Introduction

The rapid progress of graphene research has stimulated interest in other several types of two-dimensional layered materials. Recently, transition metal dichalcogenides (TMDs) have attracted considerable attention since the observation of remarkable electronic and optical properties [1,2,3]. These TMD crystals can be grown or mechanically exfoliated to monolayer thickness, similar to the exfoliation of graphene. However, in contrast to graphene, monolayer TMDs consist of more than one element, which makes their physical properties more complex than graphene. Among the TMDs, MoS2 is the most extensively studied, where one Mo plane is sandwiched between two S planes usually with a 2H-structure [4]. In contrast to these high-symmetry hexagonal structures such as MoS2, another kind of TMDs such as ReS2 is attracting much interest, which exhibits a distorted 1T’-structure [5]. The upper and lower S atoms sandwich the middle layer of Re atoms with a hexagonal structure having an additional Peierls twist [5]. This is because the rhenium atom possesses one extra valence electron, leading to the formation of additional Re–Re bonds in ReS2 (atomic structure diagram of a single-layer MoS2 and ReS2 is shown in Additional file 1: Figure S1.) The reduced symmetry in ReS2 induces significant in-plane anisotropy and therefore adds an additional degree of freedom, which makes ReS2 an interesting material for the fabrication of FETs and polarization-sensitive photodetectors [5, 6]. In this paper, we probed the polarization properties of single-layer (abbreviated as the notation SL) MoS2 and ReS2 flakes by angular-dependent reflection spectra measurements on SiO2/Si and quartz substrates. Our results will shed light on the new effects in those strongly anisotropic layered materials and can empirically be used to identify the crystal orientation.

Materials and Methods

The MoS2 and ReS2 flakes with different numbers of layers in this paper were exfoliated from bulk MoS2 and ReS2 crystals by micromechanical cleavage method and were prepared on substrates. The interaction between samples and substrates was different and the influence of substrates on experimental results should be considered. Thus, we selected two kinds of substrates: one is the Si {100} substrate covered with an 89 nm SiO2 and the other is the quartz crystal with a thickness of 1 mm, to support MoS2 and ReS2 flakes (the optical microscopic images of SL MoS2 and SL ReS2 flakes supported on SiO2/Si substrate and supported on quartz substrate are shown in Additional file 1: Figure S2.) The SL dichalcogenides have a thickness between 0.6 and 0.7 nm which are extremely sensitive to measurement accuracy of measuring instruments. We used ultra-low frequency Raman spectroscopy [7, 8] (the ultra-low frequency Raman spectra of SL MoS2 and SL ReS2 flakes supported on SiO2/Si substrate and supported on quartz substrate are shown in Additional file 1: Figure S3.) and photoluminescence (PL) spectroscopy [8, 9] (the PL spectra of SL MoS2 and SL ReS2 flakes supported on SiO2/Si substrate and supported on quartz substrate are shown in Additional file 1: Figure S4.) to accurately determine the SL MoS2 and ReS2 flakes.

Reflection spectrum measurements were performed in a backscattering geometry using a Jobin-Yvon HR800 micro-Raman system. The tungsten halogen lamp was used as a light source with the spot size below 2 μm. The objective of × 100 (NA = 0.9) was used to ensure the accuracy of tests with the size of samples above 5 μm. The best reflected light signal was achieved by focusing the microscope to get maximum peak intensity. The reflection spectra were measured from the samples and bare substrates in the broad wavelength range of 400–800 nm. The 600 lines per millimeter grating was used, which enables one to have each CCD pixel to cover 1 nm. A polarizer was placed on the light path in front of the sample. By continuously rotating the polarizer from 0 to 360°, polarization directions of incident and reflected light were simultaneously varied with polarization angles from 0 to 360°. When the polarizer was rotated to an angle, the reflection spectra of the sample (SL MoS2 or SL ReS2) and the substrate (SiO2/Si or quartz) were measured once. All of the polarization reflection spectra were measured under the condition of keeping the lamp intensity unchanged. We used R (sam + sub) and R (sub) to respectively indicate the reflection intensities of samples and bare substrates and used the optical contrast method to normalize the data by the formula of ROC = 1 − R (sam + sub)/R (sub) (the substrate is SiO2/Si) or ROC = R (sam + sub)/R (sub) − 1 (the substrate is quartz). In the following studies, the angular-dependent optical contrasts of SL MoS2 and ReS2 on different substrates were demonstrated respectively.

Results and Discussion

SL MoS2 on SiO2/Si Substrate

We firstly measured the polarization reflection spectra of SL MoS2 supported on SiO2/Si substrate by continuously rotating the polarizer from 0 to 360°. The polarizer was rotated once every 30°. Figure 1 a shows the variation of optical contrasts with polarization angles from 0 to 180°. The original curves were overlapping each other and the processed curves were offset for clarity. There are two peaks at ~ 611 nm and ~ 658 nm due to A and B exciton emission [10, 11]. We selected them as references and showed their intensities with the polarization angles from 0 to 360° in Fig. 1 b and c by pink and red circles, respectively. The intensities of two peaks are basically unchanged, which is we should predict since the SL MoS2 is hexagonally symmetrical.

Fig. 1
figure 1

a The polarization optical contrast curves of SL MoS2 flakes supported on SiO2/Si substrate. b The intensity variation at ~ 611 nm from 0 to 360°. c The intensity variation at ~ 658 nm from 0 to 360°

Full size image

SL ReS2 on SiO2/Si Substrate

The polarization reflection spectra of SL ReS2 supported on SiO2/Si substrate were measured as followed. The optical contrast curves of SL ReS2 flake with varying polarization angles from 0 to 180° are shown in Fig. 2 a and are offset for clarity. There is a valley at ~ 457 nm and a peak at ~ 629 nm [12] suggesting that SL ReS2 crystallizes in a different crystal structure from SL MoS2. The intensities at ~ 457 nm and ~ 629 nm changed as the polarization angle changed. Taking them as references, we showed their intensities with the polarization angles from 0 to 360° in Fig. 2 b and c by pink and red circles, respectively. Both of the intensities at two positions show polarization dependence on the polarization angles, which is directly resulted from the low crystal symmetry in SL ReS2. The in-plane distortion of SL ReS2 lattice is expected to affect profoundly the interlayer coupling in multilayer ReS2 crystals because the similar polarization dependence has been found in the optical contrast curves of anisotropic-like-stacked 2 L ReS2 flakes supported on SiO2/Si substrate [12] and even in the ultra-low frequency Raman spectra and PL spectra of isotropic-like-stacked 2 L ReS2 flakes [8].

Fig. 2
figure 2

a The polarization optical contrast curves of SL ReS2 flakes supported on SiO2/Si substrate. b The intensity variation at ~ 457 nm from 0 to 360°. c The intensity variation at ~ 629 nm from 0 to 360°

Full size image

We fitted the function of the intensities at ~ 457 nm and ~ 629 nm as the polarization angles by a first-order Fourier formula: f(θ) = a0 + a1 × cos(θ × w) + b1 × sin(θ × w), where θ is the polarization angle; a0, a1, and b1 are the amplitudes; and w is the frequency. The positions of minimum and maximum intensities were read as 20° and 110°, respectively, at both of ~ 457 nm and ~ 629 nm. The fitted curves were also plotted in Fig. 2 b and c with blue lines. At ~ 457 nm, a0 = 8.269, a1 = − 4.878, b1 = − 4.585, and w = 0.0348, and at ~ 629 nm, a0 = 34.27, a1 = − 5.99, b1 = − 4.747, and w = 0.03525. They have the basically identical change period with the polarization angles due to the nearly equal w. It should be derived from the distorted structure in the SL ReS2.

SL MoS2 on Quartz Substrate

Because SiO2/Si substrate is opaque, the incident light passed through interfaces of air/sample and sample/substrate and finally was absorbed by the substrate. Meanwhile, the reflected light was collected from each interface and finally transmitted into the air. The optical interference occurred in the multilayered structures and physical properties of the substrate were included in the outgoing-reflected signals in addition to the sample [12]. The SiO2/Si substrate was a polarized substrate although we used the optical contrast method to normalize the data by the formula of ROC = 1 − R (sam + sub)/R (sub). In order to eliminate the disturbance of polarized properties from the substrate, we then measured the polarization reflection spectra of SL MoS2 and ReS2 on the quartz substrate due to the transparency and isotropy of the quartz substrate.

Since the quartz substrate is transparent, the sample stage should be placed suspended to ensure transparency during measuring. The incident light passed through interfaces of air/sample, sample/substrate, and substrate/air and finally was absorbed by the air to avoid disturbing the collecting of reflected light. We used the formula of ROC = R (sam + sub)/R (sub) − 1 to normalize the data. Figure 3 a shows the polarized optical contrast curves of SL MoS2 flake on the quartz substrate with varying polarization angles from 0 to 180°. As can be seen, there are also two peaks related to A and B exciton at ~ 615 nm and ~ 665 nm, respectively. Their position has some shift towards long wavelength than that supported on SiO2/Si substrate due to interference effects on different substrates [11]. We plotted their intensities with the polarization angles in Fig. 3 b and c. The intensities of two peaks are almost no change as the polarization angle changes, which indicates that in-plane isotropic properties of SL MoS2 are unchangeable when they are attached to whatever substrates.

Fig. 3
figure 3

a The polarization optical contrast curves of SL MoS2 flakes supported on quartz substrate. b The intensity variation at ~ 615 nm from 0 to 360°. c The intensity variation at ~ 665 nm from 0 to 360°

Full size image

SL ReS2 on Quartz Substrate

Figure 4 a shows the polarized optical contrast curves of SL ReS2 flake on the quartz substrate, in which there are two valleys at ~ 477 nm and ~ 641 nm, respectively. The difference of features between supported on the quartz substrate and supported on SiO2/Si substrate is also due to interference effects on different substrates [11]. Figure 4 b and c show the intensities of two valleys with the polarization angles. Both of them show polarization dependence on the polarization angles, which indicates that SL ReS2 is in-plane anisotropic regardless of substrates. We fitted the relation of the intensities at ~ 477 nm and ~ 641 nm with the polarization angles by a first-order Fourier formula: f(θ) = a0 + a1 × cos(θ × w) + b1 × sin(θ × w), where a0 = 0.3168, a1 = − 0.02215, b1 = − 0.0004139, and w = 0.03422 at ~ 477 nm and a0 = 0.2941, a1 = − 0.06608, b1 = − 0.005685, and w = 0.0349 at ~ 641 nm. The positions of minimum and maximum intensities were read as 0° and 90°, respectively, at both of ~ 477 nm and ~ 641 nm. The fitted curves were also plotted in Fig. 4 b and c with blue lines. The w is basically identical at both ~ 477 nm and ~ 641 nm and nearly equal to that at ~ 457 nm and ~ 629 nm of SL ReS2 flakes supported on SiO2/Si substrate, which means that the polarized properties in SL ReS2 flakes exhibit a change tendency in the sin or cos function as the polarization angle changes from 0 to 360° and the period is uniform when they are attached to whatever substrates.

Fig. 4
figure 4

a The polarization optical contrast curves of SL ReS2 flakes supported on quartz substrate. b The intensity variation at ~ 477 nm from 0 to 360°. c The intensity variation at ~ 641 nm from 0 to 360°

Full size image

Conclusions

In conclusion, SL MoS2 and ReS2 on SiO2/Si substrate and on quartz substrate have been studied by polarization reflection spectra, which identify a significant in-plane isotropy in SL MoS2 due to a hexagonal structure and in-plane anisotropy in SL ReS2 due to an additional distorted structure with a hexagonal structure. According to the polarized optical contrast curves with the polarization angles, there are some wavelength-dependent peaks or valleys in SL MoS2 and ReS2 predicted by different crystal structures. The variation of intensities at peaks or valleys with the polarization angles confirms the existence of different angle-dependent properties in SL MoS2 and ReS2. The same properties exist in some SL 2D materials having a similar structure with MoS2 such as WS2, MoSe2, and WSe2, and having a similar structure with ReS2 such as ReSe2 and WTe2. There are many other SL 2D materials that have other types of asymmetric lattice structures, such as BP and SnSe, which have strongly buckled honeycomb sheets with “troughs” running along the y-axis. These samples might also show anisotropic features. It implies that some new polarization-dependent electronic devices may soon be realized and promoted considering the wide variety of samples.