PtS2, a group-10 transition metal dichalcogenide, has prominent layer-depended band structure, and can enable extremely high phonon-limited mobility at room temperature. Here, we demonstrate the theoretical study on the electronic band structures of PtS2 with different thickness by using density functional theory (DFT), as well as experimental realization of large-area synthesis of few-layer PtS2 film by direct sulfurization of pre-deposited Pt. The synthetic process suggested that the reaction pressure is a key factor in the formation of high-quality PtS2 semiconducting films. Characterizations with atomic force microscopy (AFM), Raman spectroscopy and X-ray photoelectron spectroscopy (XPS) have indicated that good film stoichiometry and uniformity have been achieved. Furthermore, field-effect transistor (FET) arrays were fabricated based on the large-scale PtS2 film, exhibiting well-uniform electrical performance with p-type transport behavior. These results can open up an attractive approach to promote the large-scale applications of PtS2 in advanced nanoelectronics and optoelectronics devices and systems.
Two-dimensional (2D) layered materials such as graphene,1 transition metal dichalcogenides (TMDCs),2–4 and black phosphorus (BP)5 have attracted considerable attention in recent years due to their unique structures and novel physical properties. Among them, TMDCs are of great interest owing to their sizable bandgap and rich materials types, exhibiting huge potential for a variety of applications, including nanoelectronics6–8 and optoelectronics.9,10 Generally, TMDCs have the formula of MX2, where M is a transition metal, and X denotes a chalcogen (S, Se, or Te). To date, most research have been focusing on group-6 TMDCs such as MoS2 and WSe2. Another interesting subcategory is the group-10 TMDCs which can exhibit very high phonon-limited mobility at room temperature as demonstrated by theoretical calculation.11 Although the material properties of these TMDs have been studied decades ago,12,13 their 2D nature and electronic device applications have not been experimentally investigated.
As a typical group-10 TMDC with intrinsic layered structure, PtS2 has been reported with the prominent layer-depended bandgap ranging from 1.6 eV in monolayer film to 0.25 eV in bulk.14 Such property is originated from the strong layer dependence and interlayer interaction. In this way, promising optoelectronics devices can be expected without extra materials engineering to manipulate the semiconductor band structure. For example, Li et al. have demonstrated a phototransistor based on few-layer PtS2 which exhibits excellent responsibly and detectivity, suggesting PtS2 as a competitive candidate for future optoelectronics applications.15 In addition, these group-10 TMDCs have ultra-high stability in air.16 These fascinating properties of PtS2 have made it intriguing to build electronic devices for various functionality. However, the material properties have only been studied on exfoliated samples, and no effective synthesis approach for large-area PtS2 haven been reported. Although mechanical exfoliation method facilitates simple device fabrication and the prepared 2D materials have low structural defects, the flake size, location and thickness are largely uncontrollable, which hinders their large-scale integration and system-level applications.17
Recently, multiple synthetic approaches have been proposed and implemented in the growth of large-area TMDCs including sulfurization/selenization of pre-deposited metals or oxides on insulating substrates.18–21 TMDC films fabricated by this method have shown notable performance in various applications.22,23 In this work, we naturally extended this synthesis approach into the preparation of large-scale PtS2 films by directly sulfurizing pre-deposited Pt on SiO2/Si substrate and studied the impact of sulfurizing vapor pressure on the chemical process and film quality. Back-gate field-effect transistors (FETs) have been further fabricated based on the homogeneous film. Their electrical performance has been characterized with the assistance of analyzing the electronic band structures of PtS2 with different thickness by first-principle calculations.
Computational Method: First-principles study was based on density functional theory (DFT) as implemented in Cambridge Serial Total Energy Package (CASTEP) code.24 The generalized gradient approximation (GGA) with Perdew-Burke-Ernzerhof (PBE)25 was used to describe exchange-correlation potential. In addition, the Van der Waals interactions were treated via DFT-D correction (TS method). The ultrasoft pseudopotential method was employed for the interactions between the ions and electrons. A plane-wave basis set with an energy cutoff of 380 eV was adopted and the convergence criterion of self-consistent field (SCF) was set to 5×10-6 eV/atom. The k-point samplings within the Monkorst–Pack scheme for integrations of the reduced Brillouin zone26 were set as 5×5×2 for monolayer and bilayer, and 5×5×1 for trilayer to quinquelayer, and 5×5×4 for bulk, ensuring sufficient computational accuracies for all the studied structures. Before the calculations of electronic properties, geometry optimizations were performed until the energy was less than 5×10-6 eV per atom and the maximum forces were less than 0.01 eV per angstrom. The optimized primitive lattice constants for the bulk PtS2 are as follows: a=b=3.6236 Å, c= 4.8051 Å. The layered structures were obtained by cutting out from the bulk in the direction of the layer stacking (001). A large vacuum space of 20 Å was introduced to hinder interactions between the adjacent structures. It should be noted that spin-orbit coupling was not considered in this study.
Film synthesis: Fig. 1(a) schematically shows the process flow of the synthesis of large-area PtS2. The synthesis was carried out on a Si substrate in size of 2 cm × 2 cm covered with 300 nm SiO2 grown by dry oxidation. A thin layer (3.4 nm, characterized by X-ray reflectivity) of Pt meal film was first sputtered (Kurt J. Lesker PVD 75) on the SiO2/Si substrate. Immediately after the Pt deposition, the substrate was loaded into the center of a chemical vapor deposition (CVD) furnace, and sulfur powder (3 g, 99.999%) was placed at the upstream end, which was about 20 cm away from the sample. Before the sulfurization process, the chamber was pumped to 0.5 Pa. Then, the substrate and sulfur powder zones were heated to different temperatures. Fig. 1(b) shows the temperature profile during the growth of few-layer PtS2 film. The region of sulfur powder was heated to 280 °C in 10 min and the region of the Pt/SiO2/Si sample was heated to 800 °C in 25 min. The pressure of the chamber was about 10 Pa in the temperature ramping process, and 10 sccm Ar was used as carrier gas. When the temperature of the substrate zone reached 800 °C, the vacuum pumping valve was adjusted to increase the pressure and maintain at 280 Pa. This was to provide sufficiently high sulfur vapor pressure for the reaction between of Pt and S (other pressure conditions were also studied for further comparison). The reaction was kept at 800 °C for 60 min before cooling the furnace down to room temperature.
FET device fabrication: Back-gate FET device arrays were fabricated based on the above synthetic process. Pt metal strips with different size were first patterned and defined by photolithography and lift-off processes. After the sulfurization of the Pt strips, Cr/Au source and drain electrodes were formed by electron-beam evaporation and lift-off processes. The heavily doped Si substrate was used as the control gate electrode.
RESULTS AND DISCUSSION
Fig. 2(a) shows the atomic structure of the crystal PtS2 in thermodynamically stable 1T phase, which is the AA arrangement for stacked layers. Each Pt atom is coordinated by six S atoms in a tilted octahedral structure. This is different from the local configuration of the most widely studied 2H-MoS2, in which each Mo atom is arranged in the site of a triangular prismatic structure. Each single PtS2 layer has a thickness of ∼0.5 nm and consists of one Pt atomic layer sandwiched by two S atomic layers through strong Pt-S covalent bonds. The PtS2 layers are constructed by vertically stacking the single layers through weak van der Waals interactions.14 The electronic band structures of 1T-PtS2 with different thickness from monolayer to bulk were investigated by first-principle calculations based on GGA method. As shown in Fig. 2(b)–2(g), 1T-PtS2 exhibits an indirect band gap for all the thickness ranging from monolayer to bulk. This is different from the traditional group-6 TMDCs such as 2H-MoS2 and 2H-WS2 which has a direct band gap in monolayer form. The monolayer PtS2 shows an indirect band gap of 1.77 eV, which is close to the reported theoretical and experimental results.14,27
We then prepared the large-scale PtS2 films by sulfurizing the patterned Pt film in a CVD furnace. Since Pt is a type of relatively stable metal material, temperature and pressure will be the key factors during the reaction between Pt and S. Our experimental results also confirmed that the partial pressure of sulfur during the CVD process is critical for the formation of high-quality crystalline PtS2 film, and we have made great efforts in optimizing the proper value. Fig. 3(a) shows atomic force microscopy (AFM) measurement results on the PtS2 film synthesized under 280 Pa. The film exhibited a relatively rough surface morphology with the root-mean-square (RMS) roughness of 2.17 nm. On the one hand, it can be attributed to the non-perfect film uniformity during the Pt sputtering process; on the other hand, the surface roughness is also due to the relatively high pressure in the sulfuration process, which can affect the transport of S atoms on the film surface. The film thickness was estimated to be 12.7 nm from the AFM height profile as shown in Fig. 3(b), which is about four times of the thickness of the as-deposited Pt layer due to the introduction of S atoms through the bonding reaction. Fig. 3(c) shows the Raman spectra obtained from different locations on the PtS2 film with the excitation laser wavelength of 532 nm. Peaks at the 333.6 cm-1 were observed in three different positions, corresponding to the A1g1 mode (out-of-plane vibration).14 It is should be noted that the Eg1 and A1g2 modes were not observed, which may be due to the non-ideal crystallization property of the film. Although the film quality of our CVD PtS2 film is not as ideal as the reported flakes prepared by using chemical vapor transport (CVT) and mechanical exfoliation,14,15 our synthesis approach demonstrated here still provides an alternative way towards large-area PtS2 for device integration as well as material property studies.
To study the impact of the reaction pressure on the growth of few-layer PtS2 film, several different pressure conditions have been tested (10-5 Pa, 10-2 Pa, 10 Pa and 280 Pa). The material components and chemical bonding of the films grown under different pressures were characterized by X-ray photoelectron spectroscopy (XPS). The spectra have been summarized and compared in Fig. 4. The energy calibrations were carried out against C 1s peak (284.6 eV) to eliminate the effects of charge accumulation and surface contamination during the analysis. Fig. 4(a) and 4(b) show the Pt 4f and S 2p core level peaks of the film prepared under the pressure of 10-5 Pa, respectively. The peaks at 71.03 eV and 74.38 eV have been observed in Fig. 4(a), corresponding to the Pt 4f7/2 and Pt 4f5/2 core level spectra of Pt element,28,29 respectively. As shown in Fig. 4(b), the S 2p core level spectra exhibits two peaks at 161.15 eV and 163.35 eV, which are ascribed to the S 3p3/2 and S 3p1/2 core level spectra of S element, respectively. Additionally, it should be noted that the detection of elemental sulfur was due to the trace amount of residual sulfur covered on the sample during cooling down. These results suggest that Pt did not chemically react with sulfur at 10-5 Pa.
The Pt 4f and S 2p core level peaks of the film grown under the pressure of 10-2 Pa are shown in Fig. 4(c) and 4(d), respectively. Two pairs of doublets have been observed in Fig. 4(c). Except the two peaks at 71.03 eV and 74.38 eV corresponding to the Pt 4f core level spectra of elemental Pt, a new doublet with the peaks of 72.15 eV and 75.00 eV is also observed, which corresponds to the Pt 4f7/2 and Pt 4f5/2 core level spectra of the Pt-S bonding of PtS.30 As for the S 2p core level spectra shown in Fig. 4(d), two peaks at 162.85 eV and 164.05 eV have been observed, which are also ascribed to the S-Pt bonding of PtS. In addition, the calculated stoichiometric ratio of Pt/S is 1:0.74. These results suggest that part of the Pt film has reacted with sulfur to form PtS under 10-2 Pa.
Fig. 4(e) and 4(f) show the Pt 4f and S 2p core level spectra of the film synthesized at 10 Pa. As shown in Fig. 4(e), the Pt 4f core level spectra exhibit two pairs of doublets. The peaks of one doublet are at 72.53 eV and 75.88 eV, which correspond to the Pt-S bonding of PtS. The small shift in peak positions of the same chemical bonding as compared with the results shown in Fig. 4(c) is due to the calibration error. The other doublet with the peaks at 74.13 eV and 77.48 eV is ascribed to the Pt-S bonding of PtS2.30 In Fig. 4(f), two groups of doublets associated with the S 2p core level spectra have been observed. The two peaks at 163.13 eV and 164.33 eV correspond to the S-Pt bonding of PtS, and the doublet with the peaks at 163.73 eV and 164.93 eV is ascribed to the S-Pt bonding of PtS2. The calculated stoichiometric ratio of Pt/S is 1:1.32. Since no feature of elemental Pt has been observed from the spectra, we can conclude that the synthesized film is composed of PtS and PtS2.
We further increased the reaction pressure to 280 Pa in order to fully sulfurize the Pt film into PtS2. The spectra showing Pt 4f and S 2p core level peaks of the film are demonstrated in Fig. 4(g) and 4(h). Only two peaks at 73.32 eV and 76.67 were observed in Fig. 4(g), corresponding the Pt 4f7/2 and Pt 4f5/2 core level spectra of the Pt-S bonding of PtS2. As for the S 2p core level spectra shown in Fig. 4(h), the two peaks at 164.01 eV and 165.21 eV are ascribed to the S 3p3/2 and S 3p1/2 core level spectra of the S-Pt bonding of PtS2. No component other than PtS2 is observed. In addition, the stoichiometric ratio of Pt/S of the synthesized film is calculated as 1:1.87, which is very close to that of ideal PtS2. So, sufficient reaction pressure is the prerequisite to initiate the reaction between Pt and S and in forming a stoichiometric PtS2 film, although the film surface morphology and crystallinity might be degraded with increasing pressure.
To investigate the electrical performance of the synthesized PtS2 film, back-gate FET arrays have been fabricated on SiO2/Si substrate. The advantage of utilizing the synthetic large-area PtS2 film is that it can enable batch fabrication of homogeneous and reproducible devices, and is very attractive for future device integration applications. Fig. 5(a) shows the cross-sectional schematic of a few-layer PtS2 back-gate FET and the FET device arrays. All the electrical measurements were performed at room temperature under ambient conditions. The transfer and output characteristics of the PtS2 FET with 10 μm channel length and 30 μm channel width are shown in Fig. 5(b) and 5(c), respectively. Typical p-type transport behavior has been observed. However, the on/off ratio is less than 100 which is largely due to the relatively high off-state current. The possible remaining Pt (although in very small amount) in the film can lead to high off-state leakage. Fig. 5(d) depicts the drain-source current (Id) vs. drain-source voltage (Vds) at small Vds level. The linear characteristics suggest ohmic contact at source and drain. Fig. 5(e) compares the transfer curves of PtS2 FETs with different channel width, but with the same channel length of 10 μm. With increasing the channel width from 10 μm to 30 μm, the FET driving capability has been improved with higher on-state current. The field-effect mobility (μ) was extracted from the linear regime of Id-Vbg (Vbg is the back-gate voltage) curve using the equation μ=(ΔId/ΔVg) × L/(WCoxVd), where L is the channel length, W is the channel width, and Cox=1.3×10-4 F/m2 is the gate oxide capacitance per unit area (Cox=ε1ε0/d, ε1=3.9, ε0=8.85×10-12 F/m is the dielectric constant of vacuum, d= 300 nm for SiO2 dielectric). The statistical study of the hole mobility of PtS2 transistors was carried out on a same batch of devices fabricated simultaneously. As shown in Fig. 5(f), the majority of devices exhibit mobility at range of 0.0015-0.0025 cm2/V·s, and the average mobility of all calculated transistors is 0.0036 cm2/V·s. Although our work shown here cannot unambiguously demonstrate the optimized electrical device built on synthetic large-area PtS2 films, the results have already demonstrated the good uniformity and homogeneity in the electrical performance which can be very attractive for future device integration. Further materials engineering and theoretical simulations on the transport properties of PtS2 thin film will shed more light on the enhancement of the device electrical performance.
In summary, the band structure of 1T-PtS2 has been theoretically studied with first-principle calculation. Large-area synthesis has also been achieved by direct sulfurization of pre-deposited Pt films. The impact of sulfurization pressure on the film quality, specifically the film composites and stoichiometry has been investigated by testing various pressure conditions. It has been found that higher pressure is necessary to induce the reaction between Pt and S and form stoichiometric PtS2. The synthesized large-area PtS2 film has provided a good platform for FET device fabrication and integration. P-type FET behavior has been observed in our PtS2 devices. Although the carrier mobility and on/off ratio is still not satisfactory, our results open up an attractive approach to implement the synthetic large-area PtS2 films in homogeneous device integration and potential system-level applications. Further materials engineering can be expected to improve the film morphology and crystallinity under high pressure conditions, as well as the device electrical performance.
This work was sponsored by the NSFC (61522404, 61474029, 61427901 and 61704030), the Shanghai Pujiang Program (17PJ1400500), the National High Technology Research and Development Program (2015AA016501), and the National Major Projects of Science and Technology (2017ZX02315005).