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

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.

www.nature.com/scientificreports/ deteriorating the electrical and optical properties. We systematically investigate the optoelectrical properties and photodetection performance of the plasmonic MoS 2 device at the NIR wavelengths of 980 and 1,550 nm. We report that the plasmonic MoS 2 device yields a sensitivity approximately four times that of the bare MoS 2 device.

Results
The schematic fabrication process of the MoS 2 device attached with the plasmonic AgNPs-tape film is briefly illustrated in Fig. 1a 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 MoS 2 and plasmonic-tape-MoS 2 (i.e., AgNPs/MoS 2 ) films. Multilayered MoS 2 films were transferred onto glass substrates to measure absorption spectra. In Fig. 2c, the AgNPs/MoS 2 film shows absorption enhancement over a broad band of spectra at all wavelengths, including the visible range; thus, the MoS 2 film becomes less transparent upon attaching the plasmonic tape. In the inset of Fig. 2c, the bare MoS 2 film exhibits peaks A and B at 672 nm and 612 nm, respectively, corresponding to the two MoS 2 direct band gap transitions [25][26][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/MoS 2 sample exhibits nearly four times stronger absorption than the bare MoS 2 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 MoS 2 without and with the plasmonic tape, where a bare Scotch-tape-attached sample is used for comparison. The Raman spectrum of multilayered MoS 2 film displays an in-plane active mode E 2g 1 at ~ 384 cm −1 and an out-of-plane mode A 1g at 405 cm −1 . In addition, it exhibits additional peaks, 156 (J 1 ), 226 (J 2 ), and 333 (J 3 ) cm −1 , featuring 1 T-MoS 2 [28][29][30] . After attaching the plasmonic tape on top of MoS 2 , 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.  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 ph = I illumination − I dark without and with the plasmonic film. It can be observed that the photocurrents of the plasmonic-MoS 2 photodetector are four times higher than those of the MoS 2 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 ) and detectivity (D * ) of the bare and plasmonic MoS 2 photodetectors are defined as R = I ph P Light and D * = R (2qI dark /A) 1/2 , respectively 6,8 , where I ph = I illumination − I 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 × 10 −19 C), R λ is the responsivity measured in units of AW −1 , and D* is the detectivity measured in units of Jones.
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 × 10 −3 AW −1 and ~ 2 × 10 −3 AW −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-MoS 2 photodetector showed photoresponsivities of ∼ 3.1 × 10 −3 AW −1 , but the bare MoS 2 photodetector did not exhibit a photoresponse. The corresponding D * value of the plasmonic device (i.e., ∼ 1.2 × 10 6 Jones) is increased by ~ 3.8 times with respect to that of the bare device (i.e., ∼ 3.1 × 10 5 Jones). The enhancement of the D * 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-MoS 2 photodetectors are significantly enhanced under the same power illumination compared with that of the bare MoS 2 photodetectors. As shown in Fig. 3e, we also confirmed  Figure S5. Therefore, it was confirmed that the plasmonic tape strongly contributed to the increase in the MoS 2 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 postprocessing 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 currents 31 , the plasmonic-tape-MoS 2 device exhibits similar noise level to the bare MoS 2 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.
We employed a three-dimensional finite-difference time-domain (FDTD) simulation for the dimensions of 3 × 3 μm 2 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 MoS 2 used for simulation is approximately extracted from a literature 32 and the index is 4 + 0.01i for the NIR wavelengths.
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-MoS 2 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-MoS 2 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 www.nature.com/scientificreports/ NIR range. Thus, the plasmonic-tape-MoS 2 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 MoS 2 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 MoS 2 , in Fig. 5c. As expected, the strongest increase in the electric field is observed at the interface between MoS 2 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  www.nature.com/scientificreports/ film for the bare MoS 2 . Furthermore, the integrated values of the squared E-field over the total MoS 2 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 MoS 2 . Therefore, the localized electric field leads to enhanced NIR absorption in MoS 2 , which is mainly responsible for the increased photocurrent. In addition to the enhanced NIR absorption in plasmonic tape-MoS 2 , the plasmon excitation may sensitize NIR light, enabling the AgNPs to inject hot carriers into MoS 2 by absorbing NIR light, because hot electrons can be generated from plasmon decay of Ag nanostructures in the visible and NIR range [33][34][35] . And, according to previous literatures [21][22][23]35 , when metals interface with MoS 2 layers, they may function as localized sources of additional carriers because hot electrons induced from plasmon decay are rapidly transferred to MoS 2 . Therefore, there is a possibility that the generation of hot electrons from the plasmonic tape and the transfer to MoS 2 contributes in part to the generation of photocurrent.

Discussion
In this study, we fabricated a multilayer MoS 2 device attached with a plasmonic tape to detect NIR wavelengths. The NIR photoresponse of the MoS 2 device was strongly enhanced by directly and simply attaching a Scotch tape with AgNPs onto the MoS 2 flakes. The plasmonic MoS 2 device exhibited strongly enhanced photoresponse up to the wavelength of 1,550 nm. We confirmed that the plasmonic-tape-attached MoS 2 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. SiO 2 /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 MoS 2 photodetector. MoS 2 photodetectors were fabricated by transferring the chemically exfoliated MoS 2 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 MoS 2 photodetector was fabricated by attaching the tape with AgNPs carefully onto the MoS 2 film. The chemically exfoliated MoS 2 films were fabricated according to a previously reported method 12,24 .