmaximum (VBM) ofMoS2. This simple analysis translates into an estim-ated value of the bandgap of 2.4 eV. The bandgap inprevious reports varies between 2 and 2.5 eV [14, 30,32, 33]. Nevertheless, an exact determination of thebandgap value is difficult and requires a more care-ful analysis as described, for example, by Murray et al[32]. This is due to the fact that the tunneling currentnear the Γ -point is typically enhanced compared tothe K-point [34], because states with finite in-planemomentum decay faster into the vacuum [35]. TheFermi level is not mid-gap but shifted towards theCBM, which we assign to the charge transfer from thegraphene [16]. When decoupling theMoS2 with hBN(figure 1(d)), we observe two effects. First, the onsetof the CBM shifts systematically to higher voltages(on average we find +0.4 eV), suggesting either a22D Mater. 12 (2025) 015023 K Nisi et alFigure 1. Sample scheme and tunneling process. (a) hBN and ML MoS2 are placed on epitaxial graphene (EG) on SiC. Thedifferent heterointerfaces and sulfur vacancies (Vacs) in the MoS2 layer are probed by STS at T = 4 K using a tungsten tipprepared on a gold substrate. (b) False-colored helium ion microscope image of the MoS2/hBN/EG heterostructure revealing thedifferent layer configurations (beam current 170 fA, acceleration voltage 30 kV, FOV 40× 40 µm2, scan dwell time 2 µs). (c)Low-temperature scanning tunneling spectroscopy measurements of ML MoS2 directly on EG/SiC and (d) ML MoS2 decoupledfrom the EG/SiC via a hBN tunneling barrier. The dashed line marks the median of the noise level around 0 V. The solid black linemarks the median signal below−2 V. The noticeable increase of the signal at voltages below−2 V is interpreted as the onset of thevalence band in the MoS2/hBN heterostructure. Tunneling spectra are slightly offset to avoid negative values in the log-plot.reduced charge transfer from the graphene or a pos-sible combination of strain and screening resulting inoverall band gap renormalization [36]. Second, tun-neling currents into the valence band of the MoS2appear to be suppressed below the noise level ofour measurement (dashed black lines in figure 1(d)).Only at bias voltages below −3.68 V, which we tent-atively assign to the VBM onset in hBN (cf. sup-porting information figure S2), the tunneling currentincreases noticeably. Averaging the signal below (solidblack line in figure 1(d)) and above (dashed blackline in figure 1(d)) the expected value of the VBM ofMoS2, the spectrum reveals a small, but finite increaseinside the gap reminiscent of the VBM in MoS2 asindicated by the black solid line in figure 1(d). Thefact that the gap of MoS2 is not properly resolvedpoints towards the possibility that we probeMLMoS2on bulk (∼22 nm) hBN, rather than on ML hBN, viathe side-contact. However, we note that it is difficultto identify the thickness of the underlying hBN basedon STS alone.To get a better understanding of the underlyingcharge transport processes, we performed comple-mentary cAFM measurements at room temperatureunder inert gas atmosphere (N2) on the above hetero-structure and on different flakes of varying thicknesslying in close vicinity. We determined the positionof each measurement and corresponding layer thick-ness by AFM maps and height profiles (cf. figuresS3 and S4). Areas of interest for cAFM (2 × 2 µm2)were cleaned with an AFM tip using contact modeto remove residual adsorbates from the exposure toambient (cf. figures S5 and S6 and Methods) [37].For the IV-spectroscopy, the contact force was keptas small as possible, but large enough to enable goodelectric contact. Nevertheless, when comparing STSat low temperature and cAFMmeasurements at roomtemperature, one needs to keep in mind that for thelatter, the tip is in physical contact with the samplesurface, i.e. the tip enters the repulsive regime of tip-sample interaction during the current measurement.Hence, the IV-curves recorded in cAFM typically donot reveal the density of states as in STS, unless a thin,insulating tunneling barrier is introduced betweenthe AFM tip and the material under test [38].Figure 2(a) depicts layer-dependent IV-curves(semi-log scale) of insulating hBN on EG. As expec-ted, the overall current amplitude at a given voltagedrops with increasing barrier height or hBN layernumber [10]. Large tunneling currents (nA-regime)are observed for barrier heights up to three layers,most probably dominated by the conductive char-acter of the underlying graphene [38]. For six lay-ers of hBN, direct tunneling is suppressed and onlyat very high bias magnitudes, we observe small cur-rents (pA-regime). For better comparison to STS,figure 2(b) shows the numerically derived ∆I/∆Vspectra measured by cAFM. The spectra for ML and32D Mater. 12 (2025) 015023 K Nisi et alFigure 2. Conductive AFM spectroscopy. (a) Absolute value of the cAFM current for three hBN/EG heterointerfaces with amonolayer (light gray), trilayer (dark gray), and hexalayer (black) hBN barrier. Significant tunneling currents through the gap ofhBN are resolved for monolayer and trilayer hBN. (b) Normalized conductance∆I/∆V numerically derived from the data in (a).(c) Current at a bias of 1 V (Current1V) extracted from the data in (a) as a function of hBN layer number. The dashed line is thetheoretical scaling of the exponential decay assuming an ideal tunneling process through a hBN barrier with given layer number.(d) Sketch of the vertical and lateral current pathways (arrows) in the side-contacted MoS2 heterostructure. (e) Absolute value ofthe cAFM current for different heterostructure configurations: MoS2 supported directly on EG/SiC (orange), MoS2 with anadditional ML hBN barrier (light red), and MoS2 on bulk hBN (dark red). Lateral transport is dominant for MoS2 on bulk hBN(dark red), whereas it is negligible for MoS2 directly on EG (orange). (f) Normalized conductance∆I/∆V numerically derivedfrom the data in (c).trilayer hBN are virtually identical after normaliz-ation, pointing towards a predominant direct tun-neling current through the thin barrier layer. Thesuppression of the tunneling current, extracted at afixed bias of 1 V, with barrier thickness is shown infigure 2(c). The dashed line is the naively expectedexponential scaling law assuming an ideal tunnelingprocess through a hBN barrier with given layer num-ber. Tentatively, we fit the data with an exponentialdecay, giving a decay constant of 1.1 Å−1, similar toprevious literature [39]. Note that when very largecontact forces are used (on the order of hundreds ofnN), the applied pressure can further modify the tun-neling current due to the local indentation of the vanderWaals stack [40]. To ensuremaximum comparab-ility within this work, we kept the applied force small(∼1 nN) such that any local indentations are expectedto be negligible.Next, cAFM measurements were conducted onthe side-contacted MoS2 monolayer with differentthicknesses of hBN below. Depending on the hBNthickness, vertical tunneling throughMoS2 and hBN,lateral current flow through the MoS2 or a combina-tion of both have to be considered (figure 2(d)). ForMoS2 probed in an area where it is directly suppor-ted on EG (orange curve, figures 2(e) and (f)), weobserve high currents (nA-regime) already close tozero bias, suggesting direct tunneling through theMLMoS2 barrier into graphene [40]. For MoS2 on bulkhBN (dark red curves in figures 2(e) and (f)), thevertical tunneling is expected to be completely sup-pressed with barrier widths far above the tunnelinglimit (cf. figure 2(c)). Then, the current flow fromMoS2 to graphene is expected to proceed exclusivelyvia a lateral drift-diffusion process. The resulting IV-curves can be interpreted as a lateral device with aAu/MoS2 Schottky-contact and an efficient, verticalMoS2/graphene junction in series (figure 2(d)). MoS2onML hBN seems to be in-between the direct tunnel-ing and the lateral transport scenario (red curves in infigures 2(e) and (f)). In other words, lateral transportand vertical tunneling provide two competing currentpathways in the investigated samples with electricalside-contacts. Consequently, the difference in currentbetween the three different configurations at identicalbias gives an idea of lateral and tunnel contributionat room temperature. As a control experiment, weremoved the side-contact of MoS2 on bulk hBN bynano-mechanicalmanipulation using the AFM tip. Inagreement with our hypothesis of predominant lat-eral current flow, no current was measurable withoutthe side-contact (figure S7).42D Mater. 12 (2025) 015023 K Nisi et alThe cAFM results demonstrate that direct tun-neling is possible at least up to three layers of hBNbarriers (figure 2(b)). Therefore, vertical tunnelingshould be possible and dominant for MoS2 on MLhBN, as well, and we conclude that the STS spectrumin figure 1(c) was in fact measured with a bulk hBNbarrier below the ML MoS2. In our heterostructureconfiguration, the MoS2 on bulk hBN still has a lat-eral side-contact with graphene, which provides thenecessary conductive connection at room temperat-ure. For STS at low temperatures, where the intrinsicconductivity is frozen out, the current is rather car-ried by the diffusion of injected carriers, which arealso called hot carriers. For positive (negative) biasabove the CBM (below the VBM), electrons (holes)are injected into the CB, that need to diffuse later-ally to the contact. The observed current suppres-sion in STS (cf. figure 1(d)) on MoS2/hBN at neg-ative bias (below the VBM of MoS2) points towardsa less efficient diffusion of injected holes. Only at ahigh bias (−3.5 V), the hole currents increase signific-antly because vertical tunneling into the VBMof hBNbecomes available as an additional channel. A possiblescenario is that at a finite residual electron density,which is consistent with the Fermi level being locatedcloser to the CBM, the injected holes will be minor-ity carriers and their diffusion is limited by recom-binationwith electrons. Another possible explanationare shallow defect states, which can effectively trapcarriers at low temperatures, while being ineffectiveat room temperature because of thermal activation[41]. For example, oxygen passivated sulfur vacanciesare abundant in MoS2 [14], and they were shown tointroduce a shallow defect resonance in the VBM ofother structurally similar transition metal dichalco-genides, in particular MoSe2 [42], without any deepin-gap state.There is a noticeable effect of the bulk hBNdecoupling layer on the defects in MoS2 as well.Figure 3(a) shows a large area STM image of variouspoint defects in MoS2 on hBN. In line with previousreports on semiconducting transition metal dichal-cogenides directly on graphene [16, 20], we observedifferent defects (panel insets) including negativelycharged point defects of unknown type (Defect A),oxygen-passivated sulfur vacancies in the top and bot-tom sulfur sub-lattice (Defects B) and unpassivatedsulfur vacancies in the top sulfur lattice (Defect C). Inthe following, we focus on unpassivated sulfur vacan-cies. The latter have a clear topographic fingerprintwith a characteristic orbital shape, and, generally,they exhibit a deep in-gap state due to the unsatur-ated bonds at the defect site [14, 20, 43]. Interestingly,we find that the vacancies are charged, as sugges-ted by the dark halo, even after decoupling MoS2with hBN [14, 18]. Figure 3(b) depicts the expec-ted energy diagram for a negatively charged vacancyincluding splitting due to the Jahn–Teller effect [18,20]. We expect three in-gap states, one unoccupiedstate close to the CB, one occupied state in the middleof the gap, and two degenerate occupied states nearthe valence band. As a reference, figures 3(c) and(d) show STS spectra across a negatively charged sul-fur vacancy defect (HOMO) in MoS2 directly ongraphene. The pronounced upwards band bending ofthe CBM (indicated by the white line as guide to theeye in figure 3(d)) is consistent with the net negativecharge on the defect. In agreement with other recentSTM studies [18, 20], three in-gap states, assigned asHOMO−1, HOMO and LUMO, are visible in the STSspectrum. Moreover, the defect states exhibit charac-teristic broadened side bands [18].Figures 3(e) and (f) show STS spectra across anunpassivated sulfur vacancy defect in the decoupledMoS2/hBNheterostructure. Asmentioned before, thedark halo of the defect points towards a charged state.Indeed, the CB exhibits an upwards band bending(indicated by the white line as guide to the eye infigure 3(f)), suggesting that the investigated defectis negatively charged in this configuration as well.However, the onset of the CB is shifted to higherbiases compared to MoS2 directly on EG, indicat-ing a more intrinsic nature of the MoS2. Similarly,the lack of band bending for the valence band inMoS2/hBN suggests that the defect can actually bebrought into an overall neutral configuration, poten-tially due to tip-induced discharging at a large neg-ative bias [19]. Similar to charged sulfur vacancies inMoS2 directly on EG, the vacancy defect in MoS2 onbulk hBN exhibits one state above the Fermi level,i.e. a LUMO state (black triangle figure 3(e)). Thespectra suggest a finite broadening of the LUMOstate, possibly due to phonons [18]. However, no fur-ther states are visible below the Fermi level, i.e. noHOMO states are resolved. Only at−3.8 V, we detectagain the onset of the hBN VBM (cf. Figure 1(d)for pristine MoS2/hBN). A possible explanation isthat the HOMO states are experimentally not access-ible because there is simply no current pathway. Thebulk hBN barrier inhibits vertical tunneling and theMoS2 band gap inhibits lateral transport because theHOMO level is a bound state deep inside the bandgap. In this picture, it is interesting to ask if the LUMOstate is in fact a strictly bound state (i.e. a deep defect[12, 44]) or if it has the character of a resonant defectstate hybridizing with the CB [45]. Since the LUMOstate is experimentally accessible in STS even withbulk hBN below (figure 3(f)), we conclude that thecharge carriers injected from the tip escape from thelocalized state into the delocalizedMoS2 CB at a suffi-cient rate. Yet, the STS spectra suggest that the LUMOstate (black triangle in figure 3(e)) is in fact separatedfrom the CBM onset by a finite gap. A likely explan-ation is therefore that tip-induced band bending at52D Mater. 12 (2025) 015023 K Nisi et alFigure 3. Defects in monolayer MoS2. (a) Constant current STM image (Vbias = 1.5 V, I = 30 pA) of ML MoS2 on hBNshowcasing several point defects. Defect A is a negatively charged point defect, defects B are oxygen-passivated sulfur vacancies inthe top and bottom layer, and defect C is a negatively charged sulfur vacancy. (b) Energy level scheme of the MoS2/hBN/EGheterostructure with a charged sulfur vacancy exhibiting Jahn–Teller splitting∆JT. (c), (d) STS line scans across a charged sulfurvacancy defect within a MoS2/EG heterostructure. Inset: close-up STM image of the defect (Vbias = 0.45 V and I = 100 pA). Threedefect states (HOMO−1, HOMO and LUMO) are visible. (e), (f) STS line scans across a sulfur vacancy within a MoS2/hBN/EGheterostructure. Inset: close-up STM image of the defect (Vbias = 1.5 V and I = 30 pA). Only the defect state above the Fermilevel is visible (black triangle). The white solid lines in (d) and (f) are guides to the eye tracing the upward band bending in theCBM, which indicates a negative charge at the defect. The spatially dependent tunneling spectra are vertically offset for clarity.positive bias enables a lateral tunneling process of theelectron localized on the defect into the CB states[46, 47].2. Discussion and conclusionWe investigated tunneling through monolayerMoS2/few-layer hBN/graphene heterostructures bymeans of LT-STS and room temperature cAFM. InSTS, we find indications that bulk hBN reduces thecharge transfer from graphene to MoS2 resulting ina Fermi level shift of −0.4 eV, such that the valenceband and CB onsets of MoS2 are approximately loc-ated at −1.5 V and 1.2 V, respectively, resulting ina close to intrinsic Fermi level. Somewhat counter-intuitively, spectroscopy on individual sulfur vacan-cies still reveals negative charging of the investigateddefects. This can be understood in terms of a very longlifetime of the trapped charge at low-temperaturesenabled by the thick hBN decoupling barrier, whichprevents any trapped charge from tunneling into thesubstrate representing an effective Coulomb block-ade of the defect state [48]. For the investigated side-contacted geometry, current flow in MoS2 with athick hBN tunneling barrier below occurs thereforedominantly via lateral transport, which leads to adiminished visibility of states near the valence bandvs. states near the CB due to suppressed transport ofholes injected from the tip. At room temperature, wefind that the finite conductivity of MoS2 can providea sufficient side-contact, similar to other STS studiesconducted at room temperature [30] and 77 K [49].Therefore, the lateral current flow competes with thevertical tunneling conductivity down to few-layerhBN. Consequently, the interplay between verticaltunneling and lateral transport plays an importantrole for device structures with both a thin tunnelingbarrier and a side-contact to the semiconducting lay-ers. This configuration is typical for gated structures,except for configurations where the semiconductinglayer is electrostatically floating. For the STS experi-ments, one should also consider that the tip typicallyapproaches the sample at a specific set bias until aset current is reached, which determines the distancebetween tip and sample. This distance is kept con-stant, and the bias is ramped. To favor tunnelingcurrents over lateral transport in low-temperatureSTS, it may thus be advantageous to use higher cur-rent setpoints and thus smaller tip-sample distancesor to eliminate the lateral contact between the semi-conductor and the substrate in the heterostructureassembly.In conclusion, the layer dependent cAFM meas-urements suggest that 2–3 layers of hBN may present62D Mater. 12 (2025) 015023 K Nisi et alan optimal regime to interrogate the defect states byelectronic scanning probe methods. In this regime,the MoS2 is already decoupled from the under-lying graphene, yet direct tunneling through thehBN is still possible, which is necessary to preventCoulomb blockade of the defect state. In STM, theunambiguous identification of hBN barrier thick-ness with large exfoliated samples presents a prac-tical experimental challenge. A recent study of defectsin multilayer WSe2 demonstrated that the lifetimeof charged defects depends on the underlying tun-neling thickness [50]. Then the width of the tunnel-ing barrier can be extracted from the current satur-ation behavior at small tip-sample distances in STS[48]. The latter approach therefore provides a suitablemethod to determine the thickness of the hBN barrierdirectly in future STS experiments towards probingboth electronic and many-body optical properties ofthe defects at their native length scale.3. Methods3.1. Sample preparationBulk crystals of hBN and MoS2 were exfoliatedusing mechanical cleavage and the monolayers weretransferred with micrometer precision onto epitaxi-ally grown graphene on (6H)-SiC substrate. To thisend, graphene was grown on (6H)-SiC wafers basedon the following [51]. SiC-wafers (purchased fromUniversity Wafer Inc.) were diced, sonicated in meth-anol, and flash annealed under vacuum at 750 ◦Cin an inductively heated reactor tube held within agraphite susceptor. Samples were then etched with aslow flow of hydrogen (0.5 SLM) in the presence ofargon (p = 1 bar) at 1400 ◦C for 5 min to preparethe surface for graphene growth.Monolayer graphenesamples were then grown by heating the sample to1620 ◦C for 20 min under argon (p = 1 bar) inaddition to a slow flow of argon (0.5 SLM). Sampleswere left to cool to room temperature after eachstep. Heating and cooling rates were 2 ◦C–3 ◦C persecond. hBN was exfoliated onto a Si/SiO2 substrateusing a clear adhesive tape (Nitto Lensguard7568),lifted with a PC/PDMS stamp at 70 ◦C and placedon the sample at 180 ◦C in ambient atmosphere.The hBN/EG substrate was cleaned in chloroform(rt, 10 min). To achieve large exfoliated monolayers,MoS2 was exfoliated using Nitto clear tape and trans-ferred directly from the tape onto a PDMS stamp. TheMoS2 was subsequently released at 40 ◦C from thePDMS onto the hBN flake in ambient atmosphere.Finally, the whole heterostructure was cleaned inchloroform (rt, 10min). The layer thicknesses of eachmaterial were determined afterwards. Prior to scan-ning probemeasurements, the samples were annealedat approximately 500 K for 24 h in the preparationchamber (in ultra-high vacuum, at approximately10−9 mbar). The annealing further reduces surfacecontaminants and residue from the viscoelastic trans-fer. The samples were stored in N2-gas atmosphereafter heterostructure preparation and in between thedifferent measurements.3.2. STM and spectroscopySTM and STS measurements were conducted usinga scanning probe microscope (Createc GmbH) atliquid helium temperatures (T < 6 K) and UHV(p < 10−10 mbar) conditions. The heterostructurewas located utilizing two optical ports and previ-ously acquired optical images. The etched tungstentip was prepared on a gold substrate. Tip stabilitywas further verified on gold substrates for all biasvoltage ranges presented. Specifically, after formingthe tip, we trained the tip at a high voltage rangefrom −4 V to 4.5 V. A priori, it is unknown atwhich exact position on the heterostructure the tiplands. Therefore, it is necessary to approach at a biasvoltage above the gap of hBN (approximately 4.5 V).After a successful approach, we incrementally reducedthe bias voltage in several steps. For materials witha bandgap, such as MoS2 and hBN, the voltage isreduced until the smallest possible value that allowsa stable tunneling current. For graphene, the biascan be reduced to values around 0 V at the set-point of 150 pA. All STM images were acquired inconstant-current mode with a bias voltage appliedto the sample. Constant-height STS measurements(setpoint = 150 pA) were recorded utilizing a lock-in amplifier (frequency 683 Hz, amplitude 5 mV).Band gaps from STS were determined by applyinga linear fit to the valence band edge, CB edge, andthe bottom of the band gap in log(dI/dV). All STMimages are presented as measured, with the exceptionof the inset in figure 3(e), which is filtered using aFourier-filter.3.3. Conductive atomic force microscopycAFMmeasurements were done at room temperaturein an inert N2-atmosphere with a commercial AFMsystem (Bruker Dimension Icon). The area of interestwas first cleaned from all residues via AFM ironingby scanning over the area in contact mode with anapplied force of approximately 300 nN. The min-imum force sufficient to clean the surfacewas determ-ined by scanning on a different flake and increment-ally increasing the contact force until all residues wereremoved. The tip was calibrated by using a standardthermal tune tip calibration procedure. cAFM meas-urements were conducted in PeakForce TUNA modeusing cantilevers with a nominal spring constant of0.4 N m−1, a force of about 1 nN and a conductivemetal Cr/Au coating (Tip: HQ:NSC19/Cr-Au, nom-inal radius <35 nm). The bias voltage was appliedat the sample. IV-curves were measured by sweepingfrom negative to positive bias voltages with a sweeprate of 2–6 Vs−1.72D Mater. 12 (2025) 015023 K Nisi et alData availability statementAll data that support the findings of this study areincluded within the article (and any supplementaryfiles).AcknowledgmentWe thank Johannes Figueiredo for his help with thehelium ion microscopy. Work at TUM was sup-ported by the Deutsche Forschungsgemeinschaft(DFG) via e-conversion—EXC 2089/1–390776260,the Munich Center for Quantum Science andTechnology (MCQST)—EXC 2111–390814868 aswell as the Munich Quantum Valley K6 and theOne Munich Strategy Forum—EQAP. Work at theMolecular Foundry was supported by the Officeof Science, Office of Basic Energy Sciences, of theU.S. Department of Energy under Contract No. DE-AC02-05CH11231. C K and A H acknowledge sup-port through TUM International Graduate School ofScience and Engineering (IGSSE). C K acknowledgesfunding through the European Union’s HorizonEurope research and innovation programme underGrant No. 101076915 (2DTopS). K W and T Tacknowledge support from the JSPS KAKENHI(Grant Nos. 21H05233 and 23H02052) and WorldPremier International Research Center Initiative(WPI), MEXT, Japan.Conflict of interestThe authors have no conflicts to disclose.ORCID iDsJ C Thomas https://orcid.org/0000-0002-2151-7725S Levashov https://orcid.org/0009-0001-4244-8554K Watanabe https://orcid.org/0000-0003-3701-8119T R Kuykendall https://orcid.org/0000-0003-1362-3285J Eichhorn https://orcid.org/0000-0003-2413-6079A W Holleitner https://orcid.org/0000-0002-8314-4397A Weber-Bargioni https://orcid.org/0000-0003-2986-1819C Kastl https://orcid.org/0000-0001-5309-618XReferences[1] Kang S, Lee D, Kim J, Capasso A, Kang H S, Park J-W,Lee C-H and Lee G-H 2020 2D semiconducting materials forelectronic and optoelectronic applications: potential andchallenge 2D Mater. 7 022003[2] McCreary A, Kazakova O, Jariwala D and Al Balushi Z Y2021 An outlook into the flat land of 2D materials beyondgraphene: synthesis, properties and device applications 2DMater. 8 013001[3] Sarkar D, Xie X, Liu W, Cao W, Kang J, Gong Y, Kraemer S,Ajayan P M and Banerjee K 2015 A subthermionic tunnelfield-effect transistor with an atomically thin channel Nature526 91–95[4] Wang J, Sciarrino F, Laing A and Thompson M G 2020Integrated photonic quantum technologies Nat. Photon.14 273–84[5] Kianinia M, Xu Z Q, Toth M and Aharonovich I 2022Quantum emitters in 2D materials: emitter engineering,photophysics, and integration in photonic nanostructuresAppl. Phys. Rev. 9 011306[6] Britnell L et al 2012 Electron tunneling throughultrathin boron nitride crystalline barriers Nano Lett.12 1707–10[7] Román R J P et al 2021 Band gap measurements ofmonolayer h-BN and insights into carbon-related pointdefects 2D Mater. 8 044001[8] Hattori Y, Taniguchi T, Watanabe K and Nagashio K 2015Layer-by-layer dielectric breakdown of hexagonal boronnitride ACS Nano 9 916–21[9] Ranjan A, Raghavan N, Holwill M, Watanabe K, Taniguchi T,Novoselov K S, Pey K L and O’Shea S J 2021 Dielectricbreakdown in single-crystal hexagonal boron nitride ACSAppl. Electron. Mater. 3 3547–54[10] Lee G-H, Yu Y-J, Lee C, Dean C, Shepard K L, Kim P andHone J 2011 Electron tunneling through atomically flat andultrathin hexagonal boron nitride Appl. Phys. Lett.99 243114[11] Michaelis De Vasconcellos S, Wigger D, Wurstbauer U,Holleitner A W, Bratschitsch R and Kuhn T 2022Single-photon emitters in layered van der Waals materialsPhys. Status Solidi b 259 2100566[12] Khalid S, Medasani B, Lyons J L, Wickramaratne D andJanotti A 2024 The deep-acceptor nature of the chalcogenvacancies in 2D transition-metal dichalcogenides 2D Mater.11 021001[13] Hötger A et al 2023 Photovoltage and photocurrentabsorption spectra of sulfur vacancies locally patterned inmonolayer MoS2 Nano Lett. 23 11655–61[14] Mitterreiter E, Schuler B, Cochrane K A, Wurstbauer U,Weber-Bargioni A, Kastl C and Holleitner A W 2020Atomistic positioning of defects in helium ion treatedsingle-layer MoS2 Nano Lett. 20 4437–44[15] Micevic A et al 2022 On-demand generation of opticallyactive defects in monolayer WS2 by a focused helium ionbeam Appl. Phys. Lett. 121 183101[16] Thomas J C et al 2024 A substitutional quantum defect inWS2 discovered by high-throughput computationalscreening and fabricated by site-selective STM manipulationNat. Commun. 15 3556[17] Gupta S, Yang J-H and Yakobson B I 2019 Two-levelquantum systems in two-dimensional materials for singlephoton emission Nano Lett. 19 408–14[18] Xiang F et al 2024 Charge state-dependent symmetrybreaking of atomic defects in transition metaldichalcogenides Nat. Commun. 15 2738[19] Jansen D, Tounsi T, Fischer J, Krasheninnikov A V,Michely T, Komsa H-P and Jolie W 2024 Tip-inducedcreation and Jahn-Teller distortions of sulfur vacancies insingle-layer MoS2 Phys. Rev. B 109 195430[20] Aliyar T et al 2024 Symmetry breaking and spin-orbitcoupling for individual vacancy-induced in-gap states inMoS2 monolayers Nano Lett. 24 2142–8[21] Tan A M Z, Freysoldt C and Hennig R G 2020First-principles investigation of charged dopants anddopant-vacancy defect complexes in monolayer MoS2 Phys.Rev. Mater. 4 114002[22] Akkoush A, Litman Y and Rossi M 2024 A hybrid-densityfunctional theory study of intrinsic point defects in MX2(M=Mo,W; X= S, Se) monolayers Phys. Status Solidi a221 23001808https://orcid.org/0000-0002-2151-7725https://orcid.org/0000-0002-2151-7725https://orcid.org/0000-0002-2151-7725https://orcid.org/0009-0001-4244-8554https://orcid.org/0009-0001-4244-8554https://orcid.org/0009-0001-4244-8554https://orcid.org/0000-0003-3701-8119https://orcid.org/0000-0003-3701-8119https://orcid.org/0000-0003-3701-8119https://orcid.org/0000-0003-1362-3285https://orcid.org/0000-0003-1362-3285https://orcid.org/0000-0003-1362-3285https://orcid.org/0000-0003-2413-6079https://orcid.org/0000-0003-2413-6079https://orcid.org/0000-0003-2413-6079https://orcid.org/0000-0002-8314-4397https://orcid.org/0000-0002-8314-4397https://orcid.org/0000-0002-8314-4397https://orcid.org/0000-0003-2986-1819https://orcid.org/0000-0003-2986-1819https://orcid.org/0000-0003-2986-1819https://orcid.org/0000-0001-5309-618Xhttps://orcid.org/0000-0001-5309-618Xhttps://doi.org/10.1088/2053-1583/ab6267https://doi.org/10.1088/2053-1583/ab6267https://doi.org/10.1088/2053-1583/abc13dhttps://doi.org/10.1088/2053-1583/abc13dhttps://doi.org/10.1038/nature15387https://doi.org/10.1038/nature15387https://doi.org/10.1038/s41566-019-0532-1https://doi.org/10.1038/s41566-019-0532-1https://doi.org/10.1063/5.0072091https://doi.org/10.1063/5.0072091https://doi.org/10.1021/nl3002205https://doi.org/10.1021/nl3002205https://doi.org/10.1088/2053-1583/ac0d9chttps://doi.org/10.1088/2053-1583/ac0d9chttps://doi.org/10.1021/nn506645qhttps://doi.org/10.1021/nn506645qhttps://doi.org/10.1021/acsaelm.1c00469https://doi.org/10.1021/acsaelm.1c00469https://doi.org/10.1063/1.3662043https://doi.org/10.1063/1.3662043https://doi.org/10.1002/pssb.202100566https://doi.org/10.1002/pssb.202100566https://doi.org/10.1088/2053-1583/ad2108https://doi.org/10.1088/2053-1583/ad2108https://doi.org/10.1021/acs.nanolett.3c03517https://doi.org/10.1021/acs.nanolett.3c03517https://doi.org/10.1021/acs.nanolett.0c01222https://doi.org/10.1021/acs.nanolett.0c01222https://doi.org/10.1063/5.0118697https://doi.org/10.1063/5.0118697https://doi.org/10.1038/s41467-024-47876-3https://doi.org/10.1038/s41467-024-47876-3https://doi.org/10.1021/acs.nanolett.8b04159https://doi.org/10.1021/acs.nanolett.8b04159https://doi.org/10.1038/s41467-024-47039-4https://doi.org/10.1038/s41467-024-47039-4https://doi.org/10.1103/PhysRevB.109.195430https://doi.org/10.1103/PhysRevB.109.195430https://doi.org/10.1021/acs.nanolett.3c03681https://doi.org/10.1021/acs.nanolett.3c03681https://doi.org/10.1103/PhysRevMaterials.4.114002https://doi.org/10.1103/PhysRevMaterials.4.114002https://doi.org/10.1002/pssa.202300180https://doi.org/10.1002/pssa.2023001802D Mater. 12 (2025) 015023 K Nisi et al[23] Trishin S, Lotze C, Krane N and Franke K J 2023 Electronicand magnetic properties of single chalcogen vacancies inMoS2/Au(111) Phys. Rev. B 108 165414[24] Cao Q, Kreßler M, Hußmann M, Hu Y, Kusch P and Eigler S2024 Photoluminescence modulation of graphene/mos2heterostructures separated by laser-inducedfunctionalization Chem. Mater. 36 3267–76[25] Lorchat E et al 2020 Filtering the photoluminescence spectraof atomically thin semiconductors with graphene Nat.Nanotechnol. 15 283–8[26] Zhang L et al 2017 Electrically driven single-photon emissionfrom an isolated single molecule Nat. Commun. 8 580–7[27] Wu S W, Ogawa N and HoW 2006 Atomic-scale coupling ofphotons to single-molecule junctions Science 312 1362–5[28] Geng H, Tang J, Wu Y, Yu Y, Guest J R and Zhang R 2024Imaging valley excitons in a 2D semiconductor withscanning tunneling microscope-induced luminescence ACSNano 18 8961–70[29] Kaiser K, Lieske L A, Repp J and Gross L 2023 Charge-statelifetimes of single molecules on few monolayers of NaCl Nat.Commun. 14 4988[30] Klein J et al 2019 Impact of substrate induced band tail stateson the electronic and optical properties of MoS2 Appl. Phys.Lett. 115 261603[31] Li H et al 2024 Imaging moiré excited states withphotocurrent tunnelling microscopy Nat. Mater.23 633–9[32] Murray C et al 2019 Comprehensive tunneling spectroscopyof quasifreestanding MoS2 on graphene on Ir(111) Phys. Rev.B 99 115434[33] Liu X, Balla I, Bergeron H, Campbell G P, Bedzyk M J andHersam M C 2016 Rotationally commensurate growth ofMoS2 on epitaxial graphene ACS Nano 10 1067–75[34] Zhang C, Chen Y, Johnson A, Li M-Y, Li L-J, Mende P C,Feenstra R M and Shih C-K 2015 Probing critical pointenergies of transition metal dichalcogenides:surprising indirect gap of single layer WSe2 Nano Lett.15 6494–500[35] Zhang Y, Brar V W, Wang F, Girit C, Yayon Y, Panlasigui M,Zettl A and Crommie M F 2008 Giant phonon-inducedconductance in scanning tunnelling spectroscopy ofgate-tunable graphene Nat. Phys. 4 627–30[36] Pielíc B, Muževíc M, Novko D, Cai J, Bremerich A,Ohmann R, Kralj M, Šrut Rakíc I and Busse C 2024 Probingthe interplay of interactions, screening and strain inmonolayer MoS2 via self-intercalation npj 2D Mater. Appl.8 61[37] Rosenberger M R, Chuang H-J, McCreary K M,Hanbicki A T, Sivaram S V and Jonker B T 2018Nano-”squeegee” for the creation of clean 2D materialinterfaces ACS Appl. Mater. Interfaces 10 10379–87[38] Li R, Taniguchi T, Watanabe K and Xue J 2020 Detectingband profiles of devices with conductive atomic forcemicroscopy Rev. Sci. Instrum. 91 073702[39] Velický M et al 2020 Electron tunneling through boronnitride confirms marcus-hush theory predictions forultramicroelectrodes ACS Nano 14 993–1002[40] de Araújo D B et al 2020 Controlling the electronic bands ofa 2D semiconductor by force microscopy 2DMater. 7 045029[41] Goodman A J, Willard A P and Tisdale W A 2017 Excitontrapping is responsible for the long apparent lifetime inacid-treated MoS2 Phys. Rev. B 96 121404[42] Barja S et al 2019 Identifying substitutional oxygen as aprolific point defect in monolayer transition metaldichalcogenides Nat. Commun. 10 3382[43] Schuler B et al 2019 Large spin-orbit splitting of deep in-gapdefect states of engineered sulfur vacancies in monolayerWS2 Phys. Rev. Lett. 123 76801[44] Khalid S, Janotti A and Medasani B 2024 Role of chalcogenvacancies and hydrogen in the optical and electricalproperties of bulk transition-metal dichalcogenides 2DMater. 11 031003[45] Aghajanian M et al 2020 Resonant and bound states ofcharged defects in two-dimensional semiconductors Phys.Rev. B 101 081201[46] Teichmann K, Wenderoth M, Loth S, Ulbrich R G,Garleff J K, Wijnheijmer A P and Koenraad P M 2008Controlled charge switching on a single donor with ascanning tunneling microscope Phys. Rev. Lett. 101 076103[47] Schofield S R, Studer P, Hirjibehedin C F, Curson N J,Aeppli G and Bowler D R 2013 Quantum engineering at thesilicon surface using dangling bonds Nat. Commun. 4 1649[48] Zhang R, Clark G, Xu X, Darancet P T and Guest J R 2021Observation of single-electron transport and charging onindividual point defects in atomically thin WSe2 J. Phys.Chem. C 125 14056–64[49] Tilak N, Li G, Taniguchi T, Watanabe K and Andrei E Y 2023Moiré potential, lattice relaxation, and layer polarization inmarginally twisted MoS2 bilayers Nano Lett. 23 73–81[50] Bobzien L et al 2024 Layer-dependent charge state lifetime ofsingle Se vacancies in WSe2 (arXiv:2407.04508v1) pp 1–17[51] Emtsev K V et al 2009 Towards wafer-size graphene layers byatmospheric pressure graphitization of silicon carbide Nat.Mater. 8 203–79https://doi.org/10.1103/PhysRevB.108.165414https://doi.org/10.1103/PhysRevB.108.165414https://doi.org/10.1021/acs.chemmater.3c03166https://doi.org/10.1021/acs.chemmater.3c03166https://doi.org/10.1038/s41565-020-0644-2https://doi.org/10.1038/s41565-020-0644-2https://doi.org/10.1038/s41467-017-00681-7https://doi.org/10.1038/s41467-017-00681-7https://doi.org/10.1126/science.1124881https://doi.org/10.1126/science.1124881https://doi.org/10.1021/acsnano.3c12555https://doi.org/10.1021/acsnano.3c12555https://doi.org/10.1038/s41467-023-40692-1https://doi.org/10.1038/s41467-023-40692-1https://doi.org/10.1063/1.5131270https://doi.org/10.1063/1.5131270https://doi.org/10.1038/s41563-023-01753-4https://doi.org/10.1038/s41563-023-01753-4https://doi.org/10.1103/PhysRevB.99.115434https://doi.org/10.1103/PhysRevB.99.115434https://doi.org/10.1021/acsnano.5b06398https://doi.org/10.1021/acsnano.5b06398https://doi.org/10.1021/acs.nanolett.5b01968https://doi.org/10.1021/acs.nanolett.5b01968https://doi.org/10.1038/nphys1022https://doi.org/10.1038/nphys1022https://doi.org/10.1038/s41699-024-00488-3https://doi.org/10.1038/s41699-024-00488-3https://doi.org/10.1021/acsami.8b01224https://doi.org/10.1021/acsami.8b01224https://doi.org/10.1063/5.0008412https://doi.org/10.1063/5.0008412https://doi.org/10.1021/acsnano.9b08308https://doi.org/10.1021/acsnano.9b08308https://doi.org/10.1088/2053-1583/aba5cbhttps://doi.org/10.1088/2053-1583/aba5cbhttps://doi.org/10.1103/PhysRevB.96.121404https://doi.org/10.1103/PhysRevB.96.121404https://doi.org/10.1038/s41467-019-11342-2https://doi.org/10.1038/s41467-019-11342-2https://doi.org/10.1103/PhysRevLett.123.076801https://doi.org/10.1103/PhysRevLett.123.076801https://doi.org/10.1088/2053-1583/ad4720https://doi.org/10.1088/2053-1583/ad4720https://doi.org/10.1103/PhysRevB.101.081201https://doi.org/10.1103/PhysRevB.101.081201https://doi.org/10.1103/PhysRevLett.101.076103https://doi.org/10.1103/PhysRevLett.101.076103https://doi.org/10.1038/ncomms2679https://doi.org/10.1038/ncomms2679https://doi.org/10.1021/acs.jpcc.1c02791https://doi.org/10.1021/acs.jpcc.1c02791https://doi.org/10.1021/acs.nanolett.2c03676https://doi.org/10.1021/acs.nanolett.2c03676https://arxiv.org/abs/2407.04508v1https://doi.org/10.1038/nmat2382https://doi.org/10.1038/nmat2382 Scanning probe spectroscopy of sulfur vacancies and MoS2 monolayers in side-contacted van der Waals heterostructures 1. Results 2. Discussion and conclusion 3. Methods 3.1. Sample preparation 3.2. STM and spectroscopy 3.3. Conductive atomic force microscopy References