Molecular Doping of MoS$_2$ with N-Heterocyclic Carbenes: A Raman Study of Interfacial Charge Transfer

Introduction

The field of two-dimensional (2D) materials has undergone rapid development, driven by their distinctive physical properties and broad potential for technological applications. Among these materials, semiconducting transition metal dichalcogenides (TMDs), such as $MoS_{2}$, have emerged as particularly promising systems. In the monolayer limit, TMDs exhibit a direct bandgap [1, 2], large exciton binding energies [3, 4], and strong light-matter interactions [5, 6], making them attractive for applications in photovoltaics, sensing, and optoelectronics at room temperature [7, 8, 9, 10].

A key feature of $MoS_{2}$ is the high degree of tunability of its electronic and optical properties. This can be achieved through a variety of approaches, including strain engineering [11, 12], electrostatic or chemical doping [13, 14, 15], and the assembly of mono- and few-layer structures into vertical hetero- and hybrid structures [16, 17]. In addition, molecular deposition has recently gained attention as a versatile strategy for modifying the properties of $MoS_{2}$ and other semiconducting TMDs in a controlled manner [18, 19, 20]. Depending on the molecular structure and the interfacial electronic alignment, molecular deposition can enable several mechanisms for tailoring material behavior. These include defect passivation or modification of defect-related states [20, 21, 22], charge doping mediated by interfacial charge transfer [15, 23, 24], and, under resonant conditions, non-radiative energy transfer processes such as Förster resonance energy transfer (FRET) [20, 25].

Recently, N-heterocyclic carbenes (NHCs) have emerged as a particularly suitable class of molecules for the controlled electron doping of TMDs such as MoS$_2$ [26, 27]. NHCs offer strong $\sigma$-donor character, molecular tunability, and the ability to form self-assembled monolayers on gold [28, 29, 30, 31, 32] and other coinage metals [33, 34, 35, 36, 37], base metals [38, 39, 40, 41, 42], and selected semiconductors [43, 44, 45, 46]. On such surfaces, covalent binding models [35, 47, 48, 49, 50], self-assembly [47, 51, 52], stability [28, 53, 54, 55], competing adsorption geometries [31, 37, 56, 57, 58] and interfacial electron transfer [45, 59, 60, 61] have been well explored. In contrast, the binding mode of NHCs on semiconducting TMDs is more complex. Proposed interaction mechanisms range from covalent binding between the carbene and the chalcogenide atoms [27, 62], to frustrated Lewis acid-base pairing between the carbene and the metal atoms at defect sites [26], to physisorption of so-called masked carbenes [63].

Recent studies [27, 28, 63, 64] indicate that NHC-derived species produce substantial interfacial electron doping in TMDs; however, the dominant binding mode and the charge transfer mechanism may depend on the molecular precursor, preparation conditions, defect density, and surface structure. Consequently, molecular deposition can affect the optical, electronic and vibrational properties, in particular the optical phonon modes through various contributions, including charge transfer, strain, and defect- or bonding-related perturbations. A quantitative comparison of as-exfoliated and NHC-covered MoS$_2$ samples is therefore required to distinguish these effects.

From an application perspective, controlled n-type doping of two-dimensional semiconductors is essential for reducing the contact resistance, adjusting the threshold voltage of field-effect transistors, and realizing complementary 2D logic and efficient photodetectors. Substitutional doping, however, introduces impurities into the channel lattice and thereby degrades the transport properties, whereas electrostatic gating requires additional device layers. In conventional semiconductors, modulation doping, in which the donors are spatially separated from the conducting channel, resolved this conflict and enabled high-mobility two-dimensional electron systems [65]. Molecular surface doping is a promising analogue for 2D semiconductors, because the donor molecules reside outside the lattice of the atomically thin channel and leave it unperturbed.

The central question addressed in this study is whether NHC deposition produces a reproducible electron-doping response in exfoliated MoS$_2$ that can be distinguished from strain, defect-related perturbations, and changes in the surrounding atmosphere.

In this work, we employ micro-Raman spectroscopy to investigate the impact of NHC deposition on exfoliated MoS$_2$ samples of varying thickness, ranging from monolayers to bulk-like flakes. From hyperspectral Raman maps acquired on three several micrometer sized samples containing mono-, bi-, few-layer and bulk-like regions, we perform a comparative statistical analysis of the optical phonon modes before and after NHC deposition. By quantitatively evaluating the mode frequencies and their relative shifts, we disentangle the contributions of changes in carrier density, strain, and defect- or covalent-bonding-related perturbations, building on previously established Raman-based analyses [5]. Our analysis reveals a pronounced n-type doping response after NHC deposition, which is strongest in monolayer and bilayer MoS$_2$ and negligible in bulk-like samples. The absence of Raman signatures of strain modification or defect generation after NHC deposition is consistent with a weak chemical perturbation of the MoS$_2$ lattice, although the binding mode cannot be determined from the present data. Comparison of the evolution of the A$_{1g}$ Raman-mode frequency in as-exfoliated and NHC-covered MoS$_2$ bilayers with literature values used for calibration [5, 64, 66] indicates an increase in the two-dimensional electron density by about two-orders of magnitude (up to $10^{13}$ cm$^{-2}$) due to interfacial charge transfer. This behavior resembles a molecular analogue of modulation doping [65], in which donor species located outside the MoS$_2$ lattice supply carriers to the conducting channel while leaving the lattice of the channel unperturbed. NHC deposition may therefore provide a route toward molecular doping in semiconducting TMDs for fundamental studies and optoelectronic applications.

EXPERIMENTAL SECTION

Sample preparation

MoS$_2$ samples were cleaved by micromechanical exfoliation [67], dry-transferred onto Si/SiO$_2$ (300 nm) substrates, and characterized by micro-Raman measurements at room temperature using a commercial Raman microscope (Witec Alpha 300).

Prior NHC deposition, the MoS$_2$ samples were annealed in a vacuum chamber ($10^{-5}$ mbar) at 190°C for approximately one hour. The vacuum chamber and the physical vapor deposition (PVD) chamber were located in a glove bag in nitrogen atmosphere to ensure a controlled environment. After annealing, the MoS$_2$ samples were not exposed to air and were transferred to the PVD chamber within 5 minutes (the setup is described in the Supporting Information, Figure S11). The NHC precursor IMe·CO$_2$ was heated to 120 °C, and the vapor deposition time was about 30 minutes.

Along with the MoS$_2$ sample, a thin-film gold substrate was treated simultaneously as a control sample (Au-NHC). The Au-NHC sample was characterized via XPS (Figure S12 and Table S1). The successful deposition of NHCs on the gold surface proves the presence of free NHCs in the gas phase and, hence, the exposure of theMoS$_2$ samples to free NHC.

	Pass energy [eV]	Dwell time [ms]	Step size [eV]	No. of scans
Survey	200	100	0.90	2
Au4f	50	100	0.05	5
N1s	100	100	0.05	10
C1s	50	50	0.05	5
O1s	50	50	0.05	5

The NHC-covered MoS$_2$ samples were placed inside a small sample chamber (Linkam Cell) filled with dry nitrogen gas directly evaporated from the liquid phase for highest purity. In this way, the Raman measurements of the NHC-covered samples were done in inert nitrogen atmosphere.

Optical measurements

All optical measurements were conducted on a commercial Raman confocal microscope (Witec Alpha 300) with 488 nm (2.54 eV) laser excitation and a diffraction-limited spot size. For Raman measurements, a grating with 1800 lines per mm was used. For as-exfoliated MoS$_2$ (NHC covered MoS$_2$) samples, a Neofluar 100x (50x) objective with 1 (9.1) mm working distance was employed, with laser power up to 1 mW (2 mW) and integration time up to 2 s (1.5 s). The laser power was increased for the NHC-deposited MoS$_2$ samples to maintain a comparable power density, accounting for the different objective and optical losses introduced by the cell window. No degradation or laser-induced modifications were observed in any of the samples, consistent with the low power densities used. Since the analysis focused exclusively on the Raman peak frequencies and full widths at half maximum (FWHMs), rather than on peak intensities, variations in laser power do not affect the reported parameters. For the as-exfoliated MoS$_2$ sample, additional measurements conducted in ambient and N$_2$ atmospheres were performed with a Neofluar 50x objective with a 9.1 mm working distance, a laser power of 2 mW and integration time of 2 s. All measurements were conducted at room temperature. Because the as-exfoliated and NHC-covered samples were measured with different objectives and laser powers, the bulk-like region of each flake serves as an internal reference. For simplicity, the bulk notation A$_{1g}$ and E$_{2g}$ is used for all thicknesses.

RESULTS AND DISCUSSION

A schematic illustration of NHC molecules deposited on the surface of few-layer MoS$_2$ sample is shown in Figure 1a. Figure 1b shows typical Raman spectra of exfoliated monolayer (1L), bilayer (2L), and few-layer (FL) MoS$_2$ samples before and after NHC deposition (sample shown in Figure 3). Here, the in-plane E$_{2g}$ phonon mode is centered at a Raman shift of (386.19 ± 0.03) cm$^{-1}$ and the out-of-plane A$_{1g}$ phonon mode at a Raman shift of (405.26 ± 0.05) cm$^{-1}$ for the as-exfoliated MoS$_2$ monolayer [68, 69, 70]. As depicted in Figure 1b and elaborated in detail later on, the frequency separation between A$_{1g}$ and E$_{2g}$ phonon modes was found to decrease after NHC deposition due to softening of the $A_{1g}$ mode, while the $E_{2g}$ mode shows no shift relative to the as-exfoliated sample. This $A_{1g}$ softening after NHC deposition depends on layer number.

To quantify this energy shift, all Raman spectra were fitted to a sum of two Lorentzian functions, from which information about their frequency shift, intensity, and full width at half maximum was extracted. With increasing number of layers, the frequency of the $A_{1g}$ mode increases due to stiffening of the lattice, while the frequency of the $E_{2g}$ mode decreases due to the combined effect of enhanced dielectric screening of the long-range Coulomb interaction and interlayer hybridization effect [71]. The opposite trend in their thickness-dependent frequency makes the frequency separation of these two modes a reliable indicator of the MoS$_2$ layer number [5, 70, 72]. Additional layer-number dependent Raman data are as shown in Figure S1,Figure S1). In the monolayer limit, the frequency difference between these two modes is below 20 cm$^{-1}$, whereas for bilayers it increases to 20-22 cm$^{-1}$; for bulk-like samples, the frequency difference is $\ge25$ cm$^{-1}$. These values provide useful boundaries, particularly for distinguishing monolayers and bilayers, although the exact values depend on the charge-carrier density [5, 64, 66] and other parameters such as strain and defects. It is worth mentioning that this method for determining the number of layers is valid for exfoliated samples, whereas for as-grown samples the frequency difference has been reported to depend on the synthesis parameters [73].

Beyond their thickness dependence, the frequencies and FWHM of these two phonon modes are also sensitive to external perturbations including strain [74, 75], defects [66, 76], and changes in the electron density induced by electronic or structural dopants [20, 64, 77, 78, 79], as illustrated in Figure S1 and Figure S2. In particular, changes in electron density can modify the phonon frequencies of the $A_{1g}$ mode through electron–phonon coupling [66] (see also Figure S2). Figure 2 a) shows an optical micrograph of an exfoliated multilayer MoS$_2$ flake, in which regions with different layer thicknesses can be identified from their optical contrast arising from the interference effects [80]. Regions with 2L and FL thickness are outlined with dashed white lines. Most of the flake exhibits a bulk-like thickness and appears yellow. Figure 2 b) and Figure 2c) show spatial maps of the frequency separation $\Delta\nu= \nu(\mathrm{A}_{1g}) - \nu(\mathrm{E}_{2g})$ between $\mathrm{A}_{1g}$ and $\mathrm{E}_{2g}$ phonon modes for the MoS$_2$ flake (sample 1) before and after NHC deposition, respectively. Through careful analysis of the $\mathrm{E}_{2g}$ and $\mathrm{A}_{1g}$ mode frequencies in the pristine and NHC-deposited flake, we identify a clear redshift for the $\mathrm{A}_{1g}$ mode after NHC deposition, while the $\mathrm{E}_{2g}$ mode shows no systematic shift. The absence of a measurable shift in the $\mathrm{E}_{2g}$ mode together with the absence of the defect-activated LA(M) mode ($\sim227$ cm$^{-1}$[66]) largely rules out strain- and defect-induced effects and indicates that the observed changes predominantly originate from the electronic modulation. Accordingly, we attribute the observed reduction in $\Delta\nu$ to the enhanced electron doping resulting from charge transfer from the NHC molecules to MoS$_2$. This interpretation is consistent with the opposite Raman response reported for adsorbed water molecules, which deplete MoS$_2$ of electrons via molecular gating [64]. This interfacial charge-transfer mechanism is conceptually related to modulation doping, in which spatially separated donor species transfer carriers to a two-dimensional conducting channel, while the lattice of the channel remains unperturbed. The modulation-doping approach was originally demonstrated in GaAs/AlGaAs heterostructures, where it enabled the formation of high-mobility two-dimensional electron systems [65, 81].

Following NHC deposition, the bulk-like region of the MoS$_2$ sample remains largely unchanged, whereas the thinner regions exhibit a reduction in optical phonon frequency separation $\Delta\nu$ of between 1-2 cm$^{-1}$, depending on the flake’s thickness. This observation is consistent with the expected decrease in the molecular-doping effect with increasing layer number $N$, since the maximum amount of interfacial charge transfer per unit area $n$ is limited by the molecular coverage [64]. Consequently, the transferred electron-density per layer approximately follows the relation $n/N$, therefore reducing the doping-induced phonon-frequency shift and making it unresolvable in thicker flakes [64]. We emphasize that a decrease in the frequency separation $\Delta\nu$ is indicative for n-type doping [20, 64, 77, 82]. This interpretation is further supported by the mode-resolved maps in Figure S3: only the $\mathrm{A}_{1g}$ mode exhibits a pronounce softening [83, 84, 85].

It is also worth noting that, in addition to the frequency softening, the FWHM of the $\mathrm{A}_{1g}$ mode increases after NHC deposition, from approximately 7 cm$^{-1}$ to 15 cm$^{-1}$ in the bilayer region, while the FWHM of $\mathrm{E}_{1g}$ mode shows no systematic changes. The broadening of $\mathrm{A}_{1g}$ mode is consistent with substantial n-type doping, although inhomogeneous broadening may also contribute [64, 77].

To verify the reproducibility of the effect of NHC deposition on the vibrational modes of MoS$_2$, we performed measurements on two additional exfoliated samples (samples 2 and 3) each containing regions of flakes with different thicknesses, as shown in Figure 3 and Figure 4. These measurements confirm that NHC deposition results in shifts of the Raman frequencies with a thickness dependence similar to that observed in the initial analysis: monolayer MoS$_2$ regions show a larger change in the frequency separation $\Delta\nu$ between as-exfoliated and NHC-deposited flakes than thicker regions (Figure 3c and Figure 4c). Based on the change in the frequency separation $\Delta\nu$, the resulting change in charge-carrier density can be estimated using an established calibration [64]. For mono- and bilayer MoS$_2$, changes in $\Delta\nu$ of approximately 2 cm$^{-1}$ and 1.2–1.5 cm$^{-1}$, respectively, are observed. For monolayers, the calibration in Figure S1 yields an increase in electron density of $\Delta n \approx9 \times10^{12}$ cm$^{-2}$, or $\approx7 \times10^{12}$ cm$^{-2}$ after subtracting the atmosphere contribution (Figure 5). Such shifts are consistently observed in all investigated MoS$_2$ samples after NHC deposition. The frequency changes of the E$_{2g}$ and A$_{1g}$ modes before (as-exfoliated) and after NHC deposition for three samples are shown in Figure S3, Figure S5, Figure S7 in the SI. Finally, the changes in the FWHMs of both modes before and after NHC depositions are shown in Figure S4, Figure S6 and Figure S8, with a significant broadening of A$_{1g}$ mode after NHC deposition.

To assess whether the surrounding atmosphere influences the intralayer vibrational modes of $\mathrm{MoS}_{2}$ and to distinguish environmental effects from NHC-induced doping, we performed a complementary Raman experiment under ambient and nitrogen atmospheres. In this comparison, a separate as-exfoliated $\mathrm{MoS}_{2}$ reference sample (not NHC-treated) was measured first under ambient conditions, which include exposure to water vapor and oxygen, and subsequently under an inert $\mathrm{N}_{2}$ gas atmosphere [64] using the same measurement configuration as for the NHC-covered flakes. As shown in Figure 5, switching from ambient air to $\mathrm{N}_{2}$-gas decreases the frequency separation $\Delta\nu$ by only 0.5 $\mathrm{cm}^{-1}$ in the monolayer and bilayer region, while the frequency separation of the bulk-like region remains unchanged. The reduced frequency separation observed under an $\mathrm{N}_{2}$ atmosphere is attributed to a slight increase in the charge-carrier density resulting from reduced depletion, which is predominantly caused by molecular gating by water molecules present in the ambient atmosphere [64]. These comparatively small changes are substantially less pronounced than those observed after NHC deposition. Since the NHC-covered samples were measured in $\mathrm{N}_{2}$ and the as-exfoliated samples in ambient air, this contribution of $\approx0.5\ \mathrm{cm}^{-1}$ is contained in the shifts discussed above; the NHC-induced change thus amounts to $\approx1.5\ \mathrm{cm}^{-1}$ for monolayers. The control experiment therefore indicates that changing the atmosphere from ambient air to $\mathrm{N}_{2}$ has only a minor effect on the Raman response and that the dominant contribution to the $\mathrm{A}_{1g}$ softening, indicative for electron doping, arises from interfacial charge transfer between the NHC molecules and $\mathrm{MoS}_{2}$. Thus, NHC deposition provides an effective route for electron doping of $\mathrm{MoS}_{2}$ through molecularly induced interfacial charge transfer.

Conclusions

In summary, a comparative Raman analysis of exfoliated MoS$_2$ samples before and after NHC deposition with MoS$_2$ thicknesses ranging from monolayers to bulk-like flakes reveals pronounced n-type doping after NHC deposition. The dominant response is a red-shift and broadening of the A$_{1g}$ mode, whereas the E$_{2g}$ mode shows no systematic shift before and after NHC deposition. Together with the absence of substantial changes attributable to strain or defect generation, these observations are consistent with electron transfer from the deposited NHC molecules to MoS$_2$ and with limited chemical perturbation of the lattice. The effect is strongly thickness dependent: frequency shifts of up to approximately 2 cm$^{-1}$ are observed in monolayer regions, whereas the A$_{1g}$–E$_{2g}$ frequency separation in bulk-like regions remains essentially unchanged. For monolayer MoS$_2$, the change in the A$_{1g}$–E$_{2g}$ frequency separation corresponds, based on established Raman calibrations, to an increase in the electron density on the order of 10$^{13}$ cm$^{-2}$ (typically two-orders of magnitude). This trend is reproduced across multiple samples. Measurements of an as-exfoliated MoS$_2$ reference sample under ambient and N$_2$ atmospheres show substantially smaller changes, indicating that the dominant contribution to the observed doping arises from NHC-induced interfacial charge transfer rather than from the change in the surrounding atmosphere. The NHC/MoS$_2$ system therefore provides a molecular charge-transfer-doping scheme that is conceptually related to modulation doping, although the donor layer is directly adjacent to the MoS$_2$ rather than separated by a conventional spacer layer. Because the donors reside outside the MoS$_2$ lattice, NHC deposition offers a promising route for molecular control of the carrier density in semiconducting TMDs that leaves the channel lattice unperturbed, which is relevant for contact engineering, field-effect transistors, and optoelectronic devices.

Associated Content

Supporting Information: Layer-dependent Raman reference data and Raman spectra of 1L MoS$_2$ in different environments; Raman frequency and FWHM maps of samples 1–3; synthesis and NMR characterization of IMe·CO$_2$; physical vapor deposition setup; XPS of IMe on Au (PDF)

Acknowledgements

The authors gratefully acknowledge financial support by the German Science Foundation (DFG) via Grants 443274199 and 556436549 (WU 637/7-2,-8-1), No.426726249 (DE 2749/2-1 and DE 2749/2-2), and the Priority Program 2244 (2DMP). A.S. acknowledges the Alexander von Humboldt Foundation for a research fellowship. During the preparation of this work, the authors used large language models to refine the manuscript regarding grammar, readability, and clarity; the authors reviewed and edited the content and take full responsibility for the publication.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Molecular Doping of MoS$_2$ with N-Heterocyclic Carbenes: A Raman Study of Interfacial Charge Transfer

Ana Senkić; Arne Nalop; Xenia Sieverding; Ariane Ufer; Sebastian Roling; Frank Glorius; Ursula Wurstbauer

Supporting Information

◇ equally contributed

• corresponding: [email protected]; [email protected];

Additional layer-dependent Raman spectra

Raman spectra on $1\mathrm{L}\,\mathrm{MoS}_{2}$ in different environments: impact of adsorbed environmental molecules

Additional Raman maps of sample 1 (Figure 2) before and after NHC deposition

Additional Raman maps of sample 2 (Figures 1 and 3) before and after NHC deposition

Additional Raman maps of sample 3 (Figure 4) before and after NHC deposition

Chemical Synthesis

Reactions were stirred using oven-dried PTFE-coated magnetic stir bars with a rotation speed of 750 rpm. Solids were weighted under ambient conditions except otherwise mentioned, liquids handled with pipettes or syringes. Reaction temperatures are reported as the temperature of metal heating blocks or cooling baths surrounding the reaction vessel. The following solvents were purchased from ACROS Organics, Sigma-Aldrich and Fischer Scientific and used from a solvent purification system, which dried HPLC grade solvents over alumina under a positive pressure of argon according to Grubb’s procedure: Tetrahydrofuran (THF), toluene and dichloromethane (CH$_2$Cl$_2$). Other solvents were also purchased from ACROS Organics, Sigma-Aldrich and Fischer Scientific and stored over activated molecular sieves. Solvents for standard work-up or purification procedures were simply purified by distillation ($n$-pentane, CH$_2$Cl$_2$, EtOAc, Et$_2$O).

Starting materials were bought from commercial suppliers such as ABCR, Acros Organics, Alfa Aesar, BLD Pharm, Carbolution, Combi-Blocks, Fisher Scientific, Fluka, Fluorchem, Merck, Sigma-Aldrich, Strem Chemicals, or TCI and used as received. Air or moisture sensitive compounds were stored and handled in an argon filled glove-box with oxygen and water levels below 1 ppm.

$^{1}$H- and $^{13}$C-NMR spectra were recorded at room temperature on an Avance II 400 MHz instrument. Chemical shifts ($\delta$) are given in ppm. Residual solvent signals were used as references for $^{1}$H-NMR spectra: D$_2$O ($\delta_{\mathrm{H}}$ 4.79 ppm). Spectra were analyzed with MestReNova from Mestrelab Research S.L. and multiplicity described with standard abbreviations.

ESI mass spectra (HRMS, high resolution mass spectrometry) were recorded on a Thermo Fisher Scientific Orbitrap LTQ XL or a Thermo Fisher Scientific Orbitrap Velos Pro.

1,3-dimethyl-1H-imidazolium iodide (IMeH·I) was synthesized according to a known literature procedure.[1, 2] 1,3-dimethyl-1H-imidazol-3-ium-2-carboxylate (IMe·CO$_2$) was synthesized according to adapted literature procedures:[1, 2] IMeH·I (10 g, 44.6 mmol, 1 eq.) and potassium tert-butoxide (6.0 g, 53.6 mmol, 1.2 eq., weighed under inert argon atmosphere) were placed in a oven dired Schlenk flask under argon. At -78 °C, 250 ml dry and degassed tetrahydrofuran (THF) were added. Whilst warming slowly to rt., the mixture was stirred for 2.5 hours. The mixture was filtered over Celite under argon. The filter was washed with THF (two times 250 ml). Still under argon, CO$_2$ was infiltrated into the yellow carbene solution and white precipitate was formed. After one hour, the white precipitate was filtered under ambient condition, washed with 100 ml THF and 400 ml CH$_2$Cl$_2$, sonicated in MeOH, washed again with CH$_2$Cl$_2$ and $n$-pentane, dried under reduced pressure and stored inside a glovebox fridge (4.8 g, 34.0 mmol, 76 % yield).

$^{1}$H NMR (400 MHz, D$_2$O) $\delta$ = 7.36 (s, 2H), 3.98 (s, 6H).

$^{13}$C NMR (101 MHz, D$_2$O) $\delta$ = 160.53, 158.36, 123.05, 36.73.

Pushrod (thermal evaporation coupled to EI): $m/z$ calculated for [IMe]$^{+\mathord\cdot}$ 96.1, found 96.2.

HRMS (ESI): $m/z$ calculated for [IMe·H]$^{+}$ 97.0760, found 97.0760.

Thermal Annealing and Physical Vapor Deposition

The annealing chamber and the deposition chamber were placed side by side in a glove-bag, which was filled with dry nitrogen (evaporated from liquid phase) and operated under constant nitrogen flow. Annealing was conducted at $10^{-5}$ mbar pressure at 190°C for approximately one hour. The annealed substrates were transferred from the annealing chamber into the deposition chamber without leaving the inert nitrogen atmosphere.

Physical vapor phase deposition (PVD) was carried out with a house-made deposition chamber (see Figure S11) which was kept inside the glovebag in nitrogen atmosphere during operation. In order to clean the chamber, its valve was connected to a vacuum pump (0.1–1.0 mbar) whilst the setup was heated to $T > 150$ °C for one hour. After cooling down to room temperature, the chamber was flooded with nitrogen, 2 mg of IMe$\cdot$CO$_2$ was loaded in the crucible and two substrates were placed on the sample stage. One substrate was gold (Au) and the other Si/SiO$_2$ with exfoliated MoS$_2$ flakes of varying thickness. The chamber was flushed with nitrogen, evacuated (0.1 mbar) and the valve was closed. Now, the setup was heated to 120 °C for 30 minutes ($p < 5$ mbar). During cooling down, vacuum was restored upon reaching $T < 80$ °C. At room temperature, the chamber was flushed with nitrogen to take out the samples.

XPS Analysis

The XPS measurement parameters and spectra are shown in Table S1 and Figure S12, respectively. All spectra were referenced relative to the Au $4f_{7/2}$ peak at a binding energy (BE) of 84 eV. An Al K$\alpha$ X-ray source was used (12 kV filament, $h\nu$ = 1486.7 eV), the spot size was 100 µm (72 W max), X-ray incidence and take-off angle amount to 60° and 0°, respectively, relative to surface normal. For charge compensation metallic clips on thin film connected to grounded sample stage and an argon flood gun were used.

The presence of the N 1s signal is a clear indication of chemisorbed IMe on Au since it's N 1s BE of 401.1 eV agrees well with previous reports.[1] However, the successful binding of IMe to Au is solely an indication for the general working principle of the experimenal PVD procedure. It solely proofs that free carbenes have been present in the gas phase. At this state, no assumption can be made about IMe binding to $\mathrm{MoS}_{2}$.

We also attempted XPS measurements of pristine and NHC-modified $\mathrm{MoS}_{2}$ flakes. However, the small lateral dimensions of the flakes, combined with the intrinsically low sensitivity factor of the N 1s core level, resulted in an insufficient signal-to-noise ratio. In addition, the Mo 3p emission overlaps with the N 1s region, preventing reliable separation and quantification of the two contributions. To avoid overinterpretation, the corresponding XPS data and analysis are therefore not included.

[1] A. Bakker, A. Timmer, E. Kolodzeiski, M. Freitag, H. Y. Gao, H. Mönig, S. Amirjalayer, F. Glorius, H. Fuchs, Journal of the American Chemical Society 2018, 140, 11889.

[2] M. Franz, A. Das, S. Chandola, M. Kubicki, M. Das, A. Sette, D. Corona, M. Palummo, L. Chiodo, R. Schöder et al., Angewandte Chemie (International ed. in English) 2025, 64, e202511094.