Enhancement of Photodetective Properties on Multilayered MoS2 Thin Film Transistors via Self-Assembled Poly-L-Lysine Treatment and Their Potential Application in Optical Sensors

Photodetectors and display backplane transistors based on molybdenum disulfide (MoS2) have been regarded as promising topics. However, most studies have focused on the improvement in the performances of the MoS2 photodetector itself or emerging applications. In this study, to suggest a better insight into the photodetector performances of MoS2 thin film transistors (TFTs), as photosensors for possible integrated system, we performed a comparative study on the photoresponse of MoS2 and hydrogenated amorphous silicon (a-Si:H) TFTs. As a result, in the various wavelengths and optical power ranges, MoS2 TFTs exhibit 2~4 orders larger photo responsivities and detectivities. The overall quantitative comparison of photoresponse in single device and inverters confirms a much better performance by the MoS2 photodetectors. Furthermore, as a strategy to improve the field effect mobility and photoresponse of the MoS2 TFTs, molecular doping via poly-L-lysine (PLL) treatment was applied to the MoS2 TFTs. Transfer and output characteristics of the MoS2 TFTs clearly show improved photocurrent generation under a wide range of illuminations (740~365 nm). These results provide useful insights for considering MoS2 as a next-generation photodetector in flat panel displays and makes it more attractive due to the fact of its potential as a high-performance photodetector enabled by a novel doping technique.


Introduction
For achieving the ultimate scaling of Si devices below a sub-nanometer, direct tunneling of leakage current in the channel is a critical problem to be resolved, because it can lead to large power consumption [1,2]. In this regard, transition metal dichalcogenides (TMDCs) with a sizable and tunable energy bandgap, depending on the number of layers, have actively been studied as a potential candidate for the next generation of semiconductors. In addition, the versatile properties of TMDCs, including outstanding electrostatic gate coupling, no dangling bonds, high effective mass of electrons leading to less direct tunneling leakage, and unique optical/chemical properties, have generated plenty of research interest [3][4][5][6]. Among the varieties of the TMDC family, molybdenum disulfide (MoS 2 ) is a representative of TMDC material with easy preparation via CVD and/or exfoliation, high electron mobility (>100 cm 2 /Vs), and an abundant form of MoS 2 as natural minerals [7,8]. Until now, enormous research activities, involving basic transport mechanisms, circuit level demonstration for flat panel displays and FinFETs, new concepts of devices, and gas sensors, have been extensively performed [9][10][11][12][13][14][15][16][17]. In particular, among the numerous research fields of MoS 2 thin film transistors (TFTs), applications for photodetectors and display backplane transistors have been regarded as important and promising topics for the future technology with its novel electrical/optical properties [18][19][20]. Those studies for MoS 2 photodetectors have reported remarkable performances including broad band detection from ultraviolet to near-infrared (NIR), ultra-high photoresponsivity, flexible application, polarization sensitive photodetection, high-speed photoresponse, high spatially resolved imaging [21][22][23][24][25][26][27]. However, most of the studies have focused on the performances of the MoS 2 photodetector itself or emerging applications using them because of the easy preparation of the devices via exfoliated or CVD layers, their relatively stable device performance, and reproducible device implementation as compared with other TMDC layers. In order to suggest a better insight into photodetective properties for MoS 2 thin film transistors (TFTs) quantitatively as photosensors, possibly integrated with a display system, in this study, we carried out a comparative study on the photoresponse of the MoS 2 TFTs and hydrogenated amorphous silicon (a-Si:H) TFTs. This was due to the immediate availability of highly accumulated information in the literature in terms of the optical properties of a-Si:H-based TFTs and their understanding on device behaviors. a-Si:H has been used as conventional TFTs in the display industry, and photodetectors integrated with active matrix are conjugated as imaging scanners, optical feedback systems, or/and touch panels for interactive display operation [28][29][30][31]. Compared to a-Si:H TFTs, MoS 2 TFTs have demonstrated potential for both display backplane TFTs and photodetectors [18][19][20][21][22][23][24][25][26][27]. Furthermore, even with comparison of amorphous indium-gallium-zinc-oxide (a-IGZO) TFTs, which are presently one of the most promising display backplane TFTs with high mobility and maturity of material, MoS 2 TFTs possess advantages of detecting spectrum and negligible photo persistent current (PPC) effects [31,32]. Thus, this comparative study was expected to provide intuitive criteria for researchers in terms of the key aspects that are required for MoS 2 layers as one of the next-generation semiconductors for the future photodetectors in display systems.
Furthermore, as a strategy to improve the field effect mobility and photoresponse of the MoS 2 TFTs, we adopted a doping treatment in this work. Until now, various doping methods have been applied including high-k-based fixed-charge induction, moleculeinduced charge transfer, organic solvents-based doping, electrical doping via local electrode configuration, electrothermal doping, and gas-based doping. Among these techniques, molecular doping has been predominantly utilized owing to the immediately accessible approach with several advantages such as post process compatibility, unnecessity to deposit additional layers, and low temperature during the treatment on devices. In this regard, molecular doping by poly-L-lysine (PLL) treatment was applied to the MoS 2 TFTs. In general, PLL with amino group has been frequently used to make graphene and carbon nanotubes positively functionalized, which results in high affinity to induce negatively charged biomolecules [33][34][35]. This technique leads to the improvement on uniformity of nanomaterials coating and its corresponding control of sequence order associated with electrical polarity of molecule ending groups. On the other hand, PLL-based doping for TMDC and its corresponding evaluation on optical detectivity control, compared with conventional TFTs, for example, a-Si:H TFTs, has been reported to be limited. In this study, PLL was coated on the surface of MoS 2 layers to enhance the electron charges in the MoS 2 TFTs. Transfer and output characteristics of the MoS 2 TFTs clearly showed improved photocurrent generation under a wide range of illumination (740~365 nm). In addition, its chemical and physical properties were analyzed by X-ray spectroscopy (XPS), Raman spectroscopy, and atomic force microscopy (AFM) to confirm the doping's effects on the MoS 2 TFTs. As a result, along with the extracted photo responsivity and detectivity, PLL treatment demonstrated not only that it induced n-doping effects but also that it improved the photoresponse of the MoS 2 TFTs.
These results will provide useful insights for considering MoS 2 as a next-generation photodetector in flat panel displays and make it more attractive as a potential highperformance photodetector enabled by a novel doping technique.

Materials and Methods
Two kinds of photosensitive TFTs were fabricated as shown in Figure 1: (i) multilayered MoS 2 (m-MoS 2 ) TFT and (ii) a-Si:H TFT. Figure 1a shows an inverted staggered a-Si:H TFT composed of a 360 nm thick Al/Mo gate metal on a glass substrate, a 430 nm thick SiN x gate insulator, a 210 nm thick a-Si:H/n+a-Si:H layer, and a 430 nm thick Mo/Al/Mo source and drain contacts, which were based on one of the conventional process protocols for the implementation of a-Si:H fabrication in the TFT-LCD mass production lines. Figure  1b shows m-MoS 2 TFTs that were implemented on the thermally oxidized Si wafers with 10 nm thick SiO 2 . The heavily phosphorus-doped silicon wafer (ρ~0.005 ohm·cm) was initially used as a global gate. The multilayers of MoS 2 were mechanically exfoliated from bulk MoS 2 crystals (SPI Supplies, 429ML-AB, West Chester, PA, USA) and transferred onto Si substrates, with thermal oxide (~10 nm) as a gate insulator, using adhesive poly dimethylsiloxane (PDMS) elastomer. Then, immediate annealing was performed in mixed gas (~Ar/H 2 ) at a temperature of 400 • C for 1 h to remove organic residues and surface treatment on the MoS 2 films that might contaminate them during the transfer process [36]. Thereafter, 35 nm Au was evaporated using e-gun evaporators, followed by lifting off on a photo lithographically patterned area, forming the source/drain electrodes. Figure 1c displays an optical microscope image of fabricated a-Si:H TFT with a 4.6 µm channel length and a 40 µm channel width. Figure 1d exhibits the height profile of the m-MoS 2 by atomic force microscopy (AFM, Bruker, MULTIMODE-8-AM, Billerica, MA, USA), confirmed as 9.5 nm (corresponding to~15 layers). The insets display optical microscope images of fabricated m-MoS 2 TFT with a 10 µm channel length and a 30 µm channel width.

Materials and Methods
Two kinds of photosensitive TFTs were fabricated as shown in Figure 1: (i) multilayered MoS2 (m-MoS2) TFT and (ii) a-Si:H TFT. Figure 1a shows an inverted staggered a-Si:H TFT composed of a 360 nm thick Al/Mo gate metal on a glass substrate, a 430 nm thick SiNx gate insulator, a 210 nm thick a-Si:H/n+a-Si:H layer, and a 430 nm thick Mo/Al/Mo source and drain contacts, which were based on one of the conventional process protocols for the implementation of a-Si:H fabrication in the TFT-LCD mass production lines. Figure 1b shows m-MoS2 TFTs that were implemented on the thermally oxidized Si wafers with 10 nm thick SiO2. The heavily phosphorus-doped silicon wafer (ρ ~ 0.005 ohm·cm) was initially used as a global gate. The multilayers of MoS2 were mechanically exfoliated from bulk MoS2 crystals (SPI Supplies, 429ML-AB, West Chester, PA, USA) and transferred onto Si substrates, with thermal oxide (~10 nm) as a gate insulator, using adhesive poly dimethylsiloxane (PDMS) elastomer. Then, immediate annealing was performed in mixed gas (~Ar/H2) at a temperature of 400 °C for 1 h to remove organic residues and surface treatment on the MoS2 films that might contaminate them during the transfer process [36]. Thereafter, 35 nm Au was evaporated using e-gun evaporators, followed by lifting off on a photo lithographically patterned area, forming the source/drain electrodes. Figure   For n-type doping, poly-L-lysine (PLL) was purchased from Sigma-Aldrich (Seoul, Korea) and used as charge enhancing molecules for the m-MoS 2 TFTs. For the treatment, the fabricated m-MoS 2 TFTs were dipped in a water solution containing PLL (0.1% w/v) in ambient air, at room temperature. The dipping time was 2.5 h for the PLL treatment, followed by baking at 100 • C on a hot plate for 10 min. A Raman spectrometer (WITEC alpha300, Ulm, Germany) was used to analyze the Raman spectra and PL intensity of the TMDC flakes with a 532 nm laser excitation and 1 µm of beam size. Light for comparison of the photo response between a-Si:H and m-MoS 2 TFTs was illuminated by the laser (wavelength: 630 nm, 530 nm, and 450 nm; power: 1 µW~20 mW) and a multi-wavelength fiber-coupled LED source (Mightex, Inc., Pleasanton, CA, USA) with various wavelength (365 nm, 455 nm, 530 nm, 656 nm, and 740 nm), and 3 mW cm −2 of power density was employed to measure the photocurrent of m-MoS 2 TFTs before and after PLL treatment. All the electrical characterizations were measured with a semiconductor impedance analyzer (Agilent 4155C, Seoul, Korea) in ambient air at room temperature.  , where g m is the transconductance, C OX is the dielectric capacitance, L and W indicate channel length and width, and V DS ·is drain-to-source voltage, respectively. The subthreshold swing (SS) of the m-MoS 2 transistor was~200 mV/dec, while it was 1 V/dec with the a-Si:H TFTs. The extracted electrical properties indicate that the quality of fabricated TFTs was reasonable compared with previous results in the literature [31,[37][38][39]; thus, the photoresponse of both TFTs were ready to be compared without consideration of noticeable defects or flaws.

Results and Discussion
For a fair comparison, both TFTs had similar channel dimensions, the distance between light source and device was fixed, and the duration of illuminated light was identically set. Then, to investigate the photoelectrical properties of the a-Si:H and m-MoS 2 TFTs, the transfer characteristics under illumination with different wavelength and intensities were measured. Figure Figure 2d-f. Power intensity varies from 1 µW to 20 mW for 660 nm laser, from 1 µW to 5 mW for 530 nm laser, and from 10 µW to 5 mW for 450 nm laser. Measured data, as shown in Figure 2, revealed the generation of photocurrent (I photo = I light − I dark ) of a-Si:H TFTs under illumination, corresponding to all wavelengths in this test. The generation of a photocurrent can be explained according to the following scenario: when the light illuminated the channel materials, photon energy of light produced the electron-hole pairs in the channel region. Then, the excited electron-hole pairs drifted along the channel by the applied lateral E-field associated with V DS , resulting in the increase of drain-to-source current. In addition, the photocurrent of a-Si:H was enhanced as the light intensity increased, that is, a stronger optical power generates more electron-hole pairs from the channel materials. Figure 2 presents a proper photo-transistor operation of a-Si:H TFTs, and the linear scale graphs indicate a gradual photocurrent increase at the on-state (V GS > 0 V) without degradation of the field-effect mobility. With the same procedures for the evaluation of the photoresponse, Figure 3 shows the transfer characteristics of the m-MoS 2 TFTs under illumination with different wavelength and intensities. The m-MoS 2 TFTs also exhibit proper photo-transistor operation, and a parallel shift of V th was observed in the linear scale graphs. When we visually compared the photoresponse for both TFTs, it can be directly seen that the photocurrent of the a-Si:H TFTs was at an nA level, whereas the m-MoS 2 TFTs was at a µA level. Furthermore, the m-MoS 2 TFTs had a sensitive dependency on the wavelength of light, and these photoresponses can be potentially tailored by engineering the layer of light, and these photoresponses can be potentially tailored by engineering the layer thickness [40,41]. Thus, these results elucidate a better potential of MoS2 for the versatile and high-performance photodetectors.  of light, and these photoresponses can be potentially tailored by engineering the layer thickness [40,41]. Thus, these results elucidate a better potential of MoS2 for the versatile and high-performance photodetectors.  One of the most important figures of merit for a photodetector is its external photoresponsivity and detectivity. To quantitatively compare the performance of photoresponse, both photoresponsivity and detectivity were extracted and shown in Figure 4. The photoresponsivity and detectivity were calculated from equations R = I photo /P light , D = R × S 1/2 (2qI dark ) 1/2 , where P is a total incident optical power, S is the effective illuminated area, q is the electron charge, and I dark is dark current. I photo and I dark are the current levels in the off regime, for the bias condition, at V GS = −1.2 V for m-MoS 2 TFTs and V GS = −1.5 V for a-Si:H TFTs, respectively, under either light illumination or dark condition. Figure 4a shows extracted photoresponsivity of m-MoS 2 and a-Si:H TFTs. The photoresponsivities of the m-MoS 2 TFTs ranged from 10 2 to 10 4 for RGB light, and those of the a-Si:H TFTs were within 10 −1 , wherein the responsivities of the a-Si:H TFTs were similar regardless of the wavelength as reported in the literature [31]. In the all of wavelengths and optical power ranges, the m-MoS 2 TFTs exhibited 2~5 orders larger responsivities than that of the a-Si:H TFTS, which is attributed to different photocurrent value for both TFTs. In addition, Figure 4b presents the detectivity of the m-MoS 2 and a-Si:H TFTs, from which detectivities of the m-MoS 2 TFTs were 10 11~1 0 13 and those of the a-Si:H TFTs were extracted as 10 8~1 0 9 . Due to the noticeable current ratio between I photo and I dark , the m-MoS 2 TFTs had clearly higher detectivities than a-Si:H TFTs. As a result, in the all of wavelength and optical power range, the m-MoS 2 TFTs exhibited 2~4 orders larger detectivities. The overall quantitative comparison of responsivities and detectivities for the m-MoS 2 and a-Si:H TFTs are summarized in Figure 4c,d, which obviously confirms a much better performance of MoS 2 photodetector. both photoresponsivity and detectivity were extracted and shown in Figure 4. The photoresponsivity and detectivity were calculated from equations R = Iphoto/Plight, D = R × S 1/2 (2qIdark) 1/2 , where P is a total incident optical power, S is the effective illuminated area, q is the electron charge, and Idark is dark current. Iphoto and Idark are the current levels in the off regime, for the bias condition, at VGS = −1.2 V for m-MoS2 TFTs and VGS = −1.5 V for a-Si:H TFTs, respectively, under either light illumination or dark condition. Figure 4a shows extracted photoresponsivity of m-MoS2 and a-Si:H TFTs. The photoresponsivities of the m-MoS2 TFTs ranged from 10 2 to 10 4 for RGB light, and those of the a-Si:H TFTs were within 10 −1 , wherein the responsivities of the a-Si:H TFTs were similar regardless of the wavelength as reported in the literature [31]. In the all of wavelengths and optical power ranges, the m-MoS2 TFTs exhibited 2~5 orders larger responsivities than that of the a-Si:H TFTS, which is attributed to different photocurrent value for both TFTs.  In addition, as a circuit level photodetector, photosensitive inverters were implemented and compared using a-Si:H and m-MoS2 TFTs in Figure 5. These photoinverters are key component of light-to-frequency conversion circuits (LFCs), which are practically beneficial for the future IoT systems required for a high level of security [37][38][39]42]. To  In addition, as a circuit level photodetector, photosensitive inverters were implemented and compared using a-Si:H and m-MoS 2 TFTs in Figure 5. These photoinverters are key component of light-to-frequency conversion circuits (LFCs), which are practically beneficial for the future IoT systems required for a high level of security [37][38][39]42]. To demonstrate the capabilities of a-Si:H and m-MoS 2 TFTs in a circuit-level photodetector, the photo response of depletion load enhancement driver (DLED) inverters was measured. In the configuration of an a-Si:H DLED inverter, the channel width of the load TFT (W load ) was varied to change the effective illuminated area of the a-Si:H channel under light illumination. Figure 5a displays the voltage transfer characteristics (VTCs) of an a-Si:H DLED inverter with 40 µm of W load . In dark conditions, the switching behavior of the inverter was poor, whereas it is improved under light illumination. This is because sufficient load currents are required to obtain reasonable swing performance in VTCs for a DLED inverter. Thus, under illumination, increased currents yielded improved switching behaviors. However, in this case, VTC curves were not distinguished depending on different wavelength of light (i.e., R, G, B), so that the requirements of photosensitive inverters were not satisfied. To secure sufficient load currents, a-Si:H DLED inverters with large W load of 4000 and 8000 µm were utilized. As shown in Figure 5b,c, in dark conditions, decent switching properties were obtained with 4000 µm W load and full swing characteristics were achieved with 8000 µm W load . Furthermore, minimum V OUT (V OL ) values did not reach the value of zero due to the fact of too much photocurrent stemming from a large illuminating area; thus, there waw a tradeoff between the stability of the inverter operation and photoresponse. On the contrary, the DLED inverter composed of m-MoS 2 TFTs with 30 µm W load exhibited stable full-swing characteristics under dark and illumination at all wavelengths, which is attributed to better mobility and photoresponse than the a-Si:H TFTs. Consequently, Figure 5 indicates that the a-Si:H TFTs require at least 8000 µm of channel width for the application of photosensitive inverters, compared to the 30 µm of the m-MoS 2 TFTs. These results elucidate that m-MoS 2 TFTs have robust advantages with regard to the level of device integration per area, chip density, and sensitive modulation properties in variations of wavelength from dark to blue compared to the a-Si:H TFTs. demonstrate the capabilities of a-Si:H and m-MoS2 TFTs in a circuit-level photodetector, the photo response of depletion load enhancement driver (DLED) inverters was measured. In the configuration of an a-Si:H DLED inverter, the channel width of the load TFT (Wload) was varied to change the effective illuminated area of the a-Si:H channel under light illumination. Figure 5a displays the voltage transfer characteristics (VTCs) of an a-Si:H DLED inverter with 40 μm of Wload. In dark conditions, the switching behavior of the inverter was poor, whereas it is improved under light illumination. This is because sufficient load currents are required to obtain reasonable swing performance in VTCs for a DLED inverter. Thus, under illumination, increased currents yielded improved switching behaviors. However, in this case, VTC curves were not distinguished depending on different wavelength of light (i.e., R, G, B), so that the requirements of photosensitive inverters were not satisfied. To secure sufficient load currents, a-Si:H DLED inverters with large Wload of 4000 and 8000 μm were utilized. As shown in Figure 5b,c, in dark conditions, decent switching properties were obtained with 4000 μm Wload and full swing characteristics were achieved with 8000 μm Wload. Furthermore, minimum VOUT (VOL) values did not reach the value of zero due to the fact of too much photocurrent stemming from a large illuminating area; thus, there waw a tradeoff between the stability of the inverter operation and photoresponse. On the contrary, the DLED inverter composed of m-MoS2 TFTs with 30 μm Wload exhibited stable full-swing characteristics under dark and illumination at all wavelengths, which is attributed to better mobility and photoresponse than the a-Si:H TFTs. Consequently, Figure 5 indicates that the a-Si:H TFTs require at least 8000 μm of channel width for the application of photosensitive inverters, compared to the 30 μm of the m-MoS2 TFTs. These results elucidate that m-MoS2 TFTs have robust advantages with regard to the level of device integration per area, chip density, and sensitive modulation properties in variations of wavelength from dark to blue compared to the a-Si:H TFTs.

Improvement in the Photoresponse of m-MoS 2 TFTs by Molecular Doping Technique
In addition to revealing the high optoelectrical performance of m-MoS 2 TFTs by comparison with a-Si:H TFTs, the photoresponsivity can be enhanced by molecular doping treatment [43][44][45]. Although there are different techniques to improve the photodetector's performance, the molecular doping method has several advantages, as it does not need a change in the device's structure or the addition of different channel materials or layers but consist of a low temperature process, ultra-thin thickness, and post-process compatibility after device fabrication. In this regard, poly-L-lysine (PLL) was adopted to enhance the photoresponsivity via n-doping effect for m-MoS 2 TFTs from PLL molecules. where L and W are the length and width of channel and q is the electron charge. As a consequence of charge enhancement, µ FE was extracted as 13.0 and 18.3 cm 2 /V·s before and after PLL treatment, respectively, which reveals a 40% increase in µ FE . In parallel, n 2D was calculated as 2.0 × 10 12 and 2.4 × 10 12 before and after PLL treatment, respectively, from which 20% enhancement of n 2D was obtained. These results clearly show the improvement of electrical properties by treatment of PLL.  Thereafter, for better understanding on electrically observed doping effects by PLL, chemical and physical properties and their analysis on m-MoS 2 TFTs were examined. First, optical properties were examined via photoluminescence (PL) spectroscopy. Figure 7a shows PL spectra of m-MoS 2 TFTs after PLL treatment, from which peak intensity after PLL treatment was reduced because donor-like doping effects enhanced the formation of tightly bound trions of MoS 2 [45][46][47], leading to a decrease in the radiative recombination of excitons. With the extracted optical properties, electrical behaviors, corresponding to each optical excitation, were investigated in the following section. As reported elsewhere [45,[47][48][49][50][51][52][53][54], Raman spectroscopy also has been dominantly used as a reliable tool to confirm the doping effects on TMDCs. Figure 4a plots the Raman spectra of the m-MoS 2 TFTs before and after the treatment of PLL. In the Figure 4a, two characteristic vibrational modes (E 1 2g and A 1g ) were observed near 383 and 409 cm −1 in the bare m-MoS 2 flakes, where E 1 2g mode was attributed to the in-plane vibration between Mo and S atoms, whereas the A 1g mode was due to the out-of-plane vibration between Mo and S atoms. After the PLL treatment, the E 1 2g and A 1g peaks were shifted left by 0.25 cm −1 and 0.63 cm −1 for m-MoS 2 TFTs, respectively. Obviously observed left shifts of two characteristics mode in Raman peaks describe that n-doping effects of PLL treatment result from the increase of the electron-phonon scattering due to the higher electron concentration [45,[47][48][49]51,52,55]. Thus, PLL treatment would enhance the electron density, leading to increase of electronphonon scattering, which lends phonon frequency decreased [54]. The Raman results support the n-doping effects on m-MoS 2 TFTs, which are attributed to the PLL treatment, and their chemical analysis via Raman nicely matched with the entire trend of electrical properties as shown in Figure 6. flake [51,56], resulting in possible tuning of electrical properties for m-MoS2 FFTs.  To validate the doping effects on improvement of photoresponse, variations of currents under illumination were monitored before and after PLL treatment. Figure 8a presents transfer characteristics (e.g., 3-terminal operation) of the m-MoS2 TFTs without any To validate the doping effects on improvement of photoresponse, variations of currents under illumination were monitored before and after PLL treatment. Figure 8a presents transfer characteristics (e.g., 3-terminal operation) of the m-MoS 2 TFTs without any doping treatment. Under illumination from the visible to UV range (740~365 nm), an increase in currents was observed in the bare m-MoS 2 TFTs, which are commonly reported behaviors in the literature [37][38][39][40]. Furthermore, an increase in the current was gradually enhanced as the wavelength of the light decrease (i.e., as photon energy increase) [37][38][39][40]. Thereafter, time-dependent photoresponse in two-terminal operation was measured. Figure 8b shows photocurrents of bare m-MoS 2 TFTs at V DS = 1 V. As similar with transfer characteristics in Figure 8a, increases in the photocurrent are gradually shown as the wavelengths of the light decreased. Then, after PLL treatment for bare the m-MoS 2 TFTs, the photoresponses were measured with same condition. Figure 8c displays the transfer characteristics of the PLL treated the m-MoS 2 TFTs. It is shown that after PLL treatment, the transfer curves shifted left in dark conditions, compared to the bare m-MoS 2 TFTs. Moreover, the currents of the PLL-treated m-MoS 2 TFTs under illumination were also higher than those of the bare m-MoS 2 TFTs, which might be attributed to enhancement of the photoresponse. However, the increased currents of the PLL-treated m-MoS 2 TFTs under illumination than the bare m-MoS 2 TFTs might be regarded as only the results of V th shift. Therefore, to clearly confirm the improvement in the photoresponse, photocurrents were extracted in Figure 8d. The larger photocurrents of the PLL-treated m-MoS 2 TFTs were evidently observed under all wavelength ranges. In addition, the µ FE of the bare m-MoS 2 TFTs was augmented at 33% under illumination of 365 nm compared to dark conditions, whereas the PLL-treated m-MoS 2 TFTs had a 92% enhanced µ FE under illumination of 365 nm. Lastly, responsivity and detectivity were examined as quantitative evaluations of photoresponse. Figure 8e presents 2~6 times higher responsivities and 1.3~3.5 times enlarged detectivities of the PLL-treated m-MoS 2 TFTs, compared to the bare m-MoS 2 TFTs. These behaviors are well matched with previous results [43][44][45]47]. As a result, Figure 8 validates the improved photoresponse of the m-MoS 2 TFTs after PLL treatment, which is possibly due to the increased number of tightly bound trions and their lifetime as shown in Figure 7a [46,57].  [37][38][39][40]. Furthermore, an increase in the current was gradually enhanced as the wavelength of the light decrease (i.e., as photon energy increase) [37][38][39][40]. Thereafter, time-dependent photoresponse in two-terminal operation was measured. Figure 8b shows photocurrents of bare m-MoS2 TFTs at VDS = 1 V. As similar with transfer characteristics in Figure 8a, increases in the photocurrent are gradually shown as the wavelengths of the light decreased. Then, after PLL treatment for bare the m-MoS2 TFTs, the photoresponses were measured with same condition. Figure 8c displays the transfer characteristics of the PLL treated the m-MoS2 TFTs. It is shown that after PLL treatment, the transfer curves shifted left in dark conditions, compared to the bare m-MoS2 TFTs. Moreover, the currents of the PLL-treated m-MoS2 TFTs under illumination were also higher than those of the bare m-MoS2 TFTs, which might be attributed to enhancement of the photoresponse. However, the increased currents of the PLL-treated m-MoS2 TFTs under illumination than the bare m-MoS2 TFTs might be regarded as only the results of Vth shift. Therefore, to clearly confirm the improvement in the photoresponse, photocurrents were extracted in Figure 8d. The larger photocurrents of the PLL-treated m-MoS2 TFTs were evidently observed under all wavelength ranges. In addition, the μFE of the bare m-MoS2 TFTs was augmented at 33% under illumination of 365 nm compared to dark conditions, whereas the PLL-treated m-MoS2 TFTs had a 92% enhanced μFE under illumination of 365 nm. Lastly, responsivity and detectivity were examined as quantitative evaluations of photoresponse. Figure 8e presents 2~6 times higher responsivities and 1.3~3.5 times enlarged detectivities of the PLL-treated m-MoS2 TFTs, compared to the bare m-MoS2 TFTs. These behaviors are well matched with previous results [43][44][45]47]. As a result, Figure 8 validates the improved photoresponse of the m-MoS2 TFTs after PLL treatment, which is possibly due to the increased number of tightly bound trions and their lifetime as shown in Figure 7a [46,57]. In addition, the response and recovery time for the m-MoS 2 and a-Si:H TFTs were extracted at 90% and 10% of maximum photo current. Overall, bare m-MoS 2 TFTs had a rise (or decay) time in the range from 3.5 (or 6.1) to 7.4 (or 6.9) s, respectively, under wavelengths from 365 to 740 nm, whereas PLL-treated m-MoS 2 TFTs possessed rise (or decay) time from 2.6 (or 7.9) to 6.1 (or 9.8) sec of rise time, respectively. The shortened response time and prolonged recovery time might be attributed to the enhanced formation of tightly bound trions with PLL treatment in the m-MoS 2 TFTs [46], leading to a decreased recombination rate of excitons. Reduced PL intensities of the m-MoS 2 TFTs with PLL treatment supports the decreased recombination rate of excitons as shown in Figure 7a. Therefore, the prolonged lifetime of carriers can induce faster response times of photoexcited carriers and their slower recovery time with PLL treatment. However, response and recovery time of a-Si:H TFTs were extracted within a 10 ms range, which is consistent with other literature [31]. Thus, response and recovery time of the m-MoS 2 TFTs remained as a problem to be resolved yet. This should be accompanied with an improvement of material, insulator, and passivation, and interface engineering.
Overall, the performance of m-MoS 2 photodetector itself is not overwhelmed, as compared with the previously reported data in the literature. However, in this study, it is meaningful that the development of generally applicable doping methodology for performance improvement of TMDC photodetectors and its comparative study with conventional products of a-Si:H TFTs in order to understand the figures of merit quantitatively. The present works could provide the intuitive criteria from the perspective of researchers in terms of key aspects that are required for MoS 2 layers as one of the next-generation semiconductors for the future photodetectors in display systems.

Conclusions
In this study, a comparative study of the photoresponse of m-MoS 2 and a-Si:H TFTs was conducted to provide better insight into the photodetector performances of m-MoS 2 TFTs as photosensors integrated with display systems. With extracted 2~4 orders larger photo responsivity and detectivity of m-MoS 2 TFTs, it revealed better performance of m-MoS 2 photodetectors in the various wavelengths and optical power ranges. Furthermore, the photosensitive m-MoS 2 DLED inverter showed robust advantages with regard to level of device integration per area, chip density, and sensitive modulation properties in variations of wavelength from dark to blue as compared to an a-Si:H DLED inverter. In addition, as a strategy to improve the field effect mobility and photoresponse of the m-MoS 2 TFTs, molecular doping by PLL treatment was applied to the m-MoS 2 TFTs. Transfer and output characteristics of the m-MoS 2 TFTs clearly showed improved photocurrent generation under a wide range of illuminations (740~365 nm), and its chemical and physical properties were analyzed by XPS, Raman spectroscopy, and AFM to confirm the doping effects on m-MoS 2 TFTs. As a result, along with extracted photo responsivity and detectivity, PLL treatment demonstrated not only the ability to induce n-doping effects but also to improve the photoresponse of the m-MoS 2 TFTs. This comparative study will provide an intuitive criteria for m-MoS 2 TFTs to be utilized as future photodetectors in flat panel displays, and it will make them more attractive with their potential for high-performance photodetection enabled by a novel doping technique.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.