This is the accepted version of the journal article: Schaefer, Christian M.; Caicedo Roque, Jose Manuel; Sauthier, Guillaume; [et al.]. «Carbon incorporation in MOCVD of MoS2 thin films grown from an organosulfide precursor». Chemistry of materials, Vol. 33, Issue 12 (June 2021), p. 4474-4487. DOI 10.1021/acs.chemmater.1c00646 This version is available at https://ddd.uab.cat/record/266839 under the terms of the license Carbon incorporation in MOCVD of MoS2 thin films grown from or- ganosulfide precursor Christian M. Schaefer1, José M. Caicedo Roque1, Guillaume Sauthier1, Jessica Bousquet1†, Clément Hébert1#, Justin R. Sperling1, Amador Pérez-Tomás1, José Santiso1, Elena del Corro1, Jose A. Gar- rido1,2* 1 Catalan Institute of Nanoscience and Nanotechnology (ICN2), CSIC and BIST, Campus UAB, 08193 Bellaterra, Spain 2 ICREA, 08010 Barcelona, Spain ABSTRACT: With the rise of two-dimensional (2D) transition-metal dichalcogenide (TMD) semiconductors and their pro- spective use in commercial (opto)electronic applications, it has become key to develop scalable and reliable TMD synthesis methods with well monitored and controlled levels of impurities. While metal-organic chemical vapor deposition (MOCVD) has emerged as the method of choice for large-scale TMD fabrication, carbon (C) incorporation arising during MOCVD growth of TMDs has been a persistent concern – especially in instances where organic chalcogen precursors are desired as a less hazardous alternative to more toxic chalcogen hydrides. However, the underlying mechanisms of such unintentional C incorporation and the effects on film growth and properties are still elusive. Here, we report on the role of C-containing side-products of organosulfur precursor pyrolysis in MoS2 thin films grown from molybdenum hexacarbonyl Mo(CO)6 and diethyl sulfide (CH3CH2)2S (DES). By combining in situ gas-phase monitoring with ex situ microscopy and spectroscopy analyses, we systematically investigate the effect of temperature and Mo(CO)6:DES:H2 gas mixture ratios on film morphology, chemical composition, and stoichiometry. Aiming at high-quality TMD growth, that typically requires elevated growth temperatures and high DES:Mo(CO)6 precursor ratios, we observed that temperatures above DES pyrolysis onset (≳600ºC) and excessive DES flow result in the formation of nanographitic carbon, competing with MoS2 growth. We found that by introducing H2 gas to the process DES pyrolysis is significantly hindered, which reduces carbon incorporation. The C content in the MoS2 films is shown to quench the MoS2 photoluminescence and influence the trion-to-exciton ratio via charge transfer. This finding is fundamental for understanding process-induced C impurity doping in MOCVD-grown 2D semiconductors and might have important implications for the functionality and performance of (opto)electronic de- vices. I. Introduction into commercial devices has resulted in a pressing demand for obtaining high-quality, electronic-grade TMDs with homogeneous film continuity on large surfaces.11 In recent years, two-dimensional (2D) materials beyond graphene have attracted great attention in fundamental Both top-down (e.g. mechanical or chemical exfoliation and application-oriented research. Layered transition- from natural bulk crystals) and bottom-up (e.g. synthetic metal dichalcogenides (TMDs), in particular, have become vapor phase methods) have been used to produce TMDs.12 a forefront of 2D material research1 due to their unique op- While bulk exfoliation via the scotch tape method has been toelectronic properties, such as layer-dependent tunable widely used in laboratories to isolate and integrate pristine bandgap with indirect-to-direct bandgap transition in the TMD layers into proof-of-concept devices with high elec- monolayer limit,2 and intense photoluminescence.3 The tronic performance,4 this method is not suitable for com- semiconducting TMDs (e.g. MoS2, WS2, and WSe2) are es- mercial batch fabrication, which requires a scalable, con- pecially suitable for (opto)electronic applications, like low- trollable, and reproducible manufacturing process. There- power transistors with high ON/OFF-ratios,4 logic-in- fore, extensive efforts have led to the development of a va- memory devices,5 memtransistors for neuromorphic com- riety of vapor phase techniques, as reviewed elsewhere.1 puting,6 and neural network photodiode arrays.7 Further- Metal-organic chemical vapor deposition (MOCVD), an more, TMDs have been used in a growing number of flexi- established manufacturing method for traditional com- ble devices, like rectennas,8 photodetectors,9 and wearable pound semiconductors,13 is one of the most promising can- active-matrix light-emitting diodes.10 The drive to fully didates to meet the aforementioned industrial criteria for take advantage of TMDs by successfully integrating them large-area and high-throughput TMD growth.14-16 MOCVD 1 relies on a precise and independent supply of gaseous pre- only on post-growth film characterization as a qualitative cursors into a heated reaction zone, which enables strict guide for process optimization. At present, little effort has control over vapor-phase chemistry. This process can yield been devoted to include in situ gas-phase monitoring and continuous and uniform deposition of atomically thin a quantitative approach that systematically relates C con- TMD films at wafer-scale17 with layer-by-layer thickness tent with MOCVD growth conditions. Consequently, a control.18 comprehensive view on the origin and influence of C incor- Yet, there are still open questions about optimal poration is still elusive. Moreover, the role of different re- MOCVD synthesis conditions depending on precursor actor setups, e.g. hot-wall and cold-wall configurations, chemistry,19 and, in particular, the chosen chalcogen pre- and their impact on C formation rate is still under debate.24 cursor (i.e. the S or Se source).14, 20, 21 While carbon-free Therefore, it is crucial to understand the underlying or- chalcogen hydrides (H2S or H2Se) are viable precursors for ganosulfur precursor pyrolysis to control impurity intro- TMD growth according to thermodynamic predictions22, 23 duction, which potentially results in doping effects,47, 48 and experimental studies,11, 24, 25 cost and safety considera- and, eventually, to find appropriate process windows for tions26 have motivated further research into their low-cost films with desired properties. and less harmful organic counterparts, such as dimethyl In this work, we investigate the growth of MoS2 on sulfide (CH3)2S,18, 27, 28 diethyl sulfide (CH3CH2)2S,17, 24, 29-31 di- SiO2/Si by MOCVD in a hot-wall reactor from molyb- ethyl disulfide (CH3CH2)2S2,15, 16 di-tert-butyl sulfide denum hexacarbonyl Mo(CO)6 and the organic chalcogen ((CH3)3C)2S,32-34 and dimethyl selenide (CH3)2Se.35, 36 precursor diethyl sulfide [DES; (CH3CH2)2S], exploring the Among these chalcogen sources, diethyl sulfide (DES) has impact of the growth parameters on the C incorporation. become a popular precursor choice after the pioneering By employing in situ gas-phase analysis with a mass spec- work of Kang et al,17 and, as such, is used in our work as trometer, we systematically study synthesis conditions for representative for the variety of organic chalcogen precur- single and multi-source DES, Mo(CO)6, DES:Mo(CO)6 and sors. However, the use of DES brings new challenges as car- DES:Mo(CO)6:H2 gas mixtures and varied temperatures. bon (C) can be introduced as an unintentional film impu- We reveal that DES pyrolysis, via an unimolecular dissoci- rity due to pyrolysis side-products from organic ligands.24, ation pathway, releases carbonaceous gas-phase constitu- 36, 37 Currently, there is an ongoing discussion about the lo- ents, resulting in co-deposited C as a competing side-prod- cation and chemical nature of such C impurities and their uct to MoS2 growth. Using extensive ex situ characteriza- influence on film growth and properties. Some researchers tion, such as scanning electron microscopy (SEM), atomic observed co-deposited ‘amorphous carbon’38 or ‘defective force microscopy (AFM), semi-quantitative Raman spec- graphene’36 to form simultaneously with the TMD layer at troscopy, X-ray photoelectron spectroscopy (XPS), and the substrate/TMD interface, sensitively impeding the lat- photoluminescence (PL) spectroscopy, we elucidate the ef- eral 2D growth and TMD film quality characteristics like fect of C on the MoS2 film morphology, chemical composi- grain size, continuity, phase purity and stoichiometry.24, 36 tion, stoichiometry, and electronic doping. Our work iden- Other researchers describe C incorporation into synthetic tifies these film properties to be highly interrelated with C TMD crystals, such as by substitutional C doping39 or CH incorporation, which underlines the significance of con- functionalization40, 41 at chalcogen sites of the TMD basal trolling C impurity levels during MOCVD. plane, or by carbide transformation of TMD edges.42 The latter edge carbidization of MoS2 catalysts is a known phe- II. Experimental Section & Methods nomenon in the field of hydrodesulfurization.43 The edge termination dictates grain morphology and is sensitively modified by the H2/H2S gas-phase environment44, 45 and MoS2 synthesis. MoS2 films were grown by metal-or- the sulfiding agent.20, 21 While sulfidation with H2S is re- ganic chemical vapor deposition (MOCVD) in a home- ported to yield thermodynamically favored triangular built, hot-wall reactor (Figure S1). First, 285 nm SiO2 grown MoS2 grain shape,20, 24 the use of organosulfides can disturb on Si wafers were diced into 1x1 cm2 chips and sonicated in crystallinity,20, 24 although the role of carbonaceous species 3:1 piranha solution (conc. H2SO4: H2O2 35 wt%) for 15 min, has not been fully understood. Additional to their effect on thoroughly rinsed with deionized water, and blow-dried film growth and morphology, C impurities were reported with N2. Substrates were placed face-down onto a quartz to alter the (opto)electronic properties of TMD films; if sample holder bar that was introduced into the vertical used in a controlled manner, C doping might be a route to quartz tube reactor chamber. A thermocouple and temper- tailor electronic TMD devices by synthetic means.40 On the ature controller were used to measure and set the temper- other hand, unintentionally introduced C impurities might ature at the sample, referred to as “growth temperature”. have unwanted consequences, such as quenched photolu- Following the growth process outlined in Figure S2a, the minescence,24, 39 and shift of film properties from semicon- reactor was ramped to the desired growth temperature at ducting to conducting resulting in impaired ON/OFF tran- a rate of around 40 °C/min under inert purge gas flow of sistor performance.38, 46 Thus, the C impurity concentration 100 sccm high-purity Ar (Air Liquide, AlphagazTM 1, in TMD films is an important quality metric that needs to 99.999%). Then, Ar was switched off and MoS2 growth was be controlled. Although several strategies have been ap- initiated by vapor draw of molybdenum hexacarbonyl plied to mitigate C incorporation, e. g. using alkali-assisted Mo(CO)6 (Sigma Aldrich 577766, >99.9 % trace metal ba- synthesis38 and adding reductive H2 gas to the growth pro- sis) and (C2H5)2S (DES, Sigma Aldrich 107247, 98 %) from cess,22-24, 32, 35, 38 common practice in these studies relied 2 dedicated canisters without any carrier gas. Mo(CO)6 pow- strategy we used a boundary condition to exclude the con- der bedded on glass beads for high surface area12, 49 and tribution of a peak appearing in the low frequency shoul- DES precursors were held at 30°C and 12 °C, respectively, der of the D peak and that we were not able to correlate to and their flow was independently regulated by needle me- the growth conditions. Crystallite size La of graphitic car- tering valves (SS-SS4-VH, Swagelok) and estimated based bon was estimated from the intensity ratio of D and G on equilibrium vapor pressure data, as described in the peaks by the relation La = 560/El4 (ID/IG)-1, where El is the Supporting Information (Figure S3). Nominal Mo(CO)6 excitation laser energy 2.54 eV.52 For precise analysis of A1g flows of 0.02 sccm and DES flows between 0.3 and 13.2 and E12g peak positions and determination of the frequency sccm were used. Additionally, high purity H2 gas (Air difference Δ(A1g-E12g) a high resolution 1800 grooves/mm Liquide, AlphagazTM 1, 99.999%) with flows between grating was used. 0 – 30 sccm was injected with a mass flow controller from For PL measurements a 600 grooves/mm grating was a separate line. Typical working pressures were in the used. After background subtraction, Lorentzian peak fit- range of 10-2 to 10-1 torr. Growth was stopped by cutting the ting was performed to account for the charged trion A-, Mo(CO)6, DES and H2 flow; then, the reactor was cooled neutral exciton A0, and B exciton contributions. Fitting down to room temperature under 100 sccm Ar flow before constraints for peak positions were chosen according to lit- sample removal. After each run, the reactor was annealed erature data (see Table S2). To ensure comparability be- at 800 °C in flowing Ar/H2 flow to remove residual reaction tween samples the PL intensity is defined by the relative products (Figure S2b), and between runs was held under (A-+A0)/B intensity ratio,53 which is the peak integral ratio vacuum at around 1×10-3 torr base pressure. During sample of summed A- and A0 contributions with respect to the B loading/unloading 150 sccm Ar was flowed to minimize exciton. ambient exposure of the tube. Samples were stored in a XPS analysis. The chemical composition and stoichi- vacuum desiccator. ometry of the films was investigated by X-ray Photoelec- SEM analysis. The film morphology was analyzed by tron Spectroscopy (XPS) in ultra-high vacuum of 5×10-10 high-resolution field-emission Scanning Electron Micros- mbar with a SPECS PHOIBOS 150 hemispherical analyzer copy (SEM) in immersion mode with a FEI Magellan 400L by using monochromatic Al Kα radiation with an energy of XHR at low accelerating voltages of 2 kV and 100 pA beam 1486.6 eV. To compensate for sample charging all spectra current at a working distance of 3.8 mm. A sample bias of were energy corrected against the Si 2p reference peak of 4 kV was used to compensate charging of the insulating the thermally oxidized SiO2/Si substrate. Details about the substrate and a vCD-detector was used for high contrast fitting procedure and constraints are given in the Support- imaging of back-scattered electrons at low voltages. SEM ing Information (Table S1). images were analyzed with the open-source program Im- Mass spectroscopy. A Hiden Analytical HPR-30 resid- ageJ 1.51j8.50 An image analysis algorithm was used that ual gas analyzer equipped with a triple filter quadrupole converts the SEM to binary images from the grain-sub- mass spectrometer was used for in situ gas-phase analysis. strate contrast and applies the “analyze particles” The gas analyzer was connected downstream from the tool to extract film coverage and nucleation den- MOCVD tube furnace via a sampling needle valve (Figure sity. Film coverage θ was determined as the percentage of S1b) and the pressure inside the mass spectrometer was set covered substrate area, and the nucleation density was de- to 10-6 torr. Soft electron ionization at low electron emis- termined as the counted MoS2 grains per unit area of the sion energy of 12 eV and 100 µA emission current was used SEM image. More details are given in the Supporting Infor- to minimize electron-ionization-induced fragmentation of mation (Figure S4). the DES molecular ion and for selective detection of DES Raman and PL spectroscopy. As-grown films were pyrolysis products. A secondary electron multiplier detec- characterized under ambient conditions with a confocal tor was used for high sensitivity detection at acquisition Witec alpha300R A300M+ microscope at 100x magnifica- range of 10-11 torr. For the temperature dependent DES py- tion by using a laser with 488 nm excitation wavelength rolysis study the reactor temperature was ramped with 10 and spot size of around 1 μm. A low laser power of 0.2 mW °C/min from room temperature to 800 °C while introduc- was used to minimize laser heating effects and avoid opti- ing different DES:Mo(CO)6:H2 mixtures and continuously cal doping and multiexciton dynamics in the PL measure- monitoring the partial pressure of gas-phase constituents, ments.51 Integration times of 5 s and two accumulations e.g. the unfragmented DES molecular ion (m/z=90) and were used for all measurements. For each sample three main pyrolysis products, such as ethyl radicals (m/z=29) spots were measured along the sample diagonal for aver- and ethylene (m/z=28). Selective ionization of ethylene is aged data and corresponding standard deviations as statis- ensured by the low ionization energy, which avoids for- tical errors. Witec PROJECT 5.0 software was used for peak mation of isobaric ions of N2 and CO that require higher fitting and analysis. ionization energies than used in this experiment. This also E12g and A1g Raman modes of MoS2, Si and carbon D and means that H2 could not be detected under these condi- G bands were measured with a 600 grooves/mm diffraction tions due to its higher ionization energy. grating within the same spectral window. For semi-quanti- tative Raman integral analysis MoS2 bands were fitted with double Lorentzians, Si bands with single Lorentzians, and the carbon D and G bands with Double-Gaussians after first order background subtraction. Within our fitting 3 As it is difficult to capture the complex mechanisms in a holistic growth model, Figure 1b illustrates a simplified ver- sion of the MOCVD process, involving (1) thermal decom- position of Mo(CO)6 and DES at high temperature, (2) ad- sorption and diffusion of Mo and S species on the SiO 2/Si substrate, (3) S desorption and (4) nucleation and lateral growth of atomically thin MoS2 domains, with simultane- ous incorporation of graphitic C(sp2), and (5) release of gaseous by-products, such as carbon monoxide CO, eth- ylene C2H4, hydrogen H2, and hydrogen sulfide H2S. In dis- cussion of the growth model, we note that MOCVD reac- tions occur both in the gas-phase and on the surface. Surface reactions involve physisorption and chemisorp- tion processes,55 and substrate-mediated precursor decom- position, such as the decarbonylation of Mo(CO)6 on SiO2.56 The sulfidation of Mo adatoms results in the nucle- ation of MoS2 domains, which grow laterally through edge attachment of additional Mo and S species.16, 17 During sul- Figure 1. a) Schematic of the hot-wall MOCVD reactor. Con- fidation a sufficiently high chalcogen-to-metal ratio is an trolled precursor flows of molybdenum hexacarbonyl important parameter to avoid Mo accumulation and for- Mo(CO)6 and diethyl sulfide [DES; (CH3CH2)2S] are injected mation of Mo metal-rich clusters.35 In our MOCVD setup, from external canisters together with H2 into the reaction the DES:Mo(CO)6 precursor ratio can be tuned by inde- zone at growth temperatures from 550°C to 800°C and pres- pendent source control to achieve desired chalcogen-to- sure of ca. 10-2 to 10-1 torr. In situ gas-phase monitoring of the metal ratio on the surface. For stoichiometric TMD reaction and pyrolysis products is enabled by a gas analyzer growth, a high chalcogen-to-metal ratio is typically re- (mass spectrometer) attached downstream the reaction zone. quired to compensate for the large discrepancy in vapor b) Simplified illustration of the MOCVD process, involving (1) pressures, sticking coefficients, and diffusion kinetics of precursor decomposition, (2) surface adsorption and diffu- the chalcogen and the transition-metal species.24, 35 sion, (3) desorption, (4) growth of MoS2 and graphitic C(sp2) domains on the SiO2/Si substrate surface, and (5) release of Gas-phase reactions include thermal decomposition gaseous by-products. (pyrolysis) of the precursors Mo(CO)657 and DES37, and also gas-phase reactions between them might be possible.58 III. Results & Discussion Our setup allows to monitor these gas-phase processes with a mass spectrometer connected to the reactor, and in This study focuses on the MOCVD growth of atomically- this study we will particularly focus on the role of DES py- thin MoS2 films from gas-source precursors in a vertical rolysis. DES pyrolysis becomes important at comparatively hot-wall reactor, as depicted in Figure 1a. Low pressure high growth temperatures of ≥600ºC and increased chalco- conditions (10-2 – 10-1 torr) were chosen for vapor draw gen-to-metal ratios, which have been reported to be favor- transport of the Mo and S precursors into the reaction able conditions for high-quality TMD growth with en- chamber without carrier gas and to promote surface reac- hanced grain sizes.14, 35, 49 However, these conditions can tions on the SiO2/Si substrate. SiO2/Si has been chosen be- lead to disturbed MoS2 growth due to the formation of gra- cause it is a technologically relevant substrate of the semi- phitic C(sp2), that results from the carbonaceous by-prod- conductor industry, has been widely used for TMD synthe- ucts of DES pyrolysis. sis due to its chemical inertness at high growth tempera- tures, and provides an ideal platform for optical character- ization by means of Raman and PL spectroscopy.54 4 Figure 2. MoS2 grown at 700ºC by co-injection of 0.3 sccm DES and 0.02 sccm Mo(CO)6 precursor flow without reductive H2 gas flow. a), b) SEM images of MoS2 nuclei after 15 min, and of coalesced 1L-MoS2 thin film after 60 min growth time, respectively. c) Raman spectrum of 1L-MoS2 from b). (d-f) show the three regions of interest for semi-quantitative Raman analysis: d) MoS2-related modes fitted with a double-Lorentzian, e) Si mode fitted with a single-Lorentzian, and f) C-related D and G bands fitted with a double-Gaussian and magnified by a factor of 80. The peak integrals for the Raman modes associated to MoS 2, Si and C are defined as IMoS2 (red), ISi (green) and IC (blue), respectively. 3.1 Assessment of MoS2 film properties and C incor- full-width-at-half maximum (FWHM) of E12g and A1g peaks poration are 5.7 cm-1 and 5.0 cm-1, respectively. The frequency differ- ence Δ = 19.8 cm-1 between the E12g and A1g modes, com- monly used as a layer-thickness indicator, confirms the In a first experiment MoS2 was synthesized by co-injec- monolayer thickness of the film.18, 61 Figure 2e shows the Si tion of 0.3 sccm DES and 0.02 sccm Mo(CO)6 at 700°C. Fig- band at 521 cm-1 from the growth substrate and Figure 2f ure 2a and Figure 2b correspond to the SEM images of two reveals broad D and G modes at around 1360 cm-1 and 1590 films with growth times of 15 min and 60 min, respectively. cm-1 due to the presence of C incorporation in the film.24, 36 The short 15 min growth allows assessment of the nuclea- Based on the D and G peak positions and an integrated tion and lateral growth of monolayer MoS2 domains (grey ID/IG intensity ratio of around 1.7, this Raman signature is contrast) on SiO2 (black contrast). The 60 min growth re- assigned (as confirmed by the XPS analysis below), to sp2 veals a continuous monolayer (1L) film with minor second carbon, such as pyrolytic graphite,62 with an estimated layer nucleation (white contrast spots). By increasing the crystallite size below 10 nm.52 The nanographitic film can growth time, few layer MoS2 growth (>1L) was also possibly contain a low fraction of amorphous C(sp3), ac- achieved (Figure S5). However, in this paper we will focus cording to the amorphization trajectory of disordered C on 0 – 1L films defined by 0 – 100% coverage θ. The high proposed by Ferrari et. al.63 Notably, this result is con- nucleation density of the first layer (~103 μm-2), which has sistent with a previous report on nanoscale nuclei of highly been observed consistently throughout this work (Figure defective graphene as side product of TMD growth using S9), produces a nanocrystalline MoS2 monolayer with an organic chalcogen source.36 small grain sizes (20 – 30 nm). This can be explained by high Mo flux controlled growth rates16, 17, 59, high gas-phase For further evaluation of Raman spectra, peak fits were supersaturation at low reactor pressure,22, 35 and low diffu- applied to the MoS2, Si, and C bands. The respective peak sivity of surface species due to hydroxyl anchoring sites on integrals in Figure 2(d-f) are specified as 𝐼𝑀𝑜𝑆2 = 𝐼𝐸2𝑔 1 + 𝐼𝐴 , 1𝑔 the SiO2 substrate.60 While large grain sizes (>μm scale) are 𝐼𝑆𝑖 , and 𝐼𝐶 = 𝐼𝐷 + 𝐼𝐺 . By using the Raman integral ratios typically desired for high performance (opto)electronic ap- 𝐼𝑀𝑜𝑆2 /𝐼𝑆𝑖 and 𝐼𝐶 /𝐼𝑆𝑖 as the indicators for monitoring the plications, the focus of our investigation was on the origin content of MoS2 and co-deposited C(sp2), respectively, we and effect of C incorporation instead of grain size, as pre- can employ Raman spectroscopy as a non-destructive tool vious works have already addressed the suppression of nu- for semi-quantitative characterization and comprehensive cleation density, e. g. by alkali-catalyzed synthesis.38, 49 monitoring of trends in the growth parameter space (Fig- Figures 2(c-f) show an exemplary Raman spectrum of the ure S18). In the following, it will be employed to elucidate coalesced 1L-MoS2 film and the three Raman regions of in- the effects of DES:Mo(CO)6 ratio, growth temperature, and terest. Figure 2d represents the prominent Raman-active H2 gas flow on MoS2 growth and C incorporation. This E12g and A1g bands at 382.6 cm-1 and 402.4 cm-1 of the in- and method, which relies on a prominent substrate reference out-of-plane phonon modes, respectively, corresponding peak, makes SiO2/Si an ideal choice for the growth surface to the thermodynamically stable 2H-phase of MoS2. The in this study. 5 Figure 3. Effect of DES flow on MoS2 thin films grown at 700°C (60 min) from 0.02 sccm Mo(CO)6 and DES flow varying between 0.3 and 13.2 sccm. a(i-v) SEM images for different DES flows, as indicated in the figures. The insets show magnified images for better visualization of the grain morphology. b) Raman spectra normalized to Si peak (521 cm -1, not shown) for 0.3 sccm (black) and 13.2 sccm (red) DES flow. The C-related D and G bands were magnified by a factor of 8.3. c(i-iii) XPS spectra for 0.3 sccm (top) and 13.2 sccm (bottom) DES flow of the C 1s, Mo 3d/S 2s and S 2p core levels; colored areas under singlet/doublet peaks represent fits assigned to the same chemical compound and pink lines show the fitting envelopes of summed peaks. d) Summary of film properties as a function of DES flow, coverage θ from SEM, Raman I MoS2/ISi and IC/ISi integral ratios, Δ(A1g-E12g) frequency differ- ence) and XPS sp2-C, MoS2-x and CxSy fractions (normalized to MoS2-related MoIV 3d5/2 peak). 3.2 Effect of DES:Mo(CO)6 ratio hinders the lateral MoS2 growth, which explains the inter- rupted film morphology. Moreover, the MoS2 layer thick- ness indicator Δ(A1g-E12g), increases despite the reducing To study the effect of S:Mo ratio, MoS2 films were grown MoS2 surface coverage (Figure 3d, ‘Raman’, bottom), sug- at a constant temperature of 700ºC, growth time of 60 min gesting a transition from a lateral growth of MoS2 domains and 0.02 sccm Mo(CO)6 flow. The DES flow was varied to a vertical, layer-stacked MoS2 growth on existing do- from 0.3 sccm to 13.2 sccm, which corresponds to a varia- mains. Topographical analysis by atomic force microscopy tion of the DES:Mo(CO)6 ratio from ~15 to ~660. SEM im- confirms this tendency by revealing particle formation ages (Figure 3a) and analysis (Figure 3d, ‘SEM’) reveal that with maximum heights up to ~4 nm and a monotonic in- as DES flow increases, the MoS2 film coverage on the crease of the root-mean-square roughness from 0.3 nm to SiO2/Si substrate monotonically decreases from 98% (i.e. a 1.1 nm as the DES flow increases (Figure S10). As previously virtually closed 1L- MoS2 film) to 19% at roughly constant suggested, we speculate that the high energy SiO2 surface nucleation density. Further, the shape of the MoS2 grains preferred for lateral MoS2 growth becomes increasingly becomes more irregular and rugged with increasing DES covered by the co-deposited, low surface energy graphitic flow, as revealed by the insets in Figure 3a. The correspond- C, impeding the 2D lateral growth of MoS2 domains and ing Raman spectra of these films (Figure 3b) show vanish- disturbing film coalescence. Thus, the presence of carbon ing MoS2 Raman E12g and A1g modes and growing D and G can lead to interrupted TMD film continuity, in agreement carbon bands with increasing DES flow (see Figure S17 for with literature.36, 59 the complete set of Raman spectra of all DES flows). The Raman-based quantification method described in Section The morphological analysis by SEM and AFM might sug- 2.1 allows us to monitor how the DES flow increase leads to gest the formation of an amorphous agglomerate of MoSx a decreasing 𝐼𝑀𝑜𝑆2 /𝐼𝑆𝑖 , and an increasing 𝐼𝐶 /𝐼𝑆𝑖 (Figure 3d, phase instead of crystalline 2H-MoS2 for the high DES ‘Raman’, top and middle). This suggests that the C contam- flows. However, such an amorphous phase, like MoS3, is re- ination introduced at high DES flow competes with and ported to be thermally unstable and known to transform to crystalline 2H-MoS2 above ~400°C64, 65 The high growth temperature of 700°C used in this study should therefore 6 favor crystallization. Furthermore, the XPS data do not re- to identify the growth-induced C incorporation with its veal any MoSx-related peaks,66 but rather suggests that asymmetric C 1s peak at 284.2±0.2 eV as graphitic C(sp2) there are changes in stoichiometry of the 2H-MoS2 crystal- (black colored area, Figure 3c(i)). The increase of DES flow line phase and film composition, as will be discussed in the from 0.3 sccm to 13.2 sccm results in an almost 20-fold in- following. crease of the fraction of co-deposited C(sp2) (Figure 3d, For comparison purposes, Figure 3c shows the XPS spec- ‘XPS’, top), resulting in heavily contaminated films. Fur- tra corresponding to 0.3 sccm (top) and 13.2 sccm (bottom) ther investigating the chemical nature of the C incorpora- DES flow (see Figure S12 for the complete set of XPS spectra tion, we do not find evidence for the formation of possible of all DES flows). The high-resolution core level spectra of C-Mo bondings, which would be apparent in the C 1s re- Mo 3d in Figure 3c(ii) reveal three types of Mo species of gion at lower binding energies of around 282.8 eV.42 There- different oxidation states; each has been fitted with a dou- fore, within the detection limit of our XPS study, we pos- blet to obtain a satisfying deconvolution of the peak shape. tulate that the C incorporation in our films is mainly com- The most prominent doublet, the blue colored area in Fig- posed of a graphitic C(sp2) layer co-deposited with MoS2 ure 3c(ii), is attributed to the MoIV oxidation state and cor- rather than substitutional doping of C into the MoS2 responds to stoichiometric MoS2, with the Mo 3d5/2 and sheet,39 Mo2C carbide formation,28, 42 or CH groups at chal- Mo 3d3/2 spin-orbit components positioned at 229.3±0.1 eV cogen sites previously reported in synthetic TMDs.41 This and 232.5±0.1 eV, respectively. These binding energies are agrees with the theoretical prediction that substitution of in agreement with literature values for MOCVD-grown S with C atoms at the MoS2 edge is thermodynamically un- MoS2.15 The higher binding energy doublet observed at favorable20 and is further supported by the experimental 232.5±0.2 eV and 235.6±0.2 eV, attributed to the MoVI oxi- observation that carbide conversion of MoS2 does not oc- dation state ascribed to MoO3, is present in all samples; it cur below 800°C.42 has been previously discussed in terms of post-growth ox- Regarding the sulfur environment, we observe the char- idation of as-grown, chalcogen-deficient TMD films upon acteristic MoS2 fingerprints attributed to the S-II oxidation air exposure,24 and it was also encountered in aged, natural state in accordance with literature15: a S 2p doublet at MoS2 crystals.67 Although the growth was carried out in 162.1±0.1 eV and 163.3±0.1 eV, (Figure 3c(iii)) and a corre- vacuum conditions, the observed low background levels of sponding S 2s singlet peak at 226.5±0.1 eV (Figure 3c(ii)). water in the reaction chamber, as monitored with mass Another S 2p doublet emerges at higher binding energies spectroscopy (Figure S14), could eventually result in a par- for DES flows above 3.9 sccm, positioned at 163.6±0.2 eV tial oxidation during the growth.68 The third doublet, ob- and 164.8±0.2 eV, as shown in Figure 3c(iii) for the 13.2 served at lower binding energy (within slight shifts at sccm DES flow. A similar S 2p doublet was previously ob- 229.0±0.2 eV and 232.2±0.2 eV), is attributed to the Mo IV-y served in WS2 growth using DES and was attributed to ei- oxidation state related to a defective, substoichiometric ther SxOy oxysulfide or CxSy organosulfur compounds.24 MoS2-x phase, where the exact position of the doublet de- Analyzing our films we discard the presence of an interme- pends on MoS2-x stoichiometry.69 The presence of this sub- diate MoSxOy oxysulfide phase, as the S 2p spectra do not stoichiometric phase can be explained by S-vacancies due show evidence of the distinct sulfur oxidation states char- to incomplete sulfidation during the deposition process, a acteristic of MoSxOy.71 Instead, we assign the doublet to a common observation in CVD-grown MoS2.70 The S:Mo CxSy organosulfur compound, whose content increases for stoichiometry of this S-deficient phase fraction can be es- high DES flow (Figure 3d, ‘XPS’, bottom, and Figure S12ef). timated to lie between 1.6 and 1.9 based on the Mo 3d 5/2 Such organosulfur compounds were reported at binding peak position.69 Upon increasing the DES flow from 0.3 energies of 163.7±0.3 eV72 for sulfurized carbon black, acti- sccm to 13.2 sccm the fraction of the substoichiometric vated carbon and charcoal, and in S-doped graphene.73 MoS2-x (MoIV-y 3d5/2 area) in relation to ideal, stoichiometric Therefore, it is possible that these additional sulfur func- MoS2 (MoIV 3d5/2 area) is reduced from 0.17 to 0.07, as tionalities are incorporated into the carbonaceous film. shown in Figure 3c(iii). To make this evident, we present a The overall trends seen in Figure 3d allow us to draw two magnified view of the MoS2-x peak evolution in the Sup- significant conclusions. On the one hand, high porting Information (Figure S12e). This result indicates DES:Mo(CO)6 ratios are beneficial due to their high sulfi- that a higher DES flow (i.e. higher sulfidation potential) dation potential, which helps reducing the fraction of the can lead to the deposition of MoS2 films closer to the ideal substoichiometric MoS2-x phase (Figure 3d, ‘XPS’, middle). stoichiometry relation. On the other hand, the high DES exposure conditions lead However, as previously discussed, SEM and Raman anal- to a C-rich reaction environment that results in the depo- ysis revealed that high DES flow results in co-deposited C, sition of graphitic C(sp2), as well as in the incorporation of which hampers MoS2 lateral growth, and, eventually, acts CxSy organosulfur compounds (Figure 3d, ‘XPS’, top and as an impurity. This conclusion is further supported by bottom), at the cost of MoS2 growth. Therefore, the use of XPS, which shows an increase of the overall C 1s peak area organic chalcogen precursors imposes a limitation on the (Figure 3c(i)) and a concurrent decrease of the Mo 3d peak maximum chalcogen-to-metal ratios as there is a clear area (Figure 3c(ii)) for increasing DES flows. XPS control trade-off between film stoichiometry and C incorporation. experiments with in situ annealing (Figure S11) were per- formed to differentiate between adventitious C from post- 3.3 DES pyrolysis as origin of C incorporation and ef- growth ambient exposure (grey colored peak areas, Figure fect of temperature 3c(i)) and the growth-related C contributions, allowing us 7 To better understand the decomposition mechanism of To clarify the origin of C incorporation into the films, DES, we have investigated the onset and degree of pyroly- SiO2/Si samples were exposed only to a single source, ei- sis via in situ mass spectroscopy, by monitoring the partial ther Mo(CO)6 or DES, for 60 min at 700 ºC (Figure S13). pressure of DES and its main pyrolysis fragments as a func- While exposure to only Mo(CO)6 yielded no D or G Raman tion of growth temperature (Figure S15). Figure 4b illus- bands characteristic for C(sp2), samples exposed to only trates unimolecular dissociation as an important pathway DES had intense D and G Raman signatures (𝐼𝐶 /𝐼𝑆𝑖 =2.1). of the DES pyrolysis mechanism.37 In the mass spectros- The lack of D and G Raman bands for the Mo(CO)6 exposed copy fragmentation pattern (Figure S15) the unfragmented, samples was expected as the C-containing decomposition single charged molecular ion of DES appears at mass-to- product of this precursor, CO, is regarded as an inert and charge ratio of m/z=90. The dissociation is initiated at the thermally stable gas-phase compound; thus, it is not ex- C-S bond and involves the production of ethyl radicals pected to participate in side-reactions that could lead to C (m/z=29), which further decompose to the more thermally incorporation under the used temperature conditions.22, 32, stable ethylene (m/z=28) after H abstraction.74 The partial 37 Therefore, DES is unequivocally identified as the source pressure ratios of these pyrolysis products relative to the of C incorporation, which is consistent with previous re- unfragmented DES species are displayed in Figure 4c (left ports of TMD growth processes using organic chalcogen axis) and can be used to monitor the pyrolysis degree. The precursors.28, 32, 36 To further explore the role of DES as a ethyl radical/DES and ethylene/DES ratios start to expo- source of C incorporation, single-source DES exposure of nentially increase at around 600ºC, marking the onset of the SiO2 surface was studied as function of growth temper- DES pyrolysis initiated by H effusion around this tempera- ature. As evident from Figure 4a, there is a rising intensity ture.75 Importantly, the pyrolysis onset coincides with the of the D and G bands with increasing temperature. We ex- onset of the deposited C amount, as extracted from ex situ plain this observation by thermally activated decomposi- Raman analysis of the 𝐼𝐶 /𝐼𝑆𝑖 integral ratio (Figure 4c, right tion (pyrolysis) of the organosulfide source as the underly- axis). This correlation gives strong support to the hypoth- ing mechanism for C incorporation, in agreement with pre- esis that C-containing products, resulting from the ther- vious studies;24, 36 and we go one step further to explicitly mally activated gas-phase decomposition of DES along the assess this occurrence in an MOCVD environment. hot-wall reactor tube, lead to the formation of graphitic C exhibiting a temperature-dependent rate. In contrast, Choudhury et al.,24 while using a cold-wall reactor for WS2 growth from DES, found a rather constant C formation rate over a wide range of temperatures, which is most probably due to reduced, premature DES pyrolysis upstream the sample in the case of a cold-wall system.24 Moreover, Choudhury et al.24 speculated that the reason of C-reduced MoS2 growth from DES using a hot-wall reactor in the work of Kang et al.17 was a result of early depletion of C radicals and C deposition on upstream tube walls. However, our in- vestigation suggests that Kang et al. avoided C incorpora- tion by choice of a comparatively low growth temperature of 550ºC and additional use of reductive H2 (as discussed in Section 2.4, below). This highlights the importance of growth temperature and reactor design when an organic chalcogen precursor is used. After the DES single source experiment, we then moni- Figure 4. DES pyrolysis and C deposition for varying tempera- tored the temperature dependent DES pyrolysis degree for ture and H2 content. a) Raman spectra for single source expo- a gas mixture of 1.2 sccm DES co-injected with 0.02 sccm sure of SiO2 substrates with 13.2 sccm DES for 60 min at tem- Mo(CO)6, as used for MoS2 growth (Figure 4d, black line). peratures ranging from 550ºC to 800ºC. b) Unimolecular dis- We observed that the DES pyrolysis onset shifted to lower sociation pathway of DES pyrolysis, involving production of temperatures (around 500ºC). This suggests the Mo(CO)6 ethyl radicals and ethylene. Mass-to-charge ratios (m/z) as in- catalyzed DES pyrolysis, as previously reported.36 Our find- dicated. The asterisks (*) mark radical species. c) Temperature ing is further confirmed by a systematic increase in C dep- dependence of partial pressure ratio of ethyl radicals (m/z=29) osition on SiO2 substrates exposed to both DES and and ethylene (m/z=28) to unfragmented DES (m/z=90) mon- Mo(CO)6 when compared to those only exposed to DES itored by in situ mass spectroscopy (left axis) and IC/ISi Raman (Figures S17 and S18). integral ratio as indicator of deposited carbon (right axis), as extracted from a). d) Effect of H2 on the partial pressure ratio of ethylene to unfragmented DES. 8 Figure 5. Effect of H2 on MoS2 thin films grown at 700°C (60 min) from 0.02 sccm Mo(CO)6, 1.2 sccm DES flow and H2 flow varying between 0 and 30 sccm. a(i-v) SEM images for different H2 flows, as indicated in the figures. The insets show magnified images for better visualization of the grain morphology. b) Raman spectra normalized to Si peak (521 cm-1, not shown) for 0 sccm (red), 5.5 sccm (green), and 10 sccm (black) H2 flow. C-related D and G bands are magnified by a factor of 8.3. c(i-iii) XPS spectra for 0 sccm (top) and 10 sccm (bottom) H2 flow of the C 1s, Mo 3d/ S 2s and S 2p core levels; colored areas under singlet/doublet peaks represent fits assigned to the same chemical compound and pink lines show the fitting envelopes of summed peaks. d) Summary of film characteristics as a function of the H2 flow, extracted from SEM (coverage θ), Raman spectroscopy (I MoS2/ISi and IC/ISi integral ratios, and Δ(A1g -E12g) frequency difference) and XPS (sp2-C, MoS2-x and CxSy fractions normalized to MoS2-related MoIV 3d5/2 peak). 3.4 Effect of H2 on DES pyrolysis, MoS2 growth and C (0 sccm H2) to a maximum of 77% (5.5 sccm), and subse- reduction quently declines monotonically for H2 flows above 5.5 sccm, while maintaining a roughly constant nucleation density across all H2 flows (Figure S9). This shows im- While low growth temperatures are a viable option to re- proved layer coalescence for an optimized H2 flow. The duce C incorporation, high growth temperatures have been 𝐼𝑀𝑜𝑆2 /𝐼𝑆𝑖 Raman integral ratio (Figure 5d, ‘Raman’, top), fol- demonstrated as ideal for the deposition of high crystalline lows the coverage dependence with H2, corroborating with quality TMDs.11, 35, 49 To this end, the use of H2 gas has been the SEM analysis. The initial increase in MoS2 film coverage proposed to reduce C incorporation in TMDs grown at also correlates well with the decrease of 𝐼𝐶 /𝐼𝑆𝑖 (Figure 5d, high temperature regimes from organic chalcogen ‘Raman’, middle), which can be explained by the enhanced sources.17, 22, 24, 35 We first investigated the effect of H2 added lateral growth of MoS2 domains due to removal of the to the Mo(CO)6:DES mixture on the temperature-depend- growth-hindering C contamination. This promoted lateral ent DES pyrolysis via in situ gas-phase monitoring. Using 2D growth mode is confirmed by the decrease of the fre- ethylene as the main indicator, it was observed that H2 in- quency shift Δ(A1g -E12g) (Figure 5d, ‘Raman’, bottom), indi- troduction significantly lowered the partial pressure of cating a decrease in MoS2 layer thickness with increasing DES pyrolysis products (Figure 4d, Figure S15), indicating H2; Δ(A1g -E12g) drops from about 20.1 cm-1 to around 19.6 a reduced DES pyrolysis degree even above 700ºC. This re- cm-1, as lateral MoS2 grain growth becomes thermodynam- sult suggests that a H2-rich gas phase effectively suppresses ically favored over vertical few layer nucleation on the C DES dissociation, possibly by counterbalancing the H-ab- contamination-reduced high-energy SiO2 surface.59 The straction mechanisms that initiate DES pyrolysis.37 SEM images also indicate the reduction of the second layer Once the effect of H2 on DES pyrolysis was assessed, we nucleation: the white contrast spots on top of the grey pri- performed a series of MoS2 growths varying H2 flow from 0 mary layer MoS2 domains disappear upon adding H2 (see to 30 sccm, while keeping constant the DES flow at 1.2 Figure S7 for the full set of SEM images). sccm, the Mo(CO)6 flow at 0.02 sccm, temperature at Additionally, the MoS2 grain shape was observed to 700°C, and growth time at 60 min. From SEM images (Fig- change gradually from irregular/roundish to a triangular ure 5a) and analysis (Figure 5d, ‘SEM’), it is observed that shape upon increasing H2, as revealed by the insets in Fig- the MoS2 surface coverage initially increases, from 46% ure 5a and by the complete set of SEM images for the H2 9 series in the Supporting Information (Figure S7). The irreg- and B excitons at around 1.89 eV and 2.04 eV, respectively. ular grain shape in absence of H2 hints at disturbed crys- These excitons arise from direct optical transitions from tallinity due to the use of an organic sulfiding agent.20, 21 the highest spin-split valence bands due to spin-orbit cou- Although our post-growth XPS analysis (Figure 5c) did not pling (SOC) within the electronic bandgap E g at the K (K’) confirm C-Mo bondings (i.e. incorporation of C into the point of the Brillouin zone.2 They can be understood as MoS2 crystal domains), it is possible that intermediate bound electron-hole pairs with a binding energy EbE in the MoSxCy edge compounds21, 42, 43 play a role during attach- quasi-particle model. Additionally, negatively charged tri- ment of S species at the catalytically active edges of the lat- ons A-, three-body quasi-particles consisting of a hole and erally growing MoS2 crystal. This could in turn hinder the two electrons with trion binding energy EbT, may occur as ideal S dimer saturation of the Mo-terminated edge that is a consequence of increased electron density in MoS2.77, 78 preferred in triangular MoS2 crystals.45 As known from hy- drodesulfurization catalysts, such “deactivation” of edge sites through carbonaceous species can be recovered in H2 ambient76 and might explain our observed shape evolution towards preferred triangular grain growth at increased H2 flow. On the other hand, excessive H2 flow above 5.5 sccm results in decreasing grain coverage at constant nucleation density. This implies reduced lateral MoS2 growth rate, which we attribute to H2-induced etching of domain edges.17, 25, 29 Indeed, Raman analysis of the IMoS2/ISi ratio for growth time studies with and without H2 confirm lowered growth rates in presence of H2 (Figure S17a(ii)). Thus, by increasing H2 flow, longer growth times are required for similar coverages, which was demonstrated in a growth pa- rameter study for two sets of H2 flows (growth times) of 5.5 sccm (60 min growth) and 12.5 sccm (180 min), which is presented in Figure S8 and Figure S9d. Figure 5c(ii) shows XPS analysis of films grown without and with 10 sccm H2, which present MoIV 3d doublets and S-II 2p doublets ascribed to MoS2, as well as MoVI 3d dou- blets assigned to MoO3. A lower fraction of substoichio- metric MoS2-x phase is observed for the 10 sccm H2 sample (Figure 5d, ‘XPS’, middle) and might be due to an opti- mized edge sulfidation mechanism. Further, compared to the H2-free growth, the 10 sccm H2 sample exhibits a lower Figure 6. MoS2 photoluminescence and charge transfer dop- C incorporation, confirmed by a decreased C(sp2) fraction ing. a) Schematic representation of the electronic band struc- (as seen in the C 1s spectra in Figure 5c(i) and analyzed in ture of MoS2 at the K point with electronic bandgap Eg be- Figure 5d, ‘XPS’, top). However, in contrast to the fully di- tween the conduction band (CB) and valence band (VB). The minished Raman D and G band signature (Figure 5b), a re- VB is spin-split due to spin-orbit coupling (SOC). Laser irrad- sidual XPS C 1s fraction is noticed for the 10 sccm H 2 con- tion excites neutral A0 and B excitons with binding energy EbE dition. This observation suggests that even though gra- and negative trions with binding energy EbT. Negative charge phitic C(sp2) can be significantly reduced in H2-ambient, transfer from graphitic C(sp2) in contact with MoS2 induces residual C impurities, which need further investigation, trion formation. b)(i-v) PL spectra of MoS2 thin films grown at might still remain in the film. 700°C (60 min) from 0.02 sccm Mo(CO)6, 1.2 sccm DES flow and different H2 flows varying between 0 and 30 sccm, as in- dicated. Lorentzian peak deconvolution marks contributions 3.5 MoS2 photoluminescence and C doping of trions A- (blue), neutral excitons A0 (green) and B excitons (purple). The fitting envelopes are shown by the pink lines. In the previous sections we have demonstrated how fine Spectra are normalized to the B exciton intensity. c) PL inte- tuning of growth parameters (e.g. DES or H 2 flow) allows gral ratio (A-+A0)/B as function of coverage θ for MoS2 films precise control over film morphology (i.e. coverage θ) and grown with different H2 flows, as indicated next to each meas- C impurity levels (i.e. IC/ISi). To understand how these pa- urement point. d) Trion-to-exciton ratio A-/A0 as function of graphitic C content, represented by IC/ISi Raman integral ratio, rameters affect the (opto)electronic properties of these for MoS2 films grown with different H2 flows, as indicated next films, we have performed PL spectroscopy. As illustrated in to each measurement point. Figure 6a, laser excitation (488 nm, 2.54 eV) above the op- tical bandgap of monolayer MoS2 results in two pro- Figure 6b shows the PL spectra of MoS2 films grown with nounced PL absorption features associated to neutral A0 different H2 flows from the previous Section 2.4, fitted with 10 their respective A0, B and A- contributions. First, we ob- electrostatic point-of-view, the doping of MoS2 can be ex- served that the PL intensity normalized to the B exciton plained by a negative charge transfer from donor states of intensity, considering the (A-+A0)/B integral ratio as a nanographitic C(sp2) located within the bandgap of MoS2, quality factor,53 correlates with the varying film coverage as as proposed in the band diagram model in Figure 6a.46, 48 function of H2 flow (Figure S20), and reaches a maximum This band line-up is supported by X-ray and ultra-violet for the highest coverage of the 5.5 sccm H2 sample. This photoemission spectroscopy (XPS/UPS) studies; further overall PL intensity behavior governed by the surface cov- discussion is provided in the Supporting Information (Fig- erage (grain size) dependence79 was also observed for films ure S23). grown with varied DES flow (Figure S21); larger monolayer coverages (larger grain sizes) generally result in stronger A 3.6 Discussion of growth model exciton peak intensity due to enhanced A exciton recom- bination lifetimes and reduced non-radiative recombina- tion channels.79 Interestingly, when plotting the (A-+A0)/B As summary of our results from the previous sections and intensity ratio as a function of coverage (Figures 6c), we do to explain the interplay between the DES pyrolysis-in- not observe a solely monotonic coverage dependence. In- duced reaction pathway and MoS2 growth (and C co-depo- stead, two branches appear for films grown with H2 flows sition), we propose the following, simplified reaction below and above 5.5 sccm, which belong to films with high scheme: and reduced amount of C impurities, respectively, as dis- cussed in Section 2.4. For example, films grown with 0 sccm and 10 sccm show similar coverages, but differ in 𝑇≳600°C (A-+A0)/B intensity ratio by almost an order of magnitude. (1) (𝐶𝐻3 − 𝐶𝐻2 )2 𝑆 (𝑔) ↔ 𝐻2 𝑆(𝑔) + C𝐻2 = 𝐶𝐻2 (𝑔) + Therefore, the reduced PL intensity in the sample grown 𝐻2 (𝑔) without H2 can be explained by its high C(sp2) content. These results confirm that the presence of C(sp2) results in (2) 𝑀𝑜(𝐶𝑂)6 (𝑔) + 2𝐻2 𝑆(𝑔) ↔ 𝑀𝑜𝑆2 (𝑠) + 6𝐶𝑂(𝑔) + strong PL quenching under illumination, as has been ob- 2𝐻2 (𝑔) served in MoS2/graphene heterostructures.48, 80 This phe- (3) 𝐶𝐻2 = 𝐶𝐻2 ↔ 2𝐶(𝑠) + 2𝐻2 nomenon has been explained by charge transfer at the MoS2/C(sp2) heterointerface81 and hints at electronic dop- __________________________________________________ ing of MoS2 in presence of C impurities.48 However, we (4) 𝑀𝑜(𝐶𝑂)6 +(𝐶𝐻3 − 𝐶𝐻2 )2 𝑆 ↔ 𝑀𝑜𝑆2 (𝑠) + 6𝐶𝑂(𝑔) + note that in the literature both n-type48, 81 and p-type80 dop- 2𝐶(𝑠) + 5𝐻2 (𝑔) ing of MoS2 in interface with C(sp2) have been discussed controversially. To confirm the hypothesis of the doping effect, a closer For simplicity, we assume full conversion of DES to H2S and ethylene (1), which were main gas-phase constituents look is taken at the A exciton peak. The Lorentzian peak deconvolution reveals that the A exciton peak consists of detected by mass spectroscopy (Figure S15). We note that the neutral A0 exciton and charged A- trion, which we find there are also other intermediate, radical fragments from DES pyrolysis (e.g. ethyl radicals) present in the gas-phase. separated by the trion binding energy EbT of around 20 – 30 meV consistent with literature.51, 77 As the appearance of H2S is consumed in the sulfidation reaction (2) for the syn- the trion is induced by an increased negative charge den- thesis of MoS2; and ethylene decomposition (3) acts as a source of C.82 If H2 is added to the gas phase, the equilib- sity within the MoS2 layer, a prominent A- weight and in- creased A-/A0 ratio are indicators for electron doping.47, 48, rium of all reactions will shift to the left side, thus resulting 77, 78 All our films in this H series show A-/A0 ratios above in suppressed DES pyrolysis in reaction (1) and reduced C 2 around 0.5 (Figure 6d), which can be attributed to various incorporation in reaction (3). However, high H2 concentra- trion-inducing doping effects, such as S vacancies, strain, tion in reaction (2) can also lower the growth rate of MoS2, as discussed in Section 3.4. The reaction pathway (1), which or the pronounced effect of charged impurities at the MoS2/SiO2 interface.78 However, it is remarkable that the normally requires temperatures above the DES pyrolysis A-/A0 ratio is increasing with increased IC/ISi ratio as a con- onset ≳600°C, see Figure 4c, can even occur at lower tem- peratures in presence of Mo(CO)6 (Figure 4d), possibly due sequence of reducing the H2 flow (Figure 6d). In other words, as the C content increases, the MoS2 monolayer be- to surface-templated reactions on the SiO2 substrate83 and comes more n-type and the trion-related PL dominates the hydrodesulfurization of DES at the catalytically active edge sites of growing MoS2,44 as discussed in Section 3.4. Reac- spectra. This trend was also found for films with increasing C content controlled by the DES flow (Figure S21d). More- tion (4) is the sum of reactions (1), (2) and (3). over, a Stokes redshift of PL emission peaks in the range of While this study focuses on the growth of monolayer 10 – 30 meV was observed with increasing C content (Fig- MoS2, our results are aligned with previous works on the ure S22), further supporting the hypothesis of the n-type MOCVD of WSe235, 36 and WS2.24 Current understanding of doping of MoS2 controlled by C impurity levels.77 From an the MOCVD growth of 2D TMDs typically highlights the importance of high chalcogen-to-metal ratios for optimal 11 sulfidation,24, 35 which is also suggested by our XPS stoichi- growth process, which can have wider implications for the ometry study (Figure 3(c-d)). However, in the case of or- performance of (opto)electronic devices. ganic chalcogen precursors a high chalcogen-to-metal ra- tios can lead to co-deposition of C(sp2), which in turn im- ASSOCIATED CONTENT pedes the lateral growth of 2D TMD crystals, in agreement The Supporting Information is available free of charge at [link] with literature.24, 36 So far, it has been widely neglected to MOCVD reactor setup and growth process; Precursor flow study the root-cause mechanisms and effects of C incorpo- estimation; SEM image analyses of growth parameter series; ration during MOCVD growth. Previous works speculated AFM analysis of lateral-to-vertical growth transition in pres- that C incorporation is caused by organic chalcogen pre- ence of C; XPS analysis and peak fitting, differentiation be- cursor pyrolysis, but did not show experimental evidence. tween adventitious C and growth-induced sp2-C; DES and Our work confirms this hypothesis by combining in situ Mo(CO)6 single source exposure experiments; Vacuum gas-phase monitoring together with post-growth, semi- background diagnosis; DES fragmentation pattern and tem- quantitative Raman analysis. We have been able to corre- perature dependent pyrolysis, suppression of DES pyrolysis late the degree of thermally activated chalcogen precursor upon adding H2; Raman spectra of growth parameter stud- pyrolysis with the C impurity level in as-grown films. More- ies (temperature, growth time, DES flow, Mo(CO)6 flow, H2 over, we have discussed this phenomenon in the context of flow) and analyses of IMoS2/ISi and IC/ISi ratios and frequency difference Δ(A1g-E12g); PL analysis and peak fitting, PL spec- the growth temperature and the reactor design (i.e. hot- tra of DES flow series and analysis of (A-+A0)/B and A-/A0 wall compared to cold-wall configuration).24 When using ratios; XPS/UPS analysis of MoS2/C/SiO2 heterointerface H2 for mitigation of C incorporation, as proposed in previ- band structure. ous works,17, 24, 35, 36 we emphasize on the careful adjustment of H2 flow to limit H2-induced etching25 and to control AUTHOR INFORMATION MoS2 growth rate and surface coverage for improved film coalescence and continuity. Corresponding Author *E-mail: joseantonio.garrido@icn2.cat IV. Conclusions Present Addresses † J. B.: DiamFab SAS, 38042 Grenoble Cedex 9, France The influence of synthesis parameters (i.e. temperature, # C. H.: Inserm and Université Grenoble Alpes, 38400 Saint chalcogen-to-metal ratio, H2 flow) during MOCVD of pol- Martin d'Hères, France ycrystalline, monolayer MoS2 thin films on SiO2 from the Author Contributions precursors Mo(CO)6 and DES was investigated in a hot- wall reactor by in situ gas-phase monitoring (mass spec- The manuscript was prepared through contributions of all troscopy) combined with ex situ film analyses (SEM, AFM, authors: Raman, XPS, PL). It was revealed that DES pyrolysis is in- volved in the production of C-containing gas-phase species C.M.S. – Investigation, conceptualization, MoS2 growth, SEM, AFM, Raman, PL, mass spectroscopy measurements, via an unimolecular dissociation mechanism; this causes formal analysis, writing - original draft; J.M.C.R. – Resources co-deposition of nanographitic C(sp2) at growth tempera- (technical support in design and installation of home-built tures above DES pyrolysis onset and high chalcogen-to- reactor setup); G.S. – XPS measurements, review of XPS anal- metal ratios. The C incorporation was found to compete ysis; J.B. Writing - review & editing; C.H. – Writing - review with and hinder the lateral growth of MoS2 domains, which & editing; J.R.S. – Conceptualization, writing - review & edit- results in interrupted film morphology and imposes a lim- ing; A.P.T. Writing - review & editing, formal analysis; J.S. – itation on the use of high chalcogen-to-metal ratios, which Supervision, writing - review & editing; E.d.C. – Conceptual- are desirable for the growth of stoichiometric MoS 2 films. ization, writing, review & editing, formal analysis; J.A.G. - Furthermore, the introduced C impurities lead to n-doping Supervision, writing - review & editing, project administra- of MoS2, suggesting charge transfer at the MoS2/C(sp2) het- tion, funding acquisition. erointerface. By adding H2 to the gas-phase we demon- strate that this reducing agent is effective in suppressing All authors have given approval to the final version of the DES pyrolysis, thus, allowing precise control of C impurity manuscript. incorporation, morphology, and photoluminescence of the MoS2 film. Notes The authors declare no competing financial interest. In summary, this work contributes important insights to gas-phase chemistry and synthesis conditions in the scala- ble MOCVD of 2D semiconductors. Our findings underline ACKNOWLEDGMENT the challenges of high-quality 2D TMDs from low-cost, low This work was funded by the European Union’s Horizon 2020 toxicity organic chalcogen precursors. We highlight oppor- research and innovation programme (BrainCom, grant No. tunities to engineer 2D semiconductors by a fine-tuned 73203, Nanosmart, grant No. 825213 and Graphene Flagship Core 3, grant No. 881603). We also acknowledge funding from Generalitat de Catalunya (2017 SGR 1426), and the 2DTecBio 12 (FIS2017-85787-R) funded by the Spanish Ministry of Science, 11. Lin, Y. C.; Jariwala, B.; Bersch, B. M.; Xu, K.; Nie, Y. F.; Innovation and Universities, the Spanish Research Agency Wang, B. M.; Eichfeld, S. M.; Zhang, X. T.; Choudhury, T. H.; (AEI) and the European Regional Development Fund Pan, Y.; Addou, R.; Smyth, C. M.; Li, J.; Zhang, K. H.; Haque, M. (FEDER/UE). The ICN2 is funded by the CERCA programme / A.; Folsch, S.; Feenstra, R. M.; Wallace, R. M.; Cho, K.; Fullerton- Generalitat de Catalunya and supported by the Severo Ochoa Shirey, S. K.; Redwing, J. M.; Robinson, J. A., Realizing Large- Centres of Excellence programme, funded by the Spanish Re- Scale, Electronic-Grade Two-Dimensional Semiconductors. ACS search Agency (AEI, grant no. SEV-2017-0706). Nano 2018, 12 (2), 965-975. 12. Backes, C.; et.; al., Production and processing of graphene and related materials. 2d Materials 2020, 7 (2), 282. C. M. S. acknowledges that this work has been done within the 13. Thompson, A. G., MOCVD technology for framework of the Ph.D. program in Materials Science of the semiconductors. Mater. Lett. 1997, 30 (4), 255-263. Autonomous University of Barcelona. Further, C.M.S. wants 14. Lee, D. H.; Sim, Y.; Wang, J.; Kwon, S. Y., Metal-organic to thank Anh Tuan Hoang for fruitful discussions. chemical vapor deposition of 2D van der Waals materials-The challenges and the extensive future opportunities. APL Mater. ABBREVIATIONS 2020, 8 (3), 18. MOCVD, metal-organic chemical vapor deposition; 2D, two- 15. Kalanyan, B.; Kimes, W. A.; Beams, R.; Stranick, S. J.; dimensional; TMD, transition-metal dichalcogenide; DES, di- Garret, E.; Kalish, I.; Davydov, A. V.; Kanjolia, R. K.; Maslar, J. E., ethyl sulfide; FWHM, full-width-at-half-maximum; SEM, Rapid Wafer-Scale Growth of Polycrystalline 2H-MoS2 by Pulsed Metal-Organic Chemical Vapor Deposition. Chem. Mat. 2017, 29 scanning electron microscopy; AFM, atomic force microscopy; (15), 6279-6288. PL, photoluminescence; XPS, X-ray photoelectron spectros- 16. Seol, M.; Lee, M. H.; Kim, H.; Shin, K. W.; Cho, Y.; copy; CB, conduction band; VB, valence band; SOC, spin-orbit Jeon, I.; Jeong, M.; Lee, H. I.; Park, J.; Shin, H. J., High- coupling; UPS, ultraviolet photoelectron spectroscopy. Throughput Growth of Wafer-Scale Monolayer Transition Metal Dichalcogenide via Vertical Ostwald Ripening. Advanced REFERENCES Materials 2020, 32 (42), 8. 1. Das, S.; Robinson, J. A.; Dubey, M.; Terrones, H.; 17. Kang, K.; Xie, S. E.; Huang, L. J.; Han, Y. M.; Huang, Terrones, M., Beyond Graphene: Progress in Novel Two- P. Y.; Mak, K. F.; Kim, C. J.; Muller, D.; Park, J., High-mobility Dimensional Materials and van der Waals Solids. Annual Review three-atom-thick semiconducting films with wafer-scale of Materials Research 2015, 45 (1), 1-27. homogeneity. Nature 2015, 520 (7549), 656-660. 2. Mak, K. F.; Lee, C.; Hone, J.; Shan, J.; Heinz, T. F., 18. Shinde, N. B.; Francis, B.; Rao, M. S. R.; Ryu, B. D.; Atomically Thin MoS2: A New Direct-Gap Semiconductor. Phys. Chandramohan, S.; Eswaran, S. K., Rapid wafer-scale fabrication Rev. Lett. 2010, 105 (13), 4. with layer-by-layer thickness control of atomically thin MoS2 3. Splendiani, A.; Sun, L.; Zhang, Y. B.; Li, T. S.; Kim, J.; films using gas-phase chemical vapor deposition. APL Mater. Chim, C. Y.; Galli, G.; Wang, F., Emerging Photoluminescence in 2019, 7 (8), 8. Monolayer MoS2. Nano Lett. 2010, 10 (4), 1271-1275. 19. Maury, F., Recent trends in the selection of metal- 4. Radisavljevic, B.; Radenovic, A.; Brivio, J.; Giacometti, organic precursors for MOCVD process. J. Phys. IV 1995, 5 (C5), V.; Kis, A., Single-layer MoS2 transistors. Nature Nanotechnology 449-463. 2011, 6 (3), 147-150. 20. Tuxen, A.; Gobel, H.; Hinnemann, B.; Li, Z. S.; 5. Marega, G. M.; Zhao, Y. F.; Avsar, A.; Wang, Z. Y.; Knudsen, K. G.; Topsoe, H.; Lauritsen, J. V.; Besenbacher, F., An Tripathi, M.; Radenovic, A.; Kis, A., Logic-in-memory based on atomic-scale investigation of carbon in MoS2 hydrotreating an atomically thin semiconductor. Nature 2020, 587 (7832), 72-77. catalysts sulfided by organosulfur compounds. J. Catal. 2011, 281 6. Sangwan, V. K.; Lee, H.-S.; Bergeron, H.; Balla, I.; Beck, (2), 345-351. M. E.; Chen, K.-S.; Hersam, M. C., Multi-terminal memtransistors 21. Fuchtbauer, H. G.; Tuxen, A. K.; Li, Z. S.; Topsoe, H.; from polycrystalline monolayer molybdenum disulfide. Nature Lauritsen, J. V.; Besenbacher, F., Morphology and Atomic-Scale 2018, 554 (7693), 500-504. Structure of MoS2 Nanoclusters Synthesized with Different 7. Mennel, L.; Symonowicz, J.; Wachter, S.; Polyushkin, Sulfiding Agents. Top. Catal. 2014, 57 (1-4), 207-214. D. K.; Molina-Mendoza, A. J.; Mueller, T., Ultrafast machine 22. Kumar, V. K.; Dhar, S.; Choudhury, T. H.; vision with 2D material neural network image sensors. Nature Shivashankar, S. A.; Raghavan, S., A predictive approach to CVD 2020, 579 (7797), 62-66. of crystalline layers of TMDs: the case of MoS2. Nanoscale 2015, 7 8. Zhang, X.; Grajal, J.; Vazquez-Roy, J. L.; Radhakrishna, (17), 7802-7810. U.; Wang, X. X.; Chern, W.; Zhou, L.; Lin, Y. X.; Shen, P. C.; Ji, 23. Dhar, S.; Kumar, V. K.; Choudhury, T. H.; X.; Ling, X.; Zubair, A.; Zhang, Y. H.; Wang, H.; Dubey, M.; Shivashankar, S. A.; Raghavan, S., Chemical vapor deposition of Kong, J.; Dresselhaus, M.; Palacios, T., Two-dimensional MoS2- MoS2 layers from Mo-S-C-O-H system: thermodynamic modeling enabled flexible rectenna for Wi-Fi-band wireless energy and validation. Phys. Chem. Chem. Phys. 2016, 18 (22), 14918- harvesting. Nature 2019, 566 (7744), 368–372. 14926. 9. Schneider, D. S.; Grundmann, A.; Bablich, A.; Passi, V.; 24. Choudhury, T. H.; Simchi, H.; Boichot, R.; Chubarov, Kataria, S.; Kalisch, H.; Heuken, M.; Vescan, A.; Neumaier, D.; M.; Mohney, S. E.; Redwing, J. M., Chalcogen Precursor Effect on Lemme, M. C., Highly Responsive Flexible Photodetectors Based Cold-Wall Gas-Source Chemical Vapor Deposition Growth of on MOVPE Grown Uniform Few-Layer MoS2. ACS Photonics WS2. Cryst. Growth Des. 2018, 18 (8), 4357-4364. 2020, 7 (6), 1388–1395. 25. Chiappe, D.; Ludwig, J.; Leonhardt, A.; El Kazzi, S.; 10. Choi, M.; Bae, S. R.; Hu, L.; Hoang, A. T.; Kim, S. Y.; Mehta, A. N.; Nuytten, T.; Celano, U.; Sutar, S.; Pourtois, G.; Ahn, J. H., Full-color active-matrix organic light-emitting diode Caymax, M.; Paredis, K.; Vandervorst, W.; Lin, D.; De Gendt, S.; display on human skin based on a large-area MoS2 backplane. Sci. Barla, K.; Huyghebaert, C.; Asselberghs, I.; Radu, I., Layer- Adv. 2020, 6 (28), 6. controlled epitaxy of 2D semiconductors: bridging nanoscale 13 phenomena to wafer-scale uniformity. Nanotechnology 2018, 29 molybdenum disulfide using sodium chloride. Nanotechnology (42), 9. 2017, 28 (46), 8. 26. Beauchamp, R. O.; Bus, J. S.; Popp, J. A.; Boreiko, C. J.; 39. Liang, T.; Habib, M. R.; Xiao, H.; Xie, S.; Kong, Y. H.; Andjelkovich, D. A.; Leber, P., A Critical Review of the Literature Yu, C.; Iwai, H.; Fujita, D.; Hanagata, N.; Chen, H. Z.; Feng, Z. on Hydrogen Sulfide Toxicity. CRC Critical Reviews in Toxicology H.; Xu, M. S., Intrinsically Substitutional Carbon Doping in CVD- 1984, 13 (1), 25-97. Grown Monolayer MoS2 and the Band Structure Modulation. ACS 27. Hoang, A. T.; Katiyar, A. K.; Shin, H.; Mishra, N.; Forti, Appl. Electron. Mater. 2020, 2 (4), 1055-1064. S.; Coletti, C.; Ahn, J. H., Epitaxial Growth of Wafer-Scale 40. Zhang, F.; Lu, Y. F.; Schulman, D. S.; Zhang, T. Y.; Molybdenum Disulfide/Graphene Heterostructures by Metal- Fujisawa, K.; Lin, Z.; Lei, Y.; Elias, A. L.; Das, S.; Sinnott, S. B.; Organic Vapor-Phase Epitaxy and Their Application in Terrones, M., Carbon doping of WS2 monolayers: Bandgap Photodetectors. Acs Applied Materials & Interfaces 2020, 12 (39), reduction and p-type doping transport. Sci. Adv. 2019, 5 (5), 8. 44335-44344. 41. Cochrane, K. A.; Zhang, T.; Kozhakhmetov, A.; Lee, J. 28. Fan, Y.; Nakanishi, K.; Veigang-Radulescu, V. P.; H.; Zhang, F.; Dong, C.; Neaton, J. B.; Robinson, J. A.; Terrones, Mizuta, R.; Stewart, J. C.; Swallow, J. E. N.; Dearle, A. E.; Burton, M.; Bargioni, A. W.; Schuler, B., Intentional carbon doping reveals O. J.; Alexander-Webber, J. A.; Ferrer, P.; Held, G.; Brennan, B.; CH as an abundant charged impurity in nominally undoped Pollard, A. J.; Weatherup, R. S.; Hofmann, S., Understanding synthetic WS2 and WSe2. 2d Materials 2020, 7 (3), 6. metal organic chemical vapour deposition of monolayer WS2: the 42. Jeon, J.; Park, Y.; Choi, S.; Lee, J.; Lim, S. S.; Lee, B. H.; enhancing role of Au substrate for simple organosulfur Song, Y. J.; Cho, J. H.; Jang, Y. H.; Lee, S., Epitaxial Synthesis of precursors. Nanoscale 2020, 12 (43), 22234-22244. Molybdenum Carbide and Formation of a Mo2C/MoS2 Hybrid 29. Cun, H.; Macha, M.; Kim, H.; Liu, K.; Zhao, Y.; Structure via Chemical Conversion of Molybdenum Disulfide. LaGrange, T.; Kis, A.; Radenovic, A., Wafer-scale MOCVD growth ACS Nano 2018, 12 (1), 338-346. of monolayer MoS2 on sapphire and SiO2. Nano Research 2019, 43. Kelty, S. P.; Berhault, G.; Chianelli, R. R., The role of (10), 2646-2652. carbon in catalytically stabilized transition metal sulfides. Appl. 30. Zhao, R.; Lo, C. L.; Zhang, F.; Ghosh, R. K.; Knobloch, Catal. A-Gen. 2007, 322, 9-15. T.; Terrones, M.; Chen, Z. H.; Robinson, J., Incorporating 44. Bruix, A.; Fuchtbauer, H. G.; Tuxen, A. K.; Walton, A. Niobium in MoS2 at BEOL-Compatible Temperatures and its S.; Andersen, M.; Porsgaard, S.; Besenbacher, F.; Hammer, B.; Impact on Copper Diffusion Barrier Performance. Adv. Mater. Lauritsen, J. V., In Situ Detection of Active Edge Sites in Single- Interfaces 2019, 7. Layer MoS2 Catalysts. ACS Nano 2015, 9 (9), 9322-9330. 31. Jin, G.; Lee, C. S.; Liao, X.; Kim, J.; Wang, Z.; Okello, 45. Mom, R. V.; Louwen, J. N.; Frenken, J. W. M.; Groot, I. O. F. N.; Park, B.; Park, J.; Han, C.; Heo, H.; Kim, J.; Oh, S. H.; M. N., In situ observations of an active MoS2 model Choi, S. Y.; Park, H.; Jo, M. H., Atomically thin three-dimensional hydrodesulfurization catalyst. Nature Communications 2019, 10, membranes of van der Waals semiconductors by wafer-scale 8. growth. Sci. Adv. 2019, 5 (7), 6. 46. Leonhardt, A.; Chiappe, D.; Afanas'ev, V. V.; El Kazzi, 32. Marx, M.; Grundmann, A.; Lin, Y. R.; Andrzejewski, D.; S.; Shlyakhov, I.; Conard, T.; Franquet, A.; Huyghebaert, C.; de Kummell, T.; Bacher, G.; Heuken, M.; Kalisch, H.; Vescan, A., Gendt, S., Material-Selective Doping of 2D TMDC through AlxOy Metalorganic Vapor-Phase Epitaxy Growth Parameters for Two- Encapsulation. Acs Applied Materials & Interfaces 2019, 11 (45), Dimensional MoS2. J. Electron. Mater. 2018, 47 (2), 910-916. 42697-42707. 33. Grundmann, A.; Andrzejewaki, D.; Kummell, T.; 47. Mouri, S.; Miyauchi, Y.; Matsuda, K., Tunable Baoher, G.; Heuken, M.; Kalisch, H.; Vescan, A., H2S-free Metal- Photoluminescence of Monolayer MoS2 via Chemical Doping. Organic Vapor Phase Epitaxy of Coalesced 2D WS2 Layers on Nano Lett. 2013, 13 (12), 5944-5948. Sapphire. MRS Adv. 2019, 4 (10), 593-599. 48. Li, Z.; Ye, R.; Feng, R.; Kang, Y.; Zhu, X.; Tour, J. M.; 34. Cohen, A.; Patsha, A.; Mohapatra, P. K.; Kazes, M.; Fang, Z., Graphene Quantum Dots Doping of MoS2 Monolayers. Ranganathan, K.; Houben, L.; Oron, D.; Ismach, A., Growth-Etch Advanced Materials 2015, 27 (35), 5235-5240. Metal–Organic Chemical Vapor Deposition Approach of WS2 49. Kim, H.; Ovchinnikov, D.; Deiana, D.; Unuchek, D.; Atomic Layers. ACS Nano 2020, 15 (1), 526-538. Kis, A., Suppressing Nucleation in Metal-Organic Chemical Vapor 35. Eichfeld, S. M.; Hossain, L.; Lin, Y. C.; Piasecki, A. F.; Deposition of MoS2 Monolayers by Alkali Metal Halides. Nano Kupp, B.; Birdwell, A. G.; Burke, R. A.; Lu, N.; Peng, X.; Li, J.; Lett. 2017, 17 (8), 5056-5063. Azcatl, A.; McDonnell, S.; Wallace, R. M.; Kim, M. J.; Mayer, T. 50. Schneider, C. A.; Rasband, W. S.; Eliceiri, K. W., NIH S.; Redwing, J. M.; Robinson, J. A., Highly Scalable, Atomically Image to ImageJ: 25 years of image analysis. Nature Methods 2012, Thin WSe2 Grown via Metal-Organic Chemical Vapor Deposition. 9 (7), 671-675. ACS Nano 2015, 9 (2), 2080-2087. 51. Lee, H. S.; Kim, M. S.; Kim, H.; Lee, Y. H., Identifying 36. Zhang, X. T.; Balushi, Z. Y.; Zhang, F.; Choudhury, T. multiexcitons in MoS2 monolayers at room temperature. Phys. H.; Eichfeld, S. M.; Alem, N.; Jackson, T. N.; Robinson, J. A.; Rev. B 2016, 93 (14), 6. Redwing, J. M., Influence of Carbon in Metalorganic Chemical 52. Cancado, L. G.; Takai, K.; Enoki, T.; Endo, M.; Kim, Y. Vapor Deposition of Few-Layer WSe2 Thin Films. J. Electron. A.; Mizusaki, H.; Jorio, A.; Coelho, L. N.; Magalhaes-Paniago, R.; Mater. 2016, 45 (12), 6273-6279. Pimenta, M. A., General equation for the determination of the 37. Zheng, X.; Fisher, E. M.; Gouldin, F. C.; Zhu, L.; crystallite size L-a of nanographite by Raman spectroscopy. Appl. Bozzelli, J. W., Experimental and computational study of diethyl Phys. Lett. 2006, 88 (16), 3. sulfide pyrolysis and mechanism. Proc. Combust. Inst. 2009, 32, 53. McCreary, K. M.; Hanbicki, A. T.; Sivaram, S. V.; Jonker, 469-476. B. T., A- and B-exciton photoluminescence intensity ratio as a 38. Song, J. G.; Ryu, G. H.; Kim, Y.; Woo, W. J.; Ko, K. Y.; measure of sample quality for trancition metal dichalcogenide Kim, Y.; Lee, C.; Oh, I. K.; Park, J.; Lee, Z.; Kim, H., Catalytic monolayers. APL Mater. 2018, 6 (11), 9. chemical vapor deposition of large-area uniform two-dimensional 54. Qin, B.; Ma, H.; Hossain, M.; Zhong, M.; Xia, Q.; Li, B.; Duan, X., Substrates in the Synthesis of Two-Dimensional 14 Materials via Chemical Vapor Deposition. Chem. Mat. 2020, 32 68. Kastl, C.; Chen, C. T.; Kuykendall, T.; Shevitski, B.; (24), 10321-10347. Darlington, T. P.; Borys, N. J.; Krayev, A.; Schuck, P. J.; Aloni, S.; 55. Ahn, C.; Park, Y.; Shin, S.; Ahn, J.-G.; Song, I.; An, Y.; Schwartzberg, A. M., The important role of water in growth of Jung, J.; Kim, C. S.; Kim, J. H.; Bang, J.; Kim, D.; Baik, J.; Lim, monolayer transition metal dichalcogenides. 2d Materials 2017, 4 H., Growth of Monolayer and Multilayer MoS2 Films by Selection (2), 10. of Growth Mode: Two Pathways via Chemisorption and 69. Baker, M. A.; Gilmore, R.; Lenardi, C.; Gissler, W., XPS Physisorption of an Inorganic Molecular Precursor. ACS Applied investigation of preferential sputtering of S from MoS2 and Materials & Interfaces 2021, 13 (5), 6805-6812. determination of MoSx stoichiometry from Mo and S peak 56. Jiang, Z. Q.; Huang, W. X., Comparative Investigation of positions. Appl. Surf. Sci. 1999, 150 (1-4), 255-262. Mo(CO)(6) Adsorption on Clean and Oxidized Si(111) Surfaces. 70. Kim, I. S.; Sangwan, V. K.; Jariwala, D.; Wood, J. D.; Chin. J. Chem. Phys. 2011, 24 (6), 729-734. Park, S.; Chen, K. S.; Shi, F. Y.; Ruiz-Zepeda, F.; Ponce, A.; Jose- 57. Usoltsev, I.; Eichler, R.; Wang, Y.; Even, J.; Yakushev, Yacaman, M.; Dravid, V. P.; Marks, T. J.; Hersam, M. C.; Lauhon, A.; Haba, H.; Asai, M.; Brand, H.; Di Nitto, A.; Dullmann, C. E.; L. J., Influence of Stoichiometry on the Optical and Electrical Fangli, F.; Hartmann, W.; Huang, M.; Jager, E.; Kaji, D.; Kanaya, Properties of Chemical Vapor Deposition Derived MoS2. ACS J.; Kaneya, Y.; Khuyagbaatar, J.; Kindler, B.; Kratz, J. V.; Krier, Nano 2014, 8 (10), 10551-10558. J.; Kudou, Y.; Kurz, N.; Lommel, B.; Miyashita, S.; Morimoto, 71. Levasseur, A.; Vinatier, P.; Gonbeau, D., X-ray K.; Morita, K.; Murakami, M.; Nagame, Y.; Nitsche, H.; Ooe, K.; photoelectron spectroscopy: A powerful tool for a better Sato, T. K.; Schadel, M.; Steiner, J.; Steinegger, P.; Sumita, T.; characterization of thin film materials. Bull. Mat. Sci. 1999, 22 (3), Takeyama, M.; Tanaka, K.; Toyoshima, A.; Tsukada, K.; Turler, 607-614. A.; Wakabayashi, Y.; Wiehl, N.; Yamaki, S.; Qin, Z., 72. Chang, C. H., Preparation and characterization of Decomposition studies of group 6 hexacarbonyl complexes. Part carbon-sulfur surface compounds. Carbon 1981, 19 (3), 175-186. 1: Production and decomposition of Mo(CO)(6) and W(CO)(6). 73. Abdelkader-Fernandez, V. K.; Domingo-Garcia, M.; Radiochim. Acta 2016, 104 (3), 141-151. Lopez-Garzon, F. J.; Fernandes, D. M.; Freire, C.; de la Torre, M. 58. Xuan, Y.; Jain, A.; Zafar, S.; Lotfi, R.; Nayir, N.; Wang, D. L.; Melguizo, M.; Godino-Salido, M. L.; Perez-Mendoza, M., Y.; Choudhury, T. H.; Wright, S.; Feraca, J.; Rosenbaum, L.; Expanding graphene properties by a simple S-doping Redwing, J. M.; Crespi, V.; van Duin, A. C. T., Multi-scale methodology based on cold CS2 plasma. Carbon 2019, 144, 269- modeling of gas-phase reactions in metal-organic chemical vapor 279. deposition growth of WSe2. J. Cryst. Growth 2019, 527, 125247. 74. Ivin, K. J.; Wijnen, M. H. J.; Steacie, E. W. R., Reactions 59. Shang, S. L.; Lindwall, G.; Wang, Y.; Redwing, J. M.; of ethyl radicals. J. Phys. Chem. 1952, 56 (8), 967-972. Anderson, T.; Liu, Z. K., Lateral Versus Vertical Growth of Two- 75. Conway, N. M. J.; Ferrari, A. C.; Flewitt, A. J.; Dimensional Layered Transition-Metal Dichalcogenides: Robertson, J.; Milne, W. I.; Tagliaferro, A.; Beyer, W., Defect and Thermodynamic Insight into MoS2. Nano Lett. 2016, 16 (9), 5742- disorder reduction by annealing in hydrogenated tetrahedral 5750. amorphous carbon. Diam. Relat. Mat. 2000, 9 (3-6), 765-770. 60. Sitek, J.; Plocharski, J.; Pasternak, I.; Gertych, A. P.; 76. Vogelaar, B. M.; Steiner, P.; van Langeveld, A. D.; McAleese, C.; Conran, B. R.; Zdrojek, M.; Strupinski, W., Eijsbouts, S.; Moulijn, J. A., Deactivation of Mo/Al2O3 and Substrate-Induced Variances in Morphological and Structural NiMo/Al2O3 catalysts during hydrodesulfurization of thiophene. Properties of MoS2 Grown by Chemical Vapor Deposition on Appl. Catal. A-Gen. 2003, 251 (1), 85-92. Epitaxial Graphene and SiO2. ACS Applied Materials & Interfaces 77. Mak, K. F.; He, K. L.; Lee, C.; Lee, G. H.; Hone, J.; 2020, 12 (40), 45101-45110. Heinz, T. F.; Shan, J., Tightly bound trions in monolayer MoS2. 61. Lee, C.; Yan, H.; Brus, L. E.; Heinz, T. F.; Hone, J.; Ryu, Nature Materials 2013, 12 (3), 207-211. S., Anomalous Lattice Vibrations of Single- and Few-Layer MoS2. 78. Dubey, S.; Lisi, S.; Nayak, G.; Herziger, F.; Nguyen, V.- ACS Nano 2010, 4 (5), 2695-2700. D.; Le Quang, T.; Cherkez, V.; González, C.; Dappe, Y. J.; 62. Lopez-Diaz, D.; Holgado, M. L.; Garcia-Fierro, J. L.; Watanabe, K.; Taniguchi, T.; Magaud, L.; Mallet, P.; Veuillen, Velazquez, M. M., Evolution of the Raman Spectrum with the J.-Y.; Arenal, R.; Marty, L.; Renard, J.; Bendiab, N.; Coraux, J.; Chemical Composition of Graphene Oxide. J. Phys. Chem. C 2017, Bouchiat, V., Weakly Trapped, Charged, and Free Excitons in 121 (37), 20489-20497. Single-Layer MoS2 in the Presence of Defects, Strain, and Charged 63. Ferrari, A. C.; Robertson, J., Interpretation of Raman Impurities. ACS Nano 2017, 11 (11), 11206-11216. spectra of disordered and amorphous carbon. Phys. Rev. B 2000, 79. Andrzejewski, D.; Marx, M.; Grundmann, A.; 61 (20), 14095-14107. Pfingsten, O.; Kalisch, H.; Vescan, A.; Heuken, M.; Kummell, T.; 64. Wildervanck, J. C.; Jellinek, F., Preparation and Bacher, G., Improved luminescence properties of MoS2 Crystallinity of Molybdenum and Tungsten Sulfides. Z. Anorg. monolayers grown via MOCVD: role of pretreatment and growth Allg. Chem. 1964, 328 (5-6), 309-318. parameters. Nanotechnology 2018, 29 (29), 10. 65. Cadot, S.; Renault, O.; Fregnaux, M.; Rouchon, D.; 80. Pierucci, D.; Henck, H.; Naylor, C. H.; Sediri, H.; Nolot, E.; Szeto, K.; Thieuleux, C.; Veyre, L.; Okuno, H.; Martin, Lhuillier, E.; Balan, A.; Rault, J. E.; Dappe, Y. J.; Bertran, F.; Le F.; Quadrelli, E. A., A novel 2-step ALD route to ultra-thin MoS2 Fevre, P.; Johnson, A. T. C.; Ouerghi, A., Large area molybdenum films on SiO2 through a surface organometallic intermediate. disulphide-epitaxial graphene vertical Van der Waals Nanoscale 2017, 9 (2), 538-546. heterostructures. Sci Rep 2016, 6, 10. 66. Benck, J. D.; Chen, Z.; Kuritzky, L. Y.; Forman, A. J.; 81. Diaz, H. C.; Addou, R.; Batzill, M., Interface properties Jaramillo, T. F., Amorphous Molybdenum Sulfide Catalysts for of CVD grown graphene transferred onto MoS2(0001). Nanoscale Electrochemical Hydrogen Production: Insights into the Origin of 2014, 6 (2), 1071-1078. their Catalytic Activity. ACS Catal. 2012, 2 (9), 1916-1923. 82. Lin, L.; Lai, M. S.; Li, H. F.; Tian, F.; Chen, Y. L.; Sun, 67. Gao, J.; Li, B. C.; Tan, J. W.; Chow, P.; Lu, T. M.; J. S.; Lin, J. M., Investigation of carbon deposition induced by Koratkar, N., Aging of Transition Metal Dichalcogenide pyrolytic decomposition of ethylene. RSC Adv. 2017, 7 (47), 29639- Monolayers. ACS Nano 2016, 10 (2), 2628-2635. 29644. 15 83. Gao, F.; Wang, Y.; Tysoe, W. T., Ethylene vacuum: From oxycarbide to carbide to MoAl alloy. Journal of hydrogenation on Mo(CO)6 derived model catalysts in ultrahigh Molecular Catalysis A: Chemical 2006, 249 (1), 111-122. 16 Table of Contents artwork 17