CVD of MoS2 single layer flakes using Na2MoO4 – impact of oxygen and temperature–time-profile

Two-dimensional (2D) materials are of great interest in many fields due to their astonishing properties at an atomic level thickness. Many fundamentally different methods to synthesize 2D materials, such as exfoliation or chemical vapor deposition (CVD), have been reported. Despite great efforts and progress to investigate and improve each synthesis method, mainly to increase the yield and quality of the synthesized 2D materials, most approaches still involve some compromise. Herein, we systematically investigate a chemical vapor deposition (CVD) process to synthesize molybdenum disulfide (MoS2) single layer flakes using sodium molybdate (Na2MoO4), deposited on a silica (SiO2/Si) substrate by spin-coating its aqueous solution, as the molybdenum source and sulfur powder as sulfur source, respectively. The focus lies on the impact of oxygen (O2) in the gas flow and temperature–time-profile on reaction process and product quality. Atomic force microscopy (AFM), Raman and photoluminescence (PL) spectroscopy, X-ray photoelectron spectroscopy (XPS), and time-of-flight secondary ion mass spectrometry (ToF-SIMS) were used to investigate MoS2 flakes synthesized under different exposure times of O2 and with various temperature–time-profiles. This detailed study shows that the MoS2 flakes are formed within the first few minutes of synthesis and elaborates on the necessity of O2 in the gas flow, as well as drawbacks of its presence. In addition, the applied temperature–time-profile highly affects the ability to detach MoS2 flakes from the growth substrate when immersed in water, but it has no impact on the flake.


Synthesis of MoS 2
Fig. S2 T-t-profile (black) and gas flows (blue) during the synthesis of MoS2 flakes, where light blue represents the pure N2 and dark blue the O2/N2 gas flow.The holding temperature Thold indicates the maximum temperature during synthesis at the center of the tube furnace and the holding time thold the time the system is at Thold.The synthesis time tsyn, which marks the time the center of the tube furnace is at a temperature higher than 650 °C, is flanked by red stars.
Tab. S1 Investigated exposure times of O2 in the gas flow (1.3 %) and corresponding temperatures in the tube furnace when the O2/N2 gas flow was turned off.Thold was 710 °C and thold was set to 6 min for all exposure times of O2.The ramp-up time was 10 minutes.The asterisks * indicate the syntheses, which were conducted a second time without sulfur present.Fig. S3 The 14-sample test was performed by running a synthesis at 790 °C for 10 min with 14 growth substrates placed in a row in the reaction tube to show how the temperature gradient in the tube furnace and therefore the T-t-profiles influences the reaction process and flake quality.

Samples
Tab. S3 Overview of samples investigated with XPS, where Thold denotes the holding temperature, thold the holding time, and Oexp the exposure time of O2.The parameters in bold represent the parameters of interest.

Survey Spectra
Fig. S6 High-resolution XP-spectra of S 2p core level peak region from the sample (a) Tt4 (710C15M) and (b) O6 (O13M), which are used as reference samples for the curve fitting for the sulfur S 2p doublet (pink) and sulfate (green).

Binding Energy Referencing of High-Resolution XP-Spectra
Due to the insulating nature of the growth substrate used (SiO2/Si), the following binding energy (BE) referencing method was adopted.First, the XP-spectrum of sample P1 (P-MoS2) was analyzed to obtain a reference spectrum of a conductive sample for the Mo 3d5/2 peak position of a MoS2 sample.The peak has its maximum at a binding energy of 229.1 eV and corresponds well to value given in literature. 1,2Afterwards, the synthesized MoS2 sample Tt4 (710C15M) was BE-referenced using the Mo 3d peak, set to the value found for P1.The main O 1s of sample Tt4, as for all other synthesized samples, is ascribed to the growth substrate (SiO2/Si) and was found at a BE of 532.5 eV, corresponding well to literature values. 3Finally, this O 1s signal of the growth substrate was used for BE referencing of all other synthesized samples.

Background and curve fitting of S 2p
The background of the S 2p signal of MoS2 has a complex shape as its peaks sit right after the Si 2s signal caused by SiO2/Si (growth substrate).The effect of the characteristic inelastic scattering losses from the Si 2s signal impacts heavily the S 2p region due to the large difference in material thickness (bulk versus atomically thin layer).To consider the prominent convex shape of the background a quadratic fitting of the background is used.
A conventional doublet is used to model the signal of MoS2 flakes (Fig. S6a), though the signal may actually include a variety of additional minor constituents, such as elemental sulfur, amorphous sulfides (MoSx with 2<x<3), partially oxidized MoS2 (MoSxOy), and different MoS2 phases such as the 2H and 1T phase, which are thus ignored.The signal ascribed to sulfates is fitted with a single conventional doublet too (Fig. S6b), after subtraction of a linear background.

Background and curve fitting of Mo 3d and S 2s
The Mo 3d signals of Na2MoO4, MoS2, and Mo-oxides overlap over a wide BE range, which also encompasses the S 2s signals of MoS2 and of sulfates, when present.
The following curve-fitting model was applied to represent the signal envelope (Fig. S7).
a.) Singlet and doublet for S 2s and Mo 3d, respectively, representing MoS2 (pink and purple).
b.) Singlet for S 2s representing sulfates (green).As this peak is entirely overlapped by the Mo 3d envelope, it is not possible to accurately see its presence.Therefore, its curve fitting is only added if a clear S 2p sulfate peak was detected in the high-resolution XP-spectra of the S 2p core level peak region.

Overview of fitting parameters
Tab. S4 All parameters and constraints used for the XPS curve fittings.
a These constraints correspond well to values given in literature.Tab.S5 The difference between the A1g and E2g 1 Raman modes of all synthesized MoS2 flakes except for the most upstream and most downstream samples (sample I and XIV) from the 14-sample test is around 20 cm -1 , indicating the presence of monolayers. 4,5For each sample, 4 to 10 different flakes were measured.

Fig. S1
Fig. S1 (a) Schematic drawing of the set-up to synthesize MoS2 flakes.(b) Picture of the tube furnace.

Fig. S5 2
Fig. S5 XP survey spectra of (a) MoS2 flakes synthesized with different exposure times of O2 and (b) all samples synthesized without sulfur present as reported in Tab.S1.
c.) Doublet for Mo 3d representing Na2MoO4/MoO3/secondary structure of the MoS2 signal (grey).d.) Doublet for Modef 3d representing defective states such as MoO3-x/MoSxOy (dark green).e.) Shirley background line with two steps at the position of the Mo 3d5/2 and Mo 3d3/2 peaks of the MoS2 signal.

Fig. S31
Fig. S31 Light microscopy image of a sample synthesized without O2 present in the gas flow.The scale bar represents 20 µm.

Fig. S32
Fig. S32 Apparent S/Mo-ratio, determined by the formula described in the section Experimental in the main manuscript of the samples synthesized at 710 °C for 6 min and different exposure times of the O2.

Fig. S33
Fig. S33 High-resolution XP-spectra of O 1s core level peak region of samples synthesized at 710 °C and various thold.The O 1s region is dominated by the signal of the growth substrate used (SiO2/Si).The inset shows a zoom-in of the shoulder of the O 1s signal to indicate the decrease in the Mo-source (Na2MoO4), in line with the appearance of the Mo 3d peaks attributed to MoS2 (see main manuscript).

Fig. S34
Fig. S34 Light microscopy images of MoS2 flakes synthesized with different holding temperatures Thold and holding times thold as indicated directly in the images.Regardless of the T-t-profiles all synthesized MoS2 flakes have a triangular shape, show a homogeneous colour, and are evenly distributed over the entire growth substrate.The scale bar represents 20 µm.

Fig. S35
Fig. S35 Light microscopy images of MoS2 flakes synthesized in the 14-sample test.Regardless of the position of the growth substrate (indicated directly in the images), all synthesized MoS2 flakes except for the most upstream and most downstream samples (sample I and XIV) have a triangular shape, show a homogeneous colour, and are evenly distributed over the entire growth substrate.The scale bar represents 20 µm.

Fig. S36
Fig. S36 Raman spectrum of the sample synthesized at 790 °C for 8 min as a representative for a typical Raman spectrum of MoS2 single layer flakes with the relevant peaks at around 384 and 404 cm -1 , representing the E2g 1 and A1g Raman mode, respectively.

Fig. S37
Fig. S37 (a) AFM height image and (b) statistical height analysis of MoS2 flakes synthesized at 750 °C for 15 min as a representative for the height measurements based on the description in [6].The left peak ( ) represents the substrate and the right peak ( ) the MoS2 flakes.The scale bar represents 5 μm.

Fig. S38
Fig. S38 AFM images of MoS2 flakes synthesized at 710 °C for 6 min and picked-up by a SiO2/Si substrate while floating on the water surface after being detached (transfer by the pick-up method, see Experimental in the main manuscript).The scale bar represents (a) 5 µm and (b) 500 nm.

Fig. S39
Fig. S39 Photoluminescence (PL) spectrum of the sample synthesized at 790 °C for 6 min as a representative for the PL spectrum of a MoS2 single layer flake with its characteristic peaks at around 1.85 and 2.0 eV, representing the A-and B-exciton, respectively.

Fig. S40
Fig. S40 Light microscopy images of MoS2 flakes synthesized with different holding temperatures Thold and holding times thold as indicated directly in the images to show the influence of the T-t-profiles on the detachability of the synthesised MoS2 flakes.The blue lines mark the immersion depth of the growth substrates into water to detach the MoS2 flakes.The scale bar represents 100 µm.

Fig. S41
Fig. S41 Light microscopy images of MoS2 flakes synthesized in the 14-sample test to show the influence of the position of the growth substrate in the tube furnace and, therefore, how the T-t-profile influences the detachability of the synthesised MoS2 flakes.The blue lines mark the immersion depth of the growth substrates in water to detach the MoS2 flakes.Samples V and VII were completely immersed in water.The scale bar represents 100 µm.

Fig. S43
Fig. S43 Light microscopy images of MoS2 flakes stored (a) at ambient conditions and (b) in vacuum (300 mbar) for different storage times indicated directly in the images to test the long-term stability of the MoS2 flakes detachability.The blue lines mark the immersion depth of the growth substrate in water.Samples without a blue line were immersed completely in water.The scale bar represents 20 μm.

Fig. S43 -
Fig. S43-Continuation Light microscopy images of MoS2 flakes stored in (c) a desiccator for different storage times as indicated directly in the images to test the long-term stability of the MoS2 flakes detachability.The blue lines mark the immersion depth of the growth substrate in water.The scale bar represents 20 μm.

Fig. S44
Fig. S44 AFM measurements of a sample synthesized at 770 °C for 12 min and after different storage time.The scale bar represents 500 nm.

Fig. S45
Fig. S45 IB/IA-ratio versus resting time of samples synthesized with different T-t-profiles, where different colors indicate the corresponding Thold.The samples in (a) are 100 % detachable (yellow circle), (b) are partial detachable (orange triangle), and (c) are not detachable at all (red cross).

Fig. S46
Fig. S46 Elemental mapping of S --and Mo + -ions of samples synthesized with different T-t-profiles measured with ToF-SIMS operated in negative (top row) and positive (bottom row) mode.

Fig. S47
Fig. S47 ToF-SIMS spectra of samples synthesized with different T-t-profiles, with ToF-SIMS operated in (a) negative and (b) positive mode.

Fig. S48
Fig. S48 Elemental mapping of Na + -ion of samples synthesized with different T-t-profiles measured with ToF-SIMS operated in positive mode.

Fig. S49
Fig. S49 ToF-SIMS spectra operated in the positive mode indicating decreased counts for Na + for samples with increasing Thold and thold.

Fig. S50
Fig. S50 High-resolution XP-spectra of (a) Mo 3d and (b) Na 1s core level peak regions of samples treated at 710 °C for 6 min in absence of sulfur and different exposure times of O2.As a reference, the sample synthesized at 710 °C for 6 min with sulfur present is shown in purple.

Fig. S51
Fig.S51(a) Photoluminescence (PL) spectra of a MoS2 flake after transfer onto a fresh SiO2/Si substrate to eliminate any influence of the growth substrate on the PL measurement.The spectrum changes depending on the laser irradiation time (561 nm, 7.6 mW) prior to PL measurement due to removal of adsorbed water and PMMA residues of the transfer. 8(b) PL spectra of a MoS2 flake on the growth substrate.The spectra only change in intensity with increasing laser irradiation, while the intensity ratio of the B-/A-exciton remains constant at 0.06, corroborating that laser irradiation for cleaning of the transferred flake has no detrimental effects.
Assumption that MoS2 is not structurally changing a lot among samples.Therefore, the relative distance on the binding energy scale, the ratio of area, and fwhm of two atoms of the same constituent must be constant.The values of the constraints are derived from the sample Tt4 (710C15M), which consists mainly of MoS2 flakes.
c Includes all defective states.d This constraint is derived from the sample O6 (O13M), which shows clearly oxidized MoS2 flakes.e f This curve fitting is only added if a clear sulfate peak is visible in the high-resolution XP-spectra of the S 2p core level peak region.g As the intensity ratio between 2s and 2p stays the same for each element, the area range of the sulfate peak is derived as +/-10 % of IS2s sul with IS2s sul = IS2s s /IS2p s * IS2p sul with I representing the intensity of the S 2s and S 2p peak of sulfur s and sulfate sul, respectively.