Electronically Tunable Transparent Conductive Thin Films for Scalable Integration of 2D Materials with Passive 2D–3D Interfaces

A novel transparent conductive support structure for scalable integration of 2D materials is demonstrated, providing an electronically passive 2D–3D interface while also enabling facile interfacial charge transport. This structure, which comprises an evaporated nanocrystalline carbon (nc‐C) film beneath nanometer‐thin atomic layer deposited AlOx, is thermally stable and allows direct chemical vapor deposition of 2D materials onto the surface. The combination of spatial uniformity, enhanced charge screening, and low interface defect concentrations yields a tenfold enhancement of MoS2 photoluminescence intensity compared to flakes on conventional Si/SiO2, while also retaining the strong optical contrast for monolayer flakes. Tunneling across the ultrathin AlOx enables facile interfacial charge injection, which is utilized for high‐resolution scanning electron microscopy and photoemission electron microscopy with no detectable charging. Thus, this combination of scalable fabrication and electronic conductivity across a weakly interacting 2D–3D interface opens up new opportunities for device integration and characterization of 2D materials.

interface would enable unprecedented control over (opto)electronic properties of 2D materials and devices. Currently, SiO 2 films on Si serve as the substrates of choice for both chemical vapor deposited and mechanically transferred 2D materials. [6][7][8][9] One major benefit of these Si/SiO 2 substrates is that strong optical contrast between mono-and few-layer 2D materials can be achieved by controlling the thickness of the SiO 2 dielectric film, thereby enabling flakes to be spatially located and the number of layers to be readily determined using nondestructive optical microscopy. [10][11][12][13] While these substrates offer advantages for numerous 2D device applications and enable back-gating of 2D transistors, [8,14] they also suffer from small gate capacitances defined by the large oxide thickness required to achieve optimal optical contrast. Furthermore, they are poorly suited for realizing electrically driven optical devices and are incompatible with several common characterization techniques, such as X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM), and Kelvin probe force microscopy (KPFM), which rely on fast carrier discharging. These critical gaps can be overcome by the development of a new interface structure that provides optical transparency, tunable gate impedance, and broad compatibility with different substrates and processes, including for direct chemical vapor deposition (CVD) of 2D materials on the surface. Since such support structures must also be chemically inert and ultra-smooth, simultaneously fulfilling all of these requirements represents a significant challenge.
Although transparent conductive films (TCFs), such as indium tin oxide (ITO) and fluorine-doped tin oxide (FTO), [15] are key components in a broad range of optoelectronic devices, [15][16][17][18] they are susceptible to high temperature instabilities, rendering them unsuitable as substrates for 2D material growth. In addition, TCFs typically have relatively high surface roughnesses (≈1 nm), [16] whereas ultra-smooth surface morphologies and potential landscapes are required for 2D semiconductors to achieve high carrier mobilities, homogeneous carrier densities, and reproducible (opto)electronic characteristics. [19] To overcome these limitations, graphene-based TCFs were recently investigated for integration into field effect transis tors and optical modulators. [20] Although these systems

Introduction
2D van der Waals (vdW) materials offer significant promise for application in advanced optoelectronics, quantum technologies, and catalysis. [1][2][3] While the extreme sensitivity of 2D materials to the surrounding environment opens unique application prospects, [4,5] device functionality is limited by the availability of substrates possessing tunable combinations of optical, electronic, and chemical properties. For example, the ability to tailor both dielectric and interfacial transport properties at the 2D-3D can exploit the high electrical conductivity, mechanical flexibility, and optical transparency of graphene, [20] single graphene layers have limited sheet carrier concentrations that can be easily screened, limiting their application as gate electrodes. Furthermore, fabrication of large-area graphene-based electronic devices is complex, often requiring transfer processes that can introduce contaminants and structural defects. [21,22] Here, we present a novel TCF structure based on conductive and transparent nanocrystalline carbon (nc-C) films that overcome the limitations of existing substrates. Although studied for many decades, nc-C remains largely underexplored as a TCF, including for 2D materials applications. We show that nc-C substrates can be produced on a large scale with tunable sheet carrier concentrations. Using atomic layer deposition (ALD), we conformally coat the nc-C layer with an amorphous aluminum oxide (AlO x ) layer, having a controlled thickness (down to 1 nm and even below [23,24] ). While the conductive (ρ < 0.01 Ω cm) and ultrasmooth (rms roughness ≈0.2 nm) carbon coating serves as the transparent electrode, the conformal alumina layer acts as a high-k dielectric (k ≈7 [25] ) spacer with tunable impedance. Of particular relevance to 2D materials, this structure is thermally and chemically stable under typical growth conditions for CVD of transition metal dichalcogenides, with the alumina layer facilitating the nucleation and growth of single layer MoS 2 flakes. Importantly, when the nc-C/ AlO x film is deposited on a Si substrate coated with 270 nm SiO 2 , the high optical contrast for discerning single and multilayer flakes is retained, thereby enabling subsequent analysis and processing. To demonstrate the utility of this structure, we show that 2D flakes can be directly characterized by electron microscopy and spectroscopy techniques, including scanning electron microscopy (SEM), laboratory-based X-ray photoelectron spectroscopy (XPS), and synchrotron-based X-ray photoemission electron microscopy (XPEEM), without charging effects that would otherwise impair such measurements on conventional Si/SiO 2 supports. This is possible because the AlO x thickness is within the length scale of tunneling, thereby enabling rapid discharging of the surface layers. Raman spectroscopy shows good crystal quality of MoS 2 and photoluminescence (PL) measurements reveal that the TCF structure yields significantly enhanced MoS 2 PL emission intensity compared to flakes on SiO 2 substrates, which is attributed to a low interfacial defect concentration and effective electronic shielding of charged defects within SiO 2 . Beyond the application of the nc-C/AlO x TCF for 2D materials, the presented scheme provides a highly versatile and flexible route to the uniform and scalable deposition of TCFs on other substrates of relevance to thin film optical and electronics applications. Figure 1a illustrates the smooth, conductive, and optically transparent nc-C/AlO x structure (hereafter also referred to as the TCF structure), which was deposited onto a Si substrate covered by a 270 nm thick layer of thermal SiO 2 . We note that the Si/SiO 2 substrate was selected here as a model support that is commonly used to provide excellent optical contrast in 2D materials research, although alternative substrates could also be utilized. The as-deposited structure was probed using atomic force microscopy (AFM) and found to have a low surface roughness (≈ 0.2 nm).

Fabrication and Optical Properties of Multi-Layer Structures
Following fabrication of the Si/SiO 2 /nc-C/AlO x structure, MoS 2 was grown on the surface via atmospheric pressure chemical vapor deposition (CVD) at 850 °C. Importantly, we found that the complete structure is thermally and chemically stable, withstanding the sulfidizing environment at elevated temperatures during CVD (see below), with the AlO x layer promoting the nucleation and growth of primarily single layer triangular MoS 2 flakes across its surface, as indicated by the optical micrograph shown in Figure 1b. In addition to confirming the successful growth of MoS 2 , this micrograph highlights that the strong optical contrast between 2D materials and the substrate is retained, despite the presence of the nc-C/AlO x TCF struc- [26] where G sub and G MoS2 are the substrate and MoS 2 intensities of the green channel of the RGB mix, respectively, yields values of C TCF = 0.28 for Si/SiO 2 /nc-C/AlO x ( Figure 1b) and C Si/SiO2 = 0.26 for standard Si/SiO 2 substrates ( Figure S1, Supporting Information). The retention of strong optical contrast between the 2D sheets and the underlying substrate is an important feature of the nc-C/AlO x TCF that arises from its optical transparency ( Figure S2, Supporting Information), which is a consequence of the wide bandgap (>6.5 eV) of the approximately stoichiometric ALD alumina layer ( Figures S3 and S4, Supporting Information), as well as the nanocrystalline and ultrathin structure of the carbon film.
Complementary AFM measurements (Figure 1c) confirm the dominant presence of monolayer, crystalline MoS 2 flakes, along with a smaller fraction of bi-and tri-layer regions. Furthermore, AFM reveals that the smooth surface morphology of the underlying TCF structure is preserved after CVD growth, with a roughness of ≈0.2 nm (Figure 1c; Figure S5, Supporting Information) that is likely determined by the height variations of the chemomechanically polished Si/SiO 2 substrate (≈0.2 nm). We emphasize that a smooth and chemically inert substrate surface is necessary to preserve the intrinsic properties of few-and mono-layer 2D materials because it reduces microscopic strain, while also facilitating the release of strain without distorting the 2D lattice during cooling from the CVD growth temperature. [27] For comparison purposes, a 3 nm thin gold coating, deposited by electron-beam evaporation onto a Si/SiO 2 substrate, provides a similar resistivity (ρ ∼ 0.01 Ω cm) to the nc-C/AlO x TCF structure, but has a much higher rms roughness (≈1.1 nm) and absorbs light more strongly, thereby reducing the optical contrast of monolayer MoS 2 to 0.17. This comparison highlights the advantages of using nc-C, rather than a metal, as the conductive layer in such a structure.

Advanced Electron Imaging and Spectroscopy Enabled by Conductive Support
In addition to facilitating CVD growth of MoS 2 , the 1 nm thin AlO x coating enables charge carriers to tunnel from MoS 2 to the conductive carbon film, as demonstrated by current-voltage measurements with a mercury probe ( Figure S6, Supporting Information). As a consequence, electron microscopy and spectroscopy methods can be readily applied for the characterization of the 2D material. Figure 1d shows a scanning electron microscopy image of individual monolayer MoS 2 flakes on top of the Si/SiO 2 /nc-C/AlO x structure. No artifacts arising from charging of the 2D material are observed, including at lower magnification (see inset) after initial high-resolution imaging. The resulting SEM images are characterized by sharp features and reveal bilayer nucleation sites within individual monolayer flakes (Figure 1d) that were not always discernable using optical microscopy. In contrast, SEM of MoS 2 on a standard Si/SiO 2 substrate displays significant charging artifacts, which greatly reduce image quality and resolution ( Figure S7, Supporting Information).
The utility of the TCF structure for facilitating nanoscale characterization of 2D materials is further highlighted by X-ray photoemission electron microscopy (XPEEM), which is a versatile spectro-microscopic method for nanoscale composition and chemical analysis of surfaces, but requires efficient discharging of the studied material. The Si/SiO 2 /nc-C/AlO x substrate presented here fulfils this requirement and enables high resolution XPEEM analysis of the local chemical composition of CVD-grown MoS 2 flakes, as shown in Figure 2a The chemical state and purity of CVD MoS 2 on the TCF is further shown by a typical Mo 3d and S 2s XPS core level spectrum obtained from a single flake (Figure 2c). The spectrum is dominated by the spin-orbit split doublet at 230.5 and 233.7 eV, corresponding to pristine Mo-S bonding within the material, along with the S 2s signal at 227.8 eV. The weak doublet having Figure 2. X-ray photoelectron maps of MoS 2 generated by analysis of spatially resolved data from the a) Mo 3d and b) Al 2p core level regions. Acquisition of these images with negligible charging was enabled by the Si/SiO 2 /nc-C/AlO x support, which allows rapid discharging of the surface layers via the conductive nc-C layer. c) A typical Mo 3d-S 2s core level spectrum extracted from a sequence of maps acquired with the photon energy ranging from 239 to 225 eV, simultaneously indicating the high chemical quality of the MoS 2 layer grown by CVD and the possibility for high spatial and energy resolution photoemission from supported flakes. Fitted components to the Mo 3d spectrum are presented in blue (Mo-S), green (Mo-O), and orange (S-Mo), respectively. components at 233.8 and 237.0 eV is attributed to Mo-O and has a concentration of <3.5%, which is well below the oxygen concentrations of intentionally doped MoS 2 films reported by Wei et al. [28] These findings reveal that the 2D MoS 2 is of good quality with minimal oxygen incorporation, doping, or contamination from other chemical species.

Structural and Optoelectronic Properties of CVD MoS 2 on Conductive Support
To further analyze the structural and optoelectronic quality of the CVD MoS 2 on the TCF, we compare Raman and photoluminescence spectra obtained from individual MoS 2 nanosheets on Si/SiO 2 /nc-C/AlO x substrates to those on standard Si/ SiO 2 surfaces. Raman measurements recorded on both substrates (Figure 3a) show that the A′ 1 and E′ Raman modes of the flakes are separated by Δ TCF = 20.94 ± 0.14 cm -1 ( Si/Sio2 ∆ = 20.60 ± 0.10 cm -1 ), which are characteristic of monolayer MoS 2 . The narrow linewidths of the two Raman spectral features, FWHM TCF,A1 ′ = 6.16 ± 0.40 cm -1 and FWHM TCF,E′ = 5.13 ± 0.30 cm -1 (FWHM Si/SiO ,A 2 1 ′ = 5.11 ± 0.25 cm -1 and FWHM Si/SiO,E′ = 4.32 ± 0.08 cm -1 ), are fully consistent with previous reports of exfoliated and CVD-grown monolayer MoS 2 . [29,30] These observations reaffirm the high crystal quality of MoS 2 grown on the TCF structure. Since comparison of the peak positions reveals no shift of the Raman modes within the error (Δ E′ = 0.03 cm -1 , A1 ∆ ′ = 0.31 cm -1 ) for MoS 2 on the TCFcoated substrate relative to MoS 2 on bare Si/SiO 2 , we conclude that an equivalent quality of MoS 2 flakes is achieved on both substrates. Finally, we note that the spectral features at ≈380 and 409 cm -1 that appear on the left-hand (right-hand) side of the E′(A′ 1 ) mode are present in both spectra with similar intensity (Figure 3a). These features are reportedly related to disorder-induced Raman modes at the M point, indicating a similar degree of structural disorder within flakes on both substrates. [31,32] We note that no spectral signatures characteristic of crystalline Al 2 O 3 were detected by Raman spectroscopy, even for 20 nm thick ALD alumina films after annealing at 800 °C. We continue our analysis of substrate-related optical properties of CVD MoS 2 with PL spectroscopy, which allowed us to probe substrate-induced effects with a higher sensitivity than Raman spectroscopy. [33] Figure 3b compares PL spectra recorded under comparable excitation conditions from MoS 2 on the TCF and SiO 2 substrates, as well as on AlO x -coated SiO 2 without the nc-C layer present. The relative intensities, along with their standard deviations from measurements on 15 different spots on each sample, are shown in Figure 3c. Remarkably, the PL intensity for monolayer MoS 2 on the TCF surface increases by more than one order of magnitude compared to the PL intensity of MoS 2 on the regular Si/SiO 2 substrate (Figure 3b,c; Figure S8, Supporting Information). In addition, even in the absence of the nc-C layer, we observed a significant (≈5×) increase of the PL intensity on AlO x . These PL intensity enhancements are accompanied by small blueshifts of the PL maxima from MoS 2 on both the TCF structure (13 meV) and the AlO x surface (9 meV) compared to the SiO 2 surface (Figure 3d). Since this blueshift is not accompanied by a shift of the Raman A′ 1 mode position (Figure 3a) or an increase of low energy PL emission due to the negatively charged trion ( Figure S8, Supporting Information), we conclude that increased electron doping is unlikely to be the dominant origin of this behavior. Taken together, these data suggest that the PL intensity enhancement for MoS 2 on AlO x relative to SiO 2 is a consequence of reduced substrate dangling bond defects and covalent interactions. [34] With the addition of the nc-C layer to form Si/SiO 2 /nc-C/AlO x /MoS 2 , we observed an even larger enhancement of the PL intensity, which is consistent with additional screening of charged defect states that are ubiquitous in SiO 2 by the conductive nc-C layer. [35] These combined effects yield a statistically significant approximately tenfold PL enhancement of MoS 2 on the TCF compared to the reference substrate SiO 2 .

Thermally Induced Transformations of Evaporated Carbon Films
Having demonstrated the applicability of this structure for the growth and characterization of high quality MoS 2 , we now turn to the analysis of the basic properties of the nc-C/AlO x structure, noting that it offers a combination of characteristics that can be broadly useful, potentially reaching beyond 2D material applications. The first fabrication step comprises room temperature deposition of a conformal 10 nm thick amorphous carbon layer, which is electrically resistive (ρ > 10 Ω cm) and smooth (0.17 nm) in its as-deposited state, as summarized in Table 1. As discussed below, the resistivity of this layer is significantly reduced via thermal annealing, including during CVD growth of MoS 2 .
Although ALD AlO x has been previously demonstrated for growth on top of van der Waals materials, [36,37] this can be challenging since a lack of binding sites could inhibit growth on the evaporated carbon. Thus, we used in situ spectroscopic ellipsometry to track the alumina thickness during ALD on the carbon film with Å-level precision ( Figure S9, Supporting Information). The lack of reactive sites on the carbon surface can explain the observed delayed film nucleation ( Figure S9, Supporting Information), requiring chemical activation during the first several ALD cycles. However, following this nucleation period, deposition of a nanometer-thin conformal AlO x coating on the as-deposited amorphous carbon layer was achieved. While charge carriers can directly tunnel through the nanometer-thin AlO x coating ( Figure S6b, Supporting Information), we determined a breakdown electric field of ≈5.7 MV cm -1 for a 5.5 nm thick alumina layer ( Figure S6c, Supporting Information), which is consistent with literature [38] and demonstrates the applicability of the nc-C/AlO x structure for field-effect transistors.
As summarized in Table 1, growth of the ALD layer has a minor impact on the high resistivity of the as-deposited carbon layer. However, we find that annealing at 800 °C in Ar atmosphere transforms the carbon to a smooth nanocrystalline layer with significantly lower resistivity of 1.2 × 10 -2 Ω cm ( Table 1). As shown in Figure S2, Supporting Information, the decreased resistivity is accompanied by a moderate increase of optical absorption, although the material remains optically transparent over a broad spectral range.
To better understand the effect of thermal processing on the structure and resistivity of the carbon film, we performed Raman spectroscopy at different stages in the fabrication of the nc-C/AlO x TCF. Figure 4 compares the two dominant Raman spectral features of the carbon film, the G peak associated with sp 2 carbon stretching modes and the D peak associated with a breathing mode of the ring-shaped sp 2 bonds present in disordered graphite. , [39,40] From analysis of the Raman peak positions and the intensity ratio of the D peak relative to the G peak, I(D)/I(G), one can assess the degree of disorder in the carbon film and estimate the cluster diameter or cluster correlation length, L a . [41,42] As illustrated in Figure 4a,b, the peak positions as well as the intensity ratio, I(D)/I(G), remained nearly constant after AlO x growth at 200 °C, suggesting that the carbon structure was unaffected by the ALD process. In contrast, after annealing in an Ar atmosphere at 800 °C for 5 min, we observe a strong blueshift of the G peak and an increase of the I(D)/I(G) ratio (Figure 4c and Table 1). Together, these changes indicate formation of graphite nanocrystals via transformation of sp 3 bonds to sp 2 bonds, [39] as well as clustering of sp 2 rich regions. [43] Given prior Raman studies of carbon thin films, [40] we conclude that the carbon layer is predominantly composed of nanocrystalline graphite (sp 2 bonds with L a ≈ 40 Å) with less than 20% sp 3 -bonded atoms. This dominant content of graphitic sp 2 carbon clusters is consistent with the significantly increased conductivities of annealed films [44,45] (Table 1), as well as the partially reduced optical transparency ( Figure S2, Supporting Information). Importantly, performing CVD growth of MoS 2 yields nearly identical Raman spectra as annealing in Ar (Figure 4d and Table 1), indicating that the conversion of the films to graphitic nc-C occurs similarly under the sulfudizing growth conditions. Advantageous for CVD of 2D materials, no pre-annealing of the substrate was required to achieve the conductive carbon layer during fabrication of the complete Si/SiO 2 /nc-C/AlO x structure.

Conclusion
In summary, we have developed a scalable support structure for 2D materials consisting of nanocrystalline carbon coated with an amorphous and conformal ALD-grown AlO x layer. The structure is smooth, conductive, and optically transparent. The ultrathin conformal AlO x coating facilitates direct CVD growth of 2D MoS 2 on the surface, while providing an electronically passive 2D-3D interface that also enables facile interfacial charge transport. While the nc-C/AlO x structure can be applied to various passive or functional substrates, its integration with conventional Si/SiO 2 supports allows the strong optical contrast between single-and multi-layer 2D materials to be retained, while also yielding a strong photoluminescence enhancement. This latter effect is a consequence of low interfacial defect concentrations and screening of charged defect states in the underlying SiO 2 by the conductive nc-C layer. Since ALD allows fabrication of conformal AlO x high-k dielectric spacers of arbitrary thickness down to 1 nm, rapid interfacial charge tunneling across the interface is possible. This feature enables physical, electronic, and chemical characterization using electron imaging and spectroscopy techniques that are otherwise complicated on oxidic supports. As two examples of this, SEM imaging and XPEEM spectromicroscopy are applied to CVD-grown MoS 2 on Si/SiO 2 /nc-C/AlO x supports that simultaneously provide high optical contrast for single layer flakes. In a wider context, the tunability of ALD dielectric coatings provides the opportunity to reliably control interfacial impedance over a wide range while retaining electronically passive and weakly interacting interfaces. Compared to commonly used transparent conductive films and encapsulants, the reported TCF strikes a unique balance of properties, while specifically offering a superior dielectric constant, surface roughness, and ease of fabrication for integration of 2D materials (Table S1, Supporting Information). Furthermore, the deposition process for scalable fabrication of nc-C/AlO x heterostructures is expected to be broadly compatible with diverse substrates and suitable for applications in numerous optoelectronic applications, extending beyond 2D materials. [46,47]

Experimental Section
Material Deposition: A Quorum Technologies Q150T system featuring pulse evaporation from a carbon rod was used to evaporate a 10 nm thick amorphous carbon (a-C) film on n + Si/SiO 2 substrates. The carbon layer was then coated with AlO x in a hot-wall plasma-enhanced atomic layer deposition (PE-ALD) reactor (Fiji G2, Veeco CNT) in continuous flow mode at 200 °C. In the first half-cycle, the process was conducted with a base pressure of ≈0.09 Torr. Here, ozone was used as the oxidant and trimethylaluminum TMA (electronic grade, 99.999%, STREM Chemicals) as the precursor for AlO x and Ar (99.9999%, Linde) as the carrier gas. In the second half-cycle, O 3 was remotely generated using oxygen (99.9999%, Linde) and supplied with 130 ms pulses and peak pressures of ≈20 Torr. Each cycle of the TMA/O 3 process followed the sequence: 80 ms TMA dose, 10 s Ar purge, 130 ms O 3 pulse, 10 s Ar purge. The parameters were implemented so that self-limiting growth occurs, which was controlled by monitoring the film thickness in situ using spectroscopic ellipsometry.
Afterward, single crystalline triangular-shaped MoS 2 was directly grown on top of the AlO x /nc-C thin film. For this process, a home-built chemical vapor deposition (CVD) system was used with a one zone horizontal tube furnace (Nabertherm RS 80/500/11) with a 5 cm inner diameter quartz tube. MoS 2 was grown at ambient pressure (1 atm) with MoO 3 (99.9995%; Alfa Aesar) and S (99.998%; Merck) solid precursors at a target temperature of 850 °C and a flow of 100 sccm of Ar (99.9999%, Linde) as the carrier gas. The TCF on Si/SiO 2 was placed next to the MoO 3 precursor about 17 cm inward from the edge of the furnace. The S precursor was placed upstream of the MoO 3 precursor to obtain a precursor temperature of about 180 °C. The growth time was set to 5 min followed by a natural cool down. For the samples with MoS 2 on the TCF, the annealing of the evaporated carbon happened during the CVD growth process. For all other annealed samples, the same CVD setup and parameters were used without the MoO 3 and S precursors.
In Situ Spectroscopic Ellipsometry: The thickness of the ALD layer was monitored in real-time during the deposition using an in situ spectroscopic ellipsometer (M-2000, J. A. Woollam) and a sampling time of 3 s. The Xenon light source (Hamamatsu, L2174-01) with a spot of ≈5 × 8 mm 2 passed through a fused silica quartz window (Lesker, VPZL-275DU) at an angle of 67°. The data were fit using a general oscillator model.
Atomic Force Microscopy and Scanning Electron Microscopy: AFM images of MoS 2 flakes were acquired using a Bruker Dimension Icon XR (Bruker, USA) AFM in ambient air with platinum silicide (PtSi) probes (Nanosensors). AFM measurements for roughness determination were performed using a Bruker Multimode V microscope (Billerica, MA, USA) in ambient air with NGS30 AFM probes (TipsNano) with a nominal tip radius of 8 nm, typical force constant of 40 N m −1 , and resonance frequency of 320 kHz. All images were acquired in tapping mode with a scan rate of 0.5 kHz and 512-point sampling for image sizes of 12 × 12 µm 2 and 3 × 3 µm 2 . SEM images were collected using a Zeiss NVISION 40 with a secondary electron detector.
Synchrotron X-Ray Photoemission Electron Microscopy: XPS maps were measured at the SIM beamline at the Swiss Light Source (SLS), Paul Scherrer Institut (PSI), by means of X-ray photoemission electron microscopy (XPEEM using a spectroscopic low energy electron microscopy (LEEM) instrument, Elmitec GmbH). During all measurements, a bias of 15 kV was applied between the sample and the objective lens. The spatially resolved X-ray photoelectron spectroscopy images were measured by collecting XPEEM image stacks as a function of photon energy for a fixed value of the energy analyzer (set to 100 eV) to be able to keep the focus settings constant. The kinetic energy of the detected electrons determined the probing depth and was chosen to analyze monolayer MoS 2 . The images were analyzed using a Python script, the Mo 3d core level spectra were extracted from the maps using ImageJ, and the spectra were fitted using CasaXPS. Thereby, the Mo 3d -S 2s core level spectrum (Figure 2c) was calibrated using the adventitious carbon reference peak at 284.8 eV.
Optical Characterization: Raman and PL measurements were performed at room temperature with a 532 nm excitation laser. For all Raman and PL measurements, a home-built Raman setup was used with a Horiba iHR 550 spectrometer (entrance slit 200 µm, 2400 grooves mm −1 grating) and liquid nitrogen cooled Horiba Symphony II CCD detector. For the Raman and PL data of MoS 2 , a 100× objective lens was used, whereas for the Raman data for the different stages of sample fabrication a 20× objective lens was used. All Raman measurements were performed using an excitation power of 5.1 mW, while PL measurements of MoS 2 on Si/SiO 2 and of the TCF were collected at 28 µW. Energy calibration of Raman spectra was carried out using the 520 cm -1 Si Raman line. Raman spectra of MoS 2 were fitted using pseudo-Voigt functions after averaging over three spectra containing an average over 60 sweeps each. The spectra of the nc-C/AlO x films were fit using the Breit-Wigner-Fano (BWF) function for all G peaks [39] and a Lorentzian lineshape for the D peaks. Here, the G peak position was calculated from the fitted peak position, ω 0 , and is smaller than the measured value because ω 0 is due to the undamped mode. [39] Electrical Characterization: For the determination of the sheet resistivity, contacts were defined via shadow masks, electrically contacted those with probes mounted on micromanipulators, and measured the voltage drop between the inner contacts while applying a constant current at the outer contacts using a Keithley 2400 source meter unit. For each given resistivity, over four measurements were averaged. A mercury (Hg) droplet (BASI, mercury electrode), operated in a nitrogen atmosphere (≤10 ppm O 2 ), was used as top electrical contact to characterize out-of-plane charge transport and determine the dielectric breakdown voltage of ALD AlO x films of the TCFs ( Figure S6, Supporting Information). A cross section of the Hg/AlO x /nc-C/SiO 2 /Si substrate structure is shown in Figure S6a (Supporting Information).

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.