1 Introduction

Two-dimensional transition metal dichalcogenides (TMDs) have rich electrical, optical, chemical properties, and their bandwidth ranges from visible to infrared. Owing to the excellent mechanical, optical, and electrical properties, TMDs are suitable materials for sensors, catalysts, storage batteries, optoelectronic devices, and excitonic devices^[1−4]. Monolayer WS₂ is an atomically thin two-dimensional material, which is a sandwich structure, and is composed of a single layer transition metal tungsten (W) and two layers of chalcogen (S)^{[5, 6]}.

One key issue in harnessing novel properties of TMDs is to understand and control the carrier dynamic processes. Because the performance of TMDs-based photoelectric devices are closely related to the carrier dynamics process, effective and continuous tuning of carrier dynamics in TMDs is highly desired. Hydrostatic pressure is pure and clean because no additional impurities are introduced by the hydrostatic pressure of the diamond anvil cell (DAC). Recent reports have indicated that the optoelectronic properties of TMDs can be effectively tuned via hydrostatic pressure, such as the interaction of interlayer, material structure, carrier density, and conductivity^[7−13]. The semiconductor to metallic phase transition in bulk WS₂ is caused by hydrostatic pressure^{[14, 15]}. In contrast, there are few studies on the carrier recombination of monolayer TMDs under pressure. Thus, it is of great value to develop an understanding of the carrier dynamics in TMDs with pressure.

Here, monolayer WS₂ is used as a sample to study the carrier dynamics via transient absorption measurements. Our results demonstrate a pressure-induced blue shift of the A-exciton resonance, and two kinds of carrier capture processes of defect states can be continuously tuned by hydrostatic pressure. This study provides basic information for understanding the dynamics of carrier in TMDs, which may be useful in the design of new photoelectric devices that are based on TMDs.

2 Synthesis and material characterization

The WS₂ monolayers were grown by CVD technique at 800 °C for 20 min onto SiO₂/Si, as shown in Fig. 1. Sulfur powder and tungsten trioxide powder with a purity of 99.9% were used as source materials for experimental preparation of samples. Meanwhile, 0.2-g sulfur powder was placed on the crucible in the first temperature zone, and 0.1-g tungsten trioxide powder was placed on the crucible in the second temperature zone. Furthermore, 40 sccm (cubic centimeters per minute) Ar and 5 sccm H₂ was used as the carrier gas to transfer the evaporated sulfur vapor to the substrate. We raised the temperature of furnace 1 from 25 °C to 200 °C for 1 h, and the temperature of furnace 2 from 25 °C to 800 °C, and maintained the temperature for 20 min. The S powder is gasified in the low temperature area, transported to the high temperature area by the carrier gas, and reacts with the gasified WO₃ on the substrate to generate WS₂. The whole CVD equipment was then naturally cooled down to room temperature.

Fig. 1. (Color online) Experimental setup of the CVD growth of monolayer WS₂ sample.

下载图片 查看所有图片

The size of the sample prepared by CVD method can reach 40 μm, as shown in Fig. 2(a). The Raman spectra and photoluminescence (PL) spectra of the sample are shown in Figs. 2(c) and 2(d), respectively. The PL peak of the sample is centered at 615 nm. The Raman peaks of the sample is centered at 350 cm⁻¹ and 418 cm⁻¹, which represents the E_2g¹ and A_1g peaks of WS₂, respectively. The difference between the E_2g¹ and A_1g peaks is 68 cm⁻¹, which is in accord with the characteristics of single layer WS₂^[16]. The schematics of these two Raman modes E¹_2g and A_1g are shown in Fig. 2(b), the A_1g mode of monolayer WS₂ originates from the transverse out-of-plane relative vibration between sulfur atoms in the interlayer, while the E¹_2g mode originates from the in-plane longitudinal vibration of tungsten and sulfur atoms in opposite directions. Therefore, E¹_2g mode vibration is sensitive to the length of W−S bond. The PL peak of the monolayer WS₂ lies at 615 nm (2.02 eV), which indicates that the valence band top and the conduction band bottom of the monolayer WS₂ are located at the transition point, which is a direct bandgap^[11]. The AFM image show that the thickness of our monolayer WS₂ sample is 0.8 nm, which is shown in Fig. 2(e) and Fig. 2(f).

Fig. 2. (Color online) The characterization of monolayer WS₂. (a) Optical image of a monolayer WS₂. (b) Schematics of Raman modes of E_2g¹ and A_1g. (c) Raman spectra of monolayer WS₂ sample. (d) PL spectra of monolayer WS₂ sample. (e) AFM image of monolayer WS₂ sample. (f) Height image intensity profile along the black line in (e).

下载图片 查看所有图片

The monolayer WS₂ was peeled from the substrate with the PVP/PVA adsorption transfer method. The pressure device used in this experiment is a four-column DAC, and the size of diamond plane is 300 μm. The T301 gasket was pre-pressed to 50 μm, and a hole with a diameter of 100 μm was punched in the center with a laser. Silicone oil was chosen as a pressure transmission medium. The silicone oil, the monolayer WS₂ sample, and a 10 μm ruby were placed in the hole. The DAC device was used to gradually increase the pressure of the monolayer WS₂ heterostructure to 2.91 GPa from 0 GPa, and the direction in which the pressure applied is perpendicular to the plane of the monolayer WS₂ sample.

3 Results and discussion

Raman measurements were performed to study the lattice vibration of the monolayer WS₂ sample when pressurized in the DAC^[17] and the pressure transmission medium (PTM) is silicone oil. The evolutions of Raman spectra show a clear pressure-induced changeover, the feature at 350 cm⁻¹ and 418 cm⁻¹ (E¹_2g and A_1g modes of monolayer WS₂) blue shift as the pressure increases (as shown in Fig. 3(a)), which is similar to other layered materials^{[18, 19]}. The Raman peak as a function of pressure is shown in Fig. 3(b), and there are two linear increases at a rate of 3.1 cm⁻¹GPa⁻¹ (A_1g) and 1.8 cm⁻¹GPa⁻¹ (E¹_2g), respectively. The A_1g modes are more sensitive to pressure than that of the E¹_2g modes, which may be the reason why the out-of-plane motion of S−S bonds of A_1g modes is more sensitive than the in-plane motion of the W−S bonds of the E¹_2g modes under pressure. Therefore, the out-of-plane compression is more favorable than in plane compression under pressure^{[20, 21]}.

Fig. 3. (Color online) Pressure-induced Raman vibration in monolayer WS₂. (a) Evolution of the Raman spectrum with pressure for monolayer WS₂ sample. (b) Raman vibrations of monolayer WS₂ as a function of pressure.

下载图片 查看所有图片

The PL spectra of monolayer WS₂ with different pressures were measured, as shown in Fig. 4(a). The PL peaks have a blue shift with an increase in pressure from 0 GPa to 2.91 GPa. The PL peaks as a function of pressure are shown in Fig. 4(b), and there is a linear increase at a rate of −2.1 nm⁻¹GPa⁻¹. The change of the bandgap can be attributed to the enhanced quantum confinement effect caused by the pressure.

Fig. 4. (Color online) (a) Evolution of the PL spectra of monolayer WS₂ sample with pressures. (b) PL peaks of monolayer WS₂ as a function of pressure.

下载图片 查看所有图片

The transient absorption (TA) spectra of monolayer WS₂ with different pressures were measured to gain more insight of the carrier dynamics process (as shown in Fig. 5(a)). For monolayer WS₂, the applied pump wavelength is 400 nm and the pump peak fluence is 10 μJ/cm². All of the samples were measured at room temperature.

Fig. 5. (Color online) (a) Schematic of material structure and TA measurements of monolayer WS₂ in the DAC. (b) Evolution of TA spectra of monolayer WS₂ under pressure from 0.30 to 3.25 Gpa. (c) A exciton peaks as a function of pressures.

下载图片 查看所有图片

The absorption peaks of monolayered WS₂ lies at 605−615 nm, which can be assigned to the A exciton resonance of WS₂, and the photo-induced bleaching signals indicate that the corresponding energy states are occupied by photo-induced carriers. The TA spectra traces of A excitons are shown in Fig. 5(b) as the pressure ranged from 0.55 to 2.91 GPa at a time delay of 3 ps. A blue shift of the A exciton and the decrease of the intensity is observed with the increase of pressure. As shown in Fig. 5(c), the peaks of photo-induced signals with different pressures can be linearly fitted with a slope of −5.93 nm/GPa. These features might be caused by the pressure-induced enlargement of the bandgap of TMDs^[22−25].

The representative TA signals of A excitons with different pressures were recorded, as shown in Fig. 6. The decay curve can be well fitted as ΔA=A1e−τ1+A2e−τ2 with two time constants. Because the vacancies of atoms and grain boundary dislocations are common in TMDs, the capture process of shallow and deep defects states may give rise to the time constants τ₁ and τ₂^[26−28].

Fig. 6. (Color online) The transient absorption signals of monolayer WS₂ with different pressures: (a) 0.55 GPa; (b) 0.89 Gpa; (c) 1.22 Gpa; (d) 1.56 GPa; (e) 1.89 GPa; (f) 2.23 Gpa; (g) 2.56 Gpa; (h) 2.91 GPa. The red lines are fitting.

下载图片 查看所有图片

It should be noted that the carrier capture process gets faster when increasing the pressure from 0.55 to 2.91 GPa, and can be divided into two distinct regions according to the slope variation of time constants upon compression, as shown in Fig. 7. For 0.55 GPa < P < 1.56 GPa, τ₁ changes with pressure at a rate of −1.12 ps/GPa and τ₂ declines with pressure at a rate of −51.78 ps/GPa. When the pressure is above 1.89 GPa, τ₁ shows a decline shift at a rate of −0.07 ps/GPa and that of τ₂ decreases to −6.38 ps/GPa.

Fig. 7. (Color online) The time constants (a) τ₁ and (b) τ₂ of A exciton as a function of pressure.

下载图片 查看所有图片

For monolayer TMDs, the time constants τ₁ and τ₂ can be assigned to the carrier trapping process of the shallow and deep defect state with different capture rates. Considering the reduced dielectric screening effect and strong quantum confinement effect of TMDs, the Auger type dynamic process can be realized at low carrier densities. The τ₁ and τ₂ may be caused by Auger type carrier capture process of defect states, which means the electrons or holes in exciton states are captured by defect states by transferring the excess energy to other electrons and holes to satisfy the energy conservation law^{[29, 30]}. According to this model, the quantum confinement effect may be enhanced by increasing the pressure, and the carriers are much easier to collide with other carriers. This feature means that the carriers are easier to release the excess energy to other electrons or holes, and this is beneficial for the Auger type carrier trapping process^[31−33]. However, there is a different model that can explain the pressure-induced acceleration of the carrier capture process—the pressure can shorten the W−S bond in the monolayer WS₂, and enhance the vibration between W and S atoms, which can increase phonon energy and be confirmed by the above Raman results. This feature is conducive to meet the requirements of energy and momentum conservation in the capturing process^{[20, 21]}.

At around 1.56−1.89 GPa, a turning point appears and changes the time constants of two decay process with the decrease of pressure, which may result from the pressure induced changeover of the electronic structure. Both of the transient components are fully recovered when the pressure is released. Thus, the results clearly demonstrate that the band structure and carrier dynamics processes of monolayer WS₂ can be accurately manipulated by hydrostatic pressure.

4 Conclusions

In summary, we have synthesized a sample monolayer WS₂ to study the optical response of the sample with pressure via Raman spectroscopy and transient absorption spectroscopy. We found that hydrostatic pressure can be an efficient method to tune the lattice vibrations. The transition spectra were analyzed by tracking the energy shift and carrier dynamics of A exciton with different pressures. The results demonstrate the efficient tuning of band structure and photo-induced carrier recombination processes by pressure engineering. Our findings provide the basic information for understanding carrier dynamics in atomically thin TMDs under external strain.