Preparation of Large-area Vertical 2D Crystal Hetero-structures Through the Sulfurization of Transition Metal Films for Device Fabrication

The Raman spectrum and the cross-sectional HRTEM images of individual MoS₂ and WS₂ fabricated using the sulfurization of pre-deposited transition metals are shown in Figure 1a-b¹⁷, respectively. Two characteristic Raman peaks are observed for both MoS₂ and WS₂, which correspond to in-plane and out-of-plane A_1g phonon vibration modes in the 2D crystals. The frequency difference Δk of two Raman peaks for the MoS₂ sample is 24.1 cm^-1, which suggests that 4 – 5 layers of MoS₂ was obtained¹⁹. However, it is difficult to determine the possible layer number directly from the large Δk value 63.4 cm^-1 of WS₂¹⁷. The cross-sectional HRTEM images of the two samples, shown in Figure 1c-d, have revealed that 5- and 4-layer MoS₂ and WS₂ are obtained for the two samples, respectively. The results have demonstrated that through sulfurizing the transition metals, large-area and uniform MoS₂ and WS₂ films can be obtained.

The cross-sectional HRTEM image of the 1.0 nm Mo film sulfurized under the sulfur deficient condition is shown in Figure 2a, wherein clusters covered with few-layer 2D crystals were observed. These results indicated that two growth mechanisms were observed during the sulfurization procedure¹⁵. Under the sulfur sufficient condition, sulfur-for-oxygen reactions took place quickly such that planar MoS₂ covered the whole sample in a short time. This planar MoS₂ film on the sample surface could prevent further material migration such that a uniform and layer-number-controllable MoS₂ film could be obtained after the sulfurization procedure. However, when the sample was sulfurized under the sulfur deficient condition, the background sulfur was not sufficient to form a complete MoS₂ film such that the Mo oxide segregation and coalescence was the dominant mechanism at the early growth stage. In this case, a sample with Mo oxide clusters covered by few-layer MoS₂ would be obtained after the sulfurization procedure¹⁵. The schematic diagram describing the model of the transition metal sulfurization is shown in Figure 2b¹⁵. Since two growth mechanisms were observed during the sulfurization procedure, there was an upper limit for the MoS₂ layer number with one-time growth.

Using sequential metal deposition with sulfurization procedures discussed above, a WS₂/MoS₂ single hetero-structure was prepared after two procedures of transition metal deposition/sulfurization. The Raman spectra and the cross-sectional HRTEM image of the sample are shown in Figure 3a-b¹⁷. Besides the characteristic Raman peaks corresponding to MoS₂ and WS₂, respectively, the identical layer number 9 with the summation of the individual 5- and 4- layer MoS₂ and WS₂ suggests that the sample was a WS₂/MoS₂ single hetero-structure. Following similar growth procedures, a WS₂/MoS₂/WS₂ double hetero-structure was prepared after three procedures of transition metal deposition/sulfurization. The Raman spectra and the cross-sectional HRTEM image of the sample are shown in Figure 3c-d. With a similar observation of MoS₂ and WS₂ characteristic Raman peaks discussed above, only three layers of 2D crystals were observed for this sample. These results have revealed that (a) good layer number controllability down to a single layer was obtained for this growth technique and (b) a vertical 2D crystal double hetero-structure can be established in three atomic layer thickness¹⁶.

Another sample with half-covering transition metal deposition was prepared to show the possibility of selective growth using the growth technique discussed in this report. By shielding half of the sapphire substrate during the 1.0 nm Mo deposition, half of the substrate could be covered with MoS₂ after sulfurization. After that, the sample was rotated 90° to deposit W to cover half of the sapphire substrate. The same sulfurization procedure was conducted again. In this case, four regions with (a) blank sapphire substrate, (b) standalone MoS₂, (c) WS₂/MoS₂ hetero-structure, and (d) standalone WS₂ were obtained within a single sapphire substrate¹⁷. The picture and the Raman spectra of the four different regions on the sample are shown in Figure 4. As shown in the figure, large-area and uniform WS₂ and MoS₂ films and their vertical hetero-structures were selectively grown on the same sapphire substrate. These results have indicated that besides the establishment of vertical hetero-structures, the growth method of transition metal sulfurization selectively grew 2D crystals on substrates. This flexibility may give more room to practical device fabrications based on 2D materials and their hetero-structures.

To compare the device performance of the transistors with the MoS₂ and the WS₂/MoS₂ vertical hetero-structure as device channels, two transistors were fabricated following the fabrication procedure described in step 3 of the protocol. The schematic diagram showing the fabrication procedure is also shown in Figure 5a. The I_D-V_GS curves of the devices at V_DS = 10 V are shown in Figure 5b. Compared with the MoS₂ transistor, a substantial drain current increase was observed for the hetero-structure device. The field-effect mobility values of the two devices with MoS₂ and WS₂/MoS₂ hetero-structure as the channels extracted from the curves are 0.27 and 0.69 cm²/V·s, respectively. Our previous prediction of electron injection from WS₂ to MoS₂ and from higher electron concentration channels under thermal equilibrium could be responsible for this phenomenon.

After the thin Mo film deposition, the sample was moved out of the sputtering chamber and exposed to air. Since the Mo film is very thin, it was oxidized and formed Mo oxides quickly under ambient conditions. The XPS curve (X-ray photoelectron spectroscopy) of the sample before the sulfurization procedure is shown in Figure 6a. As shown in the figure, the film was composed of MoO₂ and MoO₃ before the sulfurization procedure. These results suggest that the Mo film was oxidized during the transferring procedure from the sputtering chamber to the hot furnace. The other supporting evidence for the formation of the 2D crystal hetero-structure may have come from the equivalent selective etching of the 2D crystal hetero-structure. For this purpose, we have demonstrated that atomic etching can be achieved for both MoS₂ and WS₂ using low-power oxygen plasma treatment²⁰. We can achieve equivalent selective etching to the vertical hetero-structure by repeating the atomic layer etching procedure. The Raman spectrum of the etched and un-etched 4-layer WS₂/3-layer MoS₂ vertical hetero-structure are shown in Figure 6b. The atomic layer etching times were consistent with the layer number of WS₂ (4 times). The observations of both MoS₂ and WS₂ Raman peaks on the un-etched region, and MoS₂ signals only on the etched region, suggest that a vertical hetero-structure was established using the growth technique discussed in this paper.

Figure 1: Individual 2D crystals of MoS₂ and WS₂. (a, b) The Raman spectra and (c, d) the cross-sectional HRTEM images of standalone MoS₂ and WS₂, respectively¹⁷. The samples are obtained by sulfurizing 1.0 nm Mo and W films prepared by a sputtering system. As shown in the Raman spectra, two characteristic Raman peaks were observed for both MoS₂ and WS₂, which corresponded to in-plane and out-of-plane phonon vibration modes in the 2D crystals. The layer number numbers of MoS₂ and WS₂ grown using the method discussed in the current manuscript were proportional to the sputtering times of the pre-deposited Mo and W films. The determination of sputtering times to obtain MoS₂ and WS₂ with required layer numbers is based on the cross-sectional HRTEM images for samples with different sputtering times. However, if the pre-deposited Mo and W films are too thick, Mo and W oxide segregation will become the dominant growth mechanism instead of the planar MoS₂ and WS₂ film growth. Therefore, the proportionality of layer numbers with the sputtering times was limited to few-layer TMDs. With the growth conditions of MoS₂ in the current manuscript, the layer numbers will be proportional to the sputtering times when the MoS₂ film is fewer than 10 layers. The sputtering time is 30 s for the growth of 5-layer MoS₂. This figure has been modified from Wu et al.¹⁷ Please click here to view a larger version of this figure.

Figure 2: The growth model of transition metal sulfurization. (a) The cross-sectional HRTEM image of the 1.0 nm Mo film sulfurized under the sulfur deficient condition and (b) The schematic diagram describing the model for the transition metal sulfurization¹⁵. The growth conditions of the sample prepared with no sulfur powder placed in the furnace is referred to as the sulfur deficient condition. Since there is always residue sulfur accumulation near the downstream of the growth chamber after repeating growth cycles, it is expected that a small amount of sulfur will still diffuse to the sample surface and result in MoS₂ growth. However, under such sulfur deficient conditions, not all the pre-deposited Mo will be transformed into MoS₂. This figure has been modified from Wu et al¹⁵. Please click here to view a larger version of this figure.

Figure 3: MoS₂/WS₂ single- and double-hetero-structures. The Raman spectra and the cross-sectional HRTEM images of WS₂/MoS₂ (a, b) single- and (c, d) hetero-structures^16,17. As shown in the Raman spectrum, the in-plane and out-of-plane phonon vibration modes of both MoS₂ and WS₂ are observed for the 2D crystal hetero-structures. The panels have been modified from Chen et al. and Wu et al.^16,17 Please click here to view a larger version of this figure.

Figure 4: Selective growth of 2D crystals. The picture and the Raman spectra of four regions of the sample prepared with half-covering transition metal depositions on a single sapphire substrate¹⁷. Raman spectra in the (a) blank sapphire substrate, (b) standalone MoS₂, (c) WS₂/MoS₂ hetero-structure and (d) standalone WS₂ regions of the sample revealed characteristic Raman peaks. This figure has been modified from Wu et al.¹⁷ Please click here to view a larger version of this figure.

Figure 5: The device performance of MoS2 and WS₂/MoS₂ vertical hetero-structure transistors. (a) The fabrication procedure of the transistors with the MoS₂ and the WS₂/MoS₂ vertical hetero-structure as the channels and (b) the I_D-V_GS curves of the two devices at V_DS = 10 V¹⁷. The thicknesses of 1.0 nm for Mo and W films were obtained from the readings of the quartz crystal resonator. The sputtering times were 30 s for both materials. This figure has been modified from Wu et al.¹⁷ Please click here to view a larger version of this figure.

Figure 6: The oxidation of pre-deposited Mo films and the equivalent selective etching of WS₂/MoS₂ vertical hetero-structures. (a) The XPS curve of the sample with the pre-deposited Mo film before the sulfurization procedure. The film is composed of MoO₂ and MoO₃ before the sulfurization procedure. These results suggest that the Mo film was oxidized during the transferring procedure from the sputtering chamber to the hot furnace. (b) The Raman spectra of the etched and un-etched 4-layer WS₂/3-layer MoS₂ vertical hetero-structure. After four times of atomic layer etchings, only MoS₂ peaks were observed on the etched region; Panel B has been modified from Chen et al.²⁰ Please click here to view a larger version of this figure.