1. Introduction

With the development of two-dimensional (2D) materials, there is increasing interest in assembling different layered materials to form heterostructures. Van der Waals (vdW) heterostructures assemble different materials through weak vdW interactions, which include not only physical stacking of 2D-2D materials (2D/2D), but also different materials with different dimensions [1,2]. One can achieve mixed-dimensional vdW heterostructures through flexible integration of multiple material systems beyond 2D, such as 2D/zero-dimensional, 2D/one-dimensional, and 2D/three-dimensional (3D) [3,4]. In particular, the interface effects in heterostructures (including band arrangement, built-in electric field, defects, and doping), have crucial impacts on the performance of devices. For example, multifunctional 2D material/wide bandgap semiconductor heterostructures can facilitate carrier transport at the interface and engender novel physical phenomena and applications, and thus improving the performance of electronic as well as optoelectronic devices [5]. Flexible integration of different functional materials has substantial potential for applications such as light-emitting diodes, optical modulators, and photodetectors [6].

Traditional semiconductors usually have a large exciton binding energy, which substantially hinders charge transfer between layers. By constructing a hybrid structure, one can achieve effective carrier transport. The charge transfer and electronic structural optimization in a hybrid system can lead to enhanced photoluminescence (PL) and excellent catalytic performance [7,8]. In recent years, ZnS-based hybrid structures (as potential catalysts) have been an active area of research [9,10]. ZnS is a typical wide bandgap (ca. 3.7 eV) semiconductor with a large exciton binding energy (40 meV) [11]. Because of its direct and wide bandgap, ZnS exhibits excellent chemical and physical properties (e.g., polar surface, large transmittance to the visible spectrum, and thermal stability) [12]. More importantly, ZnS, as an unusually versatile and impactful material, has been widely used in light-emitting diodes [13], nonlinear optical devices [14], ultraviolet (UV) photodetectors [15], and photocatalysis [16]. However, the photogenerated electrons of ZnS tend to recombine with holes, which limits its application potential in optoelectronics. Therefore, ZnS hybrid structures (with ZnO or graphene) have attracted substantial attention [17–20], and are of substantial importance for development of imaging and sensor devices.

Graphene has a high carrier mobility and biocompatibility, which facilitate spatial charge separation in composite structures [21–25]. With a large specific surface area, graphene can act as a conductive path for electron$-$hole separation and transfer [26]. Recently, there have been some reports about the fabrication and properties of graphene-ZnS (Gr/ZnS) composites. In Gr/ZnS hybrid systems, graphene can prevent electron$-$hole recombination and facilitate charge separation in ZnS, and in so doing enhance photoelectric properties as well as achieve efficient photoelectric performance [27]. Liu et al. found that there was electron transfer between graphene and ZnS [28]. Therefore, Gr/ZnS vdW heterostructure is an excellent combination to regulate and improve the conductivity as well as sensitivity of the components. These improved properties are conducive to sensors and photocatalysis applications. However, the electrical and optical properties at a Gr/ZnS 2D/3D vdW interface have not yet been investigated in detail.

In this paper, we integrated monolayer graphene flakes with a ZnS bulk plate to form Gr/ZnS 2D/3D heterostructures, and examined the interface characteristics. To investigate the Gr/ZnS interface interactions, we explored the Raman spectroscopy, X-ray photoelectron spectroscopy (XPS), PL, and UV$-$visible absorption spectroscopy of Gr/ZnS. The XPS spectra show that the binding energy of Zn 2p and S 2p was red-shifted, which can be attributed to the carrier transfer at the Gr/ZnS interface. Moreover, the PL spectra indicate that graphene can enhance the luminescence of ZnS. The absorption spectra also reveal the strong interactions between graphene and ZnS near the band edge of ZnS. Furthermore, we theoretically evaluated the band shift and charge transfer at the Gr/ZnS heterojunction interface. These findings contribute to understanding of the interface characteristics of 2D/3D heterostructures, and provide resources for tailoring as well as modulating wide bandgap semiconductor devices.

2. Results and discussion

We transferred monolayer graphene flakes onto a sphalerite ZnS plate by wet transfer to form Gr/ZnS 2D/3D heterostructures and the schematic diagram is shown in Fig. 1(a). Specifically, we spin coated poly-(methylmethacrylate) (PMMA) layer on graphene, which was grown by chemical vapor deposition (CVD) on copper foil. The chemical etching was carried out to remove the copper foil by immersing the sample into a FeCl$_3$ solution. After the copper foil was removed, the graphene film was rinsed with deionized water several times. Then, we salvaged the film with prepared 3D ZnS, and heated it at 100 $^{\circ }$C for 30 minutes to make graphene and ZnS better adhere. Finally, the sample was soaked in acetone for 1 hour at 60 $^{\circ }$C and vacuum annealed for 3 hours at 200 $^{\circ }$C to remove the PMMA carrier, which generated continuous graphene films and a clean graphene interface, as shown in Fig. 1(a). The Gr/ZnS heterostructure with high-quality graphene can produce excellent photo-generated carriers transport at the interface. We characterized pure ZnS and Gr/ZnS by Raman spectroscopy, with a detection wavelength of 514.5 nm and a spot diameter of 1 $\mathrm{\mu}$m. It can be seen from Fig. 1(b) that the characteristic phonon modes of ZnS are mainly the transverse optical (TO) mode and the longitudinal optical (LO) mode, which are associated with horizontal and vertical phonon vibrations, respectively [29]. The first-order TO and LO modes are located at 273 and 349 cm$^{-1}$, respectively. The peaks at 522 and 667 cm$^{-1}$ are associated with the second-order TO and LO modes. Furthermore, the peak at 179 cm$^{-1}$ is caused by disorderly activated plasma resonance [30,31]. As shown in Fig. 1(c), graphene in the Gr/ZnS heterostructure is verified by the Raman spectrum (black curve). The peaks at 1580 and 2698 cm$^{-1}$ correspond to the G peak and 2D peak of graphene respectively [32]. The G mode corresponds to the vibration of sp$^{2}$ hybridized carbon. The 2D peak corresponds to the interlayer stacking mode of graphene. The intensity ratio of 2D/G indicates that the graphene is a monolayer.

 

Fig. 1. (a) Schematic of Gr/ZnS heterostructure integration and an optical micrograph of the Gr/ZnS sample. (b) Raman spectrum of ZnS. Peaks in the range of 250$-$700 cm$^{-1}$ are the TO and LO modes. (c) Raman peaks of graphene in the Gr/ZnS heterostructure; the peaks at 1580 and 2698 cm$^{-1}$ are the G and 2D peaks, respectively.

Download Full Size | PDF

We performed XPS (Axis Ultra DLD) measurements to confirm the surface chemical state and content of relevant elements. Figure 2 shows the XPS spectra of pure ZnS and the Gr/ZnS heterostructure. The XPS spectra in Fig. 2(a) indicate the presence of Zn, S, C, and O. The C peak originates from graphene and carbon impurities. The O characteristic peaks mainly come from the O hydroxyl groups and oxygen functional groups in graphene as well as impurities absorbed onto the sample. Figures 2(b) and 2(c) show high-resolution XPS spectra of Zn 2p and S 2p from pure ZnS (red curve) and Gr/ZnS (blue curve), respectively. As depicted in Fig. 2(b), the binding energies of Zn 2p$_{3/2}$ and Zn 2p$_{1/2}$ are at 1021.8 and 1044.8 eV, respectively, in the ZnS phase; the two peaks shift to 1021.5 and 1044.5 eV, respectively, in Gr/ZnS. In Fig. 2(c), the peaks at 161.8 and 163.0 eV are assigned to S 2p$_{3/2}$ and S 2p$_{1/2}$, respectively, in ZnS; these two peaks shift to 161.5 and 162.7 eV, respectively, in Gr/ZnS. We conclude that the binding energies in Gr/ZnS are both shifted to lower energies by 0.3 eV compared with the pure ZnS. These shifts can be ascribed to graphene, which lead to a much higher concentration of delocalized electrons. At the Gr/ZnS interface, electrons transfer from graphene (high Fermi level) to ZnS (low Fermi level), resulting in the decreased electron concentration in graphene and increased surface electron concentration in ZnS; thus, the binding energies of Zn 2p and S 2p decreased [33]. This change in charge concentrations will lead to enhanced luminescence and electrochemical performance, as well as much higher catalytic response [34]. Figure 2(d) shows the C 1s XPS spectra of Gr/ZnS. We decomposed the C 1s signal into three Gaussian peaks at 284.5, 285.5, and 288.6 eV (Peak 1, Peak 2, and Peak 3), which are assigned to C=C, C$-$OH, and O=C$-$OH, respectively [27,28,35,36]. Most of the carbon is in sp$^2$ hybrid state (Peak 1), whereas the content of oxygen-containing functional groups (Peak 2, Peak 3) is relatively small. Thus, a considerable quantity of sp$^2$ hybridized carbon comes from graphene in the detection depth [37]. In addition, there is no C$-$S bond peak at 286.8 eV, indicating that there is no covalent bond between graphene and ZnS [37,38]. Here, we acquired PL emission spectra of ZnS and Gr/ZnS at an excitation wavelength of 346 nm to investigate the migration, separation, and recombination processes of photogenerated electron$-$hole pairs in ambient air. From Fig. 3(a), one can see that the green emission band centered at 525 nm is evident in both ZnS (red curve) and Gr/ZnS (blue curve), which originates from Zn-vacancy states emission luminescence near the ZnS surfaces [15]. Obviously, compared with the pure ZnS sample, the Gr/ZnS heterostructure has a higher PL intensity, which is attributed to the interaction between ZnS and graphene at the heterojunction interface [28]. Such enhancement also implies that there is active interfacial charge transfer between the photo-excited ZnS and graphene. For a more concise elucidation of Gr/ZnS interactions, we theoretically estimated the charge transfer process at the Gr/ZnS interface. As shown in Fig. 3(b) and 3(c), the calculated work function of ZnS ($\mathrm{\Phi}_{\textrm{ZnS}}$) is 5.9 eV, greater than that of graphene ($\mathrm{\Phi} _\mathrm{Gr}$) 5.56 eV; thus, the Fermi level of graphene is higher than that of ZnS. This difference in the Fermi level causes charge transfer at the interface, which is consistent with the literatures [39,40].

 

Fig. 2. XPS spectra of ZnS and the Gr/ZnS heterostructure. (a) XPS survey spectrum. (b) and (c) High-resolution Zn 2p and S 2p XPS spectra of both ZnS and Gr/ZnS. (d) Peak-fitting spectrum of C 1s from Gr/ZnS.

Download Full Size | PDF

 

Fig. 3. (a) PL spectra of ZnS and Gr/ZnS samples at an excitation wavelength of 346 nm. (b) and (c) Work functions of graphene and ZnS. Band alignment of Gr-ZnS in initial contact (d), and in an equilibrium state (e). (f) Charge transfer in the Gr/ZnS sample under UV-light.

Download Full Size | PDF

In accordance with the experimental and calculated results, we revealed the probable band alignment and charge transfer process at the Gr/ZnS interface, as shown in Fig. 3(d)–3(f). When the two materials are initially in contact (Fig. 3(d)), the electrons in graphene transfer to ZnS; such that the electron concentration decreases in graphene and increases in ZnS, until their Fermi levels are equal and reach an equilibrium state. In equilibrium (in Fig. 3(e)), a built-in electric field forms at the heterojunction interface, and the ZnS surface potential decreases, resulting in band-bending downward near the interface. When UV-light is incident on the sample, the interaction between graphene and ZnS enhances the light absorption by the sample, and facilitates the charge separation of photogenerated electron-hole pairs. More electrons in the ZnS are excited from the valence band to the conduction band (in Fig. 3(f)). Some of the excited electrons recombine with holes and emit fluorescence, while others transfer from ZnS to graphene. These transferred electrons in graphene are thermalized to the Fermi level of graphene and transfer back to ZnS, and then recombine with holes and emit fluorescence [28]. Thus, the enhanced PL intensity in Gr/ZnS can be observed in the experiment. This interface charge interaction results in a much higher photocurrent compared with a corresponding monolayer-film device, and hence improves the performance of optoelectronic devices including photodetectors, phototransistors, and photocatalysts.

Furthermore, the UV$-$visible absorption spectra were measured to study the optical properties of the bulk ZnS (red curve) and the Gr/ZnS (blue curve), as shown in Fig. 4(a) and 4(b), respectively. In Fig. 4(b), there is an absorption peak and the absorbance reaches the maximum at 350 nm. Obviously, there are interface interactions between graphene and ZnS at the ZnS band edge, increasing the UV-light absorption of the sample. In addition, on the basis of the absorption spectrum, the bandgap energy can be estimated by employing the Kubelka$-$Munk function [9]. One can acquire the bandgap by a transformation of the Kubelka$-$Munk function versus the light energy, as shown in Fig. 4(c) and 4(d). Here, $\alpha$ is the absorbance and $h\nu$ is the incident light energy. It is noted that the wide bandgap of ZnS is narrowed after hybridizing with 2D graphene (the initial bandgap of ZnS is 3.72 eV, and that of Gr/ZnS is 3.42 eV). This further indicates that graphene can extend the light absorbance band edge of ZnS, which is consistent with previous reports for ZnS-graphene hybrids [41].

 

Fig. 4. UV$-$visible absorption spectra of ZnS (a) and Gr/ZnS (b). Transformed Kubelka$-$Munk function versus the light energy of ZnS (c) and Gr/ZnS (d).

Download Full Size | PDF

To verify the present experimental results, we examined the electronic band structures and density of states (DOS) with CASTEP in Material Studio, based on first principle [42]. We used the generalized gradient approximation of Perdew$-$Burke$-$Enzerhof (GGA-PBE) for the exchange-correlation function [43]. For the structural relaxation and electronic structures calculations, we set the kinetic energy cutoff to 400 eV and used $\mathrm{K}$-point meshes 7$\times$7$\times$2 to optimize the geometric structures. We chose the stress force and energy convergence criteria as 0.03 eV/Å and 10$^{-5}$ eV, respectively. Figures 5(a) and 5(b) illustrate the electronic band structures of monolayer graphene and zinc blende ZnS. It is obvious that graphene band structure crosses at the $\mathrm{K}$-point and the bandgap is zero. For ZnS, the valence band maximum (VBM) and the conduction band minimum (CBM) are located at the $\Gamma$-point, which indicates that ZnS has a direct $\Gamma -\Gamma$ gap of 2.14 eV ($\Delta E_{1}$). However, the calculated bandgap is much lower than the experimental result, which can be attributed to the GGA-PBE approximation [44]. This difference does not affect the bandgap variation tendency of ZnS caused by graphene. To gain insight into the band structure variation of ZnS caused by graphene, we evaluated the electronic band structure for the Gr/ZnS as shown in Fig. 5(c). In comparison with pure ZnS, $\Gamma -\Gamma$ gap is reduced to 1.38 eV ($\Delta E_{2}$), indicating that graphene modulates the bandgap of ZnS. To distinguish the contribution of each constituent element more clearly, we calculated the DOS from different elements. The total and partial elements’ DOS of ZnS and Gr/ZnS are both shown in Fig. 5(d) and 5(e) marked with different colors. For pure ZnS, the energy at the VBM mainly comes from S, and the energy at the CBM mainly pertains to Zn. In Fig. 5(e), one can find that the DOS of C is the main contribution to the decrease of the $\Gamma -\Gamma$ gap between the VBM and CBM. It is obvious that the Gr/ZnS bandgap narrowing is mainly caused by graphene, which is consistent with the experimental results.

 

Fig. 5. The calculated band structure for graphene (a), ZnS (b), and Gr/ZnS (c). Partial and total DOS of ZnS (d) and Gr/ZnS (e).

Download Full Size | PDF

3. Conclusions

In conclusion, we investigated the interface properties in a Gr/ZnS 2D/3D heterostructure. In our experiments, the XPS results indicate that the ZnS binding energy decreases by 0.3 eV after combining with graphene caused by charge transfer between graphene and ZnS. The absorption spectra also verified the intense band edge interactions in Gr/ZnS. Moreover, we confirmed that graphene can enhance the luminescence intensity of ZnS because of the interface interaction in Gr/ZnS. By theoretical analysis, we conclude that there was band-bending and shifting at the Gr/ZnS interface. Our results indicate that one can modulate the band structure and optoelectronic properties of ZnS with graphene, which has substantial potential for future applications.