Plasmonic-tape-attached multilayered MoS₂ film for near-infrared photodetection

Abstract

Molybdenum disulfide has been intensively studied as a promising material for photodetector applications because of its excellent electrical and optical properties. We report a multilayer MoS₂ film attached with a plasmonic tape for near-infrared (NIR) detection. MoS₂ flakes are chemically exfoliated and transferred onto a polymer substrate, and silver nanoparticles (AgNPs) dewetted thermally on a substrate are transferred onto a Scotch tape. The Scotch tape with AgNPs is attached directly and simply onto the MoS₂ flakes. Consequently, the NIR photoresponse of the MoS₂ device is critically enhanced. The proposed tape transfer method enables the formation of plasmonic structures on arbitrary substrates, such as a polymer substrate, without requiring a high-temperature process. The performance of AgNPs-MoS₂ photodetectors is approximately four times higher than that of bare MoS₂ devices.

Introduction

Two-dimensional (2D) transition metal dichalcogenides (TMDCs) show excellent optoelectronic performance because their bandgap can be adjusted by controlling the thickness. This can be advantageous for flexible optoelectronic devices on an atomic scale^1–5. Among TMDCs, molybdenum disulfide has attracted considerable attention in optoelectronic applications such as light-emitting devices and photodetectors because of its high transparency, high carrier mobility, and mechanical flexibility^6–9. However, MoS₂ is typically employed for detecting wavelengths in the visible range, and its detection performance in the near-infrared (NIR) range is limited due to its intrinsic band structure. For MoS₂ films synthesized using chemical exfoliation^10–12, the interlayers with van der Waals bonding in a bulk-layered material are broken and intercalated with small ions such as lithium¹². Chemically exfoliated MoS₂ films can partially absorb infrared light because the MoS₂ flakes chemically exfoliated using organolithium have a high content of metallic 1 T phase MoS₂^10–14. It was reported that chemically exfoliated MoS₂ films extended their absorption up to the wavelength of 1,550 nm in the NIR region.

It is well known that plasmonic nanostructures can strongly induce a localized near-field. Various low-dimensional photodetectors combined with plasmonic nanostructures have demonstrated excellent device performance^15–23. Previously, we reported that chemically exfoliated MoS₂ could generate photocurrent by extending NIR light absorption up to the wavelength of 1,550 nm, and that the responsivity could be improved by simply forming a plasmonic nanostructure as an underlayer of MoS₂²⁴. That is, a Ag thin film was deposited and annealed to form randomized silver nanoparticles (AgNPs) successively on a substrate, and finally, a MoS₂ film was layered on the AgNPs array. However, this thermal dewetting method is not applicable to polymer substrates, and the surface roughness due to the metal nanoparticles (NPs) may cause unexpected electrical and mechanical problems in the active layer of MoS₂ flakes.

In this paper, we introduce a plasmonic-tape-attached multilayered MoS₂ device to enhance NIR absorption and consequently photocurrents. In this device, chemically exfoliated MoS₂ is first transferred onto a substrate, and then a plasmonic tape is directly attached to the MoS₂ film through a Scotch tape. Herein, a plasmonic tape refers to a composite film of transparent Scotch tape and embedded metal NPs at the adhesive surface. The plasmonic tape can be reproducibly fabricated by chemically treating the surface of a substrate onto which a metal thin film is deposited, and then removing it with Scotch tape. The plasmonic tape can be reproducibly fabricated using Scotch tape by taping and peeling thermally dewetted metal NPs from the chemically treated surface of a substrate. Therefore, this plasmonic tape is suitable for thermally or chemically weak substrates that cannot directly form plasmonic NPs via thermal annealing or chemical etching. In addition, the tape film itself serves as a passivation layer that can protect a device from moisture penetration or mechanical damage without deteriorating the electrical and optical properties. We systematically investigate the optoelectrical properties and photodetection performance of the plasmonic MoS₂ device at the NIR wavelengths of 980 and 1,550 nm. We report that the plasmonic MoS₂ device yields a sensitivity approximately four times that of the bare MoS₂ device.

Results

The schematic fabrication process of the MoS₂ device attached with the plasmonic AgNPs-tape film is briefly illustrated in Fig. 1a and the detailed procedure are shown and described in Figures S1–S3 and “Methods”. Figure 1b shows the photographs of the samples of the bare MoS₂ device and the plasmonic-tape-attached MoS₂ device. Figure 2a,b show the photographs, scanning electron microscopy (SEM) images and size distribution histograms of AgNPs formed on the Si substrate and transferred onto the 3 M tape (i.e., plasmonic tape film), respectively. The AgNPs are randomly distributed and are elliptical or circular in shape. The average diameters of the bare and transferred AgNPs are ~ 100.85 nm and ~ 96.82 nm, respectively. The size of the AgNPs before and after the tape transfer is believed to be almost the same, even though the transferred AgNPs are likely to appear small due to the low contrast ratio caused by the polymer adhesive. Moreover, the transfer rate of the AgNPs from Figure S4 was estimated to be ~ 99% according to atomic force microscopy (AFM) analysis. Therefore, the transfer process was very effective without serious loss of AgNPs. A Cary 5,000 UV–VIS-NIR spectrometer was used to study the optical properties of the bare MoS₂ and plasmonic-tape-MoS₂ (i.e., AgNPs/MoS₂) films. Multilayered MoS₂ films were transferred onto glass substrates to measure absorption spectra. In Fig. 2c, the AgNPs/MoS₂ film shows absorption enhancement over a broad band of spectra at all wavelengths, including the visible range; thus, the MoS₂ film becomes less transparent upon attaching the plasmonic tape. In the inset of Fig. 2c, the bare MoS₂ film exhibits peaks A and B at 672 nm and 612 nm, respectively, corresponding to the two MoS₂ direct band gap transitions^25–27, whereas it shows slightly increased absorption and a broad absorption tail, which indicate the indirect band transition. In the NIR (900–1,600 nm) region, the AgNPs/MoS₂ sample exhibits nearly four times stronger absorption than the bare MoS₂ film, owing to the plasmonic NIR absorption in the structure. Raman measurements were obtained using a Renishaw (inVia Raman Microscope) spectrometer with an excitation wavelength of 532 nm. Figure 2d shows the Raman spectra of the multilayered MoS₂ without and with the plasmonic tape, where a bare Scotch-tape-attached sample is used for comparison. The Raman spectrum of multilayered MoS₂ film displays an in-plane active mode E_2g¹ at ~ 384 cm⁻¹ and an out-of-plane mode A_1g at 405 cm⁻¹. In addition, it exhibits additional peaks, 156 (J₁), 226 (J₂), and 333 (J₃) cm⁻¹, featuring 1 T-MoS₂^28–30. After attaching the plasmonic tape on top of MoS₂, an overall increase in the intensity of the Raman spectra is observed. Especially, for the out-of-plane A_1g modes, additional peaks appeared and became broader because of the plasmonic effect of AgNPs. However, other peaks did not appear due to the tape.

Figure 1.

(a) Fabrication process of the MoS₂ photodetectors decorated with plasmonic AgNPs. Thermally dewetted AgNPs were detached from Si substrate using a 3 M tape and attached to the surface of a MoS₂ device. (b) Schematic illustration and photograph of the plasmonic-tape-MoS₂ photodetector.

Figure 2.

Photographs, SEM images, and size distribution histograms of (a) as-prepared AgNPs on Si substrate and (b) AgNPs transferred onto 3 M tape. (c) UV–Vis–NIR absorbance spectra of multilayered MoS₂ on a glass substrate with and without plasmonic AgNPs. (d) Raman spectra of plasmonic tape/MoS₂, bare tape, and bare MoS₂ film on the glass substrate.

Figure 3a shows the I_ds–V_ds curves of bare MoS₂ and plasmonic-tape-MoS₂ devices under the illumination of wavelength 980 nm at 121.03 µW where the channel length and width are 20 µm and 200 µm, respectively. The curves show good linearity, indicating that ohmic contacts between the MoS₂ films and the electrodes are well formed. The curves for the plasmonic samples show the same distinct linear characteristics as those of the bare MoS₂ films, indicating that the attached AgNPs do not affect the interfacial electrical properties of the MoS₂ layer. The plasmonic AgNPs-MoS₂ photodetectors show an evident improvement of photocurrent at 980 nm compared with the bare MoS₂ photodetectors. As shown in Fig. 3b, the transient photoresponses of the MoS₂ photodetectors without and with AgNPs are characterized using a light pulse at V_DS = 1 V under a wavelength of 980 nm at 120 µW. We confirm that the device exhibits stable and repeatable switching characteristics under NIR laser irradiation at 980 nm. Figure 3c shows the output characteristics of the photocurrents of the devices based on $I_{\mathit{ph}}=I_{\mathit{illumination}}-I_{\mathit{dark}}$ without and with the plasmonic film. It can be observed that the photocurrents of the plasmonic-MoS₂ photodetector are four times higher than those of the MoS₂ device. The illumination power dependence shows that the photocurrent increases linearly with the illumination power, for the cases without and with AgNPs. The external responsivity ${(R}_{\lambda })$ and detectivity $(D^{\ast })$ of the bare and plasmonic MoS₂ photodetectors are defined as $R_{\lambda }=\frac{I_{\mathit{ph}}}{P_{\mathit{Light}}}$ and $D^{\ast }=\frac{R_{\lambda }}{{(2q{I}_{\mathit{dark}}/A)}^{1/2}}$, respectively^6,8, where $I_{\mathit{ph}}=I_{\mathit{illumination}}-I_{\mathit{dark}}$ is the photocurrent, P_Light is the power of the incident light applied to the channel, A is the active area of the detector, q is the absolute value of an electron charge ($1.6\times {10}^{-19}$ C), R_λ is the responsivity measured in units of AW⁻¹, and D* is the detectivity measured in units of Jones.

Figure 3.

(a) I–V characteristics of the MoS₂ photodetectors. (b) Transient photocurrent of the photodetectors at V_DS = 1 V, and (c) photocurrent with respect to illumination power of the devices at the wavelength of 980 nm. The inset shows the responsivity as a function of illumination light power. (d–f) The data acquired similarly at the wavelength of 1,550 nm.

Thus, the responsivity of the plasmonic device is enhanced by approximately four times compared with that of the bare device within the optical power range of 1–120 µW shown in Fig. 3c. Specifically, at V_DS = 1 V, the responsivity values are ~ $8\times {10}^{-3}A{W}^{-1}$ and ~ 2 $\times {10}^{-3}A{W}^{-1}$ at optical power of 100 µW for the plasmonic and bare devices, respectively. At illumination intensities lower than 1 µW (e.g., at 0.2 µW), the AgNPs-MoS₂ photodetector showed photoresponsivities of $\sim 3.1\times {10}^{-3}A{W}^{-1}$, but the bare MoS₂ photodetector did not exhibit a photoresponse. The corresponding $D^{\ast }$ value of the plasmonic device (i.e., $\sim 1.2\times {10}^6$ Jones) is increased by ~ 3.8 times with respect to that of the bare device (i.e., $\sim 3.1\times {10}^5$ Jones). The enhancement of the $D^{\ast }$ value is slightly smaller than the enhancement of the R value because the dark current increases slightly for the plasmonic device.

Similarly, we performed the above characterization for the devices without and with the plasmonic AgNPs tape under a wavelength of 1,550 nm. From Fig. 3d, it is observed that all the current–voltage curves are linear and the obtained photocurrents of the AgNPs-MoS₂ photodetectors are significantly enhanced under the same power illumination compared with that of the bare MoS₂ photodetectors. As shown in Fig. 3e, we also confirmed the reproducible time-resolved photoresponse of the devices at 1,550 nm. The photocurrent has good linearity with the irradiation powers in the measured range in Fig. 3f. Consequently, the R and D* values of the plasmonic device are enhanced by 3.5 and 3.3 times, respectively, with respect to those of the bare device, where the R values are $\sim 1.1\times {10}^{-3}A{W}^{-1}$, $\sim 3.9\times {10}^{-3}A{W}^{-1}$, and the D* values are $\sim 1.7\times {10}^5$ Jones, $\sim 5.8\times {10}^5$ Jones, for the bare and plasmonic devices, respectively at the optical power of 50 µW. And, comparative I–V curves of the MoS₂ photodetectors under the illumination of different wavelengths are provided in Figure S5. Therefore, it was confirmed that the plasmonic tape strongly contributed to the increase in the MoS₂ photocurrent, by effectively absorbing NIR radiation not only at 980 nm but also at 1,550 nm. The device exhibited robust, stable, and repeatable characteristics. And, the previously reported photodetectors based on plasmonic-2D materials are summarized in Table S1. Unlike the previous reports, our device was fabricated by a 2D material thin film of centimeter scale by chemical exfoliation method, and it was possible to fabricate the device with a simple shadow mask process without complicated lithography process, and improve the device performance by post-processing of the plasmonic tape.

In addition, we measured the dark current of both devices and analyzed its noise spectral density. As shown in Fig. 4, the dark current of the device did not critically increase with the introduction of the plasmonic tape. Furthermore, in the analysis of the noise spectral density obtained using the fast Fourier transform (FFT) of the dark currents³¹, the plasmonic-tape-MoS₂ device exhibits similar noise level to the bare MoS₂ device. This indicates that the attached plasmonic tape does not affect the low-frequency noise characteristic of the bare device, which follows the 1/f noise theory. Although the metal NPs decorated on an active layer may possibly result in an increase of the dark current in the device, the introduction of our plasmonic tape does not cause a deterioration of such electrical characteristics.

Figure 4.

(a) Dark current waveform of the bare and plasmonic MoS₂ photodetectors. (b) Analysis of the noise spectral density of the bare and plasmonic MoS₂ photodetectors.

We employed a three-dimensional finite-difference time-domain (FDTD) simulation for the dimensions of 3 × 3 μm² in the XY plane in Fig. 5 to analyze the effect of the plasmonic tape. For the 3D simulation, an unpolarized plane wave source is applied normally to the plane in the backward direction with the boundary condition of perfectly matched layer (PML). The E-field intensity, (|E_un|²), is obtained by averaging the x- (|E_x-pol|²) and y-polarized (|E_y-pol|²) profiles (i.e., |E_un|² = 1/2(|E_x-pol|² +|E_y-pol|²)). The complex refractive index of MoS₂ used for simulation is approximately extracted from a literature³² and the index is 4 + 0.01i for the NIR wavelengths.

Figure 5.

(a) Schematic illustration of the modeling geometry used in the FDTD numerical simulation. (b) E-field intensity profiles in the XY plane at the interface between AgNPs and MoS₂ (i.e., z = 0 nm) when the wavelength of the incident light is 980 nm and 1,550 nm. (c) A plot of the integrated electromagnetic field intensity in the XY plane along the z-axis (∑_z|E(x,y)²|).

The plasmonic structure was modeled using Wolfram Mathematica (ver. 11.1.1.0) based on the SEM image in Fig. 2a. In the plasmonic-tape-MoS₂ configuration, a strongly localized near-field in the proximity of the AgNPs is observed (Fig. 5b), mainly resulting from the excitation of the localized surface plasmon (LSP) modes. The plasmoic tape-MoS₂ exhibits broadband light absorption including the visible and NIR wavelengths, as shown in the simulated absorption spectra of Figure S6. The randomly arranged AgNPs array in the plasmonic tape functions as a scattering center for broadband wavelength as well as a nano-antenna inducing the localization of E-filed due to the LSPs from an individual AgNP and interparticle interaction among AgNPs. The randomly distributed AgNPs array with various particle sizes and spacing extends the LSP excitation wavelength up to NIR range. Thus, the plasmonic-tape-MoS₂ configuration can meaningfully absorb the incident NIR radiation because of the LSP mode generated around the AgNPs. Specifically, at both 980 nm and 1,550 nm, strong hot spots are observed at the interface between MoS₂ and AgNPs, enhancing the integrated values of the E-field intensity at the interface (i.e., z = 0) as much as 45.5- and 21.6-fold for the wavelengths 980 nm and 1,550 nm, respectively, compared with those of the bare MoS₂, in Fig. 5c. As expected, the strongest increase in the electric field is observed at the interface between MoS₂ and AgNPs and it decreases exponentially as one moves away from the interface. In contrast, the electric field intensity remains nearly constant on the surface and inside the film for the bare MoS₂. Furthermore, the integrated values of the squared E-field over the total MoS₂ layer are enhanced by 9.6- and 4.5-fold on average for the plasmonic tape structure at the incident wavelengths of 980 nm and 1,550 nm, respectively, compared with those of the bare MoS₂. Therefore, the localized electric field leads to enhanced NIR absorption in MoS₂, which is mainly responsible for the increased photocurrent.

In addition to the enhanced NIR absorption in plasmonic tape-MoS₂, the plasmon excitation may sensitize NIR light, enabling the AgNPs to inject hot carriers into MoS₂ by absorbing NIR light, because hot electrons can be generated from plasmon decay of Ag nanostructures in the visible and NIR range^33–35. And, according to previous literatures^21–23,35, when metals interface with MoS₂ layers, they may function as localized sources of additional carriers because hot electrons induced from plasmon decay are rapidly transferred to MoS₂. Therefore, there is a possibility that the generation of hot electrons from the plasmonic tape and the transfer to MoS₂ contributes in part to the generation of photocurrent.

Discussion

In this study, we fabricated a multilayer MoS₂ device attached with a plasmonic tape to detect NIR wavelengths. The NIR photoresponse of the MoS₂ device was strongly enhanced by directly and simply attaching a Scotch tape with AgNPs onto the MoS₂ flakes. The plasmonic MoS₂ device exhibited strongly enhanced photoresponse up to the wavelength of 1,550 nm. We confirmed that the plasmonic-tape-attached MoS₂ device yielded approximately four times higher photocurrent compared with that of the bare device mainly due to the enhanced NIR absorption, without a noticeable increase in dark current. This plasmonic tape is believed to be applicable to any substrate, including organic substrates that are too weak to be subjected to high-temperature processes.

Methods

Preparation of the plasmonic tape

SiO₂/Si or Si substrates were cleaned via ultrasonication in acetone, washed with ethanol and isopropanol, and subsequently dried. The Trichlorododecylsilane (TCS) treatment of substrates is essential to detach AgNPs easily and simply from substrates using tape. The substrates were treated with UV-ozone for 15 min. They were then immersed into a TCS solution in toluene (99.9%) with 5% volume fraction for 24 h^36,37. Subsequently, they were cleaned via ultrasonication in toluene and then dried. Subsequently, a 10-nm-thick Ag film was deposited on the substrates using a thermal evaporator. The thin Ag film was then annealed to construct a disordered array of AgNPs on the substrate using a hot plate under the air condition at 220 °C for 1 min. After the annealed substrate was cooled, a 3 M tape was placed on the substrate such that there were no bubbles, and then peeled off. Thus, the plasmonic film was obtained.

Fabrication of MoS₂ photodetector

MoS₂ photodetectors were fabricated by transferring the chemically exfoliated MoS₂ film onto a polyimide (PI) substrate (Figure S1), and then constructing Au /Cr (100 nm/10 nm) electrodes using a metal shadow mask. The channel length and width of the mask were 20 µm and 200 µm, respectively. Subsequently, a plasmon-enhanced MoS₂ photodetector was fabricated by attaching the tape with AgNPs carefully onto the MoS₂ film. The chemically exfoliated MoS₂ films were fabricated according to a previously reported method^12,24.

Author contributions

H.K. designed the research plan. M.P. conceived and developed the concept and carried out all the experiments. H.K and G.K. analyzed the results and simulations, and all authors wrote manuscript.

Competing interests

The authors declare no competing interests.

Supplementary information

is available for this paper at 10.1038/s41598-020-68127-7.